Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Nov 6.
Published in final edited form as: Neurobiol Dis. 2024 Sep 13;201:106664. doi: 10.1016/j.nbd.2024.106664

Suppression of neuronal AMPKβ2 isoform impairs recognition memory and synaptic plasticity

Nathaniel A Swift a, Qian Yang a, Hannah M Jester a, Xueyan Zhou a, Adam Manuel a, Bruce E Kemp b, Gregory R Steinberg c, Tao Ma a,d,*
PMCID: PMC11539201  NIHMSID: NIHMS2031281  PMID: 39278510

Abstract

AMP-activated protein kinase (AMPK) is an αβγ heterotrimer protein kinase that functions as a molecular sensor to maintain energy homeostasis. Accumulating evidence suggests a role of AMPK signaling in the regulation of synaptic plasticity and cognitive function; however, isoform-specific roles of AMPK in the central nervous system (CNS) remain elusive. Regulation of the AMPK activities has focused on the manipulation of the α or γ subunit. Meanwhile, accumulating evidence indicates that the β subunit is critical for sensing nutrients such as fatty acids and glycogen to control AMPK activity. Here, we generated transgenic mice with conditional suppression of either AMPKβ1 or β2 in neurons and characterized potential isoform-specific roles of AMPKβ in cognitive function and underlying mechanisms. We found that AMPKβ2 (but not β1) suppression resulted in impaired recognition memory, reduced hippocampal synaptic plasticity, and altered structure of hippocampal post-synaptic densities and dendritic spines. Our study implicates a role for the AMPKβ2 isoform in the regulation of synaptic and cognitive function.

Keywords: AMPK, Isoform, Memory, LTP, Synaptic plasticity, Mouse model

1. Introduction

At the cellular level, maintenance of energy homeostasis is tightly controlled by AMP-activated protein kinase (AMPK) (Hardie, 2014; Steinberg and Kemp, 2009). AMPK is a heterotrimeric protein complex comprising a catalytic α subunit, a scaffolding subunit β, and a regulatory γ subunit, each of which is encoded by at least two isoforms (α1, α2; β1, β2; γ1, γ2, γ3) (Dai et al., 2017; Ross et al., 2016; Steinberg and Hardie, 2023). Canonical mechanisms for AMPK activation involve the binding of AMP to the γ subunit and/or phosphorylation of the α subunit at the Thr172 site (Steinberg and Hardie, 2023). Most studies on the regulation of the AMPK have focused on manipulation of the α or γ subunit; however, accumulating evidence indicates that the β subunit is critical for sensing nutrients such as fatty acids and glycogen to control AMPK activity (Sanz et al., 2013; Steinberg and Hardie, 2023). As a molecular sensor for the maintenance of energy homeostasis, AMPK signaling has been extensively studied in the context of exercise and caloric restriction for its physiological roles. Furthermore, mounting evidence points to AMPK signaling dysregulation as a key contributor to common metabolic disorders including diabetes and obesity, and targeting AMPK has become one of the leading therapeutic strategies for such disorders (7). In comparison, the roles of AMPK in the central nervous system (CNS) under physiological and pathological conditions are not well understood.

AMPK signaling controls a number of cellular processes including mitochondrial regulation, de novo cholesterol, fatty acid, and protein synthesis (Herzig and Shaw, 2018; Oakhill et al., 2010; Zhou et al., 2001). Substantial studies have demonstrated that de novo protein synthesis is integral to the long-lasting forms of memory and synaptic plasticity such as long-term potentiation (LTP), a widely studied cellular model for learning & memory (Costa-Mattioli et al., 2009; Klann and Dever, 2004; Ma et al., 2014; Wang et al., 2019). Previous studies, mostly from non-neuronal systems, indicate that AMPK activation results in the inhibition of protein synthesis through at least two signaling mechanisms: phosphorylation of the eukaryotic elongation factor 2 (eEF2) via its kinase eEF2K and inhibition of the mammalian (mechanistic) target of rapamycin complex 1 (mTORC1) pathway (Steinberg and Kemp, 2009). Our recent studies in mice suggest an isoform-specific role of AMPKα in the regulation of LTP and cognitive function (Yang et al., 2020). Moreover, hyperactive AMPK activity (as measured by Thr172 phosphorylation) and altered expression of AMPKα isoforms have been linked to Alzheimer’s disease (AD), a devastating neurological disorder marked by synaptic failure and dementia (Ma et al., 2014; Vingtdeux et al., 2011; Zimmermann et al., 2020). While these studies have focused on the well-characterized catalytic α subunit, little is known with regard to the role of the regulatory β subunit, particularly its potential isoform-specific roles, in synaptic plasticity and learning & memory. Of the drugs currently on market which directly activate AMPK, the majority target β1, β2, or both (Steinberg and Carling, 2019); thus, understanding the roles of AMPKβ isoforms in synaptic and cognitive function could open the door to a new therapeutic target for neuronal disorders characterized by cognitive impairments using drugs which are already available.

AMPK’s β1 and β2 isoforms are encoded by Prkab1 and Prkab2, respectively, and comprise a subunit-binding sequence on the C-terminus connected via a β-linker to a carbohydrate-binding module (CBM), followed by a myristoylated N-terminus (Ross et al., 2016). The subunitbinding sequence is responsible for the stable formation of the AMPK heterotrimer, while the carbohydrate-binding module aids in the trafficking of AMPK to various carbohydrates, chiefly glycogen (Ross et al., 2016; Sanz et al., 2013). Further, the β-linker has been implicated in the protection of AMPKα from dephosphorylation, and the myristoylation of the N-terminus has been shown to be required for the AMP-assisted phosphorylation of AMPK by various upstream kinases (Oakhill et al., 2010; Yan et al., 2018). It appears that both β1 and β2 are present in neurons, although the isoform-specific cellular and subcellular distribution pattern is not well understood (Turnley et al., 1999). Moreover, to the best of our knowledge, the roles of AMPKβ isoforms in cognitive and synaptic function have not been investigated. Here, we generated transgenic mice with conditional suppression of either AMPKβ1 or β2 in neurons and conducted a series of experiments with multiple approaches to characterize potential isoform-specific roles of AMPKβ in cognitive function and underlying mechanisms. Our findings implicate the β2 isoform of AMPK as an integral factor in synaptic plasticity and recognition memory.

2. Materials and methods

2.1. Mice

All mice were housed at Wake Forest School of Medicine barrier facility under a 12-h light/dark cycle under the supervision of the Animal Research Program. Mice underwent regular cage cleaning, bedding changes, and food/water changes. AMPKβ1 and AMPKβ2 floxed mice were generated as described. B6.Cg-Tg(Camk2a-cre)T29-1Stl/J mice (stock 005359) were purchased from Jackson Laboratory. Mice harboring loxP-flanked AMPKβ1 or AMPKβ2 (O’Neill et al., 2011) were bred with mice expressing an excitatory neuron-specific Cre recombinase (Camk2a-cre) (Yang et al., 2020) to generate AMPKβ1 or AMPKβ2 conditional knockdown or knockout mice, respectively. All mice (both male and female) used in this study were 3–5 months of age.

