AMPK agonist AICAR improves cognition and motor coordination in young and aged mice

Behavior in AMPK-DN mice

To evaluate whether muscle AMPK may play a role in the enhanced performance we tested young mice overexpressing an inactive AMPK α2 catalytic subunit driven by the muscle specific MCK promoter (AMPK-DN, verified by PCR, Fig. 3A). Overexpression of the inactive AMPK α2 subunit eliminates essentially all AMPK activity in skeletal muscle (Zwetsloot et al. 2008). Female wild type (wt) and transgenic (tg) littermates (4- to 6-wk old) were injected with saline (n = 12 wt, n = 10 tg) or AICAR (n = 10 wt, n = 7 tg) for 3 d and trained 2 wk later in the water maze with four trials per day over 5 d (Fig. 3B). Two-way analysis of variance with repeated measures (Genotype × Treatment × Days) showed that there was no interaction between the groups in acquisition of the task (F_(1,35) = 0.11, P > 0.74) and no main effects of genotype or treatment (P > 0.05). Upon evaluation of retention of platform location in the probe trial 4 h after the last training trial, drug-treated wt mice (ACR–) preferred the target quadrant, as evidenced by a significant one-way ANOVA over quadrants (F_(3,36) = 4.19, P < 0.01), followed by specific post-hoc comparisons between the time spent in the target quadrant and time in each of the other quadrants (P < 0.03). In addition, a trend toward significance was observed for time spent in the target quadrant in the ACR– versus the ACR+ group (t₍₁₅₎ = 2.01, P = 0.06), supporting the notion that the compound was effective in enhancing memory retention in wt but not in tg mice (Fig. 3C).

Open in a separate window

Figure 3.

Effect of AICAR on cognition in AMPK-DN mice. (A) PCR analysis of isolated tail DNA was performed to identify AMPK-DN (tg) or wt mice. (B) Water maze performance in AMPK-DN mice (+) or wild type mice (−) treated with Saline (SAL) or AICAR (ACR, 500 mg/kg) for 3 d. Mice were trained for 5 d with four trials per day in the Morris water maze, 2 wk after injections. There was no difference between the groups in acquisition of the task. (C) In the probe trial 4 h after the last training trial drug-treated wt mice that were injected with ACR (ACR−) showed a significant preference for the platform quadrant compared to the other quadrants (*P < 0.03). The mice treated with saline (wt, SAL−, and tg, SAL+) and the tg mice treated with AICAR (ACR+) showed no retention of platform location. In addition, there was a trend toward a significant difference in time spent in the target quadrant between ACR− and ACR+ mice (^#P = 0.06), further supporting memory retention enhancement by AICAR in wt mice. Error bars indicate SEM.

Mice treated with saline (n = 8 wt, n = 6 tg) or AICAR (n = 7 wt, n = 7 tg) for 3 d were also tested in the rotarod and open field. Distance traveled in the open field over 30 min did not differ between wt and tg saline (4495 ± 251 cm, wt; 5146 ± 588 cm, tg) and AICAR (4622 ± 480 cm, wt; 4894 ± 352 cm, tg) treated mice (F_(1,24) = 0.57, P > 0.64). In addition, neither the average latency to the first fall over three trials in saline (246 ± 22 sec, wt; 251 ± 22 sec, tg) and AICAR (243 ± 21 sec, wt; 235 ± 20 sec, tg) (F_(1,24) = 0.09, P > 0.76) treated mice, nor the total number of falls (F_(1,24) = 0.03, P > 0.87), differed between the groups.

