β-Adrenoceptors, but not α-adrenoceptors, stimulate AMP-activated protein kinase in brown adipocytes independently of uncoupling protein-1
- First Online:
- Cite this article as:
- Hutchinson, D.S., Chernogubova, E., Dallner, O.S. et al. Diabetologia (2005) 48: 2386. doi:10.1007/s00125-005-1936-7
- 931 Downloads
Brown adipocytes provide a potentially important model system for understanding AMP-activated protein kinase (AMPK) regulation, where adrenergic stimulation leads to mitochondrial uncoupling through uncoupling protein-1 (UCP1) activity. AMPK is a sensor of energy homeostasis and has been implicated in glucose and lipid metabolism in several insulin-sensitive tissues. The aim of this study was to characterise the potential role of AMPK in adrenergically mediated glucose uptake and to find out whether UCP1 is involved in the adrenergic activation of AMPK.
We used primary brown adipocytes differentiated in culture and measured AMPK phosphorylation and glucose uptake following adrenergic activation.
Treatment of adipocytes with noradrenaline (norepinephrine) caused phosphorylation of AMPK via β-adrenoceptors and not α1- or α2-adrenoceptors. This effect was not β3-adrenoceptor specific, since responses remained intact in adipocytes from β3-adrenoceptor knock-out mice. These effects were also mimicked by forskolin and cAMP analogues. Treatment of cells with adenine 8-β-d-arabinofuranoside, an AMPK inhibitor, partially blocked β-adrenoceptor-mediated increases in glucose uptake. Brown adipocytes are characterised by the production of UCP1, which can uncouple the mitochondria. Using adipocytes from Ucp1+/+ and Ucp1−/− mice, we showed that noradrenaline-mediated phosphorylation of AMPK does not require the presence or activity of UCP1.
These results suggest a pathway where increases in cAMP mediated by β-adrenoceptors leads to activation of AMPK in brown adipocytes, which contributes in part to β-adrenoceptor-mediated increases in glucose uptake, an effect independent of the presence or function of UCP1.
Keywordsα-Adrenoceptor β-Adrenoceptor AMP-activated protein kinase Primary brown adipocytes Uncoupling protein-1
acetyl CoA carboxylase
AMP-activated protein kinase
Brown adipose tissue plays an important role in body temperature regulation owing to its ability to generate heat by uncoupling mitochondrial respiration, a process that is mediated by uncoupling protein-1 (UCP1) (for review, see ). It is an important organ in terms of increasing our knowledge of energy regulation and, in rodents, where UCP1 is present in significant amounts in adults, its energy-utilising capacity gives it the potential to influence energy homeostasis. Brown adipose tissue has been shown to play an important role in the regulation of glucose homeostasis and insulin secretion . Glucose uptake is significantly increased in brown adipose tissue in vivo by activation of the sympathetic nervous system independently of insulin [3, 4] and by adrenergic agonists in brown adipocytes in vitro [5, 6, 7, 8].
There has recently been much interest in AMP-activated protein kinase (AMPK), which has been suggested to act as a sensor of energy homeostasis. AMPK consists of a catalytic α and, β subunit and γ regulatory subunits, and is activated by phosphorylation at Thr172 on the catalytic subunit, while allosteric binding of AMP increases the activity of AMPK and the stability of the phosphorylated state (for review, see ). AMPK is widely produced, with high levels produced in tissues that are involved in energy homeostasis, such as liver, heart, skeletal muscle and adipose tissue. While the consensus is that AMPK is important in liver and skeletal muscle, its importance in adipose tissue is uncertain. Recently, the fat-derived hormones leptin and adiponectin, which regulate energy homeostasis, were shown to activate AMPK in skeletal muscle [10, 11], liver  and adipose tissue . AMPK has extensively been shown to increase glucose uptake in skeletal muscle (for review, see ), but its role in glucose uptake in brown adipose tissue remains unknown.
