Abstract
Here we have investigated the role of the protein caveolin 1 (Cav1) and caveolae in the secretion of the white adipocyte hormone adiponectin. Using mouse primary subcutaneous adipocytes genetically depleted of Cav1, we show that the adiponectin secretion, stimulated either adrenergically or by insulin, is abrogated while basal (unstimulated) release of adiponectin is elevated. Adiponectin secretion is similarly affected in wildtype mouse and human adipocytes where the caveolae structure was chemically disrupted. The altered ex vivo secretion in adipocytes isolated from Cav1 null mice is accompanied by lowered serum levels of the high-molecular weight (HMW) form of adiponectin, whereas the total concentration of adiponectin is unaltered. Interestingly, levels of HMW adiponectin are maintained in adipose tissue from Cav1-depleted mice, signifying that a secretory defect is present. The gene expression of key regulatory proteins known to be involved in cAMP/adrenergically triggered adiponectin exocytosis (the beta-3-adrenergic receptor and exchange protein directly activated by cAMP) remains intact in Cav1 null adipocytes. Microscopy and fractionation studies indicate that adiponectin vesicles do not co-localise with Cav1 but that some vesicles are associated with a specific fraction of caveolae. Our studies propose that Cav1 has an important role in secretion of HMW adiponectin, even though adiponectin-containing vesicles are not obviously associated with this protein. We suggest that Cav1, and/or the caveolae domain, is essential for the organisation of signalling pathways involved in the regulation of HMW adiponectin exocytosis, a function that is disrupted in Cav1/caveolae-depleted adipocytes.
Introduction
White adipose tissue, well known to be important for lipid storage, has emerged also as an important endocrine organ. One of the key hormones secreted from mature adipocytes is adiponectin. A high plasma level of adiponectin is associated with a lower risk of developing type 2 diabetes (Li et al. 2009) as well as with an improved β-cell function (Nakamura et al. 2018) and can thus be regarded as a marker of metabolic health (reviewed in Mather & Goldberg 2014). Additionally, exogenous increases of adiponectin in vivo have been shown to reduce insulin resistance in mice (Yamauchi et al. 2001). Hypoadiponectinemia is typically present in individuals with obesity-associated type 2 diabetes and this is thought to play a central role in the development of metabolic dysfunction (Arita et al. 1999, Spranger et al. 2003, Kadowaki et al. 2006). Adiponectin exists in different multimeric molecular forms that can exert tissue-specific effects. The most abundant forms in human plasma are medium-molecular weight (MMW) and high-molecular weight (HMW) adiponectin (Brochu-Gaudreau et al. 2010). Importantly, circulating HMW adiponectin has been suggested to be a better predictor of insulin resistance in humans than the total adiponectin level (Hara et al. 2006) and to be specifically reduced in type 2 diabetes (Basu et al. 2007).
Although a few studies have investigated the control of adiponectin release (Scherer et al. 1995, Bogan & Lodish 1999, Xie et al. 2008), details of the molecular regulation of adiponectin exocytosis remain elusive. However, our own previous work has shown that adiponectin exocytosis/secretion is triggered by elevated intracellular levels of cAMP and ensuing activation of Epac (exchange protein activated by cAMP). In contrast to how hormone secretion is regulated in the majority of endocrine cell types (Burgoyne & Morgan 2003), Ca2+ is not essential for adiponectin exocytosis, but an increase of intracellular Ca2+ enhances the cAMP-triggered adiponectin secretion (Komai et al. 2014, El Hachmane et al. 2015). We have moreover demonstrated that catecholamine binding to beta-3-adrenergic receptors (β3ARs) stimulates adiponectin exocytosis and that adrenergically triggered adiponectin release is blunted in obese mice in a state of catecholamine resistance (defined as lower abundance of β3ARs and Epac, isoform1; Epac1; Komai et al. 2016). Adiponectin release can also be induced by insulin (Scherer et al. 1995, Bogan & Lodish 1999, Lim et al. 2015). The regulation of insulin-stimulated adiponectin release is incompletely characterized but has been shown to depend on activation of phosphoinositide-3-kinase (Pereira & Draznin 2005, Cong et al. 2007, Blumer et al. 2008, Lim et al. 2015).
The adipocyte plasma membrane is covered with bulb-like invaginations called caveolae. Maintenance of the caveolae structure requires caveolin-1 (Cav1; Lipardi et al. 1998). Adipocyte caveolae have been demonstrated to be involved in insulin signalling (Fagerholm et al. 2009), fatty acid transport (Ortegren et al. 2006) and triacylglycerol synthesis (Ost et al. 2005), suggesting that they function as metabolic platforms. Caveolae have also been shown to comprise micro-domains for cAMP and Ca2+ signalling (Willoughby & Cooper 2007) and proteins involved in the regulation of neuroendocrine cell exocytosis are confined to lipid raft domains (Salaun et al. 2004, Lang 2007), such as caveolae. Cav1 null mice exhibit decreased circulating levels of adiponectin as well as adipose tissue metabolic and mitochondrial dysfunction. Interestingly, the HMW form of adiponectin was specifically reduced in serum, while trimeric forms of adiponectin tended to be increased instead (Asterholm et al. 2012). Knockdown of Cav1 has been demonstrated to result in disruption of the caveolae structure (Liu et al. 2002). The Cav1 proteins have, in addition, been shown to form functional scaffolds outside of the caveolae regions (Head & Insel 2007, Lajoie et al. 2009).
Here we have investigated the role of Cav1 in adiponectin secretion using primary adipocytes isolated from Cav1 null mice. We show that the secretion of adiponectin (in response to a 30 min incubation with cAMP-elevating agents or insulin) is abrogated in Cav1-deficient adipocytes and that this is associated with reduced serum HMW adiponectin in Cav1 null mice. In addition, the basal (unstimulated) release of adiponectin is increased. A role for caveolae/Cav1 is proposed for adiponectin secretion also in human adipocytes, as demonstrated by the blunted stimulated adiponectin secretion in adipocytes pre-treated with methyl-β-cyclodextrin (MβCD). Our results suggest that Cav1 and/or the caveolae structure is essential for adiponectin release and especially for the secretion of HMW adiponectin.
Materials and methods
Materials
Subcutaneous abdominal fat tissue was obtained from female patients during elective surgery under general anaesthesia at the University Hospital in Linköping or at Sahlgrenska University Hospital, Sweden. The Local Ethics Committees approved the procedures which were in concordance with the Declaration of Helsinki. The subjects had an average age of 51 (range 42–66), an average BMI of 26.6 (range 23.8–30.8) and were not diagnosed with type 2 diabetes.
Inguinal white adipose tissue (IWAT) was isolated from wildtype or Cav-1 null mice (littermates) and from wildtype mice fed chow or high-fat diet (HFD; 60% caloric intake from fat) through 16 weeks. All mice had a C57BL/6J background (The Jackson Laboratory). Animal work was approved by the Regional Ethical Review Board in Gothenburg.
Isolation of mouse and human adipocytes
Adipose tissue was isolated from mice and patients as described in Komai et al. (2016). Collagenase digestion was used to isolate the adipocytes and cells were kept in Krebs–Ringer solution containing 0.12 M NaCl, 4.7 mM KCl, 2.5 mM CaCl2, 1.2 mM MgSO4, 1.2 mM KH2PO4, 20 mM Hepes, pH 7.4, 1% (w/v) fatty acid-free BSA with 100 nM phenylisopropyladenosine, 0.5 U/mL adenosine deaminase, and 2 mM glucose at 37°C, shaking.
Adiponectin secretion
Freshly isolated human adipocytes (200 μL packed cells/mL) or mouse adipocytes (100 μL packed cells/mL) were washed with an extracellular solution (EC) containing (in mM): 140 NaCl, 3.6 KCl, 2 NaHCO3, 0.5 NaH2PO4, 0.5 MgSO4, 5 HEPES (pH 7.4 with NaOH), 2.6 CaCl2 and 5 glucose. Subsequently, test substances were added as indicated and adipocytes were incubated with gentle shaking for 30 min at 32°C. The incubation was terminated by separation of the cells from the media by centrifugation through diisononyl phthalate (Sigma-Aldrich) and instant freezing on dry ice. Tubes were cut through the diisononyl phthalate layer to separate the cells from the media. Cells were lysed in PBS containing SDS (2%) and protease inhibitor (1 tablet/10 mL; cOmplete Mini; Roche Diagnostics). Secreted total adiponectin (measured with mouse ELISA DuoSets; R&D Systems) and HMW adiponectin (measured with mouse HMW adiponectin specific ELISA: MBS742124; (www.MyBioSource.com) were expressed in relation to total protein content (Bradford protein assay; Bio-Rad).
Measurements of serum adiponectin
Blood was collected during termination from the axillary vessels and left to clot for 1 h in room temperature, thereafter spun to isolate the serum fraction which was frozen until analysis. Total adiponectin and HMW adiponectin were measured by ELISA (respectively: EZMADP-60K; EMD Millipore and MBS742124, www.MyBioSource.com) according to the manufacturers recommendations.
Blood/serum glucose and insulin
Mice were fasted for 3 h and blood or serum glucose levels were measured using Contour Next glucometer (Ascensia Diabetes Care Holdings AG, Basel, Switzerland). Serum insulin was analysed with mouse insulin ELISA kit (Mercodia, Uppsala, Sweden).
