New neurons are continuously born from a proliferating population of neural stem cells (NSCs) throughout adulthood via a mechanism known as adult neurogenesis. This process occurs mainly in two regions of the brain, the subgranular zone of the hippocampal formation, and the subventricular zone (SVZ) of the lateral ventricle wall (Alvarez-Buylla and Lim 2004). However, neurogenesis has also been reported in brain regions outside the subgranular zone and SVZ (Gould 2007), including recent data on hypothalamic neurogenesis (Kokoeva et al. 2005). Hippocampal neurogenesis contributes to cognitive plasticity in the rodent (Sahay et al. 2011), and it has been shown to occur at rates comparable in humans and mice, suggesting a similar function in both species (Spalding et al. 2013). Newly born neuroblasts in the SVZ migrate to the olfactory bulb through the rostral migratory stream, where they terminally differentiate into neurons (Whitman and Greer 2009). This process seems related to the acquisition or discrimination of new odors that are important for survival (Whitman and Greer 2009). However, the SVZ in rodents and human presents organizational differences which could also reflect functional differences (Gonzalez-Perez 2012). Hippocampal and SVZ neurogenesis is dysregulated in Huntington's, Parkinson's (PD), and Alzheimer's (AD) diseases (Winner et al. 2011), suggesting that this process could play a functional role in the development and/or response to neurodegeneration. In stroke, neural precursor cells from the SVZ can divert their migration from their normal route along the rostral migratory stream and instead migrate to the site of neural damage (Arvidsson et al. 2002). Similarly, this process has been shown to also occur in the human brain (Marti-Fabregas et al. 2010).Whether these cells can correctly integrate in brain damage areas and contribute to recovery remains to be evaluated. Diabetes and obesity are strong risk factors for pre-mature stroke (Sander and Kearney 2009) and neurodegenerative diseases such as AD (Kalaria 2009) and PD (Vanitallie 2008). Interestingly, recent studies have shown that adult neurogenesis is impaired in obese and diabetic animal models in vivo and also by a diabetic milieu in vitro (Suh et al. 2005; Zhang et al. 2008; Alvarez et al. 2009; Guo et al. 2010; Park et al. 2010; Mansouri et al. 2012).
Galanin is a 29 (rodents) or 30 (human) amino acid peptide (Tatemoto et al. 1983) with a wide-range of biological effects in both the CNS and PNS (Bartfai et al. 1993; Gundlach et al. 2001; Lang et al. 2007; Ogren et al. 2010). Galanin is involved in metabolism and reproduction (Barson et al. 2010) (Merchenthaler 2010), survival, regeneration (Hobson et al. 2010) and cognition (Crawley 1999; Ogren et al. 2010). In addition, several studies have shown that galanin plays a role in pathological conditions such as pain (Liu and Hokfelt 2002; Xu et al. 2010), AD (Counts et al. 2010), addiction (Picciotto 2010), and epilepsy (Lerner et al. 2010).
Galanin signals through three G-protein coupled receptors (GalR1, -R2 and -R3) (Lang et al. 2007; Mitsukawa et al. 2008). Both GalR1 and GalR2 are widely expressed in the rat brain (Smith et al. 1997; O'Donnell et al. 1999; Burazin et al. 2000; Waters and Krause 2000). In contrast, the GalR3 expression pattern has a more restricted distribution in rodents with transcript levels most abundant in the hypothalamus (Mennicken et al. 2002). Individual galanin receptors have been associated with certain functions, in particular GalR1 and –R2 (Mitsukawa et al. 2010; Webling et al. 2012), for example GalR2 showing neuroprotection (Elliott-Hunt et al. 2004; Pirondi et al. 2010). The physiological and pathological role of GalR3 in the brain is less well-characterized (Barreto et al. 2011; Webling et al. 2012).
The mRNA and proteins for galanin and its receptors are high in rodent neurogenic areas (Elliott-Hunt et al. 2004; Shen et al. 2005) and in rodent and human embryonic stem cells (Anisimov et al. 2002; Assou et al. 2007). In addition, a not clearly defined GalR2/GalR3 activation has been suggested to regulate the proliferation of adult hippocampal-derived NSCs (Abbosh et al. 2011). Finally, a recent study in mice has shown that galanin receptors (GalRs) are highly expressed in NSCs from the SVZ and that GalR2 and/or GalR3 activation can regulate NSC differentiation (Agasse et al. 2013).
The aim of this study was firstly to determine the potential role of galanin and its receptors to protect NSCs in response to a diabetic glucolipotoxic milieu in vitro. Furthermore, we studied apoptosis and unfolded protein response (UPR) signaling as potential mechanisms mediating such a protective effect. In addition, we performed quantitative studies on NSC regulation of galanin and GalRs in response to diabetes-like conditions in vitro and in vivo.
