Psychiatric illnesses such as depression, bipolar disorders, anxiety and schizophrenia represent an important public health problem causing not only personal suffering but also a tremendous economic burden (Wittchen et al., 2011; Vigo et al., 2016). Over the recent years, many studies have highlighted the role of genetic variations, abnormal neural development and dysregulation of neurotransmitters to explain the development of these disorders (Renoir, 2014). Even if these discoveries have led to a plethora of pharmacological treatments, their efficacy is still limited (Gould and Einat, 2007; Duman and Voleti, 2012). Therefore, identification of key cell signaling molecules and pathways are critical to uncover novel therapeutic targets and successful clinical interventions.

A family of serine/threonine kinases, the protein kinase B (PKB/Akt), has received extensive consideration in recent years for its possible involvement in psychiatric conditions. For instance, Akt activity is known to be modified by antidepressants, antipsychotics and mood stabilizers (Beaulieu et al., 2009). The Akt protein is constituted by three isoforms termed Akt1/PKBα, Akt2/PKBβ and Akt3/PKBγ. Despite the fact that they are showing strong homologies, these isoforms are encoded by distinct genes (Diez et al., 2012). The most abundant isoforms is Akt1. Akt1 is ubiquitously expressed whereas Akt2 is primarily expressed in insulin-responsive tissues and Akt3 is very much expressed in brain and testes (Hers et al., 2011). Understanding brain Akt functions has been gained by the generation of transgenic and knockout (KO) mouse models (Yang et al., 2004; Dummler et al., 2006). For instance, a partial lethality is observed in mice lacking Akt1 and the surviving mice are reduced in size at birth and remain small throughout life (Cho et al., 2001; Yang et al., 2003). Akt2 is particularly important for glucose metabolism (Cho et al., 2001; Garofalo et al., 2003) and a role in photoreceptor survival is observed (Li et al., 2007). Recently, signs of anxiety and depressive-like behaviors have been reported in mice lacking Akt2 (Leibrock et al., 2013). The role of Akt3 is less clear even though it appears to be the most highly expressed isoform in the brain. Mice lacking Akt3 exhibit small brains (Easton et al., 2005; Poduri et al., 2012). It is noteworthy that a multi-stage schizophrenia genome-wide association study has identified Akt3 as a potential contributor to schizophrenia, which underscore the importance of investigating the role of Akt3 in the brain (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014).

The goal of the present study is to define the behavioral, electrophysiological and biochemical consequences of Akt3 deletion in mice. Our findings reveal that Akt3 deletion in mice does not affect motor, cognitive and synaptic functions, but is associated with behavioral abnormalities reminiscent of psychiatric symptoms. Notably, the depressive and anxiety-like behaviors are reversed by a chronic lithium treatment. Our data provide evidence that Akt3 signaling may play a role in the pathogenesis of psychiatric conditions.

The specific function of Akt3 isoform in the brain is poorly understood. Our study on Akt3 KO mice reports that Akt3 deletion is associated with normal spatial learning and memory as revealed by the Morris water maze and novel object recognition tests. In addition, our results in the rotarod, pole, wire and stepping tests provide evidence that the Akt3 KO mice do not suffer from a generalized deficit in locomotion, motor abilities and activity levels. In fact, Akt3 deletion rather evokes changes in mouse behavior reflecting psychiatric manifestations reminiscent of schizophrenia, anxiety and depression. For instance, Akt3 KO mice exhibited reduced prepulse inhibition and social novelty, two phenotypes often seen in animal models of schizophrenia (Nestler and Hyman, 2010). In particular, the prepulse inhibition is disrupted in patients suffering from schizophrenia (Braff et al., 2001) and asociality is considered to be one of the most problematic symptoms of schizophrenia (Murphy et al., 2006). Our results in two antidepressant-sensitive tests, the force swim and TSTs, and three anxiogenic tests, the open field, the elevated plus maze and the dark-light tests, suggest that deletion of Akt3 increases susceptibility to develop symptoms related to depression and anxiety. Altogether, these comprehensive behavioral manifestations propose that Akt3 signaling contributes to the normal functioning of the neural circuitry intervening in the symptoms associated with schizophrenia, depression and anxiety.

