The central nervous system (CNS) develops from a small number of cells proliferating and interacting in a very particular manner in order to form a functional neural network with regional identities over hundreds of billions of cells in human beings. In this dynamic process, neuroepithelial cells—or the primary neural stem cells (NSC)—are converted from a simple neural plate to originate a brain and a spinal cord during a short period of time in embryogenesis.
In disagreement with the previous view of the non-neurogenic nature of adult CNS, we have learned over the past 40 years that the brain is capable of generating new neural cells and multipotent precursors that can incorporate into its complex circuitry and remain in many regions of the adult mammalian brain (1–3).
Historically, the first suggestion of the existence of dividing cells in the adult CNS was raised by Hamilton (4). Since no credits were given to him, the nonexistence of new postnatal neurons became a central dogma in neuroscience for almost a century (5). However, with the introduction of the tritiated-thymidine method based on [3H]-thymidine incorporation into DNA of dividing cell and posterior spatial detection by autoradiography (6), several studies were published reporting the generation of new neurons in postnatal mouse brains (7), including the hippocampus (8), neocortex (9) and olfactory bulb (10). The origin of these new neurons was attributed to the existence of a cluster of mitotically active undifferentiated cells in some regions of the brain.
Some years later, Nottebohm and coworkers demonstrated neurogenesis during the singing season in the vocal control nuclei of the songbirds' male brains (11, 12). Meanwhile, a series of studies from Kaplan and Hinds observed that newborn neurons in the hippocampus of 3-month-old rats survived for a long period of time and received synaptic inputs (13).
In the 90s, Reynolds and Weiss showed that cultures of adult mouse striatum proliferated and underwent differentiation in the presence of epidermal growth factor (EGF), opening new fields for in vitro studies of adult neurogenesis in view of the capability of inactive stem cells to persist inside the adult CNS (14). In this context, the Gage's group was the first to use 5-bromo-2-deoxyuridine (BrdU) labeling to study adult neurogenesis. “It occurs specifically in an area involved in memory formation and thus may contribute to the optimal functioning of this region” (15).
There are two major neurogenic regions of the adult mammalian brain including that of humans, which have received intensive attention: the subgranular zone of the dentate gyrus and the anterior part of the subventricular zone (SVZ) along the ventricle (16). However, the presence of proliferative cells in the intact adult CNS or in multiple regions of the hippocampus has been shown in many reports, even though some of these cells revealed extremely limited neurogenic potentials (17). Moreover, studies have revealed that neurogenesis occurring in the SVZ generates new neurons that migrate not only to the olfactory bulb (1), but also to neocortical areas of primates (18).
Even with a good knowledge of the proliferative regions in the adult brain, the major subject is whether these cells qualify as stem cells. Stem cells display two major features: capability of self-renewal to maintain the pool of undifferentiated stem cells and to give rise to differentiated progeny; and the multipotency to generate cells of different lineages. NSC are capable to produce neuronal and glial subtypes [for review, see (19, 20)]. In spite of its definition, the term “stem cell” is used to refer to three concepts that considerably differ: embryonic, fetal and adult stem cells. The difference among these sources lies in ethical issues, self-renewal capacity and differentiation potential.
Embryonic stem (ES) cells are undifferentiated lines established from epiblasts in the inner cell mass of an early embryo at the blastocyst stage and can be maintained in vitro. Epiblasts differentiate into three germ layers that will later form the body in the developmental process. ES cells can be induced to differentiate into various cells derived from the three germ layers under distinct culture conditions. Therefore, human ES cell lines have been established as the future of therapeutic medicine. These cells provide promising cell therapies for a wide range of diseases (21), and may also be utilized as unique source of human tissue for testing new drugs in vitro. However, there have been significant moral and ethical issues surrounding their isolation and the use of surplus embryos from reproductive medicine (22). Furthermore, because of their pluripotency, ES cells may develop into teratomas when injected into imunosupressed mice (23).
As a new approach to obtain pluripotent stem cells, in 2007, two separate teams transformed human skin cells into cells that are virtually indistinguishable from human ES cells. The team led by Shinya Yamanaka at Kyoto University showed that differentiated human cells could be reprogrammed to an embryo-like state using cultured human skin cells (fibroblasts) infected with viruses engineered to introduce four genes (oct3/4, sox2, klf4, and c-myc). The resulting colonies resembled human ES cells, and those cells were called induced pluripotent stem (iPS) cells. The iPS cells were able to differentiate into the three major cells types in vivo and in vitro (24). In the same month, James Thomson, a group leader at the University of Wisconsin, and his co-worker Junying Yu published a similar approach using fibroblasts cultured from skin cells from a newborn boy and a fetus. They also found a quartet of genes that could transform differentiated human cells into a versatile state. Instead of klf4 and c-myc, this set included the genes for nanog and lin28. The generated colonies revealed four cell lines behaving like ES cells in culture (25). Just after publication of those data, Yamanaka published an additional manuscript, showing that both human and mouse adult fibroblasts could be reprogrammed using only three of the four genes. This finding presented significant novelty, as it was possible to eliminate c-myc, a possible cancer-inducing gene (26).
As mentioned earlier, studies in both rodents and humans have shown that different adult tissues may also maintain stem cell populations (27–31), although these cells are not as easy to manipulate and grow in culture as their ES counterparts. Between these edges, we can find the NSCs or neural progenitor cells isolated from the developing brain.
