J. Neurochem. (2010) 112, 1368–1385.
The differentiation of neural stem cells toward a neuronal phenotype is determined by the extracellular and intracellular factors that form the neurogenic niche. In this review, we discuss the available data on the functional role of inflammation and in particular, pro- and anti-inflammatory cytokines, on neuronal differentiation from endogenous and transplanted neural stem/progenitor cells. In addition, we discuss the role of microglial cell activation on these processes and the fact that microglial cell activation is not univocally associated with a pro-inflammatory milieu. We conclude that brain cytokines could be regarded as part of the endogenous neurogenic niche. In addition, we propose that accumulating evidence suggests that pro-inflammatory cytokines have a negative effect on neuronal differentiation, while anti-inflammatory cytokines exert an opposite effect. The clarification of the functional role of cytokines on neuronal differentiation will be relevant not only to better understand adult neurogenesis, but also to envisage complementary treatments to modulate cytokine action that could increase the therapeutic benefit of future progenitor/stem cell-based therapies.
IL-1 expressing recombinant adenoviral vector
β-galactosidase expressing recombinant adenoviral vector
brain-derived neurotrophic factor
ectodermal dysplasia 1
fibroblast growth factor
glial fibrillary acidic protein
insulin-like growth factor 1
monocyte chemotactic protein-1
major histocompatibility complex
neural stem/progenitor cells
stem cell factor
stromal cell-derived factor
subgranular zone of the dentate gyrus of the hippocampus
substantia nigra pars compacta
subventricular zone of the lateral ventricle
transforming growth factor-β1
tumor necrosis factor-α
type III β-tubulin
Adult neurogenesis is the process of generating new neurons from neural stem/progenitor cells (NSCs) in the adult. It encompasses different processes such as adult NSCs proliferation and differentiation, migration of neuronal precursors, and survival of newborn neurons. Shortly after the development of the [H3]-thymidine incorporation technique to detect cell division in the 1960s, Altman showed that new cells with neuronal morphology could be found in the dentate gyrus (DG) of the hippocampus (Altman and Das 1965), neocortex (Altman and Das 1966), and olfactory bulb (Altman 1969). However, these results were not taken into consideration until the late 1970s, when Kaplan and Hinds (1977) showed that these new neurons could survive for long periods of time.
Adult neurogenesis is regulated by physiological and pathological conditions at different levels, including the activation and proliferation of adult NSCs, the migration, differentiation and survival of the newly born neurons and, finally, the integration of these neurons to the existing neuronal circuits (Fuchs and Gould 2000; Duman et al. 2001; Kempermann 2002; Ming and Song 2005). In the adult mammalian CNS, adult NSCs reside in two specific regions, the subventricular zone of the lateral ventricle (SVZ) and the subgranular zone of the DG of the hippocampus (SGZ; Gage 2002). In the first case, neurogenesis occurs in the olfactory bulb after migration of the NSCs through the rostral migratory stream. In the second, newborn cells are produced in the DG (Gage 2002). In the normal brain, neurogenesis outside these regions appears to be extremely limited, or nonexistent. However, in pathological conditions affecting the brain, neurogenesis can be enhanced or suppressed. It is worth noticing that in the cases where neurogenesis is enhanced, formation of new neurons is also seen in non-neurogenic areas. In addition, NSCs could be cultivated from different brain areas where neurogenesis is not physiologically observed (Palmer et al. 1999). Finally, in general, transplanted NSCs fate will depend on whether the site of transplantation promotes neuronal differentiation or not (Lie et al. 2002). These observations suggest that NSCs proliferation and differentiation toward a specific phenotype are regulated by the characteristics of the microenvironment or niche in which they reside (Gage et al. 1995; Herrera et al. 1999; Gage 2002; Dziewczapolski et al. 2003; Alvarez-Buylla and Lim 2004; Chen et al. 2007).
Although there are structural differences in the architecture of each of the neurogenic niches, the main cellular and molecular components seem to be conserved. At a cellular level, compelling evidence suggests that astroglia plays a key role not only as a structural part of the neurogenic niche (Doetsch et al. 1997), but also in regulating almost every stage of the process of neurogenesis, including self-renewal, fate specification of adult NSCs, migration, differentiation, and integration of newly born neurons (Song et al. 2002; Barkho et al. 2006; Jiao and Chen 2008). Ependymal cells seem to regulate the quiescence and proliferation of adult NSCs (Ramirez-Castillejo et al. 2006). Also mature neurons near the SGZ are thought to function as niche cells, providing spatiotemporal regulation of adult neurogenesis in response to neuronal activity (Kempermann et al. 2000; Ma et al. 2009). Recent evidence supports a role for vascular/endothelial cells in regulating the proliferation of adult NSCs as they can enhance neurogenesis in vitro (Palmer 2002; Shen et al. 2004).
Molecular signals regulating adult neurogenic niche have also been identified. Physiological factors, including Wnt (Lie et al. 2005), Sonic Hedgehog (Lai et al. 2003; Ahn and Joyner 2005), bone morphogenetic proteins (Lim et al. 2000; Bonaguidi et al. 2008), membrane-associated Notch signaling (Nyfeler et al. 2005; Androutsellis-Theotokis et al. 2006), leukemia inhibitory factor (Bauer and Patterson 2006), transforming growth factor-α (TGF-α; Tropepe et al. 1997), and vascular endothelial growth factor (Jin et al. 2002; Schanzer et al. 2004) have been shown to promote NSCs proliferation and maintenance. Also, growth factors, including fibroblast growth factors (FGFs) and neurotrophins such as brain-derived neurotrophic factor (BDNF), also significantly contribute to proliferation, survival, and dendritic development of newborn neurons in the adult brain (Yoshimura et al. 2001; Schmidt and Duman 2007).
Recently, components of the immune response have been proposed as part of the neurogenic niche, affecting NSCs proliferation, differentiation and/or survival. Specially, the innate immune response and, in particular, inflammation have attracted attention as key players in modulating NSCs biology. In addition, transplantation of NSCs or stem-derived cells into the brain will elicit, at least, an inflammatory reaction. Therefore, the information on how signals coming from this response alters the endogenous niche and, in turn, affect the transplanted cells is crucial to enhance the potential benefits of transplanted cell-based therapies. Thus, this review will focus on the effects of brain inflammation on the biology of endogenous and transplanted NSCs. In particular, we will focus on the main source of cytokines in the brain, the activated microglia. Activated microglia can promote a pro- or anti-inflammatory milieu and therefore, would not always be equivalent to brain inflammation. This concept is important to prevent the idea that phenomena related with microglial cell activation are necessarily associated with production of pro-inflammatory cytokines. In addition, inflammatory reactivity varies among different brain regions according to their cellular composition. Thus, we will examine the features of these regions and suggest possible explanations to the differences observed in the modulation of NSCs biology by the local milieu.
