A Neural Protection Racket: AMPK and the GABAB Receptor
2007; Cell Press; Volume: 53; Issue: 2 Linguagem: Inglês
10.1016/j.neuron.2007.01.004
ISSN1097-4199
AutoresD. Grahame Hardie, Bruno G. Frenguelli,
Tópico(s)Regulation of Appetite and Obesity
ResumoThe cellular energy-sensing kinase AMPK is known to be activated in neurons in response to metabolic insults, but the downstream consequences have been unclear. A study by Kuramoto and colleagues in this issue of Neuron favors the idea that AMPK activation is neuroprotective, and suggests a potential mechanism for this effect involving phosphorylation of the GABAB receptor. The cellular energy-sensing kinase AMPK is known to be activated in neurons in response to metabolic insults, but the downstream consequences have been unclear. A study by Kuramoto and colleagues in this issue of Neuron favors the idea that AMPK activation is neuroprotective, and suggests a potential mechanism for this effect involving phosphorylation of the GABAB receptor. The nervous system accounts for a high proportion of total body energy turnover, and neurons are particularly vulnerable to energy deficits due to their rather inflexible metabolism and poor capacity to store nutrients. It is therefore not surprising that 5′AMP-dependent protein kinase (AMPK), part of a signaling system that is a central player in the maintenance of energy balance at both the cellular and whole body levels, should be highly expressed in the central nervous system (Turnley et al., 1999Turnley A.M. Stapleton D. Mann R.J. Witters L.A. Kemp B.E. Bartlett P.F. J. Neurochem. 1999; 72: 1707-1716Crossref PubMed Scopus (218) Google Scholar, Culmsee et al., 2001Culmsee C. Monnig J. Kemp B.E. Mattson M.P. J. Mol. Neurosci. 2001; 17: 45-58Crossref PubMed Scopus (283) Google Scholar). Recent evidence suggests that the AMPK system plays a key role in the regulation of appetite and satiety in the hypothalamus, with the kinase responding to physiological fluctuations in glucose as well as agents such as leptin, ghrelin, and cannabinoids (Kahn et al., 2005Kahn B.B. Alquier T. Carling D. Hardie D.G. Cell Metab. 2005; 1: 15-25Abstract Full Text Full Text PDF PubMed Scopus (2191) Google Scholar). However, its wider role in the nervous system has been less clear. There is general agreement that AMPK is activated in brain tissue in response to ischemia, hypoxia, or glucose deprivation (Culmsee et al., 2001Culmsee C. Monnig J. Kemp B.E. Mattson M.P. J. Mol. Neurosci. 2001; 17: 45-58Crossref PubMed Scopus (283) Google Scholar, Gadalla et al., 2004Gadalla A.E. Pearson T. Currie A.J. Dale N. Hawley S.A. Randall A.D. Hardie D.G. Frenguelli B.G. J. Neurochem. 2004; 88: 1272-1282Crossref PubMed Scopus (114) Google Scholar, McCullough et al., 2005McCullough L.D. Zeng Z. Li H. Landree L.E. McFadden J. Ronnett G.V. J. Biol. Chem. 2005; 280: 20493-20502Crossref PubMed Scopus (287) Google Scholar), but there has been disagreement about the outcome, with one study suggesting that it is neuroprotective (Culmsee et al., 2001Culmsee C. Monnig J. Kemp B.E. Mattson M.P. J. Mol. Neurosci. 2001; 17: 45-58Crossref PubMed Scopus (283) Google Scholar), whereas another suggested that it actually exacerbates the damage caused by these insults (McCullough et al., 2005McCullough L.D. Zeng Z. Li H. Landree L.E. McFadden J. Ronnett G.V. J. Biol. Chem. 2005; 280: 20493-20502Crossref PubMed Scopus (287) Google Scholar). In addition, none of these studies identified the relevant downstream targets for AMPK in neural tissue. In this issue of Neuron, Moss and coworkers (Kuramoto et al., 2007Kuramoto N. Wilkins M.E. Fairfax B.P. Revilla-Sanchez R. Terunuma M. Tamaki K. Iemata M. Warren N. Couve A. Calver A. et al.Neuron. 2007; 53 (this issue): 233-247Abstract Full Text Full Text PDF PubMed Scopus (145) Google Scholar) report new data supporting the idea that the AMPK system is neuroprotective and suggesting that the GABAB receptor is a key mediator of this response (see Figure 1 for an overview). AMPK exists as heterotrimeric complexes composed of catalytic α subunits and regulatory β and γ subunits. Each of these are encoded by multiple genes, of which at least five (α1, α2, β1, β2, and γ1) appear to be expressed in the CNS (Turnley et al., 1999Turnley A.M. Stapleton D. Mann R.J. Witters L.A. Kemp B.E. Bartlett P.F. J. Neurochem. 