Metabolic regulation of neuronal plasticity by the energy sensor AMPK

Abstract

Long Term Potentiation (LTP) is a leading candidate mechanism for learning and memory and is also thought to play a role in the progression of seizures to intractable epilepsy. Maintenance of LTP requires RNA transcription, protein translation and signaling through the mammalian Target of Rapamycin (mTOR) pathway. In peripheral tissue, the energy sensor AMP-activated Protein Kinase (AMPK) negatively regulates the mTOR cascade upon glycolytic inhibition and cellular energy stress. We recently demonstrated that the glycolytic inhibitor 2-deoxy-D-glucose (2DG) alters plasticity to retard epileptogenesis in the kindling model of epilepsy. Reduced kindling progression was associated with increased recruitment of the nuclear metabolic sensor CtBP to NRSF at the BDNF promoter. Given that energy metabolism controls mTOR through AMPK in peripheral tissue and the role of mTOR in LTP in neurons, we asked whether energy metabolism and AMPK control LTP. Using a combination of biochemical approaches and field-recordings in mouse hippocampal slices, we show that the master regulator of energy homeostasis, AMPK couples energy metabolism to LTP expression. Administration of the glycolytic inhibitor 2-deoxy-D-glucose (2DG) or the mitochondrial toxin and anti-Type II Diabetes drug, metformin, or AMP mimetic AICAR results in activation of AMPK, repression of the mTOR pathway and prevents maintenance of Late-Phase LTP (L-LTP). Inhibition of AMPK by either compound-C or the ATP mimetic ara-A rescues the suppression of L-LTP by energy stress. We also show that enhanced LTP via AMPK inhibition requires mTOR signaling. These results directly link energy metabolism to plasticity in the mammalian brain and demonstrate that AMPK is a modulator of LTP. Our work opens up the possibility of using modulators of energy metabolism to control neuronal plasticity in diseases and conditions of aberrant plasticity such as epilepsy.

Figure 1. Metformin and 2DG activate AMPK in hippocampal CA1 neurons.

A) Schematic of the AMPK-mTOR pathway. B) AMPK is activated 30 min after exposure to 2DG (10 mM, p = 0.019, n = 9), metformin (5 µM, p = 0.005, n = 6), or phenformin (10 µM, p = 0.018, n = 6). Hippocampal slices were incubated in ACSF and drug for 30 minutes and subjected to western blot with anti-phospho-Thr172-AMPK antibody followed by βIII-tubulin as a loading control. Representative western blots of duplicate lanes are shown, together with their quantification from at least 10 samples per condition (C). D) ATP levels are reduced in the presence of 10 mM 2DG. Slices were incubated in ACSF+10 mM 2DG (n = 3) as above. Tissue was lysed and subjected to a CellTiter-Glo ATP assay (Promega). E) 2DG activates AMPK in cell bodies of the pyramidal layer (PL) and dendrites of the stratum radiatum (SR). Anti-phospho-Thr172-AMPK immunoreactivity is displayed in green. Neu-N is displayed in red. Scale bar: 10 µm.

Figure 2. AMPK activation represses the mTOR pathway.

10 mM 2DG inhibits the hippocampal mTOR pathway. A) HFS was delivered to the Schaeffer Collateral pathway of slices that had been incubated in the presence or absence of 10 mM 2DG for 30 minutes. Slices were then subjected to western blot analysis using anti-phospho-Thr389-p70S6K, anti-p70S6K, anti-phospho-Ser235/236-rpS6 or anti-rps6 antibody. Representative western blots of duplicate lanes and quantification (B) of 12 samples per condition are shown. Error bars show standard error of the mean (s.e.m). *p<0.05.

Figure 3. AMPK activation inhibits L-LTP expression.