Since AMPK activation is allosterically mediated by increases in AMP, cultured brown adipocytes are a particularly appropriate model in which to study such activation, since activation of UCP1 may elevate endogenous levels of this nucleotide at the expense of ATP, as has been reported for suspensions of mature brown adipocytes . In this study, we have used brown adipocytes in primary culture that have previously been shown to possess intact adrenergic- and insulin-signalling pathways [7, 16, 17, 18]. The aim of this study was to investigate the adrenergic phosphorylation of AMPK, focussing on delineating which adrenergic subtypes are involved in the noradrenaline (norepinephrine)-mediated AMPK phosphorylation response, and whether AMPK is involved in mediating the glucose uptake response to adrenergic agonists.
Materials and methods
Materials and reagents
Drugs and reagents were purchased as follows: insulin (Actrapid; Novo Nordisk, Bagsvaerd, Denmark); 2-deoxy-d-[1-3H] glucose (specific activity 35×104-44×104 MBq/mmol; Amersham Biosciences, Little Chalfont, UK); H89 (Calbiochem, La Jolla, CA, USA); ADP, AMP, adenine 8-β-d-arabinofuranoside (Ara-A), ATP, 8-bromo-cAMP, cirazoline, CL316243, clonidine, CoA, forskolin, (−)-isoprenaline, luciferase, d-luciferin, myokinase, nucleoside-5′-diphosphate-kinase, (−)-noradrenaline, SR59230A (Sigma Chemical Company, St Louis, MO, USA); 5-aminoimidazole-4-carboxamide 1-β-d-ribonucleoside (AICAR) (Toronto Research Chemicals, North York, ON, Canada). ICI89406 was a gift from Imperial Chemical Industries (AstraZeneca, Södertälje, Sweden).
All cell culture media and supplements were obtained from Gibco-BRL Life Technologies (Gaithersburg, MD, USA). All antibodies, except those against UCP1, were obtained from Cell Signaling Technology (Beverly, MA, USA). All other drugs and reagents used were of analytical grade.
Three-week-old FVB, Adrb3−/− (which lack the gene encoding the β3-adrenoceptor), Ucp1+/+ or Ucp1−/− mice of either sex were bred at the institute. Ucp1−/− mice were progeny of those described previously , backcrossed to at least ten generations to a pure C57BL/6J background (UCP1+/+). Adrb3−/− mice and their controls (FVB) were the offspring of a previously described strain . Experiments were conducted with permission from the North Stockholm Animal Ethics Committee.
Cell isolation and cell culture of mouse brown adipocytes
Brown fat precursor cells were isolated essentially as described previously [21, 22]. Cells were grown in DMEM (4.5 g glucose/l) containing newborn calf serum (10%, v/v), insulin (2.4 nmol/l), HEPES (10 mmol/l), l-glutamine (4 mmol/l), penicillin (50 IU/ml), streptomycin (50 μg/ml) and sodium ascorbate (25 μg/ml) under 8% CO2 at 37°C. After 5 days in culture, the brown adipocyte precursor cells spontaneously convert from displaying a fibroblast-like morphology to acquiring the typical multilocular lipid droplets seen in mature brown adipocytes; this conversion occurs at the time of cellular confluence [21, 22]. In these cells, spontaneous induction of Adrb3 mRNA reaches a steady-state level at day 5 that corresponds with the ability of noradrenaline to induce the expression of Ucp1, the gene encoding the most specific brown adipocyte differentiation marker, UCP1 .
Cells were used for all experiments following 7 days of differentiation. On day 6, cells were serum-starved overnight in DMEM/Nutrient Mix F12 (1:1) containing l-glutamine (4 mmol/l), BSA (0.5%, v/v), insulin (2.4 nmol/l), HEPES (10 mmol/l), penicillin (50 IU/ml), streptomycin (50 μg/ml) and sodium ascorbate (25 μg/ml).