Microarray analysis

Microarray analysis of hippocampal and gastrocnemius muscle tissue derived from aged mice treated with AICAR for 3 d (ACR3), 7 d (ACR7), or 14 d (ACR14) or saline (CON) was consistent with the behavioral findings. The most pronounced changes in gene expression were found in the ACR7 and ACR14 groups in both tissues. Principal component analysis (DIANE6, http://www.grc.nia.nih.gov/branches/rrb/dna/diane_software.pdf) showed a parallel dose–response pattern in both gastrocnemius muscle and hippocampus. Further analysis suggested bidirectional effects between the muscle and the brain. Muscle sample gene analysis revealed effects on brain-related genes, such as up-regulation of neuronal fate commitment. In addition, muscle-related genes, important for formation of striated muscle thin filaments, contraction, cholesterol homeostasis, and fatty acid biosynthesis, exhibited elevated expression. Furthermore, components of mitochondrion and of mitochondrial inner membrane, mitochondrial activity of NADH dehydrogenase, complex IV, and complex I were up-regulated in muscle samples. In the hippocampus up-regulation of components of the mitochondrion and endoplasmic reticulum, as well as genes involved in metabolic processes, glycolysis, oxidoreductase activity, and lipid biosynthesis were observed. However, the majority of up-regulated genes in the hippocampus were related to neuronal development (Fig. 4A–D).

Open in a separate window

Figure 4.

Microarray analysis of hippocampus and muscle tissue derived from aged mice treated with saline (CON) or AICAR for 3 d (ACR3), 7 d (ACR7), or 14 d (ACR14). (A) Gene set analysis of gastrocnemius tissue. Increased expression of mitochondrial inner membrane and metabolism related genes was observed. Each section represents the number of up-regulated genes of the respective function. (B) The heat-map illustrates the expression of the selected signature genes derived from microarray analysis. Up-regulated genes are colored in red and down-regulated genes are colored in green. (C) Top up-regulated and down-regulated genes by ACR7 and ACR14 in gastrocnemius. The tables list the most highly up-regulated genes related to energy metabolism and neuronal development; up-regulation is expressed as fold change compared to control and as z-ratio for each gene. (D) Top up-regulated and down-regulated genes by ACR7 and ACR14 in hippocampus. The tables list the most highly up-regulated genes related to energy metabolism and neuronal development; up-regulation is expressed as fold change compared to control and as z-ratio for each gene. (E) Relative gene expression level and validation. qPCR analysis of expression levels of four genes in ACR14 samples compared to CON, in gastrocnemius. (F) Relative gene expression level and validation. qPCR analysis of expression levels of four genes in ACR14 samples compared to CON, in hippocampus. N = 4 per group. Error bars indicate SEM.

In both tissues for ACR7 and ACR14 samples, the majority of the top 30 up-regulated genes are the same, albeit different for muscle and hippocampus. Specific analysis of the top 10 up-regulated genes in the muscle (Fig. 4C) revealed that eight genes were identically up-regulated in the ACR7 and ACR14 groups, in comparison to control. Increased gene expression relevant to energy metabolism in muscles was observed, for example, for NDUFA5, NDUFB9, NDUFB4 (codes for subunits of complex I of the respiratory chain, which transfers electrons from NADH to ubiquinone [Walker 1992], 3.27-, 3.07-, and 2.82-fold, respectively), SDHD (Complex II, 2.53-fold), DLD, ATP5A1 (encodes a subunit of mitochondrial ATP synthase, that catalyzes ATP synthesis [Pedersen and Amzel 1993], 2.65- and 3.11-fold, respectively), and USMG5 (integral to the mitochondrial membrane, 3.01-fold increase).

Analysis of the top down-regulated genes in muscle (Fig. 4C) revealed genes involved in fatty acid metabolism such as LPIN2 (Donkor et al. 2007), −1.53-fold, as well as genes whose inhibition increases insulin sensitivity, such as RCAN2 (Sun et al. 2011), −1.32-fold. Indeed, inactivation of the RCAN2 gene in mice also ameliorates age- and diet-induced obesity by causing a reduction in food intake (Sun et al. 2011). Down-regulation was also found for ERRFI1 (reduces cell proliferation and muscle growth by inhibiting EGFR protein [Segatto et al. 2011], −1.66-fold), SPARC (authophagy and senescence [Nakamura et al. 2012], −1.21-fold), and MYH7 (encodes for MHC β, specific for slow fibers [van Rooij et al. 2009], −1.27-fold).