2.2. Behavior assessments

2.2.1. Open field assay (OF)

All mice were handled by the experimenter for five days prior to experimentation and habituated to the experimental procedure room for one hour prior to testing. Animals were placed in an opaque, square arena (40 × 40 × 40 cm) and allowed to freely explore for 15 min. Velocity and distance traveled were measured via video tracking software (EthoVision XT). Time spent in the periphery (outer 10 cm) of the container was measured as a percentage of the total time spent in the arena.

2.2.2. Novel object recognition task (NOR)

Mice were placed in the same arena used in OF with two identical, evenly spaced objects and allowed to freely explore for 5 min, two days in a row. On the third day, one of the objects (chosen at random) was replaced with a similarly sized novel object. The amount of time spent with each object was recorded manually and expressed in the form of a discrimination index (DI=(tN+tF)÷tT, where tN = time spent with novel object, tF = time spent with familiar object, and tT = total time spent with both objects). Mice were excluded from analysis if tT < 5 s or if they displayed a location preference (≥ 90 % of tT spent with one object).

2.2.3. Morris water maze (MWM) & Visible Platform (VP)

Mice were habituated to the procedure room for one hour prior to experimentation in empty mouse cages warmed via a heating pad. Mice displaying open wounds were disqualified from use. An opaque plastic pool (135 cm diameter) was filled with water and white washable paint. A circular platform was placed in the center of one quadrant of the pool, and water level was adjusted to just obscure the platform from view of the mice. Animals were placed in the pool and allowed to swim until they found the hidden platform for a maximum of 60s (in which case they were then placed on the hidden platform for 5–10 s) and removed from the pool. Once removed, mice were briefly dried with a towel and returned to their heated cage. After 15 min, the process was repeated for a total of four trials per day, five days in a row, rotating the starting quadrant between each trial. Following the fifth day of training, mice remained in their heated cages (with access to water) for two hours before removing the platform from the pool and returning the mice to the pool (the “probe” trial). EthoVision XT software was used to track the mice as they swam for 60s and the amount of time spent in the target quadrant which previously contained the platform (TQ), as well as the frequency with which the mice entered the space previously occupied by the platform were recorded. After a one- to two-day break, a visible platform was placed in the same pool and mice were allowed a maximum of 60s to find it. Four trials were run with a similar inter-trial interval of 15 min for two consecutive days, with the platform rotated to a new quadrant between each trial.

2.2.4. Marble-burying task (MBT)

A standard rat cage was filled with 5 cm bedding and 15 marbles were evenly distributed across the surface. Mice were individually placed in the cage and allowed free reign of the arena for 30 min, after which time they were removed from the arena and the number of marbles buried was counted. A marble was considered buried if >65 % of the marble was buried under the bedding, with a higher number of buried nestlets indicating compulsive-like behavior. Their distance traveled was recorded via motion-tracking software.

2.2.5. Nestlet-shredding task (NST)

Mice were individually placed in an empty mouse cage with a thin layer of bedding and a pre-weighed square cotton nestlet placed in the center. Mice were allowed free reign of the cage for 30 min, after which time they were removed from the cage. Nestlets were lightly brushed free of shredded cotton and allowed to dry overnight before being weighed. The ratio of shredded to unshredded nestlet was recorded as a measure of compulsive-like behavior.

2.3. Western blotting

Mice were sacrificed via cervical dislocation, their brains excised, and their hippocampi dissected out. Hippocampi were immediately transferred to a microcentrifuge tube and placed on dry ice before undergoing homogenization in lysis buffer and protein concentration quantification via bicinchoninic acid assay. Protein samples were then loaded into 4–15 % Mini-PROTEAN® TGX™ Precast Gels (Bio-Rad), run at 150-200 V, and transferred to nitrocellulose membranes. After blocking with SuperBlock™ Blocking Buffer (Thermo Scientific) for 10 min, membranes were probed for various primary antibodies (Table S1) overnight. Blots were washed using a 1× Tris-buffered saline solution with 1 % Triton X-100 (TBST) and stained with horseradish peroxidase-labelled secondary antibodies before being visualized using Clarity™ ECL solution (Bio-Rad) and the Bio-Rad ChemiDoc™ MP Imaging System. The ImageJ software was used to perform densitometric analysis, and samples were normalized to either β-actin (for total protein analysis) or total protein (for phosphorylated protein analysis).

2.3.1. Surface sensing of translation assay (SUnSET)

Mice were sacrificed via cervical dislocation before their brains were harvested and sliced transversely into 400 μm-thick sections in cutting buffer (87 mM NaCl, 25 mM NaHCO3, 2.5 mM KCl, 1.25 mM NaH2PO4, 7 mM MgCl2, 0.5 mM CaCl2, 25 mM glucose, 37.5 mM sucrose) with bubbling O2. Hippocampi were dissected from slices and exposed to a 50 % cutting buffer, 50 % artificial cerebrospinal fluid (ACSF; 125 mM NaCl, 25 mM NaHCO3, 2.5 mM KCl, 1.25 mM NaH2PO4, 1 mM MgCl2,2 mM CaCl2, and 25 mM glucose) solution with bubbling O2 for 15 min. Sections were transferred to bubbling ACSF at 32 °C for two hours, at which point 0.5 μg/ml puromycin was added for one hour. Slices were flash-frozen on dry ice before being processed for western blotting. Blots were stained with puromycin (Millipore #AB3258) and normalized to β-actin.

2.3.2. Hippocampal slice preparation and synaptic electrophysiology

400 um acute transverse slices were collected from isolated hippocampus using a Leica VT1200s vibratome. Before being transferred into the chamber for recording, the slices were incubated at room temperature for at least 2 h in ACSF bubbling with 95 % O2 and 5 % CO2. After being transferred into the recording chamber, the slices were perfused with 95 % O2 / 5 % CO2-satuated ACSF at the rate 5.0 ml/min. For electrophysiology recording, all experiments were performed at 30 ± 1 °C in a submersion-type recording chamber. A monophasic, constant-current stimuli (100us, interval 30s) was delivered with concentric bipolar electrode (FHC, USA) to stimulate Schaffer collateral fibers. The initial slope of excitatory postsynaptic potentials (EPSP) was used to evaluate the amplitude of field EPSPs (fEPSPs). The basic stimulation intensity was controlled to get 30–50 % spike threshold. After 30 min stable baseline recording, long-term potentiation (LTP) was induced using high-frequency stimulation consisting of two 1 s, 100 Hz trains separated by 60s.

2.3.3. Transmission electron microscopy (TEM)

Brains were harvested and area CA1 was dissected from 1-mm thick transverse sections and fixed overnight in 1 % PFA / 2.5 % glutaraldehyde in 0.1 M Millonig’s phosphate buffer. After a one-hour exposure to 1 % osmium tetroxide in PBS, hippocampal sections were dehydrated via a series of ethanol dilutions. Samples were infiltrated with Spurr’s resin, which cured overnight at 70 °C before being sectioned at 90 nm with a Reichert-Jung Ultracut E Ultramicrotome. Tissue was stained with uranyl acetate and lead citrate, then imaged using a FEI Technai Spirit TEM (80 kV) and an Advanced Microscopy Techniques 2Vu CCD camera at x11,000. An ROI (418 nm × 275 nm) was drawn in the center of each image and the size, length, and area of each postsynaptic density (PSD) in that ROI was measured for a total of 3 mice per group, 4–6 images per mouse. PSD, polyribosome, and mitochondrial count data was obtained from 3 mice per group and 3 grid sections per mouse, with 10–12 images in each grid section analyzed and averaged.

