Metformin treatment alters memory function in a mouse model of Alzheimer's disease

Abstract

Metabolic dysfunction exacerbates Alzheimer's disease (AD) incidence and progression. Here we report that activation of the AMPK pathway, a common target in the management of diabetes, results in gender-divergent cognitive effects in a murine model of the disease. Specifically, our results show that activation of AMPK increases memory dysfunction in males but is protective in females, suggesting that gender considerations may constitute an important factor in medical intervention of diabetes as well as AD.

Keywords: AMPK; Alzheimer's disease; metformin; type 2 diabetes mellitus.

Fig. 1.

AMPK activation impairs spatial learning and memory retention in male AβPP mice. 12- to 14-month old mice received four trials per day for 10 days of training to find a hidden platform in the Morris water maze followed by a single probe trial on the final day of testing in which the platform was removed. a, b) Metformin study: a) AβPP mice in both the control group and the metformin group require more time to find the platform than their WT counterparts. Metformin-treated AβPP mice perform significantly worse than control-treated AβPP mice. b) On the probe trial, AβPP mice crossed the location where the platform had previously been located fewer times than WT mice on average, but the difference did not quite meet the criteria for statistical significance (p = 0.0696). Within the metformin-treated group, AβPP mice demonstrated significantly impaired spatial memory retention (p = 0.005). c,d) AMPK Study: c) In the presence or absence of AMPK, AβPP mice perform worse than WT mice. AβPP/AMPK mice were not statistically distinguishable from AβPP/WT mice although a slight but consistent trend toward impaired performance with AMPK expression was apparent from day four through the end of training. d) Although on average AβPP/WT mice crossed the platform location fewer times than WT/WT mice, the difference was not statistically significant (p = 0.333). With AMPK expression, AβPP mice demonstrated significantly impaired spatial memory retention. For the hidden platform (a, c): data represent the average latencies for four trials for the mice in each group to locate the escape platform; error bars represent SEM; ns, not significant; asterisks in the reference legend of each graph represent significance of the entire curve by repeated measures ANOVA: *p < 0.05, **p < 0.01, ***p < 0.001; symbols on the graph represent significance of individual points on the curve: ^#WT control versuss AβPP control; ^ϕAβPP control vrsus AβPP metformin; ^#,ϕp < 0.05, ^##,ϕϕp < 0.01, ^ϕϕϕp < 0.001. For the probe trial (b, d): data represent the average number of times the mice in each group crossed the area where the platform had previously been located; error bars represent SEM; ns, not significant; **p < 0.01. Sample sizes are identical to the training phase.

Fig. 2.

AMPK activation improves learning and memory function in female AβPP mice. 12- to 14-month old mice received four trials per day for 10 days of training to find a hidden platform in the Morris water maze followed by a single probe trial on the final day of testing in which the platform was removed. a, b) Metformin study: a) Training phase: female AβPP mice had longer latencies to find the escape platform than WT mice in the control group. Within the metformin-treated groups, AβPP mice took longer to find the platform in the early days of testing; however, metformin-treated AβPP mice were significantly improved over control-treated AβPP mice. b) Probe trial: within the control treatment group, AβPP mice crossed the platform location fewer times than WT mice. AβPP mice in the metformin treatment group, however, were indistinguishable from WT mice. c, d) AMPK study: c) Training phase: female AβPP/WT mice had longer latencies to find the escape platform than WT/WT mice in the later part of training. AβPP/AMPK mice were significantly improved over AβPP/WT mice and were in fact statistically indistinguishable from WT/AMPK mice. d) Probe trial: by the measure of platform crossings, AβPP/WT mice performed poorly, but were not statistically different from WT/WT mice (by time spent in the target quadrant, these groups were distinguishable, see Supplementary Figure 1). AβPP/AMPK mice demonstrated no memory impairment. For the hidden platform (a, c): data represent the average latencies for four trials for the mice in each group to locate the escape platform; error bars represent SEM; ns, not significant; asterisks in the reference legend of each graph represent significance of the entire curve by repeated measures ANOVA: *p < 0.05, **p < 0.01, ***p < 0.001; symbols on the graph represent significance of individual points on the curve: ^#WT control versus AβPP control; ^ϕAβPP control versus AβPP metformin; ^#,ϕp < 0.05, ^##,ϕϕp < 0.01, ^###p < 0.001. For the probe trial (b, d): data represent the average number of times the mice in each group crossed the area where the platform had previously been located; error bars represent SEM; ns, not significant; *p < 0.05. Sample sizes are identical to the training phase.