Cells were serum-starved overnight before each of the experiments performed on day 7 and were exposed to the drugs for the times and at the concentrations indicated. Extraction of cells for AMPK or acetyl CoA carboxylase (ACC) immunoblotting was performed as previously described , except that samples were electrotransferred to Hybond-P polyvinylidene difluoride membranes (pore size 0.45 μm; Amersham Biosciences, Arlington Heights, IL, USA). For UCP1 immunoblotting or measurement of total AMPK levels in primary brown adipocytes during differentiation, cells were washed twice in PBS and harvested in lysis buffer (Tris [62.5 mmol/l, pH 6.8], SDS [2%, v/v], glycerol [10%, v/v]). The lysate was sonicated for a few seconds and the concentration of the soluble proteins determined . Following determination of protein concentration, the lysis buffer was supplemented with dithiothreitol (50 mmol/l) and bromophenol blue (1%, v/v) in each sample at a ratio of 1:9. Proteins were resolved by SDS–PAGE on 10% polyacrylamide gels and electrotransferred to Hybond-P membranes (Amersham Biosciences). The primary antibodies used were phospho-AMPK (Thr172), total AMPK, phospho-ACC (Ser79) or total ACC diluted 1:1,000, or UCP1 primary polyclonal antibody (produced in rabbits by NeoMPS [Strasbourg, France] from a 12 amino acid peptide sequence at the C-terminus of mouse UCP1) used at a dilution of 1:2,000. Primary antibodies were detected using a secondary antibody (1:2,000 dilution; horseradish peroxidase-linked anti-rabbit IgG) and enhanced chemiluminescence (Amersham Biosciences). A positive control (brown adipose tissue mitochondria from a cold-exposed C57BL/6J mouse; kindly donated by I. Shabalina, Stockholm University) was used in these studies and 0.2 μg protein run on the gels to elucidate the band representing UCP1 from the primary brown adipocytes.
Results are expressed as the ratio of phosphorylated:total protein, with the ratio normalised in each experiment to that of control samples, unless stated otherwise. All experiments were performed singly or in duplicate, with ‘n’ referring to the number of independent experiments performed. In all experiments, basal phosphorylated AMPK:total AMPK levels were unchanged over the time periods examined (data not shown) and have been omitted from diagrams for clarity.
Measurement of the AMP:ATP ratio
Cells were serum-starved overnight before each experiment performed on day 7. On day 7, 1 ml of boiling water was added to each well of a six-well plate, and the cells scraped and boiled (3 min). Samples were centrifuged (12,000 g, 4°C, 10 min) to pellet cell debris, and the supernatant fractions were used for further analysis. Each sample was diluted 1:10 and three aliquots (20 μl) transferred to a white 96-well plate and 30 μl of either reagent A (aqueous tricine buffer [40 mol/l, pH 7.8], MgSO4 [8 mmol/l], EDTA [0.17 mmol/l], which measures total cellular ATP), reagent B (reagent A buffer supplemented with dCTP [100 μmol/l], nucleoside-5′-diphosphate-kinase [10 U/ml], which converts the ADP in a given sample to ATP) or reagent C (reagent B supplemented with myokinase [10 U/ml], which converts the AMP and ADP in a given sample to ATP) added for measurement of ATP, ADP and AMP, respectively. Samples were incubated for 12 h at 37°C, after which the reactions were terminated by boiling for 5 min, and then 50 μl of luciferin–luciferase reagent (aqueous tricine buffer [25 mmol/l, pH 7.8], MgSO4 [5 mmol/l], EDTA [0.1 mmol/l], d-luciferin [500 μmol/l], luciferase [10 μg/ml], dithiothreitol [2 mmol/l], CoA [0.5 mmol/l]) were added to each well. Light emission was measured with a Fujifilm LAS-1000 CCD camera (Fujifilm, Tokyo, Japan) and analysed with ImageQuantNT (Molecular Dynamics, Sunnyvale, CA, USA). AMP:ATP ratios were calculated using the following equation: (reagent C values−reagent B values)/reagent A values.
Results are presented as the mean values±SEM. Paired or unpaired t-tests were performed (where indicated) to test for significance between different treatments and/or controls. A p value less than or equal to 0.05 was considered significant.