2.3.4. Golgi-Cox staining

Brains were excised, bisected, and processed using the FD Rapid GolgiStain™ kit. 100 μm slices were created via vibratome and their hippocampi dissected out prior to mounting on gelatin-coated slides and imaged via Keyence BZ-X710 all-in-one fluorescent microscope at 100× magnification. Images were taken of dendritic processes in the stratum radiatum and their spines were classified as either immature (filopodia, long-thin, thin) or mature (mushroom, stubby, branched).

2.4. Statistical analysis

Data are presented as bar graphs with the top of the bar indicating the mean and the error bars indicating the standard error of the mean (STEM). Samples were analyzed using a one-way analysis of variance (ANOVA) with Tukey’s post-hoc comparison (where applicable). An error probability of p < 0.05 was considered statistically significant. Outliers were determined via Grubb’s test (α = 0.05) or, in the case of TEM analysis, the Robust Regression and Outlier Removal (ROUT) test (Q = 1 %). Statistics were conducted using Prism 9.5.1 (GraphPad).

3. Results

3.1.1. Generation of conditional knockout mice with suppression of neuronal AMPKβ isoforms

Briefly, mice harboring loxP-flanked Prkab1 or Prkab2 (O’Neill et al., 2011) were bred with mice expressing neuron-specific CamK2a-Cre recombinase (Yang et al., 2020) to generate heterozygous (β1+/− and β2+/−) and homozygous (β1−/− and β2−/−) AMPK mutant mice (Fig. 1A). Results from Western blot experiments demonstrated reduction in protein levels of AMPKβ isoforms in hippocampus of the mutant mice (Fig. 1B). Suppression of AMPKβ1 or AMPKβ2 had no effect on hippocampal size (Fig. S1A).

Fig. 1. AMPKβ2 suppression impairs recognition memory.

Fig. 1.

Open in a new tab

(A-B) Representative PCR (A) and hippocampal western blot (B) of the generation of a heterozygous (+/−) or homozygous (−/−) mouse model of AMPKβ1 or AMPKβ2 suppression. (C) Cre and β1+/− mice appear to spend a greater time with the NO, while β2+/− mice appear to spend equal time with both objects. (D) β2+/− mice display an impaired ability to discriminate between NO and FO, as assessed by the discrimination index (DI (tN + tF) ÷ tT). n = 14–17 per group; *p < 0.05. One-way ANOVA with Tukey’s post-hoc test, F2,43 = 4.658. (E-F) Unchanged OF performance between groups. n = 11–12 per group. (G) Unaltered MWM five-day learning curve between groups. n = 11–12 per group. (H—I) Probe trial yielded no change in time spent in target quadrant or “platform” crossings. n = 11–12 per group. (J) Mice in all groups display normal VP performance. n = 11–12 per group.

3.1.2. Suppression of AMPKβ2 impairs recognition memory

To assess the effect of isoform-specific AMPKβ suppression on cognitive function, we first performed a battery of behavioral tests on mutant mice with heterozygous AMPKβ isoform suppression and their Cre control littermates. In the novel object recognition task (NOR), which measures long-term recognition memory (Antunes and Biala, 2012; Barker et al., 2007), we found that both Cre and β1+/− mice spent more time with the novel object (NO) than the familiar object (FO), indicating normal recognition memory. Notably, the β2+/− mice spent roughly an equal amount of time with both objects, indicating memory impairments (Fig. 1C). Further analysis using the discrimination index [DI: (time spent with novel object – familiar object) / total time] confirmed impairment of recognition memory in β2+/− mice as indicated by significantly lower DI compared to either Cre or β1+/− mice (Fig. 1D). Mice in all groups had a similar amount of total time spent with both objects (Fig. S1E). In order to control for potential locomotor deficits and to assess baseline anxiety-like behavior in these mice, we employed the open-field assay (OF). Mice in all groups displayed no difference in distance traveled (Fig. 1E) or time spent in the periphery (Fig. 1F) in the OF test, suggesting unaltered locomotion and no anxiety-like phenotype associated with neuronal suppression of AMPKβ isoforms. Additionally, compulsive-like behavior was similar between groups, as measured by the marble-burying and nestlet-shredding tasks (Fig. S1B–D) (Angoa-Pérez et al., 2013).

We further tested these mice in the hidden platform Morris water maze task (MWM), which measures hippocampus-dependent spatial learning and memory (Vorhees and Williams, 2006). Similar to the NOR results, both Cre and AMPKβ1+/− mice displayed normal spatial learning and memory, as indicated by comparable day-to-day escape latency during the training phase (Fig. 1G), as well as target quadrant occupancy and “platform” crossing frequency during the probe trial phase (Fig. 1H and I). Surprisingly, AMPKβ2+/− mice also displayed normal spatial learning and memory (Fig. 1F–H), which is in contrast to the results for the NOR test (Fig. 1C and D) Subsequently, we conducted the visible platform task (VP) to control for memory-independent effects associated with suppression of AMPKβ isoforms such as vision, locomotion, and escape motivation. Escape latency during the two days was consistent among all three experimental groups (Fig. 1J), indicating no effects on any of the aforementioned qualities. Together, these data suggest AMPKβ2 is important for recognition memory, but not spatial learning & memory.

3.1.3. Effects of AMPKβ isoform suppression on hippocampal long-term potentiation and AMPKα activity

We next performed synaptic electrophysiology experiments to measure hippocampal long-term potentiation (LTP), a major form of synaptic plasticity and established cellular model of learning & memory (Bliss and Collingridge, 1993; Ma et al., 2014; Malenka, 2003). Late LTP was induced by a strong high-frequency stimulation (HFS) protocol (Yang et al., 2020) on acute hippocampal slices and the field excitatory postsynaptic potential (fEPSP) slope change, expressed as the percentage above pre-HFS baseline at 90 min. Post-HFS, was quantified. We found that AMPKβ1+/− mice exhibit normal LTP when compared to the control Cre mice (Fig. 2A and B). Importantly, LTP was significantly impaired in AMPKβ2+/− mice compared to Cre or AMPKβ1+/− mice (Fig. 2A and B). Moreover, we analyzed paired pulse facilitation (PPF), a calcium-dependent form of presynaptic plasticity (Santschi and Stanton, 2003), and input-output curve for the evaluation of basal synaptic transmission. Mice in all groups showed similar PPF and input-output curves, indicating no alterations of basal synaptic transmission with suppression of AMPKβ isoforms (Fig. 2C and D). Thus, hippocampal synaptic plasticity was impaired with repression of the AMPKβ2 isoform, but not the β1 isoform, which may contribute to the recognition memory deficits associated with the AMPKβ2+/− mice (Fig. 1A and B).