- Top of page
- Materials and methods
- Supporting Information
The incidence of type 2 diabetes (T2D) and neurodegenerative disorders is increasing in the western world accompanying the growing number of obese and elderly people. Stroke is over-represented within the diabetic population, where also the risk of developing stroke is increased 2- to 6-fold (Sander and Kearney 2009). Furthermore, there is a strong co-morbidity between T2D and AD (Kalaria 2009) sometimes referred to as T3D, the reasons for which are largely unknown. In T2D, there is a plethora of changes that alone or together can impair brain metabolism and affect neuronal viability. These include glucose toxicity, hyperlipidemia, hypertension, increased inflammation and oxidative stress and insulin resistance (Sims-Robinson et al. 2010). These events may lead to damage of cerebral microvasculature and/or neural tissue, which in turn could pre-dispose to brain disorders, including neurodegeneration.
On top of these well-characterized mechanisms, deranged adult neurogenesis in hippocampus has been proposed to play a role. In fact, pre-clinical data show dysfunctional hippocampal neurogenesis in diabetic and obese animal models (Zhang et al. 2008; Alvarez et al. 2009; Park et al. 2010) that could result in cognitive impairment (Stranahan et al. 2008). In addition, recent pre-clinical reports have indicated a role of hypothalamic adult neurogenesis in maintaining energy balance in response to environmental and physiologic insults (Pierce and Xu 2010). Thus, hypothalamic neurogenesis may represent one of the adaptive mechanisms used by the brain to limit functional impairment resulting from obesity. With regard to SVZ-neurogenesis, in vivo and in vitro experimental models have shown that this process is severely impaired by diabetes (Lang et al. 2009; Guo et al. 2010; Mansouri et al. 2012). SVZ-derived neuroblasts can, after stroke in the normal non-diabetic brain, integrate into the stroke-damaged striatum and might play a role in functional recovery (Arvidsson et al. 2002). If so, SVZ neurogenesis impaired by diabetes could decrease the endogenous brain restorative response after stroke. To understand if this hypothesis has clinical relevance, more research is needed to better characterize the role of adult SVZ neurogenesis in neurological disorders in the normal and diabetic brain as well as to understand whether functional differences between humans and rodents exist. Collectively, these results suggest that impaired adult neurogenesis could play an important role linking metabolic disorders to central neurological complications. Conversely, such a defect is an attractive target in the quest of identifying factors capable of normalizing impaired neurogenesis and, hence, preventing or limiting diabetic CNS complications.
To identify molecules with such properties, an in vitro system has been developed, where we have shown that a diabetic milieu characterized by glucolipotoxicity impairs NSCs viability (Mansouri et al. 2012). In this study, we demonstrate that galanin and the equally affinity agonist for GalR2/GalR3, AR-M 1896 (galanin 2-11) (Liu et al. 2001; Lu et al. 2005) can counteract the impaired NSC viability induced by such a diabetic milieu. This protective effect does not seem to be restricted to the diabetic challenge, since the limited, naturally occurring cell death in our primary cultures was also improved by AR-M 1896. Furthermore, our results show that the increased NSC viability conferred by AR-M 1896 in our assay occurs entirely through cell protection, since no change in proliferation was observed. Ma et al. have shown that AR-M1896 increases the length of neuritis in SVZ-derived NSCs (Ma et al. 2008), in agreement with studies on injured dorsal root ganglion neurons (Mahoney et al. 2003). However, Abbosh et al. (2011) reported that low concentrations of AR-M 1896 are both trophic and proliferative in adult hippocampal NSCs and that these effects could be blocked by the GalR2 selective antagonist M871. We note that Abbosh et al.'s studies are from hippocampal NSCs (vs. ours being SVZ-derived), and that different exposure times of AR-M 1896 to the cells were used (only 24 h in our assay vs. 5 days in the Abbosh et al. work).
The unselectivity of AR-M 1896 may explain why some previous studies, not only in CNS, have failed to discriminate between GalR2- and GalR3-mediated effects (Lu et al. 2005; Shen et al. 2005; Pirondi et al. 2010; Agasse et al. 2013). Moreover, two recent reports that attempted to address galanin-mediated NSC proliferation and differentiation were not able to clearly distinguish between a GalR2- and/or a GalR3-mediated effect (Abbosh et al. 2011; Agasse et al. 2013). By combining AR-M 1896 with a GalR2 preferring antagonist and a specific antagonist for GalR3, respectively, we now are able to show that the protective effect mediated by AR-M 1896 occurred selectively via GalR3 activation. These results are supported by the recent demonstration of GalR3 protein with immunohistochemistry and Western blot in the mouse SVZ (Agasse et al. 2013) (also see below) and our own demonstration, in vivo and in vitro, of GalR3 mRNA in NSCs and the SVZ region, to our knowledge the first to show cellular protection specifically mediated by GalR3.