1999; 72: 1707-1716Crossref PubMed Scopus (218) Google Scholar, Culmsee et al., 2001Culmsee C. Monnig J. Kemp B.E. Mattson M.P. J. Mol. Neurosci. 2001; 17: 45-58Crossref PubMed Scopus (283) Google Scholar). AMPK is activated >100-fold by phosphorylation at T172, located within the kinase domain of the α subunits, by upstream kinases. Binding of AMP to two sites on the γ subunits (Scott et al., 2004Scott J.W. Hawley S.A. Green K.A. Anis M. Stewart G. Scullion G.A. Norman D.G. Hardie D.G. J. Clin. Invest. 2004; 113: 274-284Crossref PubMed Scopus (576) Google Scholar) inhibits dephosphorylation of T172 (Sanders et al., 2006Sanders M.J. Grondin P.O. Hegarty B.D. Snowden M.A. Carling D. Biochem J. 2006; (in press. Published online December 6, 2006)https://doi.org/10.1042/BJ20061520Crossref Scopus (496) Google Scholar) and also causes allosteric activation of the phosphorylated enzyme by up to 10-fold; both effects are antagonized by high concentrations of ATP. AMP is very low in unstressed cells because the adenylate kinase reaction (ATP + AMP ↔ 2ADP) runs from left to right, driven by the high cellular ATP:ADP ratio. However, metabolic stresses that cause a fall in the ATP:ADP ratio will tend to displace the reaction in a leftward direction and cause a large increase in AMP. At least two upstream kinases that phosphorylate T172, i.e., the LKB1 complex and the calmodulin-dependent kinase kinases (especially the CaMKKβ isoform) have recently been identified (Witters et al., 2005Witters L.A. Kemp B.E. Means A.R. Trends Biochem Sci. 2005; 31 (Published online December 13, 2005): 13-16https://doi.org/10.1016/j.tibs.2005.11.009Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar), and both are highly expressed in the CNS. The LKB1 complex appears to be constitutively active and provides a constant low level of T172 phosphorylation that can be markedly increased in response to inhibition of dephosphorylation when AMP binds to the γ subunit. Phosphorylation by CaMKKβ, on the other hand, is triggered by a rise in Ca2+ and can occur in the absence of any increase in AMP (Hawley et al., 2005Hawley S.A. Pan D.A. Mustard K.J. Ross L. Bain J. Edelman A.M. Frenguelli B.G. Hardie D.G. Cell Metab. 2005; 2: 9-19Abstract Full Text Full Text PDF PubMed Scopus (1186) Google Scholar), although the latter would accentuate the effect if it also occurred. Consistent with this view, K+-induced depolarization of rat brain slices caused phosphorylation and activation of AMPK that was sensitive to the CaMKK inhibitor, STO-609, but was not associated with changes in cellular AMP:ATP ratio. Conversely, treatment with the drug phenformin, which inhibits the respiratory chain, caused phosphorylation and activation of AMPK that was insensitive to STO-609 and was associated with large increases in the cellular AMP:ATP ratio (Hawley et al., 2005Hawley S.A. Pan D.A. Mustard K.J. Ross L. Bain J. Edelman A.M. Frenguelli B.G. Hardie D.G. Cell Metab. 2005; 2: 9-19Abstract Full Text Full Text PDF PubMed Scopus (1186) Google Scholar). Although this remains to be properly tested, one can speculate that the Ca2+→CaMKK mechanism for AMPK activation may be the key player during periods of intense excitation produced by normal neural activity, while the AMP- and LKB1-dependent mechanism may become more important during pathological episodes such as hypoxia or ischemia. GABA is the major inhibitory neurotransmitter in the vertebrate brain. It exerts rapid effects via ionotropic GABAA receptors, and more prolonged, slower effects via the metabotropic GABAB class of receptor. While GABAA receptors are extremely diverse, pentameric complexes encoded by at least 17 genes, GABAB receptors are heterodimers formed from only 2 gene products of the 7 transmembrane helix family, i.e., GABABR1 and GABABR2 (Huang, 2006Huang Z.J. Neuron. 2006; 50: 521-524Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar). The R1 subunit contains the GABA binding site, while the R2 subunit provides the coupling mechanism to the heterotrimeric G proteins Gi and Go. The receptors are coupled via these G proteins primarily to activation of K+ channels at postsynaptic sites, where they produce a prolonged hyperpolarization, and are coupled to inhibition of Ca2+ channels at presynaptic sites, where they suppress neurotransmitter release. Activation of GABAB receptors thus suppresses neuronal excitation by multiple mechanisms. Moss and coworkers (Kuramoto et al., 2007Kuramoto N. Wilkins M.E. Fairfax B.P. Revilla-Sanchez R. Terunuma M. Tamaki K. Iemata M. Warren N. Couve A. Calver A. et al.Neuron. 