A) AMPK activation inhibits LTP induced by HFS. 10 mM 2DG (n = 10) reduces L-LTP to 10% of control (n = 20) (p = 0.034). 5 µM metformin (n = 8) reduces L-LTP to 40% of control (p = 0.028). B) AMPK activation inhibits LTP induced by TBS. 10 mM 2DG reduces L-LTP (p = 0.022, n = 7) to 30% of control (n = 13), 5 µM metformin reduces L-LTP (p = 0.042, n = 12) to 51% of control. 1 mM AICAR reduces L-LTP (p = 0.0025, n = 8) to 11% of control C) 2DG, metformin and AICAR do not have an effect on basic synaptic transmission. Top: input-output relationships for Schaeffer collateral stimulation and fEPSP slope measured in the presence of ACSF (n = 27), 10 mM 2DG (n = 14), 5 µM metformin (n = 17) or 1 mM AICAR (n = 8). Bottom: Paired Pulse Facilitation is not affected by the presence of 10 mM 2DG (n = 11), 5 µM metformin (n = 17) or 1 mM AICAR (n = 8) compared to ACSF alone (n = 14). Results are plotted as the ratio of fEPSP slopes (2^nd stimulus/1^st stimulus X100) as a function of interpulse interval (0–300 msec). *p = 0.0002. A and B) Inset: representative fEPSP traces shown were taken 4 minutes prior and 180 minutes after stimulation. Error bars show s.e.m.

Figure 4. AMPK inhibition rescues L-LTP expression.

AMPK inhibitors prevent TBS-induced L-LTP loss in the presence of 2DG, metformin or AICAR. A) 1 µM compound-C (n = 9) or B) 100 µM araA (n = 6) prevent 10 mM 2DG-mediated loss of L-LTP. C) 1 µM compound-C prevents 1 mM AICAR-mediated loss of L-LTP (n = 6). D) 1 µM compound-C (n = 9) or E) 100 µM araA (n = 6) prevents 5 µM metformin-mediated loss of L-LTP. Control data from Fig. 3B is reproduced in Fig. 4A and Fig. 4B for comparison. A–E) Inset: representative fEPSP traces shown were taken 4 minutes prior and 180 minutes after stimulation. F) 1 µM compound-C abolishes 10 mM 2DG-mediated AMPK activation. Slices were incubated in ACSF (n = 8), 10 mM 2DG (n = 5) or both 10 mM 2DG and 1 µM compound-C (n = 6) for 30 minutes, subjected to western blotting and quantified as in Fig. 1A. *p = 0.0002. Error bars show s.e.m.

Figure 5. AMPK activation suppresses L-LTP within a time-restricted window.

A) 5 µM metformin was added 20 minutes prior to TBS and washed out immediately prior to TBS (n = 8). L-LTP at 180 minutes post TBS was equal to control (n = 17), B) metformin was added immediately prior to TBS and washed out immediately after TBS (n = 8). L-LTP was reduced to 30% of control (p = 0.0172). C) metformin was added 5 min after stimulation for the duration of the experiment (n = 5). L-LTP at 180 minutes was indistinguishable from control. Control data in Fig. 5A is reproduced in Fig. 5B and Fig. 5C for comparison. Inset: Representative fEPSP traces shown were taken 4 minutes prior to and 180 minutes after TBS. Error bars show s.e.m.

Figure 6. AMPK regulation of L-LTP is rapamycin sensitive.

A) 1 µM compound-C (n = 8, p = 0.0069) or 100 µM ara-A (n = 8, p = 0.0045) results in heightened potentiation after TBS compared to ACSF alone (n = 14). B) 1 µM rapamycin results in suppression of L-LTP to 33% of control (n = 4, p = 0.0097). 1 µM compound-C in the presence of rapamycin fails to significantly enhance L-LTP above rapamycin alone (n = 9, p = 0.2706). Control data in Fig. 6A is reproduced in Fig. 6B for comparison. Inset: Representative fEPSP traces shown were taken 4 minutes prior to and 180 minutes after TBS. Error bars show s.e.m.