Measurements of cAMP content
cAMP levels in adipose tissue were measured using Cyclic AMP XP Kit (Cell Signaling Technology) according to the manufacturer’s instructions. Protein concentrations were determined using Bradford protein assay (Bio-Rad).
Electron microscopy and immunogold labelling
Isolated adipocytes were attached to poly-l-lysine coated and Formvar covered nickel grids. Plasma membrane sheets were obtained by flushing attached cells with 150 mM KCl, in 2 mM Tris pH 7.4 prior to fixation in 3% paraformaldehyde and 0.5% glutaraldehyde. The grids were blocked in 5% BSA-c (Aurion, Wageningen, The Netherlands) and 0.1% gelatine for 1 h at 37°C. Primary anti-adiponectin antibodies were added and incubated overnight at 4°C where after the secondary 15-nm colloidal gold-conjugated goat anti-rabbit IgG antibody (Aurion) was added for 1 h at 37°C. After washing the grids in 200 mM sodium cacodylate, 200 mM sucrose and 5 mM CaCl2 (pH 7.2) they were additionally fixed in 2% glutaraldehyde for 10 min and then washed and incubated with 4% OsO4 for 30 min. The grids were quickly frozen in liquid propanol before they were freeze-dried and covered with 2 nm tungsten by magnetron sputtering directly in the freeze-dryer. Transmission electron microscopy was performed with a Jeol EX1200 TEM-SCAN (Tokyo, Japan).
Preparation of plasma membrane fraction
Adipocytes were homogenized at room temperature in 10 mM Tris (pH 7.4), 1 mM EDTA, 0.25 M sucrose, 25 mM NaF, 1 mM Na4-PPi, 0.5 mM EGTA, 4 mM iodoacetate, 10 mM leupeptin, 1 mM pepstatin, 1 mM aprotinin, and 100 mM PMSF. All subsequent procedures were carried out at 0–4°C. Plasma membrane was prepared by centrifugation of the homogenate at 16,000 g for 20 min and the pellet was further fractionated by discontinuous sucrose density gradient ultracentrifugation to obtain the purified plasma membrane fraction (Oka & Czech 1984).
Isolation of caveolae and subfractionation of caveolae in subclasses
The total caveolae fraction was prepared without detergent form purified plasma membranes by resuspending the plasma membrane fraction in 0.5 M Na2CO3, pH 11 (Ortegren et al. 2004), and sonicate with a probe-type sonifier (Soniprep 150, MSE, Crawley, UK) three times for 20 s. This was then adjusted to 45% (w/v) sucrose in 12 mM Mes, pH 6.5, 75 mM NaCl, 0.25 M Na2CO3, with protease inhibitors, and loaded under a 5–35% (w/v) discontinuous sucrose gradient in the same solution and centrifuged at 200,000 g for 16–20 h in a SW41 rotor (Beckman Instruments, Fullerton, CA, USA). The light-scattering band at the 5–35% sucrose interface was collected (Gustavsson et al. 1999) and referred to as total caveolae fraction. To subfractionate caveolae into subclasses, the sonicated plasma membrane fraction was adjusted to 40% (w/v) sucrose in 12 mM Mes, pH 6.5, 75 mM NaCl, 0.25 M Na2CO3 and made into a linear sucrose gradient with 15% (w/v) sucrose in the same buffer and centrifuged at 200,000 g for 16–20 h in the SW-41 rotor. Fractions of 1 mL were collected from the bottom of the tube. Fractions were diluted and pelleted for 30 min at 170,000 g for further analysis.
PAGE and immunoblotting
For immunoblotting in Fig. 2, an equal volume of each caveolae fraction was subjected to PAGE and separated proteins were transferred on to polyvinylidene difluoride blotting membrane (Millipore). For analysis of multimeric adiponectin, non-reduced samples were subjected to PAGE under non-reducing conditions (Xu et al. 2005). The membrane was blocked with milk proteins and incubated with rabbit anti-HSL antibodies (generous gift from Prof C Holm, University of Lund), rabbit anti-adiponectin antibodies or mouse anti-caveolin antibodies (Transduction Laboratories, Lexington, KY, USA), as indicated, followed by detection with ECL® according to the manufacturer's instructions (Amersham Biosciences) and evaluated by chemiluminescence imaging (Las 1000, Fuji, Tokyo, Japan). For immunoblotting of serum adiponectin, 0.5 µL of serum samples were subjected to SDS-PAGE under non-reducing conditions and transferred to nitrocellulose membrane at 90 V for 3 h at ice cold condition. The membrane was blocked with 2% BSA. The membrane was probed with α-rabbit polyclonal adiponectin antibody (1:1000 dilution, Abcam) and horseradish peroxidase-conjugated secondary antibody (1:5000, Abcam), and detected using Clarity Western ECL Substrate (Bio-Rad). The densitometric analysis was done using Image Lab 5.2.1 software (Bio-Rad).
Confocal microscopy and image analysis
3T3-L1 pre-adipocytes were plated onto borosilicate glass coverslips and differentiated into adipocytes as previously described (Komai et al. 2014). Nine days following induction of differentiation, 3T3-L1 adipocytes were washed twice with KREBS + 2.6 mM CaCl2 and then incubated with KREBS + 2.6 mM CaCl2 for 30 min at 32°C. 3T3-L1 adipocytes were then incubated with either 5 µM adrenalin or 200 nM insulin, diluted in KREBS + 2.6 mM CaCl2 + 5 mM glucose, or vehicle control, for 10 min at 32°C. The adipocytes were then washed with PBS and fixed with 4% paraformaldehyde in PBS for 10 min at room temperature, washed with PBS and then permeabilised with 0.25% Triton X-100. Cells were blocked with 1% BSA for 1 h at room temperature prior to incubation with goat polyclonal α-adiponectin antibody AF1119 (R&D systems) and rabbit anti caveolin-1 antibody (Cell Signaling) diluted in PBS + 1% BSA, overnight at 4°C. The cells were then washed with PBS and incubated with Rhodamine Red donkey anti-goat secondary antibody (Jackson Immunoresearch) and Alexa 647 anti-rabbit secondary antibody (Jackson Immunoresearch) in PBS + 1% BSA for 2 h at room temperature. Coverslips were mounted with Prolong Gold (Life Technologies) onto a glass slide and sealed with nail polish. Confocal microscopy was performed on a Zeiss LSM 800 microscope using a 63 × 1.4 NA oil immersion objective and images of cells close to the plane of the coverslip were acquired using Zen Black software (Zeiss). Images were processed and analysed using ImageJ. Mander’s co-localisation co-efficients were calculated for background corrected regions of interest encompassing an individual cell footprint using the Coloc2 plugin in ImageJ.
Gene expression
Gene expression was measured using the Fast SYBR Green system (Applied BioSystems) and a QuantStudio 7Flex light cycler (Applied BioSystems). RNA was isolated (ReliaPrep™ RNA tissue Miniprep System, Promega Inc.) from adipocytes or adipose tissue from male mice. The mice had an average age of 10 or 24 weeks (16 weeks of chow or HFD diet). Gene expression was normalized against β-actin (Actb) using the relative ΔCt method (for primer sequences, see Supplementary Table 1, see section on supplementary materials given at the end of this article). Primers were used at a concentration of 500 nmol/L, and PCR efficiencies were determined from the slope of the standard curve.
Histology of white adipose tissue
Samples were fixed in 4% formaldehyde for 48 h and embedded in paraffin. Sections, 7 µm thick, were stained with hematoxylin/eosin following manufacturer’s instructions. Images of slices were taken in a random order using a light microscope (Olympus BX60).
Data analysis
The statistical significance of variance between two means was calculated using Student’s t-test, (paired or unpaired) and ANOVA when appropriate. Data are presented as mean values ± s.e.m.
Results
Serum and tissue levels of adiponectin in Cav1 null mice
An association between Cav1 knockout and lowered plasma adiponectin levels has was previously shown (Asterholm et al. 2012). In the current study, we evaluated Cav1 null mice on a different genetic background and these mice displayed a slightly reduced body weight (Fig. 1A) and moderately elevated fasting glucose levels (Fig. 1B). This was not associated with significant alterations in serum insulin levels (Fig. 1C), signifying the presence of mild insulin resistance. In agreement with (Asterholm et al. 2012), we found that serum levels of HMW adiponectin were markedly reduced in Cav1 null mice compared to wildtype controls (Fig. 1D), whereas total adiponectin levels (representing all molecular forms of the hormone) remained unaltered (Fig. 1E). Interestingly, the adiponectin protein content was higher in the Cav1 null adipocytes compared to those from wildtype mice (Fig. 1F), whereas adiponectin gene expression was reduced in Cav1-depleted adipocytes (possibly as a result of accumulation of adiponectin protein; Fig. 1G). The fraction of HMW relative to total adiponectin remained unchanged in intact adipose tissue (Fig. 1H). Our results show that adiponectin protein expression is not abrogated in the Cav1 null adipocytes, therefore suggesting that a secretion defect underlies the lowered circulating HMW adiponectin observed in mice lacking Cav1.