Fig. 2. Suppression of AMPKβ2 impairs hippocampal LTP and alters AMPKα phosphorylation.

Fig. 2.

Open in a new tab

(A) Hippocampal high-frequency stimulation (HFS)-induced long-term potentiation (LTP) in Cre mice is conserved in β1+/− mice but dampened in β2+/− mice. N = 3, n = 6–8 slices per group. (B) β2+/− mice display significantly reduced fEPSP at 90 min. Post-HFS. N = 3 mice, n = 6–8 slices per group; *p = 0.0106, **p = 0.0039. One-way ANOVA with Tukey’s post-hoc test, F2,19 = 8.486. (C) Unaltered paired pulse facilitation (PPF) between all groups. N = 3 mice, n = 6–8 slices per group. (D) Unchanged input-output response between groups. N = 3 mice, n = 6–8 slices per group. (E) Western blot indicates impaired AMPKα phosphorylation at Thr172. n = 8–11 per group; *p = 0.0137. One-way ANOVA with Tukey’s post-hoc analysis, F2,25 = 4.743. (F) Mice in all groups display similar protein translation as assessed by the SUnSET assay. n = 6 per group. Vertical line indicates a break in the gel for the representative image.

To examine the isoform-specific effects of AMPKβ suppression on AMPK signaling, we first assessed in hippocampus alterations of pan-AMPKα phosphorylation at the Thr172 site, which is indicative of overall AMPK activity (Hardie, 2014). Compared to Cre mice, β2+/− mice displayed decreased levels of p-AMPKα (Thr172), indicating impaired AMPK activity. In contrast, suppression of AMPKβ1 did not alter AMPK activity as evaluated by AMPKα phosphorylation (Fig. 2E). Assessment of isoform-specific phosphorylation sites of AMPKα1 and α2, Ser487 and Ser491, respectively (Coughlan et al., 2016; Dagon et al., 2012; Hawley et al., 2014), showed no change to inhibitory phosphorylation of either isoform in β1+/− or β2+/− mice when compared to Cre (Fig. S2A and S2B). We further analyzed the activities of two major downstream targets of AMPK involved in de novo protein synthesis regulation, eEF2 and mTORC1 (Browne et al., 2004; Figueiredo et al., 2017; Inoki et al., 2003; Steinberg and Kemp, 2009). Interestingly, despite seeing altered AMPK activity in β2+/− mice, we found no change among all three groups with regards to the activities of either eEF2 or mTORC1, as measured by phosphorylation levels of eEF2 (Thr56) and mTOR (Ser2448) (Fig. S2D and S2E). Additionally, we examined S6K1, a downstream target of mTORC1 and upstream regulator of both eEF2 and AMPK (Browne et al., 2004; Dagon et al., 2012; Hay and Sonenberg, 2004), and found no change in phosphorylation at the Thr389 site (Fig. S2F). Moreover, assessment of overall hippocampal de novo protein translation via the surface sensing of translation (SUnSET) assay revealed that the suppression of either β1 or β2 had no effect on protein translation (Fig. 2F). Taken together, these data suggest that AMPKβ2 suppression results in impairment of AMPK activity, without significant impact on the de novo protein synthesis pathway.

3.1.4. Suppression of AMPKβ2 alters post-synaptic morphology in hippocampus

We next conducted transmission electron microscopy (TEM) experiments on the CA1 region of the hippocampus to investigate the effects of AMPKβ isoform inhibition on postsynaptic densities (PSD), structures critical for synaptic function (Vyas and Montgomery, 2016). Interestingly, ultrastructural analysis revealed no change in the number of PSDs in β1+/− or β2+/− mice when compared to Cre mice, but a significant increase in the number of PSDs found in β2+/− mice when compared to β1+/− mice (Fig. 3A–B). Furthermore, measurement of total PSD area was reduced in β2+/− mice compared to Cre or β1+/− mice (Fig. 3C). Additionally, results from Western blot indicates an increase in expression of postsynaptic density protein 95 (PSD-95), a key protein involved in PSD structure and synaptic maturation (El-Husseini et al., 2000) (Fig. S3A).

Fig. 3. Altered postsynaptic morphology of CA1 dendrites in β2+/− mice.

Fig. 3.

Open in a new tab

(A) Representative TEM images. Arrows indicate PSDs. Scale = 500 nm. (B) Number of PSDs counted per 10 μm2. N = 3 mice, n = 9 grid sections per group; *p = 0.0125. One-way ANOVA with Tukey’s post-hoc analysis, F2,24 = 4.890. (C) Quantification of PSD area. N = 3 mice, n = 15 images per group; **p = 0.0048, ***p = 0.0005. One-way ANOVA with Tukey’s post-hoc analysis, F2,42 = 9.600. (D) Classification of immature (1, filopodia; 2, long-thin; 3, thin) and mature (4, mushroom; 5, stubby; 6, branched) spines and representative Golgi-Cox images. Scale = 10 μm. (E) Total density of spines per 10 μm. N = 3 mice, n = 45 images per group; *p = 0.0382, **p = 0.0071. One-way ANOVA with Tukey’s post-hoc test, F2,132 = 5.321. (F-G) Density of mature and immature spines. N = 3 mice, n = 45 images per group; *p = 0.397, **p = 0.0091, ****p < 0.0001. One-way ANOVA with Tukey’s post-hoc analysis, F2,132 = 27.39 (F), F2,132 = 13.46 (G). (H) Quantification of spine maturity as expressed by mature-to-immature ratio. N = 3 mice, n = 45 images per group; ****p < 0.0001. One-way ANOVA with Tukey’s post-hoc analysis, F2,132 = 32.32.

Using a rapid Golgi staining protocol, we also assessed the morphology of dendritic spines of hippocampus, alterations of which have been associated with synaptic plasticity and memory formation (Gipson and Olive, 2017). Spines were classified as either mature (mushroom, branched, stubby) or immature (filopodia, long-thin, thin) (Fig. 3E). Interestingly, compared to the Cre mice, the overall spine density was increased in β1+/− mice, but unaltered in β2+/− mice (Fig. 3F). Further analysis on the spine types revealed that the density of mature spines was decreased in β2+/− mice and increased in β1+/− mice when compared to the Cre group (Fig. 3G, Fig. S3I–J). In comparison, the density of the immature spines was significantly increased in β2+/− mice compared to the Cre or β1+/− mice (Fig. 3H, Fig. S3F-H). Finally, overall spine maturity (as quantified by the ratio of mature spines to immature spines) was decreased in β2+/− mice, but unchanged in β1+/− mice when compared to Cre mice (Fig. 3I). Together with the results of PSD analysis, these data show that suppression of AMPKβ2 alters PSD morphology and decreases dendritic spine maturity, which may contribute to the cognitive deficits and LTP failure observed in the β2+/− mice.