AMPK content increases during differentiation of primary brown adipocytes
Insulin fails to phosphorylate AMPK in primary brown adipocytes
The AMPK activator AICAR was able to phosphorylate AMPK in primary brown adipocytes (approximately four-fold increase relative to basal levels; Fig. 1b). In contrast, insulin, a potent activator of glucose uptake in these cells , was unable to phosphorylate AMPK (Fig. 1c).
β-Adrenoceptor, but not α1- or α2-adrenoceptor, activation stimulates AMPK phosphorylation in primary brown adipocytes
Forskolin and cAMP analogues mimic the phosphorylation of AMPK by the β-adrenoceptor activation
AMPK phosphorylation is not impaired in Ucp1−/− adipocytes
AMP:ATP ratios in Ucp1+/+ and Ucp1−/− adipocytes are identical
Since the basal level of phosphorylated AMPK was significantly reduced in Ucp1−/− adipocytes compared with Ucp1+/+ adipocytes, we measured the basal AMP:ATP ratios in unstimulated cells using a luciferin–luciferase-based assay. The AMP:ATP ratios were comparable in the two genotypes (AMP:ATP ratio 0.61±0.03 for Ucp1+/+ vs. 0.74±0.5 for Ucp1−/−; n=2 in triplicate), suggesting that the observed decrease in basal AMPK phosphorylation levels observed in the Ucp1−/− adipocytes cannot be due to an initial reduction in the AMP:ATP ratio.
β-Adrenergically mediated glucose uptake requires AMPK
UCP1 is not required for β-adrenergically stimulated glucose uptake
We examined whether UCP1 is involved in insulin- or adrenergically mediated glucose uptake using Ucp1+/+ or Ucp1−/− mice. Basal glucose uptake was similar in cells derived from the two genotypes (data not shown). Acute exposure of adipocytes to insulin or noradrenaline stimulated glucose uptake to a similar extent in Ucp1+/+ and Ucp1−/− mice, indicating that there is no impairment of the insulin- or adrenergically-mediated pathways of glucose uptake (Fig. 6c). Long-term noradrenaline treatment, which, as shown above, significantly raises UCP1 protein levels, had no influence on basal glucose uptake (data not shown), and acute exposure of adipocytes to noradrenaline increased glucose uptake to the same extent in Ucp1+/+ and Ucp1−/− adipocytes.
In this study, we have investigated whether AMPK plays a role in adrenergically mediated glucose uptake in primary brown adipocytes. Noradrenaline phosphorylated AMPK in these cells, which are known to have intact insulin and adrenergic signalling systems. Brown adipose tissue produces α1-, α2- and β-adrenoceptor subtypes, but brown adipocytes, in addition to α1-adrenoceptors, predominately produce β1- and β3-adrenoceptors, with little or no β2-adrenoceptors . Since brown adipose tissue contains cell types other than adipocytes, such as endothelial cells, blood vessels, fibroblasts and nerves, the relative contribution of the different adrenergic subtypes in each cell type differs vastly. Endothelial cells have been used to elucidate AMPK signalling and effects by several agents [29, 30, 31], which makes it difficult to elucidate the effects of noradrenaline in brown adipocytes using brown adipose tissue, because of potential interactions with other cell types in the tissue. We therefore used primary brown adipocytes in culture to examine effects on mature brown adipocytes.