Microglia and astrocytes are the main mediators of inflammation in the brain. In addition, astrocytes are also considered part of the neurogenic niche (Song et al. 2002). This review will concentrate on describing the role of microglial cells which, from our perspective, should also be considered as one of the principal cellular components of the neurogenic niche. For a comprehensive review on the role of astrocytes on neurogenesis, see (Wang and Bordey 2008).
Microglia is typically described as the resident macrophages in the brain parenchyma (Gehrmann et al. 1995). They have no neural origin, but derive from invading monocytes during development (Perry and Gordon 1988). Microglial cells comprise up to 15% of total cells in different regions of the brain (Perry et al. 1993a). Under basal conditions, microglia exhibit a quiescent phenotype as indicated by a ramified morphology, down-regulation of activation antigens such as major histocompatibility complex (MHC) II, and constitutive expression of macrophage antigens such as complement receptor 3 and ionized calcium binding adaptor protein-1 (Ladeby et al. 2005). Under CNS pathological conditions and systemic infectious processes, microglial cells undergo a phenotypic transformation characterized by up-regulation of antigen presentation molecules (MHC II, CD80, CD86, and CD40) and morphological changes. Activation proceeds through an activated ramified microglia, with long-shaped cell body and long but thick processes. Further activation leads to ameboid microglia, characterized by round-shaped cell body and few, short processes. In the final stage of activation, microglial cells are rounded cells with vacuolated cytoplasm and no visible processes which stain for ectodermal dysplasia 1 (ED1), a marker for phagocytic activity (see Fig. 1). At least four stages of activation characterized by morphology have been described (Kreutzberg 1996). These stages are dynamic and reversible (Perry et al. 2007).
Typically, microglial cell activation has been associated with production of pro-inflammatory cytokines, generation of reactive oxygen and nitrogen species, and the display of phagocytic activity (Streit et al. 1999). However, activated microglia may display a subset of these phenotypes or even anti-inflammatory effects depending on the type of activation (for a comprehensive review, see Perry et al. 2007). Microglia can enter an activation state, wherein they retain a ramified morphology, up-regulate activation antigens such as MHC II, and yet not secrete pro-inflammatory molecules (Betmouni et al. 1996; Walsh et al. 2001). For example, when microglia are activated by lipopolysaccharide (LPS; a well-known cytokine inducer), phagocytosis takes place with the release of pro-inflammatory factors (Hausler et al. 2002); in contrast when apoptosis is the trigger for microglial cell activation, anti-inflammatory cytokines are released (Magnus et al. 2001). Furthermore, LPS signaling can also promote different microglial cell activation phenotypes depending on whether it is an acute or a chronic stimulus (Cacci et al. 2008). In particular, in animal models of neurodegeneration, microglial cell activation has been defined as atypical (Depino et al. 2003; Perry et al. 2003; Godoy et al. 2008). These atypically activated microglia do not promote inflammation during neurodegeneration but are primed for a burst of secretion of pro-inflammatory molecules (Perry et al. 2003). Perry et al. (2003) developed a conceptual framework of microglia priming, in which a primary insult or disease process such as neurodegeneration induces this primed microglia phenotype. Upon subsequent induction of a peripheral or central pro-inflammatory process (secondary insult), previously primed microglia exhibit an exacerbated pro-inflammatory phenotype. This second, pro-inflammatory insult on primed microglia and the consequent increase in inflammation result in exacerbated neurodegeneration, increased behavioral or motor symptoms of the disease state as well as potentiate pro-inflammatory cytokine production in the CNS in models of prion, Alzheimer’s and Parkinson’s disease (PD; Depino et al. 2003; Perry et al. 2003; Godoy et al. 2008).
From these studies, it is also clear that whereas microglial cell activation can elicit pro- and anti-inflammatory effects, brain inflammation does not equal microglial cell activation. Our definition of inflammation includes microglial cell activation, but the reverse does not always apply. Thus, the brain inflammatory reaction that includes peripheral immune cell recruitment into the brain should be clearly differentiated from the endogenous response of the brain where only microglial cell activation takes place.
It is also worth noticing that the brain can also be classified in different compartments according to its reactivity to inflammation. There are a number of stimuli that fail to elicit an inflammatory response in the parenchyma, which can elicit a rapid response when they are placed in the ventricles or subarachnoid space (Perry et al. 1993b). This can be because of differences in the cellular composition of the different compartments. Cells of the monocyte/macrophage system can be divided into three populations: parenchymal microglia, perivascular macrophages, and macrophages located in the meninges and the choroids plexus (Lassmann et al. 1991). Macrophages in the choroid plexus and meninges show an up-regulated phenotype when compared with parenchymal microglia, expressing OX62 and MHC II (Matyszak et al. 1992). Moreover, there is a large number of mast cells in the meninges, and these cells are not observed in other regions of the brain (Perry et al. 1985). Furthermore, ventricular ependymal cells are abundant in the SVZ (20–50% of the overall cell number; Doetsch et al. 1997) and represent an important barrier to protect neuronal tissue by an innate immune response. Different receptors allow ependymal cells to sample the cerebrospinal fluid and respond to various infectious challenges. Ependymal cells also express a number of factors that sustain NSCs proliferation and differentiation (e.g. FGF-2, Notch-I, and bone morphogenetic proteins; reviewed in Hauwel et al. 2005).
With the previous information in mind, several hypotheses can be proposed. NSCs in the SVZ have contact with ventricular macrophages with an up-regulated phenotype while NSC in the SGZ of the hippocampal formation are preferentially surrounded by parenchymal microglial cells. These differences allow us to hypothesize that in the presence of the same pro-inflammatory stimulus, ventricular macrophages will elicit and inflammatory response while parenchymal microglial cells will not. Thus, NSC from the SVZ will be affected by an inflammatory milieu, while NSC from the SGZ will not. Following a similar line of thinking, a stem cell transplant will elicit a particular inflammatory response according to the local microenvironment of the host tissue, which will influence cell proliferation, differentiation and survival. For example, an exacerbated inflammatory response to transplantation is expected to happen in the ventricles in comparison to the brain parenchyma.
Thus, information on the brain compartment of study should be kept in mind to avoid drawing general conclusions from experiments performed only in one brain compartment.