Systematic behavioral studies have never been performed on the Akt3 KO mice before. However, it is established that these mice have a reduction in brain weight resulting from a decline in both cell size and number, but preserve normal glucose homeostasis and body weights (Easton et al., 2005; Tschopp et al., 2005). Based on these observations, it is believed that Akt3 could be an important regulator of postnatal brain development (Tschopp et al., 2005). Interestingly, mood disorders such as schizophrenia and major depression showed common and distinct patterns of volumetric alteration relative to healthy individuals in a number of brain areas (Wright et al., 2000; van Tol et al., 2010; Balevich et al., 2015; Wise et al., 2016). The possibility that the observed psychiatric-like behaviors may originate from the reduced brain volume induced by Akt3 deletion is interesting, but highly speculative. A small number of clinical studies suggest that a copy number variation at the Akt3 locus could be associated with abnormal brain size. For instance, individuals with Akt3 deletions show microcephaly (Boland et al., 2007), while duplications or gain-of-function Akt3 mutations are associated with megalencephaly (Rivière et al., 2012; Wang et al., 2013; Nellist et al., 2015) and focal cortical dysplasia (Chung et al., 2014). Whether Akt3 is the only responsible gene for these witnessed clinical phenotypes is unclear. Further investigations are definitively required in order to untangle the putative brain consequence of impaired Akt3 activity. It is becoming apparent, however, that Akt3 plays a critical role in brain development and psychiatric manifestations in mice.

While the link between deletion of Akt3 isoform and psychiatric behaviors has never been established before, impaired Akt activity was previously suspected in psychiatric conditions. For example, post mortem tissue samples from patients with schizophrenia showed significant reductions of phosphorylated Akt levels in hilar neurons of the dentate gyrus and the neurogenic zone of the hippocampus (Balu et al., 2012). Akt activity is decreased in some brain regions (occipital cortex and ventral prefrontal cortex) of depressed patients (Hsiung et al., 2003; Karege et al., 2007). Phosphorylated Akt levels in the ventral tegmental area are markedly decreased in mice subjected to a depression model induced by stress, an effect reversed by fluoxetine (Krishnan et al., 2008). The study of normal physiological functions of each Akt isoform in the adult brain has been made possible recently by the generation of KO mouse models. Basic behavioral profiling of Akt1 KO mice showed that there are no overall motor, anxiety and widespread cognitive behavioral deficits, although displaying abnormal working memory involving the hippocampus (Lai et al., 2006; Balu et al., 2012). On the other hand, Akt2-deficient mice exhibit normal growth, but develop a diabetes-like syndrome with elevated fasting plasma glucose levels, elevated hepatic glucose output, and peripheral insulin resistance (Cho et al., 2001; Garofalo et al., 2003). Lately, Akt2 deletion has been associated with anxiety and depressive-like behaviors (Leibrock et al., 2013). Our data in Akt3 KO mice, showing preserved normal glucose homeostasis and body weights (Easton et al., 2005; Tschopp et al., 2005), demonstrate an endophenotype reminiscent of schizophrenia, depression and anxiety. A more methodical behavioral study using all three KO mice is definitively required in order to further understand the specific role of each isoform. It is still remarkable to note that the three Akt isoforms do not exhibit identical functions. Another evidence for this contention is the fact that these KO mice studies demonstrate no compensatory upregulation of the remaining isoforms (assessed at the mRNA or protein levels). Correspondingly, our study reveals no effect of Akt3 deletion on the protein levels of Akt1 and Akt2 in the anterior cortex, striatum, hippocampus and cerebellum of Akt3 KO mice.

Our behavioral and electrophysiological observations on Akt3 KO mice suggest that Akt3 signaling may not play such an important role in hippocampal related functions. To begin with, normal spatial navigation and recognition memory are revealed in these mice using two independent tests, the Morris water maze and the object recognition. Many studies have highlighted the implication of the hippocampus as a part of the brain responsible for spatial navigation and recognition memory (Broadbent et al., 2004; Vann and Albasser, 2011). It is therefore interesting that electrophysiological measurements achieved in the hippocampal area CA1 of Akt3 KO mice are comparable to WT mice; which correlate nicely to the behavioral data. Since Akt3 KO mice exhibit normal performances at the Morris water maze, in the novel object recognition test and in four motor abilities tests, we believe that other brain functions, like motricity and motivation, are unaltered in Akt3 KO mice. This additional evidence further underscores the existence of a brain regional specificity of Akt3 function. Levels of phosphorylated GSK3, a known direct Akt molecular target, are robustly decreased in the hippocampus of Akt3 KO mice. Lower levels of GSK3 phosphorylation is well known to lead to enhanced GSK3 kinase activity and signaling (Freland and Beaulieu, 2012). A previous study performed in knockin mice with serine-to-alanine mutations of GSK3α/β, to block inhibitory serine-phosphorylation, confirms our results (Polter et al., 2010). These GSK3 knockin mice display increased susceptibility to stress-induced depressive-like behaviors, whereas normal electrophysiological recording from the hippocampal area CA1 is shown. Altogether, these results suggest that while Akt3 deletion is associated with an endophenotype reminiscent of schizophrenia, anxiety and depression, the Akt3/GSK3 signaling has no influence in hippocampal area CA1 functions.