Metabolic parameters and adiponectin in Cav1 null and wildtype control mice. Body weight (A), blood glucose (B) and serum insulin (C) in Cav1 null and wildtype control mice. HMW (D) and total (E) adiponectin measured in serum from Cav1 null (n = 5) and wildtype (n = 8) mice. (F) Total adiponectin analysed in isolated adipocytes (18–36 samples from three wildtype and three Cav1 null mice). (G) Adiponectin gene expression measured in adipocytes from wildtype or Cav1 null mice (ten animals in each group). (H) HMW to total adiponectin ratio, measured in IWAT tissue from five wildtype and five Cav1 null mice. *P < 0.05; **P < 0.01.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Metabolic parameters and adiponectin in Cav1 null and wildtype control mice. Body weight (A), blood glucose (B) and serum insulin (C) in Cav1 null and wildtype control mice. HMW (D) and total (E) adiponectin measured in serum from Cav1 null (n = 5) and wildtype (n = 8) mice. (F) Total adiponectin analysed in isolated adipocytes (18–36 samples from three wildtype and three Cav1 null mice). (G) Adiponectin gene expression measured in adipocytes from wildtype or Cav1 null mice (ten animals in each group). (H) HMW to total adiponectin ratio, measured in IWAT tissue from five wildtype and five Cav1 null mice. *P < 0.05; **P < 0.01.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Metabolic parameters and adiponectin in Cav1 null and wildtype control mice. Body weight (A), blood glucose (B) and serum insulin (C) in Cav1 null and wildtype control mice. HMW (D) and total (E) adiponectin measured in serum from Cav1 null (n = 5) and wildtype (n = 8) mice. (F) Total adiponectin analysed in isolated adipocytes (18–36 samples from three wildtype and three Cav1 null mice). (G) Adiponectin gene expression measured in adipocytes from wildtype or Cav1 null mice (ten animals in each group). (H) HMW to total adiponectin ratio, measured in IWAT tissue from five wildtype and five Cav1 null mice. *P < 0.05; **P < 0.01.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
The spatial relationship between adiponectin vesicles and caveolae/Cav1
We next investigated the localization of adiponectin in relation to caveolae. Immuno-gold labelling of human primary adipocyte plasma membrane sheets showed the typical caveolae structures (Fig. 2A), which have been extensively identified as caveolae in plasma membrane sheets of adipocytes, by us (Thorn et al. 2003, Karlsson et al. 2004) and by others (Foti et al. 2007). Adiponectin was localized to only a small fraction of the caveolae (Fig. 2A), suggesting that a subclass of caveolae (described in Ortegren et al. 2006) may be involved in adiponectin exocytosis. To study this, caveolae sub-types were separated by fractionation of human adipocytes. We have previously identified subclasses of caveolae in rat adipocytes using this technique (Ost et al. 2005, Ortegren et al. 2006). A similar set of caveolae subclasses was found in the human adipocytes (Fig. 2B). Interestingly, adiponectin was restricted to the high density (HD) caveolae as determined by its co-fractionation with hormone-sensitive lipase (HSL; Fig. 2B), part of which has been shown to be confined to HD caveolae (Aboulaich et al. 2006, Ortegren et al. 2006). This was true for total adiponectin protein (reduced monomeric adiponectin; top panel Fig. 2C) as well as for the different multimeric forms of adiponectin (lower panel Fig. 2C). The caveolae sub-fractions clearly overlap, but we have previously shown how specific proteins are separated in the different fractions using the same method (Ost et al. 2005, Aboulaich et al. 2006, Ortegren et al. 2006). The immuno-gold labeling of plasma membrane sheets and the fractionation studies jointly provide evidence for an association between adiponectin and caveolae.
Localisation of adiponectin in relation to caveolae/Cav1. (A) Plasma membrane sheets of primary human adipocytes immuno-labelled with antibodies against adiponectin and examined by transmission electron microscopy. Arrows designate immunogold-labelled adiponectin. (B) Localization of adiponectin to a triacylglycerol metabolizing subclass of caveolae. Caveolae were fractionated by density gradient ultracentrifugation. Equal volume fractions were collected from the bottom (fraction 1) of a linear 15–40% sucrose gradient. Membranes were pelleted, examined by non-denaturing PAGE, and immunoblotted against Cav1, reduced adiponectin and HSL, as indicated. The amount of each protein is presented as percent of maximal. (C) Blots displaying the molecular weights of the monomeric, dimeric, LMW, MMW and HMW adiponectin. Note that upper panel represents denatured SDS-PAGE and lower panel non-denatured (native) PAGE, thus molecular masses are approximate. The displayed results are representative of a total of seven experiments. (D) Confocal microscopy imaging of 3T3-L1 adipocytes immunostained for adiponectin and Cav1. Cells were incubated during 10 min with vehicle, 5 µM adrenalin or 200 nM insulin. Scale bar: 10 µm. The shown images are representative of three separate experiments studying a total of 15–25 cells under each condition.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Localisation of adiponectin in relation to caveolae/Cav1. (A) Plasma membrane sheets of primary human adipocytes immuno-labelled with antibodies against adiponectin and examined by transmission electron microscopy. Arrows designate immunogold-labelled adiponectin. (B) Localization of adiponectin to a triacylglycerol metabolizing subclass of caveolae. Caveolae were fractionated by density gradient ultracentrifugation. Equal volume fractions were collected from the bottom (fraction 1) of a linear 15–40% sucrose gradient. Membranes were pelleted, examined by non-denaturing PAGE, and immunoblotted against Cav1, reduced adiponectin and HSL, as indicated. The amount of each protein is presented as percent of maximal. (C) Blots displaying the molecular weights of the monomeric, dimeric, LMW, MMW and HMW adiponectin. Note that upper panel represents denatured SDS-PAGE and lower panel non-denatured (native) PAGE, thus molecular masses are approximate. The displayed results are representative of a total of seven experiments. (D) Confocal microscopy imaging of 3T3-L1 adipocytes immunostained for adiponectin and Cav1. Cells were incubated during 10 min with vehicle, 5 µM adrenalin or 200 nM insulin. Scale bar: 10 µm. The shown images are representative of three separate experiments studying a total of 15–25 cells under each condition.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Localisation of adiponectin in relation to caveolae/Cav1. (A) Plasma membrane sheets of primary human adipocytes immuno-labelled with antibodies against adiponectin and examined by transmission electron microscopy. Arrows designate immunogold-labelled adiponectin. (B) Localization of adiponectin to a triacylglycerol metabolizing subclass of caveolae. Caveolae were fractionated by density gradient ultracentrifugation. Equal volume fractions were collected from the bottom (fraction 1) of a linear 15–40% sucrose gradient. Membranes were pelleted, examined by non-denaturing PAGE, and immunoblotted against Cav1, reduced adiponectin and HSL, as indicated. The amount of each protein is presented as percent of maximal. (C) Blots displaying the molecular weights of the monomeric, dimeric, LMW, MMW and HMW adiponectin. Note that upper panel represents denatured SDS-PAGE and lower panel non-denatured (native) PAGE, thus molecular masses are approximate. The displayed results are representative of a total of seven experiments. (D) Confocal microscopy imaging of 3T3-L1 adipocytes immunostained for adiponectin and Cav1. Cells were incubated during 10 min with vehicle, 5 µM adrenalin or 200 nM insulin. Scale bar: 10 µm. The shown images are representative of three separate experiments studying a total of 15–25 cells under each condition.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
We next investigated the co-localisation of adiponectin vesicles with Cav1 in mature 3T3-L1 adipocytes (Fig. 2D). Confocal imaging of the plasma membrane of cells immuno-labelled for adiponectin and Cav1 showed the localisation of adiponectin in vesicular-like compartments, in agreement with what has been reported by others (Bogan & Lodish 1999, Lim et al. 2015, Rodiger et al. 2018). Previous studies using widefield fluorescence microscopy have shown that adiponectin vesicles do not co-localise with Cav1 in 3T3-L1 adipocytes (Khan et al. 2009). Using confocal scanning laser microscopy at the plane of the plasma membrane, we likewise observed little evidence of adiponectin vesicle overlap with Cav1 (Fig. 2D). Mander’s correlation co-efficient for adiponectin association with Cav1 (Supplementary Fig. 1A) or Cav1 association with adiponectin (Supplementary Fig. 1B) was similar for images in which the Cav1 channel was rotated relative to the adiponectin channel, compared to non-rotated images; thus, no specific association between adiponectin and Cav1 was detected using this methodology. We hypothesized that adiponectin vesicles might redistribute to interact with Cav1 following activation of signalling pathways promoting adiponectin vesicle exocytosis. However, we did not observe any change in co-localisation between adiponectin vesicles and Cav1 in 3T3-L1 adipocytes following a 10-min incubation with either adrenaline or insulin (Fig. 2D and Supplementary Fig. 1). Collectively, our results suggest that adiponectin vesicles are predominantly not co-localised with the Cav1 protein, but that some adiponectin is associated with a specific HD subclass of caveolae.