3.1.5. Homozygous knockout of AMPKβ2 isoform results in recognition memory impairment and LTP failure

We also characterized transgenic mice with a homozygous knockout of AMPKβ1 (β1−/−) and AMPKβ2 (β2−/−) using behavioral and electrophysiological approaches. Examination of AMPKβ isoform expression in whole hippocampal lysate shows an apparent exacerbation of decreased AMPKβ1 expression in β1−/− mice, but no change in AMPKβ2 expression in β2−/− mice, when compared to their heterozygous counterparts (Fig. 1B). In line with the results from the heterozygous knockdown cohort, AMPKβ2−/− (but not β1−/−) mice displayed impaired long-term recognition memory as measured by the NOR test (Fig. 4A and B). Both β1−/− and β2−/− mice exhibited normal performance in the OF test (Fig. 4C and D) and had a similar amount of total time spent with both objects (Fig. S1F). Moreover, neither β1−/− nor β2−/− mice displayed spatial learning/memory deficits as evaluated by the MWM test (Fig. 4E-4G). Analysis of VP results did not reveal any difference between groups (Fig. 4H). Results from electrophysiological experiments revealed similar results to those seen in the heterozygous knockdown group, with β2−/− (but not β1−/−) mice displaying impaired hippocampal LTP (Fig. 4I, J). Surprisingly, analysis of the Western blots experiments indicated that hippocampal levels of p-AMPKα (Thr172) in the β2−/− mice was not significantly decreased compared to the Cre group (Fig. 4K). Taken together, these data indicate that suppression of AMPKβ2, but not β1, impairs recognition memory and hippocampal LTP in mice.

Fig. 4. Homozygous knockout mice display similar recognition memory and hippocampal LTP impairments to heterozygous knockdown mice.

Fig. 4.

Open in a new tab

(A) Cre and β1−/− mice appear to spend more time with the NO than the FO. (B) β2−/− mice display an impairment in their ability to discriminate between NO and FO, as assessed by the discrimination index. n = 14—16 per group; *p < 0.05. One-way ANOVA with Tukey’s post-hoc test, F2,41 = 4.530. (C—D) Unchanged OF performance between groups. n = 11–12 per group. (E-G) Unaltered MWM learning curve, time spent in target quadrant, or “platform” crossings. n = 11–12 per group. (H) Mice in all groups display normal VP performance. n = 11 per group. (I) Hippocampal high-frequency stimulation (HFS)-induced long-term potentiation (LTP) is dampened in β2−/− mice, not β1−/− when compared to Cre. N = 3 mice, n = 8 slices per group. (J) β2−/− mice display significantly reduced fEPSP at 90 min. Post-HFS. N = 3 mice, n = 8 slices per group; *p = 0.0333, **p = 0.0048. One-way ANOVA with Tukey’s post-hoc test, F2,21 = 7.364. (K) Western blot of p-AMPKα (Thr172) in homozygous knockout mice. n = 8 per group.

4. Discussion

Integral AMPK function is critically important for the maintenance of energy metabolism balance and aberrant brain AMPK signaling has been linked to neuronal diseases characterized by cognitive impairments (Mairet-Coello et al., 2013; Marinangeli et al., 2016; Zimmermann et al., 2020). Investigation of the roles of AMPK, particularly its isoforms, in regulation of basic neuronal functions could provide insights into mechanistic understanding of such diseases and perhaps novel therapeutic strategies. Taking advantage of the unique transgenic mice we generated with conditional neuronal suppression of AMPKβ isoforms, our current study demonstrates isoform-specific roles of AMPKβ in regulation of synaptic and cognitive function.

Through a variety of behavioral tests, we found the suppression of β2 (not β1) to impair recognition memory, but not spatial learning & memory (Fig. 1). These results appear not to be due to any compulsive- or apathy-like behavior, as measured by marble-burying test (Fig. S1A-C) and total NOR interaction time (Fig. S1D). Furthermore, we found decreased LTP in the β2+/− hippocampus (Fig. 2A-B); this LTP impairment is unlikely due to any change in presynaptic plasticity (Fig. 2C) or basal synaptic transmission (Fig. 2D). It is intriguing that suppression of AMPKβ2 leads to recognition memory impairment (based on NOR results), but no impact on spatial memory (based on MWM results). Both NOR and MWM tests are considered to be dependent on hippocampus (Akirav and Maroun, 2006; Buzsáki and Moser, 2013), although other brain regions involved could be different between the two behavioral assays; for example, it was reported that the perirhinal cortex is heavily implicated in the regulation of recognition memory, but research on its role in spatial memory has yet to reach a consensus (Aggleton et al., 2004; Allen et al., 2020; Broadbent et al., 2004). Additionally, while the NOR test mostly depends on the innate ability of rodents to recognize novelty in an open-box set-up, the MWM appears to be a more stressful test (e.g. swimming for 4 trials per day for 5 days). It is intriguing that while we observed many positive results in experiments using hippocampus (e.g. imaging, biochemical assays, and LTP), the results from MWM are negative in mice with AMPKβ suppression. Future comprehensive studies on how AMPK signaling is regulated under various behavioral paradigms may help uncover the underlying mechanisms. Finally, one limitation of the study is that we use only young (3–5 month old) mice; it is possible that the effects of AMPKβ isoform suppression on murine behavior may be exacerbated or attenuated by age; future studies are planned which will investigate the impact of AMPKβ isoform suppression on the behavior of middle-age (10–12 months) and old-age (18–20 months) mice.

In situations of low cellular energy or increased cellular stress, upstream kinases including LKB1 and CaMKKβ phosphorylate AMPKα at its Thr172 site. This activates AMPK, allowing it to restore homeostasis via the downregulation of ATP-intensive anabolic pathways and the upregulation of ATP-restoring catabolic pathways (Green et al., 2011; Hardie, 2007; Hardie et al., 2012; Oakhill et al., 2010). Here, we found a downregulation of AMPKα phosphorylation at Thr172 in mice with suppression of AMPKβ2 but not β1 (Fig. 2E), suggesting that complexes containing AMPKβ2 may aid in controlling α phosphorylation. The effect seen in the heterozygous group appears to be an AMPK-specific and Thr172-specific effect: examination of two major AMPK inhibitory sites – Ser487 on α1 and Ser491 on α2 – as well as the phosphorylation of each of their main phosphorylating proteins, Akt and S6K1, respectively (Coughlan et al., 2016; Dagon et al., 2012; Hawley et al., 2014; Joseph et al., 2015), yielded no changes between groups (Fig. S2A–C and S2F). Literature suggests that this change in Thr172 phosphorylation could be due to the role of the β-linker in protecting Thr172 from dephosphorylation. Studies have shown that the linker on β2 more potently protects this site than β1 (Yan et al., 2018); thus, it’s possible that phosphatases are more readily able to deactivate AMPK in the absence of β2 than that of β1. It is worth mentioning that we did not find significantly decreased levels of phospho-AMPKα in hippocampus of the AMPKβ2 homozygous knockout mice (Fig. 4K), and the underlying mechanisms (e.g. compensation of other signaling pathways impacting AMPK activities) are unclear.