Our results also show that the protective effect mediated by galanin against glucolipotoxicity correlates with increased protein levels of Bcl-2 and decreased cleaved-caspase 3, suggesting that the protective effect mediated by GalR3 activation occurs via decreasing apoptosis. ER stress may play a fundamental role in the development and pathology of certain forms of both diabetes and neurodegenerative diseases (Lindholm et al. 2006; Ortsater and Sjoholm 2007). Therefore, in this study we have extended our previous work on the effect of glucolipotoxicity on NSCs (Mansouri et al. 2012) by determining whether ER stress is regulated by a diabetic milieu. Indeed, we show that glucolipotoxicity significantly activates JNK and UPR signaling, by increasing the ER stress markers CHOP, BIP and the alternative spliced form of XBP1. Although similar results were obtained by Li et al. in the C17.2 immortalized cell line derived from cerebellum (Li et al. 2011), our results on the activation of ER stress by a diabetic milieu are, to our knowledge, the first obtained in primary adult NSCs. By assessing the potential role of GalR3 activation on the UPR signaling, we quantified mRNA levels of CHOP after palmitate and AR-M1896 exposure, and we show that the activation of GalR3 decreased the enhanced CHOP levels induced by a diabetic milieu. As activation of GalR3 was without any effect on BIP mRNA levels, our data indicate that stimulation of the GalR3 regulates the cytotoxic pathway of the UPR by a mechanism independent of increased chaperone expression. This suggests that at least part of the protective effect mediated by GalR3 activation occurs through UPR regulation.
The functional role of galanin on NSCs in the SVZ begins to be uncovered (Shen et al. 2005; Ma et al. 2008), and the recent report by Agasse et al. showed that Gal R1, -R2, and -R3 are expressed in the SVZ not only in tissue samples but also in neurospheres from adult mouse, using qPCR and blotting [(protein data to be confirmed since not validated as outlined by Lu et al. (Lu and Bartfai 2009)] of the adult mouse (Agasse et al. 2013). Here, they promote neuronal differentiation through GalR1 and -R2 activation, but not self-renewal, proliferation or cell death (Agasse et al. 2013). In view of the emerging importance of the galanin system in SVZ neurogenesis, and of the potential relevance at the pharmacological/therapeutic level, we wanted to address whether or not galanin and GalRs expression are impacted by diabetes in this brain area. To do so, we quantified galanin and GalR expression in response to glucolipotoxicity in vitro, and by comparing young pre-diabetic and old diabetic ob/ob mice with aged matched lean littermates in vivo. Our results show that GalR3 expression remained unchanged in a diabetic milieu in vitro as well as in young pre-diabetic ob/ob mice in vivo. However, GalR3 expression was strongly down-regulated in aged diabetic ob/ob mice in comparison with their lean littermates.
Whether or not the decreased expression of GalR3 in diabetes plays a role in the decreased NSC survival and/or proliferative and differentiative capacity remains an interesting hypothesis to be further investigated in vivo. In contrast to GalR3 expression, we also show that galanin, GalR1, and GalR2 were up-regulated by diabetes, both in vitro and in vivo. Furthermore, GalR2 expression remained high in aged diabetic ob/ob mice in comparison with their age-matched lean littermates. Although galanin and GalRs have been previously reported to be highly expressed in the SVZ of the adult rodent brain (Agasse et al. 2013) (Shen et al. 2005), very little is known about their physiological role in this brain area. The fact that the expression of both galanin and its receptors is regulated by diabetic conditions in SVZ provides an impetus for future in vivo research, aimed at understanding their physiological role in this brain area under normal and diabetic conditions.
The regulation of NSCs in the SVZ has been suggested to play a regenerative role in AD, PD, and stroke (Emsley et al. 2005), disorders that are strongly over-represented in the diabetic population (Vanitallie 2008; Kalaria 2009; Sander and Kearney 2009). Our results indicate the possibility that NSC protection via GalR3 agonists may be used to prevent and treat the neurological complications of diabetes, but is so far an attractive hypothesis for further studies.
In conclusion, we found that galanin via GalR3 activation counteracts NSC glucolipotoxicity, which correlates with decreased apoptosis and modulation of the UPR signaling. Furthermore, we show that each of the three GalR subtypes is regulated in response to glucolipotoxicity both in vitro and diabetic milieu in vivo, effects that may serve to regulate neuronal differentiation, proliferation, and survival in diabetes/obesity.