br Materials and methods br Results br Discussion

Materials and methods
Results
Discussion GSK-3 overexpression diminishes neurogenesis (Fuster-Matanzo et al., 2013, Kondratiuk et al., 2013, Sirerol-Piquer et al., 2011), while the inhibition of GSK-3 with lithium or valproate enhances neurogenesis, as measured by cell proliferation, the number of immature neurons, or commitment to neural lineages (Boku et al., 2009, Chen et al., 2000, Hao et al., 2004, Hsieh et al., 2004, Silva et al., 2008, Sintoni et al., 2013, Wexler et al., 2008). While lithium is a non-selective GSK-3 inhibitor, and valproate is an indirect GSK-3 inhibitor, the effects of synthetic selective GSK-3 inhibitors have recently been analyzed, and also appear to favor neurogenesis (Mao et al., 2009, Morales-Garcia et al., 2012). In parallel, GSK-3 inhibition has been shown to exert efficient antidepressant effects in mice, as evident in the forced swim test (Du et al., 2010, Gould and Manji, 2004, Kaidanovich-Beilin et al., 2004, Rosa et al., 2008). However, it is still not known whether depletion of the neurogenic cell populations is a cause of depression or if this cell population just contributes to the action of some antidepressant drugs (Lucassen et al., 2014, Petrik et al., 2012). Moreover, as most of the studies mentioned addressed the effects of GSK-3 inhibition through a small number of behavioral tests, and given the side effects of these treatments, we considered it necessary to include a basic behavioral battery to assess our new drug, and to support the potential beneficial effects on mood and neurogenesis. We demonstrate that administration of the GSK-3 inhibitor called VP2.51 increases cell proliferation (as measured by pH3+ cell number), as well as the short- and long-term survival of newborn neurons (as assessed by the 24 h survival of BrdU+ and DCX+ neurons), while significantly increasing the commitment of Neuromedin B australia to the granule neuron lineage (as measured by Prox1 immunoreactivity). In parallel, VP2.51 induces a net antidepressant effect that was reflected by the decrease in immobility time in the forced swim test, with no significant effect on cognition, as well as a therapeutic effect after a Porsolt-induced stress evident through the immobility in the tail suspension test. Interestingly, the morphological changes were found prominently in the ventral region of the hippocampus. Region-specific morphological drug-induced effects are compatible to a putative region-specific function of hippocampus (for a recent review, see for example Tannenholz et al., 2014), considering the observed drug effects on mood and cognition. Interestingly O'Leary et al. (2012) described that treatment with lithium preferentially increased neurogenesis in the ventral hippocampus of previously stressed mice. Most relevant, we have been able to demonstrate that these effects are neurogenesis dependent. Last, but not least, VP2.51 did not provoke changes in weight or in a battery of behavioral tests, including learning and memory tests, and those of basal activity. As the effects of VP2.51 were concomitant with the expected increase in β-catenin expression and a shift towards the inactive form of GSK-3, we suggest that VP2.51 has therapeutic benefits following stress, and it may be a preventive treatment in situations where a potential depressive state and/or loss of memory is associated with diminished neurogenesis, through selective GSK-3β inhibition. As for neurogenesis (results in Fig. 2), we demonstrate a net effect of VP2.51 treatment on the survival of immature subpopulations during AHN. These subpopulations constitute a key factor for the functioning of the hippocampal dentate gyrus (Perez-Domper et al., 2013) and for the regulation of hippocampal-dependent behaviors (Dupret et al., 2007, Tashiro et al., 2007). We found a net increase in cell proliferation, as expected from the results with either chemically diverse selective GSK-3 inhibitors (Morales-Garcia et al., 2012) or following lithium/valproate administration (Chen et al., 2000, Jin et al., 2005, Lenox and Wang, 2003). This effect might be mediated by a direct effect on Wnt, a known regulator of both GSK-3 activity (McNeill and Woodgett, 2010) and of progenitor proliferation (Wexler et al., 2009). It is important to mention that we have found an increase of β-catenin levels in hippocampus after 6 h of VP2.51 treatment, which may indicate Wnt signaling modulation by the GSK-3 inhibitor. Moreover, the increase in cell proliferation found here is relevant as it occurs in parallel to the increased commitment to the granule neuron lineage (Prox1), an effect which may also be mediated through the canonical Wnt pathway, given that Prox1 expression is also regulated by Wnt signaling (Karalay et al., 2011, Karalay and Jessberger, 2011). Furthermore, for the first time we found a relevant and significant increase in the number of immature DCX-positive neurons in the dentate gyrus after treatment with a selective GSK-3 inhibitor VP2.51. A similar effect on the survival of BrdU+ cells labeled one week before sacrifice was described previously (Mao et al., 2009). This pool is heterogeneous, from type 1 cells to cells expressing only calretinin (Kempermann et al., 2004), and it is thought to be responsible for some hippocampal functions, such as pattern separation, the dysregulation of which might be a key factor in mood disorders (Leal et al., 2014, Shelton and Kirwan, 2013). This immature population is also thought to be relevant in the antidepressant and cognitive effects of several pharmacological drugs (Castren and Hen, 2013, Costa et al., 2015, Wu and Hen, 2014, Wu et al., 2014) and may be one of the reasons explaining the antidepressant effect of the GSK-3 inhibitor VP2.51 (discussion about TMZ and neurogenesis experiment below).