Effects of Cav1/caveolae depletion on adiponectin secretion
While the adipocyte function and diet-induced obesity of Cav1 null mice have previously been investigated (Razani et al. 2002, Asterholm et al. 2012), the role of Cav1 in the secretion of adiponectin from white adipocytes has never been studied. We isolated adipocytes from subcutaneous inguinal white adipose tissue (IWAT) of Cav1 null mice and littermate wildtype controls. Adiponectin release was induced by 30-min incubations with the cAMP elevating combination of forskolin and IBMX (FSK/IBMX; Komai et al. 2014), the β3AR agonist CL316.243 (CL; Komai et al. 2016), or with insulin (Lim et al. 2015). Our previous work defining the regulation of adiponectin vesicle exocytosis largely focused on IWAT as well as on human subcutaneous adipocytes (Komai et al. 2014, El Hachmane et al. 2015, Komai et al. 2016) and subcutaneous adipose tissue has been suggested to be most important for the control of circulating adiponectin levels (Lihn et al. 2004, Meyer et al. 2013). Although IWAT is a convertible adipose tissue that can undergo browning under certain conditions, such as prolonged adrenergic stimuli (Kaisanlahti & Glumoff 2019), we regard it unlikely that the short exposure to CL used here significantly affects browning. Consistent with published data (Komai et al. 2014, 2016, Lim et al. 2015), adiponectin secretion was induced in wildtype control adipocytes by all secretagogues; the stimulated release was, however, abolished in adipocytes from Cav1 null mice (Fig. 3A). To further investigate the role of caveolae/Cav1 in stimulated adiponectin secretion, IWAT mouse adipocytes were chemically depleted of caveolae structures using methyl-β-cyclodextrin (MβCD) to extract cholesterol from the cell surface (Karlsson et al. 2004, Fagerholm et al. 2009). We have previously shown that such treatment of rat adipocytes disrupts the caveolae structure, while Cav1 remains clustered in the plasma membrane (Parpal et al. 2001, Thorn et al. 2003). All secretagogues induced adiponectin secretion in non-treated cells, but the stimulated release was abolished in MβCD-treated adipocytes (Fig. 3A). As shown in Fig. 3B, basal (unstimulated) adiponectin secretion was elevated both in Cav1 null and MβCD-treated adipocytes. The magnitude of lipolysis stimulated by FSK/IBMX was similar in Cav1 null and wildtype adipocytes and glycerol release was also induced in response to CL stimulation in Cav1 null adipocytes (although less potent compared to in wildtype adipocytes; Fig. 3C). Basal lipolysis was reduced in Cav1 null compared to wildtype adipocytes (Fig. 3D), in agreement with the described blunted β3AR agonist-induced release of free fatty acids in vivo (Asterholm et al. 2012). Although the Cav1 null adipocytes release less glycerol compared to wildtype adipocytes (Fig. 3D), their ability to respond to cAMP/adrenergic stimulation with increased glycerol release verifies that the isolated cells are functional with regard to other metabolic processes. Measurements of cAMP showed elevated basal (non-stimulated) cAMP levels in Cav1 null compared to wildtype adipose tissue (Fig. 3E).
Release of adiponectin and glycerol as well as cAMP levels in wildtype and Cav1-depleted adipocytes. (A) Primary wildtype (black) or Cav1 null (red) mouse adipocytes were incubated during 30 min (32°C) with or without 200 nM insulin (I), a combination of 10 µM forskolin and 200 µM IBMX (F/B) or 1 μM CL 316,243 (CL). Separate wildtype adipocytes were pre-incubated with 8 mM MβCD for 50 min before stimulations (blue). Secreted adiponectin was measured in wildtype (16–24 samples from seven animals), MβCD-treated (14–21 samples from six animals) and Cav1 null (26–33 samples from eight animals) adipocytes. Results are normalised to total protein content and expressed as fold over average control for each treatment. (B) Basal (unstimulated) adiponectin release in wildtype (24 samples from seven animals), MβCD (21 samples from six animals) or Cav1 null (33 samples from eight animals) adipocytes. (C) Glycerol release from wildtype (22–24 samples from seven animals) or Cav1 null (21–32 samples from ten animals) adipocytes expressed as fold increase over average control. (D) Basal glycerol release from wildtype (24 samples from seven animals) or Cav1 null (32 samples from ten animals) adipocytes. (E) Absolute cAMP levels in adipose tissue lysates from wildtype or Cav1 null mice (four animals in each group). (F) Adiponectin secretion in primary human subcutaneous adipocytes incubated during 30 min (32°C) with or without the same secretagogues as in A. Adipocytes were pre-incubated with (14–22 samples from eight subjects; blue) or without (16–26 samples from eight subjects; black) 8 mM MβCD for 50 min. Results are expressed as fold over average control for each experiment. (G) Basal (unstimulated) adiponectin release during 30 min in primary human adipocytes that were pre-incubated with (22 samples from eight subjects; blue) or without (26 samples from eight subjects; black) MβCD. Secretion is normalized to total protein. *P < 0.05; **P < 0.01; ***P < 0.001.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Release of adiponectin and glycerol as well as cAMP levels in wildtype and Cav1-depleted adipocytes. (A) Primary wildtype (black) or Cav1 null (red) mouse adipocytes were incubated during 30 min (32°C) with or without 200 nM insulin (I), a combination of 10 µM forskolin and 200 µM IBMX (F/B) or 1 μM CL 316,243 (CL). Separate wildtype adipocytes were pre-incubated with 8 mM MβCD for 50 min before stimulations (blue). Secreted adiponectin was measured in wildtype (16–24 samples from seven animals), MβCD-treated (14–21 samples from six animals) and Cav1 null (26–33 samples from eight animals) adipocytes. Results are normalised to total protein content and expressed as fold over average control for each treatment. (B) Basal (unstimulated) adiponectin release in wildtype (24 samples from seven animals), MβCD (21 samples from six animals) or Cav1 null (33 samples from eight animals) adipocytes. (C) Glycerol release from wildtype (22–24 samples from seven animals) or Cav1 null (21–32 samples from ten animals) adipocytes expressed as fold increase over average control. (D) Basal glycerol release from wildtype (24 samples from seven animals) or Cav1 null (32 samples from ten animals) adipocytes. (E) Absolute cAMP levels in adipose tissue lysates from wildtype or Cav1 null mice (four animals in each group). (F) Adiponectin secretion in primary human subcutaneous adipocytes incubated during 30 min (32°C) with or without the same secretagogues as in A. Adipocytes were pre-incubated with (14–22 samples from eight subjects; blue) or without (16–26 samples from eight subjects; black) 8 mM MβCD for 50 min. Results are expressed as fold over average control for each experiment. (G) Basal (unstimulated) adiponectin release during 30 min in primary human adipocytes that were pre-incubated with (22 samples from eight subjects; blue) or without (26 samples from eight subjects; black) MβCD. Secretion is normalized to total protein. *P < 0.05; **P < 0.01; ***P < 0.001.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Release of adiponectin and glycerol as well as cAMP levels in wildtype and Cav1-depleted adipocytes. (A) Primary wildtype (black) or Cav1 null (red) mouse adipocytes were incubated during 30 min (32°C) with or without 200 nM insulin (I), a combination of 10 µM forskolin and 200 µM IBMX (F/B) or 1 μM CL 316,243 (CL). Separate wildtype adipocytes were pre-incubated with 8 mM MβCD for 50 min before stimulations (blue). Secreted adiponectin was measured in wildtype (16–24 samples from seven animals), MβCD-treated (14–21 samples from six animals) and Cav1 null (26–33 samples from eight animals) adipocytes. Results are normalised to total protein content and expressed as fold over average control for each treatment. (B) Basal (unstimulated) adiponectin release in wildtype (24 samples from seven animals), MβCD (21 samples from six animals) or Cav1 null (33 samples from eight animals) adipocytes. (C) Glycerol release from wildtype (22–24 samples from seven animals) or Cav1 null (21–32 samples from ten animals) adipocytes expressed as fold increase over average control. (D) Basal glycerol release from wildtype (24 samples from seven animals) or Cav1 null (32 samples from ten animals) adipocytes. (E) Absolute cAMP levels in adipose tissue lysates from wildtype or Cav1 null mice (four animals in each group). (F) Adiponectin secretion in primary human subcutaneous adipocytes incubated during 30 min (32°C) with or without the same secretagogues as in A. Adipocytes were pre-incubated with (14–22 samples from eight subjects; blue) or without (16–26 samples from eight subjects; black) 8 mM MβCD for 50 min. Results are expressed as fold over average control for each experiment. (G) Basal (unstimulated) adiponectin release during 30 min in primary human adipocytes that were pre-incubated with (22 samples from eight subjects; blue) or without (26 samples from eight subjects; black) MβCD. Secretion is normalized to total protein. *P < 0.05; **P < 0.01; ***P < 0.001.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
MβCD removes cholesterol from the plasma membrane and may therefore have negative effects on cell viability. However, the same protocol used here was found to perturb only metabolic and not mitogenic insulin signalling in rat adipocytes (Parpal et al. 2001). In addition, Karlsson et al. found isoproterenol-stimulated phosphorylation of perilipin to be unaffected using this same protocol, while the ability of insulin to counter this effect was abolished (Karlsson et al. 2004). To verify the viability of the MβCD-treated adipocytes, we measured secretion of the mouse adipocyte hormone resistin. Our own studies using 3T3-L1 adipocytes show that both insulin and FSK/IBMX stimulate resistin release (S Musovic & C S Olofsson, unpublished observations). FSK/IBMX stimulated the secretion of resistin in MβCD-treated mouse adipocytes, verifying the secretory capacity of the cholesterol/caveolae-depleted adipocytes. In accordance with the disrupted insulin signalling demonstrated in rat (Parpal et al. 2001) and human (Karlsson et al. 2004) adipocytes, insulin-induced resistin release was abolished in adipocytes treated with MβCD (Supplementary Fig. 2A) and the basal resistin release was elevated (Supplementary Fig. 2B).
To confirm the translational relevance of our findings in mice, we examined adiponectin secretion in primary human subcutaneous adipocytes. FSK/IBMX, CL or insulin all stimulated adiponectin release in human adipocytes and these secretory responses were abrogated by MβCD treatment (Fig. 3F). Basal adiponectin secretion was higher in the MβCD treated human adipocytes (Fig. 3G), similar to what we observed in mouse adipocytes.