Microglial cell activation and neurogenesis
In 2003, two independent studies showed that an LPS challenge in the brain led to microglial cell activation and a decline in hippocampal neurogenesis (Ekdahl et al. 2003; Monje et al. 2003). Furthermore, the administration of anti-inflammatory drugs restored hippocampal neurogenesis. In these reports, a negative association between the number of activated microglia and the number of new neurons was found. Also, proliferation of NSCs was significantly reduced only when cultured in the presence of conditioned media from activated microglia, but not from resting ones. The effect was thought to be mediated by the production of pro-inflammatory cytokines and chemokines such as interleukin (IL)-1β, tumor necrosis factor-α (TNF-α), IL-6, IL-18, and monocyte chemotactic protein-1 (MCP-1), and various reactive nitrogen and oxygen species (Rock et al. 2004). Recent evidence suggests that activated microglia are not always detrimental for neurogenesis. In fact, under certain conditions they can even be beneficial (Hanisch and Kettenmann 2007; Ekdahl et al. 2009). For example, microglial cell activation to stage 3, but not stage 4 (phagocytic) has been associated with enhanced neurogenesis and not pro- but anti-inflammatory cytokine production (Battista et al. 2006).
It is well known that the final effect of microglial cell activation upon NSCs biology will depend on the resulting profile of molecules they secrete. This review will concentrate on one of the main families of molecules derived from microglia and that has also been extensively studied in relation to NSCs proliferation and neurogenesis, i.e. pro- and anti-inflammatory cytokines.
Cytokines involved in adult neurogenesis
Cytokines are considered main components of the inflammatory/immune response. Pro- and anti-inflammatory cytokines are constitutively expressed and induced under a variety of stimuli in the brain (Schöbitz et al. 1994). These cytokines can be endogenously expressed by brain cells without the infiltration of immune cells from the periphery (Pitossi et al. 1997). Cytokines and their receptors are expressed in brain cells, both in glia and in neurons (Hopkins and Rothwell 1995; Rothwell et al. 1996). These endogenous molecules not only serve as communicators of immune signals into the brain and modulators of the glial response, but they also appear to contribute to exclusive brain function (e.g. long-term potentiation and synaptic plasticity; Besedovsky and Rey 2007). This particular fact gains relevance if cytokines derived from brain cells are to be considered as a part of the endogenous neurogenic niche without the intervention of the immune system.
It is worth to mention the differences found in the dynamics of cytokine expression in the brain when it is compared with the periphery. For example, contrary to the periphery, central IL-1 fails to induce TNF-α expression, and TNF-α does not induce either itself or IL-1 in the brain (Blond et al. 2002; Ferrari et al. 2004; Depino et al. 2005). This altered cytokine network reflects a specific level of control on the final effects of cytokines in the brain.
We here summarize the current literature showing a role of specific cytokines in modulating NSCs proliferation, differentiation, and neuronal survival. We have concentrated our analysis on prototypic pro- (IL-1, IL-6, and TNF-α) and anti-inflammatory cytokines (IL-4 and TGF-β) and not in other immune-related molecules such as interferons. We have also not intended to review the beneficial and toxic effects of cytokines in the brain, which could be found elsewhere (Liberto et al. 2004). In addition, we will not describe effects of the adaptive immune system on NSCs biology (for a review on the subject see (Rolls et al. 2009). For the considerations given before and as most cytokines have context-dependent functions, we differentiate between NSCs of SVZ and SGZ origin, and whether experiments were performed in vivo or in vitro (Table 1).
|Cytokine||Type of NSCa||Experimental model used||Role in proliferationb||Role on neuronal differentiationb||Role on survivalb||References|
|IL-1β||SGZ||In vivo||↓||=||=||Koo and Duman (2008)|
|In vitro||↓||ND||=||Koo and Duman (2008)|
|ND||=||ND||Monje et al. (2003)|
|SVZ||In vivo||=||ND||ND||Koo and Duman (2008)|
|TNF-α||SGZ||In vivo||↓ (TNF-R1)c||↓ (TNF-R1)c||(TNF-R2)d||Iosif et al. (2006)|
|In vitro||ND||ND||↓||Cacci et al. (2005)|
|ND||↓||ND||Monje et al. (2003)|
|SVZ||In vivo||↓||ND||ND||Iosif et al. (2008)|
|←||ND||ND||Wu et al. (2000)|
|In vitro||↓||=||←||Ben-Hur et al. (2003)|
|←||=||↓||Widera et al. (2006)|
|=||=||↓||Wong et al. (2004)|
|IL-6||SGZ||In vivo||↓||↓||↓||Vallières et al. (2002)|
|In vitro||=||↓||↓||Monje et al. (2003)|
|IL-4||SVZ||In vitro||ND||=||ND||Butovsky et al. (2006)|
|In vitro (microglia co-culture)e||=||←||=||Butovsky et al. (2006)|
|TGF-β1||SGZ||In vivo||=||←||ND||Battista et al. (2006)|
|↓||=||ND||Wachs et al. (2006)|
|In vivo (Tg)f||↓||NDg||↓||Buckwalter et al. (2006)|
|In vitro||↓||←||↓||Battista et al. (2006)|
|↓||ND||=||Buckwalter et al. (2006)|
|SVZ||In vivo||←||←||ND||Ma et al. (2008)|
|←||←||ND||Mathieu et al. unpublished data|
|↓||=||ND||Wachs et al. (2006)|
|In vitro||↓||=||ND||Wachs et al. (2006)|
Interleukin-1 and IL-1 receptor (IL-1R) are constitutively expressed in the hippocampus (Ban et al. 1991; Pitossi et al. 1997). In fact, the hippocampus is the brain region with the highest density of IL-1R (Kabiersch et al. 1988; Rothwell and Hopkins 1995). IL-1β alone is capable of overriding the intrinsic resistance of the CNS to leukocyte infiltration, resulting in acute cellular recruitment to the brain parenchyma (Anthony et al. 1997; Proescholdt et al. 2002; Ferrari et al. 2004; Ching et al. 2005; Depino et al. 2005). Additionally, blocking the actions of IL-1β using the IL-1R antagonist results in significant reductions in parenchymal leukocyte infiltration after injury (Garcia et al. 1995; Yang et al. 1998). IL-1β can also have autocrine effects in the CNS and has been proposed as a mechanistic link between a potential beneficial inflammatory response and detrimental glutamate excitotoxicity (Fogal and Hewett 2008).
It has been reported that IL-1 also show non-immune effects in the brain. For example, endogenous IL-1 modulates long-term potentiation (Schneider et al. 1998) and learning (Depino et al. 2004). Moreover, exogenous, high levels of IL-1β block memory formation (Pugh et al. 1999, 2001).