In the hippocampus, even if diffusely up-regulated as a gene class, mitochondria-related genes were a minority in the top 30 enriched genes (ALDOA [2.15-fold], MDH2 [1.79-fold], COX4 [1.32-fold]), whereas distinct elevations were present for neuronal development related genes (Fig. 4D): ATXN10 (neuronal development, 2.15-fold [März et al. 2004]), YWHAG (synaptic plasticity, neuronal differentiation, 2.05-fold [Cabeza-Arvelaiz et al. 2011]), MAPK3 (proliferation and differentiation, 1.25-fold), RTN4 or NOGO, a ligand for NGR1 (axonal regulation/growth inhibition, neuroprotection [Teng and Tang 2008, 2013], and dendritic spine stabilization [Mironova and Giger 2013], 2.07-fold), GOLGA2 (axonogenesis, 1.94-fold), and FGF13 (encodes for a protein member of the fibroblast growth factor [FGF] family, 2.03-fold). FGFs possess broad mitogenic and cell survival activities, and are involved in a variety of biological processes, including embryonic development cell growth (Wu et al. 2012). Moreover, we found enrichment of genes relevant to trophic factors, such as PTGDS and PTPRE (2.78- and 2.42-fold, respectively). The protein encoded by PTGDS is a glutathione-independent prostaglandin D synthase that catalyzes the conversion of prostaglandin H2 to prostaglandin D2 (Nagata et al. 1991), which functions as a neuromodulator as well as a trophic factor in the central nervous system (Nagata et al. 1991). In addition, PTPRE encodes for a member of the protein tyrosine phosphatase (PTP) family. PTPs are known to be signaling molecules that regulate a variety of cellular processes, including cell growth and differentiation (Avraham et al. 2000).

Furthermore, among the top down-regulated genes in the hippocampus, the microarray analysis revealed genes whose inhibition results in an increase in neurogenesis and neuronal development: GRIT (regulates dendritic spine morphology and strength [Nakamura et al. 2002], −1.48-fold), MEF2C (skeletal muscle myogenesis and neurogenesis, by negatively regulating transcription factor activity [McDermott et al. 1993; Ornatsky and McDermott 1996], −1.25-fold), TENM4 (neurite outgrowth, interacts with the DNA-binding transcriptional repressors, and regulates transcription factor activity, −1.32-fold). Moreover, down-regulation of Pde1a, a calmodulin-dependent phosphodiesterase with a role in learning and memory, was observed (−1.27-fold). Inhibition of Pdela is a target for treatment of cognitive disorders (Fig. 4D; Xu et al. 2011).

Behavioral testing

Mice were trained in the Morris water maze (5 ft diameter) to find a platform hidden 5 mm below the surface of water made opaque with white nontoxic paint (Morris et al. 1982). It should be noted that a 5-ft-diameter pool is likely more challenging than the 4-ft-diameter pool, typically used with mice (Terry 2009), and may be preferable for memory enhancement experiments. Starting points were changed for each of the four daily trials. A trial lasted either until the mouse had found the platform or for a maximum of 60 sec. Mice rested on the platform for 10 sec after each trial. Upon completion of training, when the average latency to the platform for at least one of the groups was 20–30 sec, a probe trial was performed. Specifically, the platform was removed for a 60-sec probe trial 4 h after the last training trial. Platform latency, swim path, and speed were recorded (Anymaze, Stoelting Co.).

In the open field test, mice were placed in the center of the open field arena and allowed to freely move while being tracked by an automated tracking system (Activity Monitor Version 4, MED Associates). Aged mice were also tested twice, 4 d apart, on a rotarod (Med-Associates), at 20 rpm for 300 sec. For the aged mice the average latency to fall for the two tests was calculated. The total number of falls within 5 min was not measured in the old mice due to their advanced age and possible frailty. Young mice were tested once, in accelerating paradigm of 2.5–25 rpm in 300 sec. The latency to the first fall in each 5-min period was recorded three consecutive times.

Tissue collection

Animals were deeply anesthetized with isofluorane and perfused transcardially with 0.9% NaCl solution. Hippocampus and gastrocnemius muscle were quickly removed, and frozen on dry ice. Tissue was stored at −80°C for microarray analysis.

