br Materials and methods br Results br Discussion

Materials and methods
Results
Discussion Our data show that the ghrelin receptor, Ghsr, links energy homeostasis with a form of adult hippocampal plasticity. The presence of Ghsr on mature DG neurons suggest that acyl-ghrelin may modulate NSPCs indirectly, possibly via soluble factors such as BDNF that support AHN (Bekinschtein et al., 2013). Previous work suggests that ghrelin-treatment increases hippocampal BDNF levels in streptozotocin-induced diabetic rats (Ma et al., 2011), however, further studies are required to determine whether this neurotrophic factor is involved in acyl-ghrelin-mediated AHN. Similarly, under stressful conditions, the elevation in circulating acyl-ghrelin (Lutter et al., 2008, Walker et al., 2014) may protect AHN by inhibiting the release of inflammatory cytokines such as interleukin-6 (Beynon et al., 2013), that are known to impair AHN (Monje et al., 2003, Vallières et al., 2002). Acyl-ghrelin is known to induce c-Fos and Egr-1 expression in mouse hypothalamus (Hewson and Dickson, 2000). We now show that c-Fos+ cells are increased in the DG following acyl-ghrelin treatment, but not following acute CR. However, we report for the first time a robust increase in Egr-1+ cells in the DG following treatment with either acyl-ghrelin or with CR. The increase in Egr-1 was similarly observed both in DG cells expressing Ghsr and in cells lacking the receptor, suggesting that ghrelin signaling induces network expression of Egr-1 within the DG. Furthermore, Egr-1 immunoreactivity was regulated in other conotoxin australia regions, including the BLA and the cingulate cortex. Ultimately, a comprehensive dissection of these regions will need to be performed to identify their contribution to CR-mediated AHN and cognition. Nonetheless, our data suggest that Egr-1 may be particularly responsive to adaptations in energetic balance. Notably, hippocampal Egr-1 expression is rapidly induced by learning and retrieval of memories, its blockade impairs memory formation (Bozon et al., 2003, Jones et al., 2001), particular the re-consolidation of hippocampal dependent contextual fear memories (Lee et al., 2004). More recently, Egr-1 expression in mature DG neurons was shown to be essential for the survival, maturation and integration of newborn adult neurons into the hippocampal circuitry. Furthermore, Egr-1-KO mice showed deficits in hippocampal-dependent long-term spatial memory (Veyrac et al., 2013). These findings suggest that the CR-mediated increase in Egr-1 within the DG may support cognition. Previous studies have demonstrated that CR increases the number of surviving newborn cells, rather than triggering proliferation, in the DG (Lee et al., 2002) in a ghrelin-dependent manner (Kim et al., 2015). However, the impact of CR on new neuron formation in the DG is unknown. These findings prompted us to ask whether a more prolonged period of CR would increase the number of new mature adult born DG neurons. Indeed, we show that a 2-week period of CR, with just a 30% reduction in daily calories, results in a significant increase in new neurons 31 days following the end of the CR period in wild-type but not Ghsr−/− mice, suggesting that ghrelin signaling mediates the neurogenic effect of CR. Whilst studies using ghrelin reporter mice suggest that the generation of acyl-ghrelin is restricted to the periphery (Sakata et al., 2009), we cannot rule out the possibility that brain-derived ghrelin may influence AHN. In addition, it is possible that currently unknown Ghsr ligands, other than ghrelin, may play a role in promoting AHN in this context. However, as we have previously shown that peripheral treatment with acyl-ghrelin increases AHN (Kent et al., 2015) and that CR is known to elevate plasma ghrelin (Lutter et al., 2008), we suggest that the CR-mediated and Ghsr-dependent increase in AHN is likely induced by acyl-ghrelin. AHN is necessary for hippocampus-dependent memory and newborn neurons contribute to spatial pattern separation (Clelland et al., 2009). Notably, dendritic synapses of newborn adult neurons show enhanced plasticity between 4 and 6 weeks of age compared with other stages (Ge et al., 2007). At this point they exhibit increased intrinsic excitability, lower activation threshold (Marin-Burgin et al., 2012, Schmidt-Hieber et al., 2004) and recruitment into circuits mediating behavior (Kee et al., 2007, Nakashiba et al., 2012, Tashiro et al., 2007). Optogenetic silencing of 28 day old adult born neurons resulted in impaired retrieval of a contextual fear memory (Gu et al., 2012). To test whether the CR-mediated increase in AHN contributes to hippocampal function our study was designed so that newborn neurons were 4–6 weeks of age during the CFC assessment. In keeping with previous studies we show that the increase in 4–6 week old neurons was associated with enhanced remote contextual fear memory. This improved retrieval of remote memory is consistent with increased re-consolidation over time; a process also associated with AHN (Kitamura et al., 2009, Pan et al., 2012). These data suggest that CR-induced new adult born neurons assume functional roles in hippocampal circuits supporting mnemonic function. However, the extent to which the CR-mediated maintenance of remote fear memory is relevant to the Ghsr-dependent increase in AHN remains to be tested using more specific approaches. In particular, as AHN is essential for accurate pattern separation, testing our experimental paradigm using behavior tests that place a high demand on discrimination is now warranted.