Interleukin-1 and its receptors are expressed in the DG (Koo and Duman 2008; Arguello et al. 2009), but IL-1R1 was not detected in progenitor cells derived from SVZ (Ben-Hur et al. 2003). Administration of exogenous IL-1β protein (20 or 100 ng, i.c.v.) significantly decreased the number of 5-bromo-2′-deoxyuridine (BrdU)-positive cells in the DG, but did not affect the number of proliferating cells in the SVZ (Koo and Duman 2008). Moreover, the administration of IL-1R antagonist protein (250 ng, i.c.v.) blocked the decrease in proliferating cells in the DG caused by acute stress. Similar results were obtained in vitro, as a 2 h treatment with 10 ng/mL of IL-1β protein reduced the proportion of proliferating NSC, and 100 ng/mL of IL-1ra protein blocked this effect (Koo and Duman 2008).
Concerning the effects of IL-1β on adult neurogenesis, we have studied the effects of the chronic expression of IL-1β in the DG of adult rats. To achieve chronic expression of human IL-1β, we administered a low dose (107 plaque formation units) of a replication-deficient, recombinant adenoviral vector expressing human IL-1β (AdIL-1β) to the right DG as described before (Ferrari et al. 2004, 2006). As control, a similar dose of β-galactosidase expressing adenovirus was used (Adβgal). At these doses, control vectors do not elicit a long-lasting inflammatory response in the brain (Ferrari et al. 2004).
At the peak of transgenic expression (14 days after the injection of the vectors; Ferrari et al. 2004), the tissue response was analyzed. Long-lasting expression of IL-1β elicited microglial cell activation accompanied by ED1 expression in the SGZ and the hilus and some tissue damage (17% of the DG; data not shown). No ED1-positive cells were observed in the contralateral hippocampus or in control animals (data not shown).
To analyze the effect of long-lasting IL-1β expression on neurogenesis, animals were injected intraperitoneally with BrdU during four consecutive days ending 3 days before adenoviral injection. In this way, we analyzed the effect of IL-1β on the fate of cells proliferating before adenovirus administration and not the population of cells such as microglia and astrocytes that are caused to proliferate by IL-1β expression. We performed double immunofluorescence analysis of BrdU and the neuronal neuron marker (NeuN), indicative of neurogenesis (Fig. 2a), or glial fibrillary acidic protein, indicative of astrogliogenesis (Fig. 2b). Cell counts were performed as described before (Battista et al. 2006).
After AdIL-1β injection, we observed a 50% reduction in the percentage of BrdU–NeuN-positive neurons among the total BrdU-positive cell population (*p < 0.05, Fig. 2a). On the contrary, we did not observe any changes in astrocytic differentiation after AdIL-1β injection (p = 0.8, Fig. 2b). When contralateral hemispheres were compared, there was no difference in the percentage of double-labeled cells for BrdU–NeuN over total BrdU-positive cells (33.9 ± 8.3% vs. 44.3 ± 11.2%, AdIL-1β vs. Adβgal, respectively; p = 0.33) or BrdU–glial fibrillary acidic protein over total BrdU-positive cells (14.9 ± 5.5% vs. 17.8 ± 6.6%, AdIL-1β vs. Adβgal; p = 0.75). This indicates that fewer cells acquired a neuronal phenotype because of long-lasting IL-1β expression in the DG.
To analyze a possible direct effect of IL-1β on NSC differentiation, primary hippocampal NSCs were cultured with 1, 10, or 100 ng/mL of rat IL-1β during 5 days. The number of type III β-tubulin (Tuj)-positive cells was quantified as indicative of neuronal lineage. There was no difference in the number of Tuj-positive cells in any of the IL-1β concentrations tested, compared with control cultures (Fig. 2c). A similar observation was made by Monje et al. (2003) using doublecortin as a marker of neuronal precursor.
Thus, IL-1 had no direct effect on neurogenesis in vitro under the experimental conditions tested and this suggests that additional factors should be present to observe the anti-neurogenic effect detected in vivo. It is interesting to stress the observation that the anti-neurogenic effects of IL-1 were observed in a DG that has also been affected by the adenoviral injection.
Tumor necrosis factor-α
Tumor necrosis factor-α is expressed at very low levels in basal conditions in the CNS (Pitossi et al. 1997). Astrocytes, microglia, and neurons have been described to produce increased TNF-α upon exposure to multiple physiological and pathological stimuli (e.g. CNS infection and brain injury; Munoz-Fernandez and Fresno 1998). TNF-α is present in two isoforms in the CNS, as a soluble molecule and as a transmembrane one. Upon released, TNF-α exerts its action through binding to TNF receptors (TNF-R1 and TNF-R2). Different actions of TNF-α can partly be explained by the combination of TNF-α isoforms and TNFR expression (McCoy and Tansey 2008).
Transgenic mice specifically over-expressing murine TNF-α in the CNS, either in neurons or in astrocytes, and at different stages of development, spontaneously develop chronic CNS inflammation. This is characterized by widespread reactive astrocytosis and microgliosis, meningeal and parenchymal infiltration by activated T cells and macrophages, loss of white matter, and neurodegeneration (Probert et al. 1995). On the contrary, mice lacking TNF-α or both receptors show exacerbated neuronal damage elicited by different neurotoxic stimuli (Bruce et al. 1996; Liu et al. 1998). Therefore, TNF-α can be neurotoxic or have neuroprotective effects depending on the level of expression, the type of neurons affected and other unknown factors.
Tumor necrosis factor-α in the brain has also been observed to modulate CNS functions such as peripheral activation of neuroendocrine axes, release of neurotransmitters or neuronal firing in a paracrine or autocrine fashion (Turnbull et al. 1997; Besedovsky and Rey 2007).
Tumor necrosis factor-α and its receptors are expressed in the DG (Harry et al. 2008) and the SVZ (Iosif et al. 2008). Hippocampal NSCs express both TNF-R1 and TNF-R2 (Cacci et al. 2005). As a neurotoxic factor, TNF-α increases the death of a hippocampal NSCs line in vitro (Cacci et al. 2005). In line with this, Monje et al. (2003) showed that exposure of hippocampal NSCs in vitro to recombinant TNF-α (20 ng/mL for 60 h) significantly decreased neurogenesis. TNF-α reduces proliferation of NSCs derived from the SVZ, without affecting their differentiation (1–100 ng/mL for 48 h; Ben-Hur et al. 2003). Moreover, these in vitro studies showed that 100 ng/mL TNF-α increases NSCs migration.
However, recent work has highlighted the role of TNF-α as a positive regulator of neurogenesis. In vitro treatment of neurospheres grown from adult rat SVZ with TNF-α increased the proliferation of these spheres, without affecting differentiation (10 ng/mL for 24–72 h; Widera et al. 2006). Moreover, an in vivo study showed an increase in SVZ progenitor cells 6 and 24 h after injection of TNF-α protein above the lateral ventricle (355 ng per rat; Wu et al. 2000).