Collectively, our results indicate that Cav1 and/or an intact caveolae structure are essential for stimulation of adiponectin release, both in mouse and human adipocytes. Moreover, Cav1 and/or intact caveolae might serve to suppress basal adiponectin release. This altered adiponectin secretion is unlikely to be due to a general functional defect in Cav1/caveolae-depleted cells, since these cells are otherwise metabolically functional.
The expression of key molecular components involved in adrenergically stimulated adiponectin secretion is not altered in Cav1 null adipocytes
To investigate how Cav1 knockdown affects the adiponectin exocytosis machinery, we measured the expression of β3AR and Epac1, two molecular components essential for mediating catecholamine-triggered adiponectin exocytosis (Komai et al. 2016). As shown in Fig. 4A and B, neither β3AR nor Epac1 mRNA levels were significantly different in Cav1 null compared to wildtype adipocytes. In agreement with previous data (Komai et al. 2016), βARs 1 and 2 were weakly expressed in the adipocytes and their mRNA levels were not significantly affected by Cav1 ablation (Fig. 4A). This suggests that the blunted secretion in Cav1 null adipocytes in response to adrenergic stimulation does not result from altered levels of β3AR or Epac1.
Gene expression of beta-adrenergic receptors and Epac1 in wildtype and Cav1 null adipocytes. (A and B) Gene expression of βARs 1 (Adrb1), 2 (Adrb2) and 3 (Adrb3) as well as Epac1 (Rapgef3) in primary adipocytes from Cav1 null (red, n = 10) or wildtype (black, n = 10) littermates.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Gene expression of beta-adrenergic receptors and Epac1 in wildtype and Cav1 null adipocytes. (A and B) Gene expression of βARs 1 (Adrb1), 2 (Adrb2) and 3 (Adrb3) as well as Epac1 (Rapgef3) in primary adipocytes from Cav1 null (red, n = 10) or wildtype (black, n = 10) littermates.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Gene expression of beta-adrenergic receptors and Epac1 in wildtype and Cav1 null adipocytes. (A and B) Gene expression of βARs 1 (Adrb1), 2 (Adrb2) and 3 (Adrb3) as well as Epac1 (Rapgef3) in primary adipocytes from Cav1 null (red, n = 10) or wildtype (black, n = 10) littermates.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Cav1 expression in adipose tissue from mice with diet-induced obesity
The maintained serum levels of total adiponectin together with a clear drop in levels of the HMW form in Cav1-ablated animals are very similar to our previous findings in mice with diet-induced obesity/diabetes. Moreover, isolated adipocytes from the obese and diabetic mice also display, like Cav1 null adipocytes, blunted stimulated adiponectin secretion concomitant with elevated basal release (Komai et al. 2016). We, therefore, decided to investigate if altered Cav1 expression is involved in obesity-associated alterations of adiponectin levels and/or secretion. We analysed the expression of Cav1 in IWAT isolated from chow-fed control or HFD-fed mice after 16 weeks. The mice were obese (average body weight 33 ± 0.8 and 51 ± 0.6 g for chow- and HFD-fed mice, respectively) and insulin resistant, as evidenced by elevated plasma glucose (9.1 ± 0.5 for chow-fed and 13.5 ± 0.5 mM for HFD-fed animals; P < 0.01) and serum insulin (4.5 ± 0.6 vs 9.7 ± 0.9 mM for chow- and HFD-fed mice, respectively; P < 0.01) levels. Adipocytes from HFD-fed mice were visually larger than those from chow-fed animals (Fig. 5A). As shown in Fig. 5B, and quantified in Fig. 5C, the HMW/total adiponectin ratio was lower in serum from obese/diabetic mice, consistent with what we have previously found in mice fed HFD through 8 weeks (Komai et al. 2016). Levels of MMW and LMW adiponectin remained unaltered between chow- and HFD-fed mice (Supplementary Fig. 3A and B). The reduction of HMW adiponectin demonstrated by protein immunoblotting was confirmed by ELISA (Supplementary Fig. 3C). We found a small (~25%) reduction in Cav1 gene expression (Fig. 5D), but Cav1 levels were not significantly altered at the protein level (Fig. 5E; representative blot shown in Supplementary Fig. 3D).
Adiponectin levels and Cav1 expression in adipocytes from lean and obese wildtype mice. (A) Characteristic images of freeze sections of IWAT. Scale bar: 100 µM. Representative immunoblot of adiponectin (B) and densitometric quantification of HMW adiponectin (C) in serum from chow- or HFD-fed mice (n = 4). Cav1 mRNA (D) and protein (E) levels in adipose tissue from chow- (black) or HFD-fed (green) mice (results from eight lean and eight obese animals) *P < 0.05.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Adiponectin levels and Cav1 expression in adipocytes from lean and obese wildtype mice. (A) Characteristic images of freeze sections of IWAT. Scale bar: 100 µM. Representative immunoblot of adiponectin (B) and densitometric quantification of HMW adiponectin (C) in serum from chow- or HFD-fed mice (n = 4). Cav1 mRNA (D) and protein (E) levels in adipose tissue from chow- (black) or HFD-fed (green) mice (results from eight lean and eight obese animals) *P < 0.05.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Adiponectin levels and Cav1 expression in adipocytes from lean and obese wildtype mice. (A) Characteristic images of freeze sections of IWAT. Scale bar: 100 µM. Representative immunoblot of adiponectin (B) and densitometric quantification of HMW adiponectin (C) in serum from chow- or HFD-fed mice (n = 4). Cav1 mRNA (D) and protein (E) levels in adipose tissue from chow- (black) or HFD-fed (green) mice (results from eight lean and eight obese animals) *P < 0.05.
Citation: Journal of Endocrinology 247, 1; 10.1530/JOE-20-0078
Discussion
Although adiponectin was identified as an adipocyte hormone two decades ago (Scherer et al. 1995), the regulation of its release is only partly understood. We have defined a cAMP/β3AR-dependent pathway for stimulation of adiponectin exocytosis (Komai et al. 2014, 2016, El Hachmane et al. 2015, Brannmark et al. 2017) and secretion of adiponectin is also induced by insulin (Pereira & Draznin 2005, Cong et al. 2007, Blumer et al. 2008, Lim et al. 2015). The knowledge that adipocyte caveolae constitute micro-domains for the organisation of cAMP and Ca2+ signalling (Willoughby & Cooper 2007) and that they are also involved in signalling via the insulin receptor (Fagerholm et al. 2009) suggests a role for Cav1 and/or caveolae in the regulation of adiponectin release.
Cav1/caveolae are essential for the secretion of HMW adiponectin
A number of studies have shown the occurrence of adiponectin-containing vesicles in adipocytes (Bogan & Lodish 1999, Komai et al. 2014, Lim et al. 2015, Rodiger et al. 2018). Our own immunofluorescence microscopy analyses in Fig. 2D confirm a vesicular-like pattern for adiponectin, located in close proximity to the cell membrane. In agreement with what has been shown before (Khan et al. 2009), we found little evidence for co-localisation of adiponectin with Cav1. Nevertheless, secretion data in Cav1/caveolae-depleted adipocytes (Fig. 3) indicate a role for Cav1 and caveolae in both basal and stimulated adiponectin release. Interestingly, our own data (Fig. 1D and E) and a previous study by Asterholm et al. (2012) demonstrate that serum levels of HMW adiponectin are reduced in the Cav1 null mice, while circulating levels of smaller adiponectin forms remain unaffected. This suggests that specifically secretion of the HMW form of adiponectin is disrupted in Cav1 null mice. HMW adiponectin constitutes only a fraction of total serum adiponectin (Hamilton et al. 2011, Asterholm et al. 2012, Komai et al. 2016) and our own unpublished data suggest that HMW adiponectin is secreted from a population of vesicles that differ from those containing only smaller forms of the hormone (S Musovic & C S Olofsson, unpublished observations). As a consequence, only a minor fraction of adiponectin-containing vesicles can be expected to contain multimeric adiponectin and to interact with Cav1, something that would be difficult to observe with the technique used here. Importantly, this does not exclude the possibility that smaller forms of adiponectin are also enclosed within the HMW-enriched vesicles, as indicated by the fractionation experiments in Fig. 2B and C (and from our own adiponectin secretion data; S Musovic & C S Olofsson, unpublished observations).
An alternative interpretation of our data is that an interaction between adiponectin vesicles and Cav1 is not required in order to regulate adiponectin secretion. Previous studies in adipocytes demonstrate that Cav1 and 2 are not situated in the caveolae bulb but rather located to the neck of the caveolae structure (Thorn et al. 2003). Moreover, G protein-coupled receptors (GPCRs) are known to localize to lipid rafts and caveolae (Barnett-Norris et al. 2005), and beta adrenergic receptors (Schwencke et al. 1999), including β3ARs (Sato et al. 2012), have been shown to physically interact with Cav1. Thus, the adiponectin vesicles themselves might not need to interact with Cav1 upon exocytosis, even if their stimulated release necessitates Cav1.