There was no change in overall hippocampal protein translation (as assessed by the SUnSET assay) in mice with suppression of either AMPKβ isoform (Fig. 2F). Consistently, we did not observe in these mice alterations of the activities of eEF2 or mTORC1 (Fig. S2D and S2E), two established (from previous in vitro studies) downstream effectors of AMPK that are associated with mRNA translation (Ma, 2020). Notably, our previous studies in mice with neuronal suppression of AMPKα isoforms also demonstrate that AMPK deactivation does not correlate with alterations of the activities of eEF2 and mTORC1 (Yang et al., 2020). Future systematic studies are warranted to elucidate distinct roles of AMPK signaling in different systems. It is worth mentioning that the current study examined only whole-cell lysate of hippocampal dissections; future studies will investigate a synaptosomal preparation of hippocampal tissue, thus providing a synapse-specific view of any potential changes to protein phosphorylation. Intriguingly, we found decreased polyribosome count in both β1+/− and β2+/− mice when compared to Cre (Fig. S3D). As polyribosome presence is indicative of active mRNA translation (Afonina and Shirokov, 2018), the decrease seen in polyribosome count in synapses may indicate alterations of “local” de novo protein synthesis in β1+/− and β2+/− mice, as opposed to the overall de novo protein synthesis assessed in the SUnSET experiments (Fig. 2F). Future studies could investigate local protein synthesis at various neuronal sites via techniques such as immunohistochemistry-based SUnSET, BONCAT, or axon-TRAP (Glenn et al., 2017; Goodman and Hornberger, 2013; Kim and Jung, 2015) to better understand the impact on AMPKβ isoform suppression on protein synthesis.

We found that, while the number of PSDs in the β2+/− mice was unchanged compared to Cre mice but significantly increased compared to β1+/− mice, the average PSD area was significantly decreased in β2+/− mice when compared to Cre or β1+/− group (Fig. 3B-C). Additionally, β2+/− mice display impaired dendritic spine density – fewer mature spines and more immature spines – compared to Cre or β1+/− group (Fig. 3F-H). As the primary factors associated with the effect of PSDs and spines on LTP are their area and head volume, respectively (Borczyk et al., 2019), rather than their abundance, this altered synaptic morphology might be the driving force behind the LTP and recognition memory deficits reported here. Surprisingly, β1+/− mice display an increased total spine density when compared to Cre and β2+/− which appears to be chiefly driven by a significant increase in mature spine density (Fig. 3E-G). This suggests a potential compensatory effect of AMPKβ1 suppression on synaptic morphology, which may account for the behavioral and electrophysiological normality of β1+/− mice. Interestingly, we also found an increase in hippocampal PSD-95 levels in β2+/− mice compared to Cre or β1+/− group (Fig. S3A). Multiple studies point to PSD-95 as a major postsynaptic scaffolding protein that increase AMPA receptor density at the synapse, strengthening that synapse (El-Husseini et al., 2000; Nelson et al., 2013; Zhang and Lisman, 2012). However, it has also been suggested that an increase in PSD-95 expression inhibits further synaptic strengthening, thus occluding the generation of LTP in mice overexpressing PSD-95 (Borczyk et al., 2019; El-Husseini et al., 2000; Stewart et al., 2005; Zhang and Lisman, 2012).

A recent study found impaired recognition memory, decreased hippocampal LTP, altered dendritic spine maturity, and impaired PSD morphology in AMPKα2 (but not AMPKα1) conditional knockout (cKO) mice. The study also found no change to hippocampal morphology or overall protein synthesis in hippocampus as assessed by SUnSET, which is consistent with our data in the current study. In contrast, AMPKα2 suppression also resulted in impaired spatial learning & memory, decreased PSD-95 expression, and altered basal synaptic transmission (Yang et al., 2020). These results suggest that, while β2 suppression appears to impact AMPK phosphorylation and therefore its activity (Fig. 2E), AMPKβ signaling may play different roles in learning & memory and the expression of synaptic proteins compared to the AMPKα signaling.

In conclusion, the data presented here support an isoform-specific role of AMPKβ2 in synaptic and cognitive function. As a cellular energy sensor and master kinase, AMPK is involved in numerous fundamental biological processes and drives significant attention from therapeutic perspectives (Dai et al., 2017; Marinangeli et al., 2016; Vingtdeux et al., 2011; Wang et al., 2022). While the results presented herein largely agree with those found in other studies investigating the α subunit (Yang et al., 2020), the contrast in the impact of AMPKβ suppression compared to AMPKα suppression indicates previously unrecognized complexity of the roles of AMPK in regulation of synaptic and cognitive function Elucidation of the distinct roles of AMPK isoforms in the brain will not only contribute to our understanding of neuronal diseases associated with AMPK signaling dysregulation but may also provide implications for AMPK-targeting therapeutics (e.g., long-term safety) for non-neuronal diseases.

Supplementary Material

1
2

Acknowledgments

This work was supported by National Institutes of Health Grants R01 AG05581, R01 AG073823, and RF1 AG082388 (T.M.). The authors have no conflicts of interest to disclose.

Footnotes

CRediT authorship contribution statement

Nathaniel A. Swift: Writing – original draft, Project administration, Investigation, Formal analysis, Conceptualization. Qian Yang: Writing – review & editing, Investigation, Formal analysis. Hannah M. Jester: Writing – review & editing, Investigation, Formal analysis. Xueyan Zhou: Writing – review & editing, Project administration, Methodology, Investigation. Adam Manuel: Investigation. Bruce E. Kemp: Writing – review & editing, Methodology. Gregory R. Steinberg: Writing – review & editing, Methodology. Tao Ma: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors have no conflicts of interest to disclose.

Appendix A. Supplementary Data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.nbd.2024.106664.

Data availability

Data will be made available on request.