Using mice in which TNF-R1 and/or TNF-R2 have been deleted, Iosif et al. (2006) showed a differential effect of these receptors on hippocampal neurogenesis. Signaling through TNF-R1 suppresses neural progenitor proliferation and neurogenesis in the adult brain in vivo, whereas binding of TNF-α to TNF-R2 increases the survival of newly formed neurons. This suggests that the effect of TNF-α would depend not only on the type of NSC studied, but also on the levels of the cytokine and the relative expression of TNF-R1 and TNF-R2 on the NSCs (see Table 1).
Interleukin-6 regulates immune responses, acute phase reactions and hematopoiesis, and plays a central role in host defense mechanisms. IL-6 is a pleiotropic cytokine that is produced in a variety of cells and acts on a wide range of tissues, exerting growth-inducing, growth-inhibitory and differentiation-inducing effects, depending on the nature of the target cells (Hirano et al. 1990). In the CNS, IL-6 and its specific receptor are both expressed on neurons and glial cells including astrocytes (Gadient and Otten 1997).
Interleukin-6 exerts pronounced effects on the CNS: stimulates neuronal growth (Satoh et al. 1988; Hama et al. 1991), reduction of food intake (Plata-Salaman 1988), induction of fever (LeMay et al. 1990) and activation of the hypothalamic-pituitary-adrenal system (Naitoh et al. 1988). Transgenic mice with high levels of IL-6 expression develop severe neurologic disease characterized by tremor, ataxia, and seizure (Campbell et al. 1993). These mice also exhibit neurodegeneration, astrocytosis, and angiogenesis. In contrast, IL-6 also supports neuroregeneration of hippocampal neurons in the CNS (Hakkoum et al. 2007).
Interleukin-6 receptor is expressed in the DG (Gadient and Otten 1994; Vallieres and Rivest 1997). Monje et al. (2003) revealed that the negative effect of activated microglial cells on neurogenesis was blocked upon addition of IL-6R antibody. Interestingly, when IL-6 is chronically expressed in astroglia of young transgenic mice, a remarkable decrease in the proliferation, survival and differentiation of NSCs was observed (Vallieres and Rivest 1997). These data suggest a clear anti-neurogenic effect of IL-6 in vitro and in vivo (see Table 1).
Functional IL-4Rs are expressed in rat hippocampal neurons and IL-4 is secreted by glial cells (Nolan et al. 2005). IL-4 can act as a neurotrophic factor increasing the survival of hippocampal neuronal cultures (Araujo and Cotman 1993). It has been proposed that IL-4 exert an anti-inflammatory effect in the brain, by down-regulating IL-1R1 and IL-1β expression (Nolan et al. 2005). Moreover, IL-4 can modulate microglial cell activation (Lyons et al. 2007), and induces a neuroprotective phenotype of microglia (Butovsky et al. 2005). This protective state appears to correlate with the down-regulation of TNF-α and with the up-regulation of insulin-like growth factor 1 (IGF-1).
Interleukin-4 had no direct effect on cultures of SVZ-derived NSCs. However, IL-4-treated microglia induces neurogenesis and oligodendrogenesis in the NSCs by different mechanisms (Butovsky et al. 2006). IL-4-treated microglia induce differentiation to oligodendrocytes via the IGF-1 signaling pathway, but the effect on neurogenesis is independent of IGF-1. In the same study it was shown that IL-4 have only limited effects on NSCs proliferation (< 10%) and co-culture of NSCs with IL-4-treated microglia had no effect on proliferation. Moreover, IL-4 treatment neither directly nor in microglia co-cultures affected NSCs survival. Thus, IL-4 seem to have a pro-neurogenic effect (see Table 1).
Transforming growth factor-β1
Transforming growth factor-β1 is normally present at low levels in the healthy adult CNS cells, and is rapidly up-regulated after injury, inducing the expression of many injury responsive genes (Finch et al. 1993). The main sources of TGF-β1 in the injured brain are astrocytes and microglia (Finch et al. 1993), but neurons can produce it as well (Flanders et al. 1998).
Transforming growth factor-β1 inhibits the synthesis of pro-inflammatory cytokines such as TNF-α, prostaglandins, nitric oxide, and eicosanoids from macrophages, adding to the protective effect of phagocytosing apoptotic cells and reducing the inflammatory response (Freire-de-Lima et al. 2006). The phagocytosis of apoptotic cells is closely linked to the down-regulation of inflammatory cytokines and also promotes the resolution of inflammation, contributing to CNS immune privilege. Microglial phagocytosis of apoptotic inflammatory T cells leads to down-regulation of microglial immune activation (Magnus et al. 2001). In line with this, it was previously proposed that TGF-β1 could contribute to the region-specific refractory response to inflammatory stimuli in specific brain regions (Depino et al. 2005). In summary, TGF-β1 generates an anti-inflammatory milieu in the brain, which contributes to the resistance of brain to inflammation.
Transforming growth factor-β1 is expressed in the DG (Bravo et al. 2006), and its expression is augmented during aging (Bye et al. 2001). Battista et al. 2006 showed in a model of enhanced neurogenesis by adrenalectomy that there was a positive correlation between increased neurogenesis, the number of activated microglia, and the levels of TGF-β1. Moreover, neutralizing antibody to TGF-β1 reduced the increase in adrenalectomy-induced neurogenesis, providing evidence for a positive role of this cytokine in this process. This same treatment did not significantly alter NSCs proliferation (Battista et al. 2006).
However, adult transgenic mice over-expressing TGF-β1 as embryonic days, exhibit decreased cell proliferation (BrdU-positive cells), lower number of NSCs (nestin-positive) and neural progenitor cells (doublecortin-positive) in the hippocampus (Buckwalter et al. 2006). The number of proliferating progenitor and newborn neurons was also decreased, most likely because of the dramatic effect of TGF-β1 expression on cell proliferation (60% reduction). In these animals, a reduction on the pool of NSCs because of effects of TGF-β1 during development that could account for these results could not be discarded. In a third study, the chronic intracerebroventricular infusion of TGF-β1 (6 ng/day for 7 days) reduced NSCs proliferation in the DG, but it did not change the differentiation fate to neurons in this region (Wachs et al. 2006).
In the SVZ, TGF-β1 transgenic expression or chronic intracerebroventricular infusion (6 ng/day for 7 days) resulted in decreased cell proliferation and had no effect on NSCs differentiation to a neuronal phenotype (Buckwalter et al. 2006; Wachs et al. 2006). However, TGF-β1 was able to increase proliferation and neuronal differentiation if the SVZ area was affected previously by a stroke lesion (intranasal delivery of 1 μg of recombinant human TGF-β1; Ma et al. 2008). In addition, adenoviral mediated delivery of TGF-β1 to the SVZ (a treatment that also results in a minor injury to the region injected) also resulted in an increased cell proliferation and differentiation toward a neuronal phenotype (Mathieu et al., unpublished observations). Thus, as for other cytokines, TGF-β1 effects on adult NSCs proliferation and differentiation in vivo are context-dependent. However, a tendency toward inhibiting cell proliferation (either during development or in the adult) and increasing neurogenesis in the intact DG and SVZ emerges (see Table 1). In addition, the contribution of other stimuli to the experimental set-up such as infarction or adenoviral vector delivery could have an opposite effect on NSCs proliferation.