Collectively, our data indicate the existence of a functional association between HMW adiponectin and Cav1/caveolae. We acknowledge that Cav1 has been demonstrated to regulate cellular processes beyond the caveolae structure (Head & Insel 2007, Lajoie et al. 2009) and that the genetic deletion of Cav1 does not provide evidence that intact caveolae are required to maintain control of adiponectin secretion. However, several factors indicate that the loss of Cav1 and thus the disruption of caveolae-dependent scaffolds for signalling proteins is connected to the disturbed adiponectin secretion in Cav1-depleted adipocytes. First, Cav1 remains clustered in the plasma membrane in adipocytes where the caveolae structure has been disrupted by MβCD (Parpal et al. 2001, Thorn et al. 2003). Secondly, the insulin receptor itself is unaffected in MβCD-treated adipocytes but its downstream signalling is inhibited (Parpal et al. 2001, Karlsson et al. 2004). Thirdly, MβCD treatment has been shown to abolish βAR agonist-induced activation of protein kinase A (PKA; Kim et al. 2014). Thus, our results are compatible with a requirement for intact caveolae structures for the stimulated secretion of HMW adiponectin, although we acknowledge that our data do not provide conclusive evidence for this. As subsequently discussed, the loss of stimulated adiponectin secretion observed in caveolae/Cav1-depleted adipocytes may be due to the disruption of signalling pathways regulating adiponectin vesicle exocytosis.
The importance of Cav1/caveolae for signalling regulating adiponectin secretion
The spatial organization of GPCRs within the caveolae of the plasma membrane (Schwencke et al. 1999, Sato et al. 2012) facilitates signalling selectivity and allows cells to modify their responses by the restricted organization of downstream target mediators, such as adenylyl cyclases (ACs), phosphodiesterases (PDEs), PKA and Epac (Ostrom & Insel 2004, Insel et al. 2005). Thus, depletion of Cav1/caveolae may disrupt the β3AR signalling complex, resulting in blunted adrenergically stimulated adiponectin release. The observation that adiponectin secretion in response to FSK/IBMX is likewise abrogated indicates that cAMP signalling is not only disrupted at the receptor level, but that the downstream pathway is also disturbed.
It could be argued that the blunted triggered adiponectin secretion is merely a consequence of elevated basal cAMP levels that enhance release of adiponectin under basal conditions. Therefore, it could be envisaged that the over-secreting, Cav1/caveolae-depleted adipocytes have reached their maximal secretion capacity. However, for several reasons it is unlikely that this alone explains the disturbed triggered adiponectin secretion. First, if the primary function of Cav1 is to suppress basal adiponectin release, then the same population of vesicles secreted under stimulated conditions in wildtype/control adipocytes should be released in an unregulated fashion from Cav1/caveolae-depleted fat cells lacking the ‘secretion break’. Consequently, the ratio between different adiponectin forms in serum would be expected to remain largely unaltered in Cav1 null mice. This is not the case, as levels of HMW adiponectin are specifically reduced in serum from Cav1 null mice. Moreover, the obese/diabetic mice in Komai et al. (2016) similarly display much reduced circulating HMW adiponectin, whereas the total serum adiponectin concentration remains unchanged. Basal adiponectin release is increased both in Cav1 null adipocytes and in adipocytes isolated from obese/diabetic mice (Komai et al. 2016). Interestingly, our own data demonstrate that obesity-associated basal over-secretion represents release of MMW/LMW adiponectin only (S Musovic & C S Olofsson, unpublished observations). It is thus likely that the enhanced basal adiponectin release in Cav1 null mice can also be explained by increased release of smaller adiponectin forms. This is supported by our observation that only serum levels of HMW adiponectin are reduced in Cav1 null mice. In this context, it is of interest that the formation and secretion of higher-order complexes of adiponectin have been shown to depend on prolonged retention of adiponectin in the endoplasmic reticulum (ER; Wang et al. 2007). It could thus be speculated that loss of Cav1 shortens the time that adiponectin resides in ER and that this favors synthesis and release of smaller adiponectin forms as well as reduced synthesis and secretion of HMW adiponectin. However, the observation that neither total nor HMW adiponectin content is altered in Cav1-depleted adipocytes (Fig. 1F and H) indicates that the secretory disturbance present in Cav1 null adipocytes is likely situated further down-stream, such as within the exocytosis process itself. It is not surprising that Cav1-depletion differentially affects secretion of different molecular weight forms of adiponectin, as constitutive and regulated vesicle exocytosis typically involve diverse signalling pathways (Morvan & Tooze 2008, Park & Loh 2008). Thus, while one function of Cav1 might be to restrict basal adiponectin release, our data strongly suggest that Cav1-depletion, in addition, disrupts regulated exocytosis of HMW adiponectin.
The finding that adiponectin is localised to the HD caveolae fraction, similar to HSL (Ortegren et al. 2006), might appear difficult to reconcile with the preserved lipolysis in Cav1 null adipocytes (Fig. 3C). However, HSL is mainly located within the central lipid droplet, where lipolysis also occurs. Moreover, lipolysis is stimulated via PKA activation (Galitzky et al. 1997, Omar et al. 2009), while adiponectin secretion depends on activation of Epac1 (Komai et al. 2016) which binds to cAMP with lower affinity. Consequently, less cAMP is required to stimulate PKA-induced lipolysis compared to that required for Epac-dependent adiponectin exocytosis. The preserved CL-triggered lipolytic response that we observe agrees with previous work demonstrating that the β3AR agonist stimulated glycerol release in isolated Cav1 null adipocytes (Cohen et al. 2004).
Insulin receptor signalling also depends on the caveolae microenvironment (Fagerholm et al. 2009, Brannmark et al. 2010) and blunted insulin-induced adiponectin release is therefore expected in caveolae-depleted adipocytes. Our own recent work identified inhibition of lipolysis occurring via phosphoinositide 3-kinase alpha and activation of protein kinase B (PKB/Akt) in response to either insulin and/or cAMP/beta-adrenergic stimuli (Jonsson et al. 2019). This suggests possible parallel mechanisms for stimulation of adiponectin secretion by insulin and/or beta-adrenergic agonists that would be interesting to explore in future studies.
Pathological consequences of altered adiponectin secretion in Cav1/caveolae-depleted adipocytes
There are some striking similarities between our findings in Cav1 null mice and in mice with diet-induced obesity (Komai et al. 2016), both with regards to adiponectin release and circulating adiponectin levels. Moreover, studies from the Scherer lab showed that in vivo treatment of mice with simvastatin, which disturbs the caveolae structure by reducing cholesterol content, likewise led to a 40% reduction in serum HMW adiponectin, whereas total circulating adiponectin levels remained unaltered. Collectively, their findings emphasize that an unregulated over-secretion of smaller adiponectin forms is present in simvastatin treated mice, while HMW adiponectin release and/or levels are reduced. This proposes disturbed Cav1/caveolae function as an underlying cause of the lowered serum HMW adiponectin (Khan et al. 2009). Khan et al. also showed that Cav1 gene expression was upregulated in simvastatin-treated mice (Khan et al. 2009).
In the current work, we were unable to show an obesity/diabetes-associated effect on Cav1 abundance. Regardless of Cav1 expression levels, it can be speculated that the lower density of Cav1 protein per membrane area in ‘obese’ adipocytes interferes with Cav1/caveolae-associated signalling and with the regulation of adiponectin secretion. The obese adipocytes described in Fig. 5A have a calculated three-fold larger membrane area compared to adipocytes from lean animals (average area 5876 ± 500 µm2 for chow vs 15,522 ± 1082 µm2 for HFD; P < 0.001; calculated from 9967 chow and 3509 HFD cells), which yields a 60% lower density of Cav1 protein per membrane area. Human abdominal subcutaneous adipocytes have been found to be almost doubled in volume and plasma membrane area, in obese (BMI >30) as compared to those from lean (BMI <25) subjects (Tchoukalova et al. 2008). A number of studies indicate that adipocyte hypertrophy is associated with metabolic dysfunction (reviewed in Tandon et al. 2018). It should be emphasized that the similarities between Cav1 null and obese/diabetic mice with regard to adiponectin secretion and levels might exist because the same signalling pathway is disrupted in both situations, although the underlying mechanistic defects differ (Komai et al. 2016). Nonetheless, it would be interesting to study the relationship between Cav1/caveolae plasma membrane density and adiponectin secretion in future work.
In summary, the observations presented here propose that the stimulated secretion of HMW adiponectin depends on Cav1 and/or intact caveolae structures. We postulate that disturbed Cav1/caveolae function may contribute to the lowered serum HMW adiponectin associated with diabetes. It is clear that, more than 20 years after its discovery, adiponectin remains an enigmatic hormone and there is still much to be discovered regarding the molecular regulation of its secretion.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/JOE-20-0078.
Declaration of interest
E I K is currently employed by AstraZeneca. The other authors have nothing to disclose.
Funding
This work was supported by the Swedish Diabetes Foundation (DIA2015-062, DIA2017-273 and DIA2018-358), the Novo Nordisk Foundation, the Knut and Alice Wallenberg Foundation (P. Rorsman), and the Swedish Research Council (Grant IDs: 2013-7107, 2017-00792 and 2019-01239). E I K was partly supported by the AstraZeneca postdoc program.