Elucidation of the receptor subtype(s) involved in noradrenaline activation using selective adrenergic agonists revealed that β-adrenoceptors, but not α1- or α2-adrenoceptors, were responsible for the noradrenaline effect. Previous studies in recombinant CHO-K1 and L6 skeletal muscle cells indicated that Gq-coupled receptors, but not Gs- or Gi-coupled receptors, phosphorylate and activate AMPK . However, our results in primary brown adipocytes, showing that Gs-coupled receptors (β-adrenoceptors) are capable of phosphorylating AMPK, are in accordance with studies in 3T3-L1 adipocytes  and in isolated rat adipocytes . In FVB adipocytes, isoprenaline (a non-selective β-adrenergic agonist that acts primarily via β3-adrenoceptors and not β1-adrenoceptors in this study) and CL316243 (a β3-adrenergic agonist) phosphorylated AMPK to the same degree as noradrenaline. In addition, isoprenaline was able to phosphorylate ACC at Ser79, which is phosphorylated in response to AMPK, and not PKA, activation (for review, see ), demonstrating that isoprenaline affects the AMPK cascade at another level. By comparing primary brown adipocytes from FVB and Adrb3−/− mice, it was clear that both β3- and β1-adrenoceptors were capable of mediating AMPK phosphorylation. In mature adipocytes, the β3-adrenoceptor is the predominant β-adrenoceptor that is coupled to intracellular signalling events [7, 17]. The β1-adrenoceptor only becomes coupled when the β3-adrenoceptor is absent in mature adipocytes . In contrast to earlier results, which indicated that α1-adrenoceptors activated glucose uptake in cells from Adrb3−/− mice , no α1-adrenoceptor stimulation of AMPK was evident, irrespective of whether the β3-adrenoceptor was present or not. Thus, the previously reported α1-adrenoceptor stimulation of glucose uptake clearly does not involve activation of AMPK. Since forskolin and 8-bromo-cAMP both phosphorylated AMPK, this indicates that β-adrenoceptors phosphorylated AMPK through increases in cAMP. The PKA inhibitor H89 inhibited AMPK phosphorylation by isoprenaline in FVB adipocytes, but these results may not be reflective of the role of PKA, since H89 potentially inhibits mammalian AMPK . H89 is also an antagonist of β1- and β2-adrenoceptors , but the action of H89 is presumably not achieved through any antagonist action at β3-adrenoceptors . Therefore, the use of PKA inhibitors that are more selective to address this issue is under investigation.
UCP1 has been hypothesised to play a role in AMPK activation, since UCP1 activity can decrease the ATP:ADP ratio in a cell and therefore increase the AMP:ATP ratio ; this means that UCP1 activation could be linked to AMPK activation. It is well established that adrenergic activation in brown adipocytes significantly increases both UCP1 production and activity. Based on this it may be hypothesised that noradrenaline increases AMPK phosphorylation via a UCP1-mediated increase in AMP. An interaction between UCP1 and AMPK activities has been suggested , following observations that ectopic overproduction of UCP1 in white adipocytes enhanced AMPK activity, and that mitochondrial uncouplers such as dinitrophenol activate AMPK in skeletal muscle [39, 40]. In our study, in adipocytes grown in culture from Ucp1+/+ or Ucp1−/− mice, where effects from alterations in the sympathetic drive from the animals are abolished, AICAR- and adrenergically mediated increases in AMPK phosphorylation levels (measured as a fold-increase relative to baseline for each of the respective genotypes) were identical. One difference was the observation that basal phosphorylation levels of AMPK were consistently lower (∼50%) in adipocytes derived from Ucp1−/− animals when results were expressed as a percentage of the Ucp1+/+ unstimulated state, a result not easily explainable in terms of UCP1 activity, since UCP1 production is negligible in non-stimulated +/+ cells and the basal AMP:ATP ratio is similar between genotypes. Under conditions (long-term adrenergic stimulation) where UCP1 protein levels are markedly enhanced and activated in Ucp1+/+, but not Ucp1−/− adipocytes, stimulation of phosphorylation of AMPK was comparable in the two genotypes, suggesting that UCP1 activity is not a prerequisite for adrenergically mediated increases in AMPK.
AMPK is extensively linked to glucose uptake in skeletal muscle (for review, see ), and the fat-derived hormones leptin and adiponectin activate AMPK in skeletal muscle, liver and adipose tissue [10, 11, 12, 13]. AICAR increases glucose uptake in skeletal muscle. Results in white adipocytes are varied, with some reports showing a positive action of AICAR on glucose uptake [41, 42] and others not . More confusion arises because of reports that the positive action of AICAR on glucose uptake in white adipocytes is due to actions other than AMPK, despite AICAR being able to phosphorylate AMPK [41, 42]. In this study we have shown that AICAR significantly increases glucose uptake in primary brown adipocytes and that this increase is correlated with the ability of AICAR to phosphorylate AMPK. Mitochondrial uncouplers and inhibitors of mitochondrial complexes are activators of glucose uptake  and AMPK activity [39, 40, 43], and it has been suggested that UCP1 activity may also promote glucose uptake as a result of AMPK phosphorylation in adipocytes. We show that UCP1 was not required for β-adrenoceptor AMPK phosphorylation or adrenergically mediated glucose uptake.