Akt3 deletion decreases GSK3α phosphorylation at serine 21 and GSK3β at serine 9 in every brain structure analyzed in Akt3 KO mice, suggesting that the observed phenotype could be a consequence of altered GSK3α/β phosphorylation. Since Akt can modulate over a hundred of known substrates (Manning and Cantley, 2007), we cannot exclude that the observed phenotype in Akt3 KO mice may be associated with other signaling pathways. Nevertheless, impaired GSK3α/β activity has been previously recognized to play a role in psychiatric conditions (Li and Jope, 2010). For instance, previous publications have shown that anxiety and depressive-like behaviors are associated with decreased brain levels of phosphorylated GSK3α/β (Polter et al., 2010; Beurel et al., 2011). Notably, in our experiments, the chronic lithium treatment known to amplified GSK3α phosphorylation at serine 21 and GSK3β at serine 9 (Mines et al., 2010; Liu et al., 2011), is able to reverse the behavioral signs of anxiety and depression observed in Akt3 KO mice. Besides, we observe that lithium is returning back to control value the decreased phosphorylated levels of GSK3α/β in the Akt3 KO mice. Since lithium plasma concentrations measured were within the therapeutic range (0.4–1.2 mEq/L) and no relevant adverse effects were observed in mice fed with lithium (Malhi and Tanious, 2011), we believe the administered doses of lithium are clinically relevant in our experimentations. The mechanism by which lithium amplified GSK3α/β phosphorylation in Akt3 KO mice was unfortunately not assessed by our study. Lithium is a competitive inhibitor of magnesium and can directly inhibit Mg²⁺-ATP-dependent catalytic activity of GSK3. Alternatively, indirect GSK3 inhibition can be induced through the inhibition of the protein phosphatase-2A or by increasing Akt activity (Pan et al., 2011; Freland and Beaulieu, 2012). In that regards, it was demonstrated that Akt1 isoform is essential for lithium to modulate mood-related behaviors in mice (Pan et al., 2011). Protein levels of Akt1 are not affected in the Akt3 KO mice but it is still possible that the observed amplified GSK3α/β phosphorylation is caused by an elevation of Akt1 activity. Additional biochemical experiments are definitively required to confirm this hypothesis. However, our data suggests that mechanisms underlying the lithium therapeutic effects are intact in Akt3 KO mice. They also confirm the implication of Akt3/GSK3α/β pathway in the depressive and anxiety-like behaviors seen in these mice.

It is noteworthy that recent human genetic studies are in accord with our data and suggest that impaired Akt3 is associated with schizophrenia, depression and anxiety endophenotypes. In particular, Akt1 was originally identified as a potential schizophrenia susceptibility gene (Emamian et al., 2004). However, a recent multi-stage schizophrenia genome-wide association study of up to 36,989 cases and 113,075 controls did not reproduce this finding (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014). They reported instead that only Akt3 represents a potential contributor to schizophrenia with a high risk factor of 64 out of 108 conservatively defined loci that meet genome-wide significance. Another argument is the fact that FXR1P, a direct molecular substrate of GSK3, has also been identified as a high schizophrenia risk factor (26/108) in this study. Notably, the study establishing FXR1P as a GSK3 substrate has also demonstrated the importance of these proteins interaction in the regulation of mood and emotion processing (Del’Guidice et al., 2015). Altogether, our data paired with these recent evidences raise the interesting possibility that Akt3 might play a pivotal role in human brain pathologies such as schizophrenia, depression and anxiety.

In conclusion, modeling of human neuropsychiatric disorders in rodents is extremely challenging. Progress in understanding key molecular target in order to develop new therapeutics would certainly benefit to this research field. Therefore, the involvement of the Akt3 isoform in mice behaviors related to schizophrenia and mood disorders is novel and highly relevant. We believe that further investigation on the Akt3 isoform could lead to a better understanding of the molecular basis implicated in neuropsychiatric diseases.

YB and MC designed the study. YB conducted this research. GB contributed for the behavioral experiments. M-ÉL-L conducted the electrophysiology experiments. YB and MC wrote the manuscript. EA provided the Akt3 KO mice. MC and GM contributed to the conceptual frame of the study and edited the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.