Microarray analysis

Microarray analysis was performed on hippocampal and muscle tissue derived from four aged mice per group (CON, ACR3, ACR7, ACR14). RNA isolation was carried out using the Qiagen RNeasy Mini Kit for animal tissues (Qiagen, Inc.) following the manufacturer's instructions and hybridized to MouseRef-8 v1 Expression beadchips (Illumina) following protocols listed on the Gene Expression and Genomics Unit of the NIA (http://www.grc.nia.nih.gov/branches/rrb/dna/index/protocols.htm). Microarray fluorescent signals were extracted using an Illumina BeadArray 500GX reader. The signals on each sample are normalized by log z-transformation to obtained z-scores and tests for distributions as previously described (Cheadle et al. 2003). Correlation analysis, sample clustering analysis, and principal component analysis include all of the probes and were performed to identify/exclude any possible outliners. The resulting dataset was next analyzed with JMP6, a spreadsheet-based microarray analysis program. After normalization by log z-transformation, the statistically significant gene list for each comparison condition was selected upon the conditions of (1) the probe signal detection P-value ≤0.02, (2) one way ANOVA over sample groups was P < 0.05, (3) pairwise z-test P < 0.05, (4) false discovery rates are <0.30, and (5) gene expression changes measured by z-ratio were not less than 1.5 the mean of their absolute values. Further clustering/correlation analyses were done based on these selection criteria. Gene set enrichment analysis used gene expression change values (z-ratio) for the average of all of the genes on the microarray. Parametric analysis of gene set enrichment (PAGE) was used (De et al. 2010) for gene set analysis. Gene Sets include the MSIG database (http://www.broadinstitute.org/gsea/msigdb/collection_details.jsp#C2, Gene Ontology Database [http://www.geneontology.org/]); GAD human disease and mouse phenotype gene sets (De et al. 2010; Zhang et al. 2010) were used to explore functional level changes. The data discussed in this article have been deposited in NCBI's Gene Expression Omnibus (Edgar et al. 2002) and are accessible through GEO Series accession number {"type":"entrez-geo","attrs":{"text":"GSE50873","term_id":"50873"}}GSE50873 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc={"type":"entrez-geo","attrs":{"text":"GSE50873","term_id":"50873"}}GSE50873).

Real-time PCR

Real-time PCR to validate the Illumina microarray data was performed on the mRNA samples previously described (n = 4 per group). Reverse transcription was carried out on 1 μg of RNA using qScript cDNA Supermix (Quanta Bioscience) according to manufacturer protocol; primers for the selected genes ATP5A1, SDHD, COX4, MAPK3, and HSP90 were purchased from Integrated DNA Technologies using IDT online primer designer. The specific sequence of each of the primers was as follows:

• ATP5A1: rev. 5′-GGGCTCCAGTTTGTCAAGATA-3′, fwd. 5′-CGTCTGACCGAGTTGCTAAA-3′;
• SDHD: rev. 5′-TCACGAATGGTCGAACCTAAC-3′, fwd. 5′-GGTGGGCAGAATGTCTTCTAA-3′;
• MAPK3: rev. 5′-CTCTTAGGGTTCTTTGACAGTAGG-3′, fwd. 5′-CTACACCAACCTCTCGTACATC-3′;
• COX4: rev. 5′-GATCGAAAGTATGAGGGATGGG-3′, fwd. 5′-TGAATGGAAGACAGTTGTGGG-3′; and
• endogenous control, HSP90: rev. 5′-CCTCTTTCTCACCTTTCTCTTCC-3′, fwd. 5′-ATTCGCAGTTCATAGGCTATCC-3′.

All primers have been tested by melting curve analysis for specificity for one single amplicon product. qPCR was performed with PerfeCTa SYBR Green FastMix (Quanta Bioscience) according to manufacturer protocol; reactions were carried out using a Chromo4 qPCR detection system (Bio-Rad Laboratories). Results were analyzed using Pfaffl method (Pfaffl 2001).

This research was supported by the Intramural Research Program of the NIH, National Institute on Aging (NIA). We thank Chunyan Yuan for assistance with initial experiments and William H. Wood III for assistance with the microarray analysis.