Ara-A, an intracellular competitive inhibitor of AMPK, inhibits AMPK-mediated glucose uptake in skeletal muscle , papillary muscle  and adipose tissue . Its use to inhibit skeletal muscle contraction-mediated glucose uptake was ‘unsatisfactory’ , since Ara-A partially inhibited contraction-mediated glucose uptake with no inhibition of contraction-mediated increases in AMPK activity. In adipose tissue, Ara-A has proved more reliable since it inhibits AICAR-mediated increases in AMPK phosphorylation and activity and does not inhibit insulin-mediated increases in glucose uptake . β-Adrenoceptor-mediated glucose uptake was partially inhibited by Ara-A, suggesting that AMPK is involved in adrenergically mediated glucose uptake. In brown adipocytes, insulin stimulation results in the translocation of vesicles containing GLUT4 (now known as SLC2A4) from intracellular compartments to the plasma membrane, resulting in increased glucose uptake . Noradrenaline has been proposed to increase glucose uptake in brown adipocytes via a mechanism that is independent of GLUT4 translocation [46, 47] and instead involves increases in GLUT1 activity . The involvement of AMPK in this process is currently under investigation.
It was recently suggested that sympathetically stimulated glucose utilisation in brown adipose tissue is due to the serial activation of UCP1 and AMPK . However, our results indicate that β-adrenoceptors activate AMPK independently of activation of UCP1 (but that there is a basal difference between cells isolated from Ucp1+/+ and Ucp1−/− animals that we are currently unable to explain but is clearly independent of UCP1 production). The differences between the results of Inokuma et al.  and the present results are most likely to be a consequence of the different systems used. Cells in culture have not previously been exposed to a sympathetic stimulus, whereas this is probably the case in intact mice at 26°C. We have indications that sympathetic activity is enhanced to a greater degree in Ucp1−/− animals; these are consistent with other studies . Thus, we suggest that the stimulatory action of sympathetic activity (noradrenaline) is probably already maximal in the Ucp1−/− mice, not that this activity is lost.
AMPK also has implications for fatty acid synthesis and oxidation, glycogen synthesis, fatty acid uptake and lipolysis in adipose tissue. Recently, β-adrenoceptors were shown to stimulate lipolysis partly via AMPK in 3T3-L1 adipocytes, using adenoviral expression of a gene encoding a dominant negative form of the α2-AMPK subunit . This is in contrast to other studies where AICAR inhibited isoprenaline-mediated lipolysis in rat white adipocytes [49, 50] and 3T3-L1 adipocytes . Isoprenaline-mediated lipolysis was reduced by expression of gene encoding a constitutively active form of AMPK and, conversely, increased when AMPK activity was decreased by either a dominant negative form of AMPK or in α1-AMPK subunit knock-out mice (the predominant subunit expressed in white adipocytes) . Hence, investigation of the role of AMPK in other β-adrenoceptor effects in adipose tissue would be of interest, considering that AMPK is a candidate target for the treatment of diabetes and obesity.
In conclusion we have shown that noradrenaline, acting via β-adrenoceptors, increases AMPK phosphorylation probably via increases in cAMP levels, a response that is independent of UCP1 expression and function. This activation of AMPK is partially responsible for the noradrenaline-mediated increases in glucose uptake seen in primary brown adipocytes. Further studies are needed to ascertain the importance of AMPK in other adrenergically mediated functions in brown adipose tissue, such as lipolysis, glycogen synthesis and fatty acid metabolism.
This study was supported by the Swedish Science Research Council, the Tore Nilssons Grant for Medical Research, Magnus Bergvall Stiftelse and Jeanssonska Stiftelser. D. S. Hutchinson is a C. J. Martin fellow from the National Health and Medical Research Council of Australia.