In vitro, all studies revealed an inhibitory effect of TGF-β1 on NSCs proliferation irrespective of their origin (SVZ or DG) and culture conditions. The proportion of cells expressing a neuronal marker increased after incubating adult NSCs from the DG or the SVZ in the presence of human TGF-β1 and low (0.5 ng/mL) FGF-2 with or without 1% Fetal calf serum and 0.5 mM forskolin (Battista et al. 2006 and unpublished observations). In another study, NSCs isolated from the SVZ did not show any change in cell differentiation to any phenotype in the presence of 10 ng/mL of TGF-β1 and 20 ng/mL of FGF-2 or 5% fetal calf serum. Thus, culture conditions are important to visualize the pro-neurogenic potential of TGF-β1 in vitro (see Table 1).
Effects of inflammation on transplanted stem cells
The knowledge on the functional role of brain cytokines on the proliferation, differentiation and/or survival of NSCs should also help to understand the effects of cytokines on transplanted NSCs.
Following cell transplantation in a neurodegenerating brain, the host recognition of the transplanted cells will activate an innate immune response and liberate cytokines, among other mediators. This response will not only be affected by the cell type, cell number and architecture of the transplant but also by the previous inflammatory characteristics of the diseased host tissue. Increasing complexity, transplanted NSCs could also have immunomodulatory effects per se. Thus, there are three basic components of the inflammatory response to transplantation: the transplanted cell, the diseased host tissue and the interactions between them. To identify and study the functional role of key inflammatory mediators of these three components is crucial to understand and try to predict this response and, in turn, imagine possible complementary treatments to transplantation that could enhance the possibility of therapeutic success. The previously reviewed effects of cytokines on NSCs biology highlight the relevance of this knowledge to optimize the functional outcome of a cell-based therapy.
Here, we will first review the basic characteristics of the inflammatory component in a well-known neurodegenerative disease, PD, then the effects of transplantation in this disease and the consequences of the inflammatory environment in transplanted cell survival, differentiation and migration. Finally we will discuss the inherent immunomodulatory components of NSCs.
Inflammation in Parkinson’s disease
Parkinson’s disease is characterized by a progressive degeneration of dopaminergic neurons in the substantia nigra pars compacta (SN). The degenerated neurons present intracytoplasmatic inclusions or protein aggregates called Lewy’s bodies, a feature characteristic of PD. The loss of dopaminergic neurons causes most of the motor symptoms of the disease, which can be alleviated by restoring neurotransmission with the dopamine precursor levodopa or with dopaminergic agonists (Lees et al. 2009).
It has been demonstrated that microglial cell activation is a primary component in the pathogenesis of the disease. In PD, robust microglial cell activation is consistently found in animal models and PD patients (McGeer et al. 1988; Hunot et al. 1999; Mogi and Nagatsu 1999; Mirza et al. 2000; Vila et al. 2001; Cicchetti et al. 2002; Depino et al. 2003; Hirsch et al. 2003). In addition, the substantia nigra is the brain region with the highest density of microglial cells in the brain (Lawson et al. 1990), neurons from the substantia nigra are particularly susceptible to microglial-mediated toxicity in vitro and in vivo (Castano et al. 1998, 2002; Kim et al. 2000; Liu et al. 2000) and this effect could be reversed by the administration of anti-inflammatory drugs (Liu et al. 2000; Castano et al. 2002; Sanchez-Pernaute et al. 2004; Hald and Lotharius 2005; Marchetti and Abbracchio 2005). In patients, microglia seem to be activated for long periods of time (Langston et al. 1999).
In animal models, we have previously shown that during neurodegeneration in the substantia nigra, microglial cells change to a ‘primed’ state in which pro-inflammatory cytokine transcription is increased but translation is blocked (Depino et al. 2003). This microglia overreacts when a central or peripheral pro-inflammatory stimulus is elicited, leading to an exacerbation of neurodegeneration, motor symptoms and increased production of IL-1β (Godoy et al. 2008). If cells are transplanted in an area with ongoing neurodegeneration and primed microglia, it can be speculated that anti-inflammatory measures had to be enforced to prevent a deleterious effect from the reactivation of primed microglia.
Transplantation in Parkinson’s disease
The first transplants performed in human brains included solid pieces of human fetal or embryonic ventral mesencephalon. One of these studies showed clinical improvement at 5 months up to at least 3 years after surgery (Li et al. 2008). Its postmortem analysis demonstrated grafted cells survival and dopaminergic differentiation. Even though symptomatic relief was observed, the grafts have shown characteristics of the disease such as α-synuclein- and ubiquitin-positive Lewy’s bodies (Lindvall et al. 1989; Li et al. 2008). In 2001, Fred et al. reported a clinical trial focus on transplanted human embryonic dopamine neurons into the brains of PD patients. The evaluation 1 year post-grafting revealed some clinical improvements in younger but not in older patients (Freed et al. 2001). It is important to comment that besides the partial benefit, five over 40 patients of the trial develop dyskinesias more than a year after the transplant.
Likewise, in a double-blind, placebo-controlled trial of fetal nigral transplantation using one or four donors, the motor features of PD patients did not significantly improve in comparison to the placebo group but there was an increase in striatal fluorodopa uptake on positron emission tomography scan and evidence of surviving transplanted cells (Olanow et al. 2003). Other studies had similar results with an initially marked improvement that gradually deteriorates. This deterioration may have been because of graft failure caused by the PD process or an immune reaction, as suggested by the extensive microglial cell infiltration of the graft (Kordower et al. 2008). However, a different study where three subjects with advanced idiopathic PD received intracerebral transplantation of fetal ventral midbrain cell suspension grafts showed reduced inflammation response, grafted dopamine and serotonin neurons survival and not significant degeneration for a long term (Mendez et al. 2008). These results from trials of human transplants in PD patients suggest that grafted cells might be affected by the disease environment and that the variables that affect the inflammatory response are still unclear.