References
Aboulaich N, Ortegren U, Vener AV & Stralfors P 2006 Association and insulin regulated translocation of hormone-sensitive lipase with PTRF. Biochemical and Biophysical Research Communications 350 657–661. (https://doi.org/10.1016/j.bbrc.2006.09.094)
Arita Y, Kihara S, Ouchi N, Takahashi M, Maeda K, Miyagawa J, Hotta K, Shimomura I, Nakamura T, Miyaoka K, et al.1999 Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity. Biochemical and Biophysical Research Communications 257 79–83. (https://doi.org/10.1006/bbrc.1999.0255)
Asterholm IW, Mundy DI, Weng J, Anderson RG & Scherer PE 2012 Altered mitochondrial function and metabolic inflexibility associated with loss of caveolin-1. Cell Metabolism 15 171–185. (https://doi.org/10.1016/j.cmet.2012.01.004)
Barnett-Norris J, Lynch D & Reggio PH 2005 Lipids, lipid rafts and caveolae: their importance for GPCR signaling and their centrality to the endocannabinoid system. Life Sciences 77 1625–1639. (https://doi.org/10.1016/j.lfs.2005.05.040)
Basu R, Pajvani UB, Rizza RA & Scherer PE 2007 Selective downregulation of the high molecular weight form of adiponectin in hyperinsulinemia and in type 2 diabetes: differential regulation from nondiabetic subjects. Diabetes 56 2174–2177. (https://doi.org/10.2337/db07-0185)
Blumer RM, van Roomen CP, Meijer AJ, Houben-Weerts JH, Sauerwein HP & Dubbelhuis PF 2008 Regulation of adiponectin secretion by insulin and amino acids in 3T3-L1 adipocytes. Metabolism: Clinical and Experimental 57 1655–1662. (https://doi.org/10.1016/j.metabol.2008.07.020)
Bogan JS & Lodish HF 1999 Two compartments for insulin-stimulated exocytosis in 3T3-L1 adipocytes defined by endogenous ACRP30 and GLUT4. Journal of Cell Biology 146 609–620. (https://doi.org/10.1083/jcb.146.3.609)
Brannmark C, Palmer R, Glad ST, Cedersund G & Stralfors P 2010 Mass and information feedbacks through receptor endocytosis govern insulin signaling as revealed using a parameter-free modeling framework. Journal of Biological Chemistry 285 20171–20179. (https://doi.org/10.1074/jbc.M110.106849)
Brannmark C, Lovfors W, Komai AM, Axelsson T, El Hachmane MF, Musovic S, Paul A, Nyman E & Olofsson CS 2017 Mathematical modeling of white adipocyte exocytosis predicts adiponectin secretion and quantifies the rate of vesicle exo- and endocytosis. Journal of Biological Chemistry 292 20032–20043. (https://doi.org/10.1074/jbc.M117.801225)
Brochu-Gaudreau K, Rehfeldt C, Blouin R, Bordignon V, Murphy BD & Palin MF 2010 Adiponectin action from head to toe. Endocrine 37 11–32. (https://doi.org/10.1007/s12020-009-9278-8)
Burgoyne RD & Morgan A 2003 Secretory granule exocytosis. Physiological Reviews 83 581–632. (https://doi.org/10.1152/physrev.00031.2002)
Cohen AW, Razani B, Schubert W, Williams TM, Wang XB, Iyengar P, Brasaemle DL, Scherer PE & Lisanti MP 2004 Role of caveolin-1 in the modulation of lipolysis and lipid droplet formation. Diabetes 53 1261–1270. (https://doi.org/10.2337/diabetes.53.5.1261)
Cong L, Chen K, Li J, Gao P, Li Q, Mi S, Wu X & Zhao AZ 2007 Regulation of adiponectin and leptin secretion and expression by insulin through a PI3K-PDE3B dependent mechanism in rat primary adipocytes. Biochemical Journal 403 519–525. (https://doi.org/10.1042/BJ20061478)
El Hachmane MF, Komai AM & Olofsson CS 2015 Cooling reduces cAMP-stimulated exocytosis and adiponectin secretion at a Ca2+-dependent step in 3T3-L1 adipocytes. PLoS ONE 10 e0119530. (https://doi.org/10.1371/journal.pone.0119530)
Fagerholm S, Ortegren U, Karlsson M, Ruishalme I & Stralfors P 2009 Rapid insulin-dependent endocytosis of the insulin receptor by caveolae in primary adipocytes. PLoS ONE 4 e5985. (https://doi.org/10.1371/journal.pone.0005985)
Foti M, Porcheron G, Fournier M, Maeder C & Carpentier JL 2007 The neck of caveolae is a distinct plasma membrane subdomain that concentrates insulin receptors in 3T3-L1 adipocytes. PNAS 104 1242–1247. (https://doi.org/10.1073/pnas.0610523104)
Galitzky J, Langin D, Verwaerde P, Montastruc JL, Lafontan M & Berlan M 1997 Lipolytic effects of conventional beta 3-adrenoceptor agonists and of CGP 12,177 in rat and human fat cells: preliminary pharmacological evidence for a putative beta 4-adrenoceptor. British Journal of Pharmacology 122 1244–1250. (https://doi.org/10.1038/sj.bjp.0701523)
Gustavsson J, Parpal S, Karlsson M, Ramsing C, Thorn H, Borg M, Lindroth M, Peterson KH, Magnusson KE & Stralfors P 1999 Localization of the insulin receptor in caveolae of adipocyte plasma membrane. FASEB Journal 13 1961–1971. (https://doi.org/10.1096/fasebj.13.14.1961)
Hamilton MP, Gore MO, Ayers CR, Xinyu W, McGuire DK & Scherer PE 2011 Adiponectin and cardiovascular risk profile in patients with type 2 diabetes mellitus: parameters associated with adiponectin complex distribution. Diabetes and Vascular Disease Research 8 190–194. (https://doi.org/10.1177/1479164111407784)
Hara K, Horikoshi M, Yamauchi T, Yago H, Miyazaki O, Ebinuma H, Imai Y, Nagai R & Kadowaki T 2006 Measurement of the high-molecular weight form of adiponectin in plasma is useful for the prediction of insulin resistance and metabolic syndrome. Diabetes Care 29 1357–1362. (https://doi.org/10.2337/dc05-1801)
Head BP & Insel PA 2007 Do caveolins regulate cells by actions outside of caveolae? Trends in Cell Biology 17 51–57. (https://doi.org/10.1016/j.tcb.2006.11.008)
Insel PA, Head BP, Patel HH, Roth DM, Bundey RA & Swaney JS 2005 Compartmentation of G-protein-coupled receptors and their signalling components in lipid rafts and caveolae. Biochemical Society Transactions 33 1131–1134. (https://doi.org/10.1042/BST20051131)
Jonsson C, Castor Batista AP, Kjolhede P & Stralfors P 2019 Insulin and beta-adrenergic receptors mediate lipolytic and anti-lipolytic signalling that is not altered by type 2 diabetes in human adipocytes. Biochemical Journal 476 2883–2908. (https://doi.org/10.1042/BCJ20190594)
Kadowaki T, Yamauchi T, Kubota N, Hara K, Ueki K & Tobe K 2006 Adiponectin and adiponectin receptors in insulin resistance, diabetes, and the metabolic syndrome. Journal of Clinical Investigation 116 1784–1792. (https://doi.org/10.1172/JCI29126)
Kaisanlahti A & Glumoff T 2019 Browning of white fat: agents and implications for beige adipose tissue to type 2 diabetes. Journal of Physiology and Biochemistry 75 1–10. (https://doi.org/10.1007/s13105-018-0658-5)
Karlsson M, Thorn H, Danielsson A, Stenkula KG, Ost A, Gustavsson J, Nystrom FH & Stralfors P 2004 Colocalization of insulin receptor and insulin receptor substrate-1 to caveolae in primary human adipocytes. Cholesterol depletion blocks insulin signalling for metabolic and mitogenic control. European Journal of Biochemistry 271 2471–2479. (https://doi.org/10.1111/j.1432-1033.2004.04177.x)
Khan T, Hamilton MP, Mundy DI, Chua SC & Scherer PE 2009 Impact of simvastatin on adipose tissue: pleiotropic effects in vivo. Endocrinology 150 5262–5272. (https://doi.org/10.1210/en.2009-0603)
Kim TJ, Sun J, Lu S, Zhang J & Wang Y 2014 The regulation of beta-adrenergic receptor-mediated PKA activation by substrate stiffness via microtubule dynamics in human MSCs. Biomaterials 35 8348–8356. (https://doi.org/10.1016/j.biomaterials.2014.06.018)
Komai AM, Brannmark C, Musovic S & Olofsson CS 2014 PKA-independent cAMP stimulation of white adipocyte exocytosis and adipokine secretion: modulations by Ca2+ and ATP. Journal of Physiology 592 5169–5186. (https://doi.org/10.1113/jphysiol.2014.280388)
Komai AM, Musovic S, Peris E, Alrifaiy A, El Hachmane MF, Johansson M, Wernstedt Asterholm I & Olofsson CS 2016 White adipocyte adiponectin exocytosis is stimulated via beta3-adrenergic signaling and activation of Epac1: catecholamine resistance in obesity and type 2 diabetes. Diabetes 65 3301–3313. (https://doi.org/10.2337/db15-1597)
Lajoie P, Goetz JG, Dennis JW & Nabi IR 2009 Lattices, rafts, and scaffolds: domain regulation of receptor signaling at the plasma membrane. Journal of Cell Biology 185 381–385. (https://doi.org/10.1083/jcb.200811059)
Lang T 2007 SNARE proteins and 'membrane rafts'. Journal of Physiology 585 693–698. (https://doi.org/10.1113/jphysiol.2007.134346)
Li S, Shin HJ, Ding EL & van Dam RM 2009 Adiponectin levels and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA 302 179–188. (https://doi.org/10.1001/jama.2009.976)
Lihn AS, Bruun JM, He G, Pedersen SB, Jensen PF & Richelsen B 2004 Lower expression of adiponectin mRNA in visceral adipose tissue in lean and obese subjects. Molecular and Cellular Endocrinology 219 9–15. (https://doi.org/10.1016/j.mce.2004.03.002)
Lim CY, Hong W & Han W 2015 Adiponectin is released via a unique regulated exocytosis pathway from a pre-formed vesicle pool on insulin stimulation. Biochemical Journal 471 381–389. (https://doi.org/10.1042/BJ20150301)
Lipardi C, Mora R, Colomer V, Paladino S, Nitsch L, Rodriguez-Boulan E & Zurzolo C 1998 Caveolin transfection results in caveolae formation but not apical sorting of glycosylphosphatidylinositol (GPI)-anchored proteins in epithelial cells. Journal of Cell Biology 140 617–626. (https://doi.org/10.1083/jcb.140.3.617)
Liu P, Rudick M & Anderson RG 2002 Multiple functions of caveolin-1. Journal of Biological Chemistry 277 41295–41298. (https://doi.org/10.1074/jbc.R200020200)
Mather KJ & Goldberg RB 2014 Clinical use of adiponectin as a marker of metabolic dysregulation. Best Practice and Research in Clinical Endocrinology and Metabolism 28 107–117. (https://doi.org/10.1016/j.beem.2013.06.008)
Meyer LK, Ciaraldi TP, Henry RR, Wittgrove AC & Phillips SA 2013 Adipose tissue depot and cell size dependency of adiponectin synthesis and secretion in human obesity. Adipocyte 2 217–226. (https://doi.org/10.4161/adip.24953)
Morvan J & Tooze SA 2008 Discovery and progress in our understanding of the regulated secretory pathway in neuroendocrine cells. Histochemistry and Cell Biology 129 243–252. (https://doi.org/10.1007/s00418-008-0377-z)
Nakamura A, Miyoshi H, Ukawa S, Nakamura K, Nakagawa T, Terauchi Y, Tamakoshi A & Atsumi T 2018 Serum adiponectin and insulin secretion: a direct or inverse association? Journal of Diabetes Investigation 9 1106–1109. (https://doi.org/10.1111/jdi.12821)
Oka Y & Czech MP 1984 Photoaffinity labeling of insulin-sensitive hexose transporters in intact rat adipocytes. Direct evidence that latent transporters become exposed to the extracellular space in response to insulin. Journal of Biological Chemistry 259 8125–8133.