2007; 53 (this issue): 233-247Abstract Full Text Full Text PDF PubMed Scopus (145) Google Scholar) initiated their project via a yeast two-hybrid screen of a rat brain library using the cytoplasmic tail (C-tail) of GABABR1 as bait, which resulted in the identification of the α1 subunit of AMPK as an interacting protein. The interaction was confirmed by coprecipitation analysis of the endogenous proteins in rat brain extracts. Both the α1 and α2 isoforms of AMPK, and the R1 and R2 subunits of the receptor, were found to be associated together in the same complexes. These proteins also appeared to colocalize by fluorescence microscopy in the neuronal processes of cultured hippocampal neurons. As well as forming a complex with the GABAB receptor, AMPK also appears to phosphorylate the protein. GST fusions of the C-terminal tails of both R1 and R2 were phosphorylated by endogenous AMPK in rat brain extracts and by exogenous purified AMPK. One major site was identified on R1 (S917) and one on R2 (S783). To study the possible functional effects of these phosphorylation events, Moss and coworkers transiently expressed R1 and R2 in HEK-293 cells stably expressing the K+ channels Kir 3.1 and 3.2. Internal perfusion via a patch pipette (containing ATP and GTP) initiated a time-dependent rundown of the K+ current activated by a submaximal concentration of GABA. Intriguingly, this rundown was markedly reduced if AMP was included in the nucleotide solution, an effect that was lost with R1/R2S783A or R1S917AR2S783A mutant receptors, but not with an R1S917AR2 mutant receptor. Similar results were obtained with endogenous wild-type GABAB receptors activated by baclofen in cultured hippocampal neurons, where AMPK was activated using the drug metformin. Overall, these results suggested that the phosphorylation of S783 on R2 by AMPK stabilizes the activation of K+ channels by the GABAB receptor. The authors went on to study this phosphorylation further by raising a phosphospecific antibody (anti-pS783) against this site. The antibody recognized the GST fusion of the R2 C-terminal tail only after phosphorylation by AMPK, and also recognized wild-type R2, but not an S738A mutant, at the cell surface of HEK-293 cells. The signal obtained with the cell surface receptor increased in both the transfected HEK-293 cells and in cultured hippocampal neurons in response to the AMPK-activating drug phenformin. The antibodies were also used in immunofluorescence studies in cultured neurons. Double-labeling using anti-S738A and anti-R2 antibodies revealed that 85% of regions labeled with the former were also labeled by the latter, although only 25% of regions labeled with anti-R2 were also labeled with anti-S783, suggesting that a large proportion of cellular R2 was not phosphorylated at S783. To assess whether AMPK was necessary for phosphorylation of S783, Moss and coworkers expressed by transfection an inactive mutant of the α1 subunit of AMPK in cultured hippocampal neurons. This acts as a dominant-negative mutant because the α subunits are unstable in the absence of the β and γ subunits, and the inactive α1 competes with the endogenous, active α subunits for binding to β and γ. Consistent with their hypothesis, the signal obtained with the anti-pS783 antibody was significantly decreased in those neurons that also expressed the transfected, inactive α1 subunit. In an attempt to put these findings into the context of brain injury, the authors transiently occluded the middle cerebral artery in rats to cause an episode of ischemia. Enhanced immunoreactivity with anti-pS783 was seen in the CA3 and dentate gyrus (DG) regions of the hippocampus, but only on the injured side of the brain, whereas immunoreactivity with a phosphorylation-independent R2 antibody was unaltered in response to ischemia or injury. Finally, they addressed whether their novel mechanism exerted a neuroprotective effect by subjecting cultured hippocampal neurons to inhibition of catabolism using deoxyglucose plus azide (inhibitors of glycolysis and oxidative phosphorylation, respectively), which, as expected, caused activation of AMPK and phosphorylation of S783 on R2. When wild-type or S783A mutants of R2 were expressed in these cells by nucleofection (which did not alter the overall level of R2 expression), there was a significant decrease in survival after a 15 min anoxic insult in the cells expressing the mutant. The effect was small (11%), although the authors argue that this may be because the efficiency of transfection was only about 35%. These new results support the previous findings (Culmsee et al., 2001Culmsee C. Monnig J. Kemp B.E. Mattson M.P. J. Mol. Neurosci. 