Transplant survival, differentiation, and migration
To assess the immunogenicity of allogeneic fetal neural dopamine tissue Widner et al. (1989) developed different donor–host transplants combinations in animal models; 15 weeks after grafting, there were no signs of rejection regardless whether grafted cells were syngeneic or allogeneic, or when different histocompatible donor–host combinations were performed. None of the animals showed considerable infiltration of immune cells and there were no difference in the amount of immune cells between the different groups. Another study showed that transplants of allogeneic or syngeneic embryonic mesencephalic rat cells in an intact or toxin-induced inflammated animal tissue displayed similar graft survival. More abundant infiltrates of macrophages, activated microglia, and astrocytes where present around both syngeneic and allogeneic grafts in toxin-lesioned striatum, than around cells transplanted in intact tissue (Duan et al. 1998). These studies proved that the immune privilege of the brain permits a different response to syngeneic and allogeneic grafts than the one expected in the periphery.
In the case of multiple transplantations, timing between interventions seems to affect the immune response and the final outcome of the treatments. A first, intracerebral allograft into unilateral 6-hydroxydopamine (6-OHDA) lesioned animals increased the host immune response against a second allograft 8 weeks later, resulting in a lower survival of the second graft (Duan et al. 1993). On the contrary, when repeated allografts were made with 14 weeks in between them, there was no significant difference between the survival of the first and the second allogeneic graft. Moreover, the immune infiltration around the second grafts was low and displayed a microglia/macrophage-like morphology (Duan et al. 1993; Widner and Brundin 1993).
The environment where cells are transplanted could have an effect on the differentiation of cells. For example, in a previous work it was demonstrated that cells obtained from adult rat hippocampus transplanted into an adult neurogenic area such as the hippocampus differentiated in much greater number into mature neurons or glia compared with in vitro assays. This finding indicates that there are signals within the adult brain that are crucial for the differentiation of the progenitors (Gage et al. 1995). In 1999, Fricker et al. demonstrated that human neuronal progenitors from embryonic brain tissue transplanted in adult rat SVZ migrated to the olfactory bulb, and expressed neural markers. When these cells were grafted in the striatum, part of them differentiated into glia, but they also underwent neurogenesis in this non-neurogenic environment and assumed neuronal phenotype similar to those normally present there. The transplanted cells remain close to the injection site in absence of suitable substrate for migration. The authors concluded that the long-term propagated embryonic neuronal progenitors are capable of migration, integration and site-specific differentiation in the adult brain (Fricker et al. 1999). Other work, where adult brain subventricular zone progenitor cells were transplanted into the striatum of adult rats showed that adult progenitor cells can also differentiate in a region-specific manner to striatal neurons (Zhang et al. 2003). In contrast to these studies, a plethora of evidence proved that adult NSCs from hippocampus or subventricular zone can only produce neurons when they are transplanted into neurogenic regions, and mainly produce glial cells when transplanted into non-neurogenic regions, such as the striatum (Gage et al. 1995; Herrera et al. 1999; Dziewczapolski et al. 2003; Chen et al. 2007). The differences could be explained in part by the neuronal compromise of the cells before transplant, as was characterized by Zhang et al. (2003) were the majority of the cells were already committed neuronal precursors (Tuj-positive-cells). It is still unclear whether NSCs and progenitor cells at different stages of differentiation elicit different inflammatory responses when transplanted in the brain. There is also evidence lacking on the differences on the inflammatory response elicited by a certain type of transplanted cell depending on the brain region transplanted.
In the same way, the environment generated by brain lesions could also dictate the fate of transplanted cells. Embryonic rat striatal cells transplanted into the adult rat striatum lesioned by excitotoxicity showed extensive migration into the surrounding host tissue and increased cellular proliferation as a response to signals from the lesion. The majority of grafted cells assumed glial-like morphology and only a very small fraction developed neuron-like characteristics (Lundberg et al. 1997). On the other hand, low numbers of transplanted mouse embryonic stem cells developed into mature ventral midbrain-like dopaminergic neurons and successfully reduced motor asymmetry when cells were transplanted into the lesioned rat striatum. The applications of low numbers of embryonic stem cells might result in dilution of epidermal growth factors and other germ layer-inducing signals, and cell–cell contacts that would facilitate neuronal differentiation (Bjorklund et al. 2002).
These evidences support the knowledge that the survival and differentiation of transplanted cells are affected by the milieu where they are transplanted. There is still systematic studies lacking to show what cytokines during an inflammatory state have a main impact on the fate or survival of transplanted cells. On the other hand, the role of chemokines on migration of transplanted cells has been better documented.
Experiments with human neural stem cells from fetal brain tissue showed that cells transplanted into the cortical region of adult uninjured rats showed no spontaneous migration, but grafts performed in injured brains by a distal middle cerebral artery occlusion showed migration toward the lesion. Although human neural stem cells can moved appreciable distances to regions of pathology and migrated within the impaired region, their migration depended on the distance between the graft and the lesion. If this distance is too large there will not be targeted migration (Guzman et al. 2007; Bjugstad et al. 2008). Migration toward the area of lesion was described in different disease models; adult NSCs migrated to sites of demyelination when they were intravenous injected into an animal model of Multiple Sclerosis (Pluchino et al. 2003), or NSCs lines transplanted in a mouse model of cerebral ischemia showed migration toward the infarcted area (Imitola et al. 2004). In contrast to these results, it was demonstrated that SVZ cells from adult rats grafted into the intact striatum migrated a distance about 1.5 mm from their site of transplant in the adult brain. The authors suggest that in addition to extrinsic cues, during culture conditioning, these cells may lose the intrinsic factors that inhibit their migration in a heterotypical region (Zhang et al. 2003). The molecules that have been implied in migration of neural stem cells are chemoattractants such as stromal cell-derived factor-1α (SDF-1) and MCP-1. Human neural stem cells express the chemokine receptor 4, the receptor for SDF-1α, and transplanted in hypoxic-ischemic cerebral injured mice showed migration that correlated with the area of SDF-1α expression. Inflammation may also activate astrocytes and endothelial cells. These glial cells produce SDF-1α that may direct the migration of NSCs toward the injure site (Imitola et al. 2004). MCP-1 is important for leukocyte traffic into the brain and might play a similar role in directing migration of progenitor cells. The up-regulation of MCP-1 synthesis is produced by neuroinflammation and together with the activation of its receptor, the chemokine receptor 2 expressed by neural progenitors, may be important in guiding the NSCs migration (Belmadani et al. 2006). Another factor that was proposed to be involved in transplant migration is stem cell factor (SCF). SCF expression is induced by brain injury in neurons surrounding the lesion. Besides, it was demonstrated that recombinant SCF triggers the migration of NSCs in vivo and in vitro, and c-kit (the SCF receptor) was detected on NSCs and was activated after SCF exposure (Sun et al. 2004).