References

  1. Afonina ZA, Shirokov VA, 2018. Three-dimensional Organization of Polyribosomes–a Modern Approach. Biochem. Mosc 83, S48–S55. 10.1134/S0006297918140055. [DOI] [PubMed] [Google Scholar]
  2. Aggleton JP, Kyd RJ, Bilkey DK, 2004. When is the perirhinal cortex necessary for the performance of spatial memory tasks? Neurosci. Biobehav. Rev 28, 611–624. 10.1016/j.neubiorev.2004.08.007. [DOI] [PubMed] [Google Scholar]
  3. Akirav I, Maroun M, 2006. Ventromedial prefrontal cortex is obligatory for consolidation and reconsolidation of object recognition memory. Cereb. Cortex N. Y. N 1991 (16), 1759–1765. 10.1093/cercor/bhj114. [DOI] [PubMed] [Google Scholar]
  4. Allen LM, Lesyshyn RA, O’Dell SJ, Allen TA, Fortin NJ, 2020. The hippocampus, prefrontal cortex, and perirhinal cortex are critical to incidental order memory. Behav. Brain Res 379, 112215 10.1016/j.bbr.2019.112215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Angoa-Pérez M, Kane MJ, Briggs DI, Francescutti DM, Kuhn DM, 2013. Marble burying and Nestlet shredding as tests of repetitive, compulsive-like behaviors in mice. J. Vis. Exp. JoVE 50978. 10.3791/50978. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Antunes M, Biala G, 2012. The novel object recognition memory: neurobiology, test procedure, and its modifications. Cogn. Process 13, 93–110. 10.1007/s10339-011-0430-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Barker GRI, Bird F, Alexander V, Warburton EC, 2007. Recognition memory for objects, place, and temporal order: a disconnection analysis of the role of the medial prefrontal cortex and Perirhinal cortex. J. Neurosci 27, 2948–2957. 10.1523/JNEUROSCI.5289-06.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bliss TVP, Collingridge GL, 1993. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361, 31–39. 10.1038/361031a0. [DOI] [PubMed] [Google Scholar]
  9. Borczyk M, Śliwińska MA, Caly A, Bernas T, Radwanska K, 2019. Neuronal plasticity affects correlation between the size of dendritic spine and its postsynaptic density. Sci. Rep 9, 1693. 10.1038/s41598-018-38412-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Broadbent NJ, Squire LR, Clark RE, 2004. Spatial memory, recognition memory, and the hippocampus. Proc. Natl. Acad. Sci 10.1073/pnas.0406344101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Browne GJ, Finn SG, Proud CG, 2004. Stimulation of the AMP-activated protein kinase leads to activation of eukaryotic elongation factor 2 kinase and to its phosphorylation at a novel site, serine 398. J. Biol. Chem 279, 12220–12231. 10.1074/jbc.M309773200. [DOI] [PubMed] [Google Scholar]
  12. Buzsáki G, Moser EI, 2013. Memory, navigation and theta rhythm in the hippocampal-entorhinal system. Nat. Neurosci 16, 130–138. 10.1038/nn.3304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Costa-Mattioli M, Sossin WS, Klann E, Sonenberg N, 2009. Translational control of long-lasting synaptic plasticity and memory. Neuron 61, 10–26. 10.1016/j.neuron.2008.10.055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Coughlan KA, Valentine RJ, Sudit BS, Allen K, Dagon Y, Kahn BB, Ruderman NB, Saha AK, 2016. PKD1 inhibits AMPKα2 through phosphorylation of serine 491 and impairs insulin signaling in skeletal muscle cells. J. Biol. Chem 291, 5664–5675. 10.1074/jbc.M115.696849. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Dagon Y, Hur E, Zheng B, Wellenstein K, Cantley LC, Kahn BB, 2012. p70S6 kinase phosphorylates AMPK on serine 491 to mediate leptin’s effect on food intake. Cell Metab. 16, 104–112. 10.1016/j.cmet.2012.05.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Dai C, Zhang X, Xie D, Tang P, Li C, Zuo Y, Jiang B, Xue C, 2017. Targeting PP2A activates AMPK signaling to inhibit colorectal cancer cells. Oncotarget 8, 95810–95823. 10.18632/oncotarget.21336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. El-Husseini AE-D, Schnell E, Chetkovich DM, Nicoll RA, Bredt DS, 2000. PSD-95 involvement in maturation of excitatory synapses. Science 290, 1364–1368. 10.1126/science.290.5495.1364. [DOI] [PubMed] [Google Scholar]
  18. Figueiredo VC, Markworth JF, Cameron-Smith D, 2017. Considerations on mTOR regulation at serine 2448: implications for muscle metabolism studies. Cell. Mol. Life Sci 74, 2537–2545. 10.1007/s00018-017-2481-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gipson CD, Olive MF, 2017. Structural and functional plasticity of dendritic spines – root or result of behavior? Genes Brain Behav. 16, 101–117. 10.1111/gbb.12324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Glenn WS, Stone SE, Ho SH, Sweredoski MJ, Moradian A, Hess S, Bailey-Serres J, Tirrell DA, 2017. Bioorthogonal noncanonical amino acid tagging (BONCAT) enables time-resolved analysis of protein synthesis in native plant Tissue 1 [OPEN]. Plant Physiol. 173, 1543–1553. 10.1104/pp.16.01762. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Goodman CA, Hornberger TA, 2013. Measuring protein synthesis with SUnSET: a valid alternative to traditional techniques? Exerc. Sport Sci. Rev 41, 107–115. 10.1097/JES.0b013e3182798a95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Green MF, Anderson KA, Means AR, 2011. Characterization of the CaMKK β-AMPK signaling complex. Cell. Signal 23, 2005–2012. 10.1016/j.cellsig.2011.07.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hardie DG, 2007. AMP-activated/SNF1 protein kinases: conserved guardians of cellular energy. Nat. Rev. Mol. Cell Biol 8, 774–785. 10.1038/nrm2249. [DOI] [PubMed] [Google Scholar]
  24. Hardie DG, 2014. AMPK – sensing energy while talking to other signaling pathways. Cell Metab. 20, 939–952. 10.1016/j.cmet.2014.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hardie DG, Ross FA, Hawley SA, 2012. AMPK: a nutrient and energy sensor that maintains energy homeostasis. Nat. Rev. Mol. Cell Biol 13, 251–262. 10.1038/nrm3311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hawley SA, Ross FA, Gowans GJ, Tibarewal P, Leslie NR, Hardie DG, 2014. Phosphorylation by Akt within the ST loop of AMPK-α1 down-regulates its activation in tumour cells. Biochem. J 459, 275–287. 10.1042/BJ20131344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hay N, Sonenberg N, 2004. Upstream and downstream of mTOR. Genes Dev. 18, 1926–1945. 10.1101/gad.1212704. [DOI] [PubMed] [Google Scholar]
  28. Herzig S, Shaw RJ, 2018. AMPK: guardian of metabolism and mitochondrial homeostasis. Nat. Rev. Mol. Cell Biol 19, 121–135. 10.1038/nrm.2017.95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Inoki K, Li Y, Xu T, Guan K-L, 2003. Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling. Genes Dev. 17, 1829–1834. 10.1101/gad.1110003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Joseph BK, Liu H-Y, Francisco J, Pandya D, Donigan M, Gallo-Ebert C, Giordano C, Bata A, Nickels JT, 2015. Inhibition of AMP kinase by the protein phosphatase 2A Heterotrimer, PP2APpp2r2d. J. Biol. Chem 290, 10588–10598. 10.1074/jbc.M114.626259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kim E, Jung H, 2015. Local protein synthesis in neuronal axons: why and how we study. BMB Rep. 48, 139–146. 10.5483/BMBRep.2015.48.3.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Klann E, Dever TE, 2004. Biochemical mechanisms for translational regulation in synaptic plasticity. Nat. Rev. Neurosci 5, 931–942. 10.1038/nrn1557. [DOI] [PubMed] [Google Scholar]