2001; 17: 45-58Crossref PubMed Scopus (283) Google Scholar) that AMPK exerts a neuroprotective effect during metabolic insults, and do not support the previous findings suggesting that it exacerbates brain injury (McCullough et al., 2005McCullough L.D. Zeng Z. Li H. Landree L.E. McFadden J. Ronnett G.V. J. Biol. Chem. 2005; 280: 20493-20502Crossref PubMed Scopus (287) Google Scholar). The latter study is now looking somewhat isolated, and its interpretation is in any case complicated by the uncertain specificity of the pharmacological agents used (C75, compound C, and AICA riboside). The new findings are also significant because, for the first time, a target for AMP specific to neural tissue, which can potentially explain the neuroprotective effect, has been identified. The authors did not specifically address whether coupling of the GABAB receptor to Ca2+ channels was also affected, but if the phosphorylation works by promoting coupling between the receptor and the G protein, it seems likely that this would be the case. If so, AMPK activation would not only accentuate hyperpolarization in response to the postsynaptic action of GABA, but it would also promote its presynaptic effects to inhibit glutamate release. Although the new results represent an important step forward, some important questions and caveats remain. The sequence around S783 on R2 is a very poor fit to the consensus recognition motif for AMPK (Scott et al., 2002Scott J.W. Norman D.G. Hawley S.A. Kontogiannis L. Hardie D.G. J. Mol. Biol. 2002; 317: 309-323Crossref PubMed Scopus (135) Google Scholar). The authors argue that the formation of a complex between AMPK and the GABAB receptor may allow a noncanonical site to become phosphorylated, but an alternative explanation is that phosphorylation is brought about by an unidentified kinase that is activated by AMPK and is a contaminant in one or the other of the preparations used. It should also be pointed out that, while the anti-pS783 antibody appears to be specific in western blotting, it is a little risky to extrapolate this to its use in immunohistochemistry. One elegant way to confirm the author's model would be to create mice with an S783A knockin mutation of R2. While this would not be a trivial undertaking, it should be feasible because all GABAB receptors appear to contain only R1 and R2, so there should be no problems of redundancy. Tissues from these mice would not only serve as an excellent control for immunohistochemistry with the anti-S783 antibody, but one would also predict that the effect of AMPK activation on GABAB receptor function, as well as the neuroprotective effects, should be lost or reduced. One would also expect these effects to be lost in mouse knockouts of the AMPK catalytic subunits. However, since both α1 and α2 were found to be associated with the receptor, it may be necessary to perform a double knockout. A global double α1/2 knockout is already known to cause an embryonic lethal effect, so it may be necessary to make neuron-specific knockouts of each catalytic subunit and then cross them. Phospho-Dependent Functional Modulation of GABAB Receptors by the Metabolic Sensor AMP-Dependent Protein KinaseKuramoto et al.NeuronJanuary 18, 2007In BriefGABAB receptors are heterodimeric G protein-coupled receptors composed of R1 and R2 subunits that mediate slow synaptic inhibition in the brain by activating inwardly rectifying K+ channels (GIRKs) and inhibiting Ca2+ channels. We demonstrate here that GABAB receptors are intimately associated with 5′AMP-dependent protein kinase (AMPK). AMPK acts as a metabolic sensor that is potently activated by increases in 5′AMP concentration that are caused by enhanced metabolic activity, anoxia, or ischemia. Full-Text PDF Open Archive
Referência(s)