One main feature to obtain a true regenerative therapy is to reestablish damaged circuits in the CNS. One main, although not exclusive, area of neurodegeneration is the SN. Loss of neurons in this area affects their target area, the caudate–putamen, resulting in a deficit in dopaminergic transmission in that area. Because of the lack of information to force transplanted cells in the SN to connect to the target area, the main site of transplanted strategies in PD is the site where dopamine is required (i.e. the caudate–putamen). Thus, most therapeutic strategies against PD are focused on replacing dopamine locally in the caudate–putamen, lacking the physiological control of dopamine release of a functional nigrostriatal system. Seminal work by Alberto Aguayo presented the proof of principle that axonal elongation is inducible under defined conditions (Benfey and Aguayo 1982). Yet, there is still not enough knowledge available to design efficient and reliable strategies to guide axons to target areas in vivo.
Immunomodulation by neural stem/progenitor cell
Stem cells from different origins were proved to reduce the inflammation response in distinct neurodegenerative models. Human NSCs lines from fetal brain and postnatal SVZ cells injected intravenously in an intracerebral hemorrhage model displayed anti-inflammatory functions that protected the tissue by interrupting the splenic response after intracerebral hemorrhage (Lee et al. 2008). The neuroprotective effect was associated with modulation of immune recognition, as these cells produced TGF-β under basal conditions. Similarly, an anti-inflammatory action was observed when mouse neural stem cells injected intravenously in a mouse model of Sandhoff’s disease migrated extensively trough the animal body and distributed mainly in systemic organs as liver, spleen and lung. However, only a small number of injected cells were found in the brain. The authors proved that NSCs reduced mRNA levels of TNF-α and IL-6 in the brain and spleen, and it a lower number of microglia in NSCs injected animals was found. So, in this model cerebral inflammation was modified by the interaction of intravenously injected NSCs with the peripheral inflammatory system (Lee et al. 2008). Moreover, the injection of adult NSCs either i.v. or i.c.v. in an animal model of multiple sclerosis, experimental autoimmune encephalomyelitis, showed that transplanted cells can reach the brain and produce functional recovery (Pluchino et al. 2003). The amelioration of the symptoms was demonstrated to be because of a double effect: a direct remyelination by the differentiation of transplanted NSCs into oligodendrocytes and by a bystander regulation of endogenous oligodendroglia and the reactive astrogliosis. The authors detected levels of mRNA for FGF-2, TGF-β, ciliary neurotrophic factor, glial-derived neurotrophic factor, nerve growth factor, BDNF, leukemia inhibitory factor, and neurotrophin-3 in cultured neural precursors, which might be responsible for the bystander effects (Pluchino et al. 2003). In another example, human NSCs transplanted in MPTP-treated, dopamine depleted monkeys produce behavioral improvements up to 60 days after transplant. The results showed prolonged cell survival (more than 7 months) but only a 5.47–8.03% of tyrosine hydroxylase (TH)-positive, dopaminergic grafted cells. This percentage of TH-positive neurons should not be enough to explain the behavioral benefit observed. Thus, human neural stem cells appeared to exert protective effects that may be displayed by the transplanted cells that acquired astrocyte phenotype and expressed glial-derived neurotrophic factor (Redmond et al. 2007). Also in an middle cerebral artery occlusion model, adult or embryonic neural stem cells transplanted differentiated into astrocytes and the therapeutic effects detected might also be because of a trophic factor secretion rather than replacement of damaged neurons (Takahashi et al. 2008). Human fetal-derived neural stem cells line transplanted after 6-OHDA injection in the middle forebrain produced behavior improvement in apomorphine-induced rotations 4 weeks after the graft. TH staining of the striatum and the SN was also significantly preserved. Even though the graft survival was 1.1% and there were no TH-positive transplanted cells, in an in vitro assay the authors observed neuroprotection by conditioned NSCs medium against 6-OHDA neurotoxicity. This effect could be explained by a reduction in apoptosis. In this conditioned medium it was demonstrated the presence of SCF and BDNF, two growth factors that could be responsible of the observed beneficial effects (Yasuhara et al. 2006).
As reviewed here, cytokines can modulate NSCs proliferation, differentiation, and survival. As brain cytokines can be constitutively expressed in neurogenic regions, and NSCs express cytokine receptors, endogenous NSCs can be subjected to the functional influence of endogenous brain cytokines. We believe that enough evidence is available to consider brain cytokines as part of the endogenous neurogenic niche. This field of research is just emerging and as most cytokine actions are context-dependent, data are not consistent for all models studied. However, when all available data are gathered together, a picture emerges where pro-inflammatory cytokines have mainly an anti-neurogenic effect and anti-inflammatory cytokines have pro-neurogenic functions (see Fig. 3). This concept will be validated or proved wrong by accumulating data obtained in future studies.
In addition, cytokines will be produced after NSCs grafting in the brain. As in the case with endogenous NSCs, cytokine production triggered by cell transplantation will undoubtedly affect the proliferation, differentiation, and survival of the grafted cells, and the final therapeutic effect of cell transplantation. The study of the inflammatory reaction expected after grafting remains largely unexplored. We believe that considerable efforts should be taken to try to elucidate main basic rules governing the inflammatory reaction after grafting. Here, we show evidences that these rules will depend upon the type of NSCs studied and their degree of differentiation, the microglial state of activation of the diseased brain area to be grafted, the timing between graftings and the dilution of the number of cells transplanted, among other factors. The discovery of the factors determining the type of inflammatory reaction after transplant is challenging as basic rules governing CNS inflammation are just starting to be elucidated. As a first step, we propose to clearly differentiate microglial cell activation, which could be anti- or pro-inflammatory to brain inflammation which also includes pro-inflammatory immune cell recruitment from the periphery. This view will help to prevent the automatic extrapolation of results showing microglial cell activation to a pro-inflammatory mileu. In addition, microglial cell activation should be studied in concert with the identification of its secreted molecules, which will define its net effect on a given phenomenon. Furthermore, the effect of the grafted NSC per se on the fate of the graft by secreting neuroprotective and growth factors should not be disregarded.
Brain transplantation of NSC at different levels of neuronal differentiation is regarded as a promising new avenue of therapeutic opportunities against many neurodegenerative diseases. Certainly, many challenges should be faced to translate this type of approach to the clinic. No doubt, that the better we understand the inflammatory response after grafting and its modulatory role on the transplant, the closer we will be to design therapeutically effective interventions. Studies to elucidate not only the role of the transplanted cell, the state of the brain tissue at the moment of grafting and the interactions between transplanted cells and host tissue, but also basic studies on the functional role of endogenous cytokines on adult neurogenesis should pave the way for a better planning of immunomodulatory treatments aiming at increasing therapeutic efficacy.
We would like to thank Fabio Fraga and Isabel Farías for their excellent technical support during the work on IL-1 and neurogenesis and Estela Lafita for the schematic drawings. The data presented in this work was obtained with the financial help of the René Barón Foundation and the Agencia Nacional de Promoción Científica y Tecnológica through its PAE program.