  33. Ma T, 2020. Dysregulation of neuronal protein synthesis in Alzheimer’s disease. In: The Oxford Handbook of Neuronal Protein Synthesis, Oxford Handbooks. Oxford University Press, Oxford, New York. [Google Scholar]
  34. Ma T, Chen Y, Vingtdeux V, Zhao H, Viollet B, Marambaud P, Klann E, 2014. Inhibition of AMP-activated protein kinase signaling alleviates impairments in hippocampal synaptic plasticity induced by amyloid β. J. Neurosci 34, 12230–12238. 10.1523/JNEUROSCI.1694-14.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Mairet-Coello G, Courchet J, Pieraut S, Courchet V, Maximov A, Polleux F, 2013. The CAMKK2-AMPK kinase pathway mediates the Synaptotoxic effects of Aβ oligomers through tau phosphorylation. Neuron 78, 94–108. 10.1016/j.neuron.2013.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Malenka RC, 2003. The long-term potential of LTP. Nat. Rev. Neurosci 4, 923–926. 10.1038/nrn1258. [DOI] [PubMed] [Google Scholar]
  37. Marinangeli C, Didier S, Vingtdeux V, 2016. AMPK in neurodegenerative diseases: implications and therapeutic perspectives. Curr. Drug Targets 17, 890–907. 10.2174/1389450117666160201105645. [DOI] [PubMed] [Google Scholar]
  38. Nelson CD, Kim MJ, Hsin H, Chen Y, Sheng M, 2013. Phosphorylation of Threonine-19 of PSD-95 by GSK-3β is required for PSD-95 mobilization and long-term depression. J. Neurosci 33, 12122–12135. 10.1523/JNEUROSCI.0131-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Oakhill JS, Chen Z-P, Scott JW, Steel R, Castelli LA, Ling N, Macaulay SL, Kemp BE, 2010. β-Subunit myristoylation is the gatekeeper for initiating metabolic stress sensing by AMP-activated protein kinase (AMPK). Proc. Natl. Acad. Sci 107, 19237–19241. 10.1073/pnas.1009705107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. O’Neill HM, Maarbjerg SJ, Crane JD, Jeppesen J, Jørgensen SB, Schertzer JD, Shyroka O, Kiens B, Van Denderen BJ, Tarnopolsky MA, Kemp BE, Richter EA, Steinberg GR, 2011. AMP-activated protein kinase (AMPK) β1β2 muscle null mice reveal an essential role for AMPK in maintaining mitochondrial content and glucose uptake during exercise. Proc. Natl. Acad. Sci. USA 108, 16092–16097. 10.1073/pnas.1105062108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Ross FA, MacKintosh C, Hardie DG, 2016. AMP-activated protein kinase: a cellular energy sensor that comes in 12 flavours. FEBS J. 283, 2987–3001. 10.1111/febs.13698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Santschi LA, Stanton PK, 2003. A paired-pulse facilitation analysis of long-term synaptic depression at excitatory synapses in rat hippocampal CA1 and CA3 regions. Brain Res. 962, 78–91. 10.1016/s0006-8993(02)03846-5. [DOI] [PubMed] [Google Scholar]
  43. Sanz P, Rubio T, Garcia-Gimeno MA, 2013. AMPKbeta subunits: more than just a scaffold in the formation of AMPK complex. FEBS J. 280, 3723–3733. 10.1111/febs.12364. [DOI] [PubMed] [Google Scholar]
  44. Steinberg GR, Carling D, 2019. AMP-activated protein kinase: the current landscape for drug development. Nat. Rev. Drug Discov 18, 527–551. 10.1038/S41573-019-0019-2. [DOI] [PubMed] [Google Scholar]
  45. Steinberg GR, Hardie DG, 2023. New insights into activation and function of the AMPK. Nat. Rev. Mol. Cell Biol 24, 255–272. 10.1038/s41580-022-00547-x. [DOI] [PubMed] [Google Scholar]
  46. Steinberg GR, Kemp BE, 2009. AMPK in health and disease. Physiol. Rev 89, 1025–1078. 10.1152/physrev.00011.2008. [DOI] [PubMed] [Google Scholar]
  47. Stewart MG, Medvedev NI, Popov VI, Schoepfer R, Davies HA, Murphy K, Dallérac GM, Kraev IV, Rodríguez JJ, 2005. Chemically induced long-term potentiation increases the number of perforated and complex postsynaptic densities but does not alter dendritic spine volume in CA1 of adult mouse hippocampal slices. Eur. J. Neurosci 21, 3368–3378. 10.llll/j.1460-9568.2005.04174. [DOI] [PubMed] [Google Scholar]
  48. Turnley AM, Stapleton D, Mann RJ, Witters LA, Kemp BE, Bartlett PF, 1999. Cellular distribution and developmental expression of AMP-activated protein kinase isoforms in mouse central nervous system. J. Neurochem 72, 1707–1716. 10.1046/j.1471-4159.1999.721707.x.. [DOI] [PubMed] [Google Scholar]
  49. Vingtdeux V, Davies P, Dickson DW, Marambaud P, 2011. AMPK is abnormally activated in tangle- and pre-tangle-bearing neurons in Alzheimer’s disease and other tauopathies. Acta Neuropathol. (Berl.) 121, 337–349. 10.1007/S00401-010-0759-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Vorhees CV, Williams MT, 2006. Morris water maze: procedures for assessing spatial and related forms of learning and memory. Nat. Protoc 1, 848–858. 10.1038/nprot.2006.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Vyas Y, Montgomery JM, 2016. The role of postsynaptic density proteins in neural degeneration and regeneration. Neural Regen. Res 11, 906–907. 10.4103/1673-5374.184481. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wang X, Zimmermann HR, Ma T, 2019. Therapeutic potential of AMP-activated protein kinase in Alzheimer’s disease. J. Alzheimers Dis 68, 33–38. 10.3233/JAD-181043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wang D, Yang L, Liu Y, 2022. Targeting AMPK signaling in the liver: implications for obesity and type 2 diabetes mellitus. Curr. Drug Targets 23, 1057–1071. 10.2174/1389450123666220429082702. [DOI] [PubMed] [Google Scholar]
  54. Yan Y, Zhou XE, Xu HE, Melcher K, 2018. Structure and physiological regulation of AMPK. Int. J. Mol. Sci 19 10.3390/ijms19113534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Yang W, Zhou X, Zimmermann HR, Ma T, 2020. Brain-specific suppression of AMPKα2 isoform impairs cognition and hippocampal LTP by PERK-mediated eIF2α phosphorylation. Mol. Psychiatry 10.1038/s41380-020-0739-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zhang P, Lisman JE, 2012. Activity-dependent regulation of synaptic strength by PSD-95 in CA1 neurons. J. Neurophysiol 107, 1058–1066. 10.1152/jn.00526.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Zhou G, Myers R, Li Y, Chen Y, Shen X, Fenyk-Melody J, Wu M, Ventre J, Doebber T, Fujii N, Musi N, Hirshman MF, Goodyear LJ, Moller DE, 2001. Role of AMP-activated protein kinase in mechanism of metformin action. J. Clin. Invest 108, 1167–1174. 10.1172/JCI13505. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Zimmermann HR, Yang W, Kasica NP, Zhou X, Wang X, Beckelman BC, Lee J, Furdui CM, Keene CD, Ma T, 2020. Brain-specific repression of AMPKα1 alleviates pathophysiology in Alzheimer’s model mice. J. Clin. Invest 130, 3511–3527. 10.1172/JCI133982. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2031281-supplement-1.docx (1.7MB, docx)
2
NIHMS2031281-supplement-2.docx (31KB, docx)

Data Availability Statement

Data will be made available on request.

RESOURCES