Impaired Spatial Learning is Associated with Disrupted Integrity of the White Matter in Akt3 Knockout Mice

He Wang,Bao-Feng Zhang,Ting-Ting Zhang,Long Wang,Xiao-Yan Zou,Yun Xu,Ling Chen,Gui-Quan Chen
DOI: https://doi.org/10.1111/cns.12647
2016-01-01
CNS Neuroscience & Therapeutics
Abstract:The PI3K/Akt signaling pathway is important for various biological processes [1], and impaired PI3K/Akt signaling has been observed in neurodevelopmental and neurodegenerative diseases [2,3]. In mammals, whereas Akt exists in three isoforms including Akt1, Akt2, and Akt3 [1], Akt3 is the dominant isoform to be expressed in the brain [4]. Recently, abundant evidence has shown that mutations on Akt3 (loss-of-function) but not Akt1 or Akt2 are associated with microcephaly, a disease displaying intellectual disability and decreased size of the brain [3]. However, it remains unknown whether Akt isoforms differentially regulate brain development and affect learning ability. In this study, two different lines of Akt knockout (KO) mice, Akt3 / [5] and Akt1 / [6], were used for behavioral and morphological analyses. To assess spatial learning and memory, Akt3 / mice and agematched littermate wild-type (WT) animals were examined by a Morris watermaze task. First, a visible cuetask was performed. Comparable performance was observed between Akt3 / mice and WTs (data not shown), suggestive of normal basic learning in Akt3 / mice. Second, a hidden platform task of the watermaze [7] was conducted. The mice were trained to learn a submerged platform, and the training protocol consisted of two blocks per day and two trials per block over 5 days. Analysis of variance (ANOVA) revealed a highly significant main genotype effect on the latency to escape (Fig. 1A: F = 13.5, df1/18, P < 0.005) during the training period, suggesting that spatial learning in Akt3 / mice was impaired. The ANOVA also showed a significant block effect (F = 18.4, df4.2/75.1, P < 0.001), suggesting that the training protocol worked well so that all the mice had improved the performance significantly after the training. Third, spatial reference memory was accessed by a probe test 24 h after the hidden platform training. We found that WT and Akt3 / mice displayed comparable performance on the time spent in the target quadrant (Fig. 1B: P > 0.2, two-tailed Student’s t-test). The target quadrant occupancy for WT and Akt3 / mice was significantly better than the chance performance (25%) (Fig. 1B: Ps < 0.05), suggesting that Akt3 / mice had formed solid spatial reference memory after the 5-day training. Whereas Akt3 is the most abundant Akt isoform to be expressed in mammals, Akt1 is the second [4,5]. To study whether Akt1 is also required for spatial learning, we trained Akt1 / mice in the watermaze using the same protocol for Akt3 / mice. The learning curve was plotted in Fig. 1C. ANOVA did not reveal any significant genotype effect on the latency to escape (F = 0.3, df1/19, P > 0.4), suggesting that learning ability of Akt1 / mice was not impaired. The ANOVA showed a significant block effect (F = 18.4, df3.8/ 73.1, P < 0.001), indicating effective learning by all the mice. Moreover, WT and Akt1 / animals exhibited equivalent level of the time spent in the target quadrant during the probe test (Fig. 1D), suggesting that spatial reference memory of Akt1 / mice was not affected. In addition, results from a cuetask showed no difference on the performance for Akt1 / and WT animals (data not shown). To understand molecular mechanism by which Akt3 regulates learning ability, we examined brain morphology of mutant mice. First, using an antibody against MBP (myelin basic protein), we conducted immunohistochemistry (IHC) [8] on MBP and found
What problem does this paper attempt to address?