Omar B, Zmuda-Trzebiatowska E, Manganiello V, Goransson O & Degerman E 2009 Regulation of AMP-activated protein kinase by cAMP in adipocytes: roles for phosphodiesterases, protein kinase B, protein kinase A, Epac and lipolysis. Cellular Signalling 21 760–766. (https://doi.org/10.1016/j.cellsig.2009.01.015)
Ortegren U, Karlsson M, Blazic N, Blomqvist M, Nystrom FH, Gustavsson J, Fredman P & Stralfors P 2004 Lipids and glycosphingolipids in caveolae and surrounding plasma membrane of primary rat adipocytes. European Journal of Biochemistry 271 2028–2036. (https://doi.org/10.1111/j.1432-1033.2004.04117.x)
Ortegren U, Yin L, Ost A, Karlsson H, Nystrom FH & Stralfors P 2006 Separation and characterization of caveolae subclasses in the plasma membrane of primary adipocytes; segregation of specific proteins and functions. FEBS Journal 273 3381–3392. (https://doi.org/10.1111/j.1742-4658.2006.05345.x)
Ost A, Ortegren U, Gustavsson J, Nystrom FH & Stralfors P 2005 Triacylglycerol is synthesized in a specific subclass of caveolae in primary adipocytes. Journal of Biological Chemistry 280 5–8. (https://doi.org/10.1074/jbc.C400429200)
Ostrom RS & Insel PA 2004 The evolving role of lipid rafts and caveolae in G protein-coupled receptor signaling: implications for molecular pharmacology. British Journal of Pharmacology 143 235–245. (https://doi.org/10.1038/sj.bjp.0705930)
Park JJ & Loh YP 2008 How peptide hormone vesicles are transported to the secretion site for exocytosis. Molecular Endocrinology 22 2583–2595. (https://doi.org/10.1210/me.2008-0209)
Parpal S, Karlsson M, Thorn H & Stralfors P 2001 Cholesterol depletion disrupts caveolae and insulin receptor signaling for metabolic control via insulin receptor substrate-1, but not for mitogen-activated protein kinase control. Journal of Biological Chemistry 276 9670–9678. (https://doi.org/10.1074/jbc.M007454200)
Pereira RI & Draznin B 2005 Inhibition of the phosphatidylinositol 3′-kinase signaling pathway leads to decreased insulin-stimulated adiponectin secretion from 3T3-L1 adipocytes. Metabolism: Clinical and Experimental 54 1636–1643. (https://doi.org/10.1016/j.metabol.2005.07.002)
Razani B, Combs TP, Wang XB, Frank PG, Park DS, Russell RG, Li M, Tang B, Jelicks LA, Scherer PE, et al.2002 Caveolin-1-deficient mice are lean, resistant to diet-induced obesity, and show hypertriglyceridemia with adipocyte abnormalities. Journal of Biological Chemistry 277 8635–8647. (https://doi.org/10.1074/jbc.M110970200)
Rodiger M, Werno MW, Wilhelmi I, Baumeier C, Hesse D, Wettschureck N, Offermanns S, Song K, Krauss M & Schurmann A 2018 Adiponectin release and insulin receptor targeting share trans-Golgi-dependent endosomal trafficking routes. Molecular Metabolism 8 167–179. (https://doi.org/10.1016/j.molmet.2017.11.011)
Salaun C, James DJ & Chamberlain LH 2004 Lipid rafts and the regulation of exocytosis. Traffic 5 255–264. (https://doi.org/10.1111/j.1600-0854.2004.0162.x)
Sato M, Hutchinson DS, Halls ML, Furness SG, Bengtsson T, Evans BA & Summers RJ 2012 Interaction with caveolin-1 modulates G protein coupling of mouse beta3-adrenoceptor. Journal of Biological Chemistry 287 20674–20688. (https://doi.org/10.1074/jbc.M111.280651)
Scherer PE, Williams S, Fogliano M, Baldini G & Lodish HF 1995 A novel serum protein similar to C1q, produced exclusively in adipocytes. Journal of Biological Chemistry 270 26746–26749. (https://doi.org/10.1074/jbc.270.45.26746)
Schwencke C, Okumura S, Yamamoto M, Geng YJ & Ishikawa Y 1999 Colocalization of beta-adrenergic receptors and caveolin within the plasma membrane. Journal of Cellular Biochemistry 75 64–72.
Spranger J, Kroke A, Mohlig M, Bergmann MM, Ristow M, Boeing H & Pfeiffer AF 2003 Adiponectin and protection against type 2 diabetes mellitus. Lancet 361 226–228. (https://doi.org/10.1016/S0140-6736(03)12255-6)
Tandon P, Wafer R & Minchin JEN 2018 Adipose morphology and metabolic disease. Journal of Experimental Biology 221 jeb164970. (https://doi.org/10.1242/jeb.164970)
Tchoukalova YD, Koutsari C, Karpyak MV, Votruba SB, Wendland E & Jensen MD 2008 Subcutaneous adipocyte size and body fat distribution. American Journal of Clinical Nutrition 87 56–63. (https://doi.org/10.1093/ajcn/87.1.56)
Thorn H, Stenkula KG, Karlsson M, Ortegren U, Nystrom FH, Gustavsson J & Stralfors P 2003 Cell surface orifices of caveolae and localization of caveolin to the necks of caveolae in adipocytes. Molecular Biology of the Cell 14 3967–3976. (https://doi.org/10.1091/mbc.e03-01-0050)
Wang ZV, Schraw TD, Kim JY, Khan T, Rajala MW, Follenzi A & Scherer PE 2007 Secretion of the adipocyte-specific secretory protein adiponectin critically depends on thiol-mediated protein retention. Molecular and Cellular Biology 27 3716–3731. (https://doi.org/10.1128/MCB.00931-06)
Willoughby D & Cooper DM 2007 Organization and Ca2+ regulation of adenylyl cyclases in cAMP microdomains. Physiological Reviews 87 965–1010. (https://doi.org/10.1152/physrev.00049.2006)
Xie L, O’Reilly CP, Chapes SK & Mora S 2008 Adiponectin and leptin are secreted through distinct trafficking pathways in adipocytes. Biochimica et Biophysica Acta 1782 99–108. (https://doi.org/10.1016/j.bbadis.2007.12.003)
Xu A, Chan KW, Hoo RL, Wang Y, Tan KC, Zhang J, Chen B, Lam MC, Tse C, Cooper GJ, et al.2005 Testosterone selectively reduces the high molecular weight form of adiponectin by inhibiting its secretion from adipocytes. Journal of Biological Chemistry 280 18073–18080. (https://doi.org/10.1074/jbc.M414231200)
Yamauchi T, Kamon J, Waki H, Terauchi Y, Kubota N, Hara K, Mori Y, Ide T, Murakami K, Tsuboyama-Kasaoka N, et al.2001 The fat-derived hormone adiponectin reverses insulin resistance associated with both lipoatrophy and obesity. Nature Medicine 7 941–946. (https://doi.org/10.1038/90984)