Since they were first described by Reynolds and Weiss (14), neural progenitor cells have been the focus of a growing number of studies because of their potential in the study of neurogenesis and development of future therapy in neurological disorders. In this study, Reynolds and Weiss isolated striatum cells from adult mouse brain and induced them to proliferate as free-floating spherical clonal expansion (neurospheres) in the presence of EGF. In their undifferentiated stage, cells originally express nestin (stemness marker) and following attachment to culture dishes, they change their morphology and antigenic properties to those of neurons and astrocytes. These new neural cells were immunoreactive for different neurotransmitters of the adult striatum, such as gamma-aminobutyric acid (GABA) and substance P. Thus, Reynolds and Weiss demonstrated for the first time that cells from the adult brain have the capacity to divide and differentiate into neurons and astrocytes.
Several new methods and protocols have been established to identify, isolate, and characterize live NSC in terms of their capability to proliferate and differentiate (32). NSC can be isolated by FACS based on physical properties such as size (forward scattering), granularity (side scattering;33, 34), and expression of surface antigens such as CD24 (35) and CD133 (36) (Table 1). Recent studies indicate that human CNS precursor cells expressing high levels of CD133 with little or no expression of CD24 had the highest frequency of initiating neurospheres (37).
Table 1. Surface markers frequently used for NSC isolation and characterization by FACS
ORIGIN CELLS OR TISSUE
Embryonic forebrain-derived cells
Enrichment of neural stem/progenitor and neurosphere-initiating cells
Panchision et al. (38) recently published an interesting overview of an optimized flow cytometry method for NSC isolation, dissociation and characterization based on differences in surface antigen expression. Many methods of cell dissociation have been successfully used for neural tissue, including mechanical (39) and enzymatic dissociation with papain (40), trypsin (41), and collagenase (42). In this context, Panchision et al. (38) established that the best balance of dissociation, antigen preservation, and viability of NSC was obtained by using Liberase-1 and TrypLE™ (Invitrogen). The work also showed that differently from NSC, neural progenitor cells in addition to exposing CD24 on their cell surfaces expressed higher mean levels of CD133 and CD15 than multipotent stem cells of the fetal mouse brain did. Subsequently, cell separation based on CD24 coexpression increased the isolation efficiency of these populations. Recently, Brewer and Torricelli (43) reviewed methods for extraction of neurons from adult rodent CNS with density gradient separation for adult neuron and neurosphere isolation and culture.
Following initial isolation NSC may be maintained in vitro and grow as monolayer or neurospheres (44). A detailed protocol for rat NSC isolation and neurospheres culture is provided in Figure 1. Neurospheres contain low numbers of “true” stem cells and many more restricted progenitors. As some of these cells are not committed to differentiate into a specific phenotype, the effects of genetic or environmental signals on their development can be assessed in vitro for the elucidation of mechanisms underlying neuronal and glial development. Furthermore, as cell therapy for neurological disorders advances, these cells may prove to be the ideal candidates for neural transplantation in diseases such as Parkinson's (45). In recent studies, the surprising observation has been made that mouse NSC are also pluripotent when exposed to the correct environmental signals. They can produce blood cells when injected into irradiated mice (46) and cells of many different mesodermal lineages following injection into mouse blastocysts or chicken embryos (47). Thus, the therapeutic potential of these interesting cells may be far wider than previously considered, making their characterization of even greater importance.
Neural differentiation of NSC has been extensively studied in vitro. After 7 days of differentiation, a radial pattern of cell migration can be visualized, and neuron-specific β-3 tubulin expression being present at low levels in undifferentiated neurospheres, significantly increased (Fig. 2). The percentage of nestin-positive cells decreases following induction of NSC to differentiation. According to Martins et al. (41), flow cytometry analysis confirms differences in expression of these proteins between undifferentiated and differentiated neurospheres. β-3 tubulin-positive cells were detected in 25% of undifferentiated and 71% of differentiated cells. Glial fibrillary acidic protein (GFAP, astrocytic marker) and nestin were expressed in 20% and 35% of undifferentiated cells, respectively. The percentage of GFAP-positive cells in differentiated neurospheres increased to 35%, while nestin-positive cells decreased to 16% of the total cell population. In this process, the number of galactocerebroside (Gal-C, oligodendrocyte marker)-positive cells slightly augmented from 5 to 8% of the total population. Moreover, McLaren et al. (33) confirmed the existence of a similar expression pattern of these specific stage marker proteins in undifferentiated and differentiated NSC by flow cytometry. In addition, their results indicated that astrocytes obtained from differentiated neurospheres expresses higher levels of MHC than undifferentiated cells or neurons do. Therefore, removal of those glial cells using physical parameters may be beneficial prior to cell transplantation.
NSC Differentiation: A Matter of Cell Fate
Neural-cell diversity arises in embryonic development from proliferative regions nearby the neural tube. In adult brain, new neurons are continuously added to the neural circuits in restricted areas. The molecular mechanisms and the cross-talk between multiple signaling pathways, transcription factors, and cell–cell interaction controlling the cells fate have lately turned into the focus of attention. These studies revealed paracrine interactions among astrocytes, radial glial cells, and neural progenitor cells. These interactions can promote as well as control neurogenesis and gliogenesis of resident precursor cells in their niches (48). Furthermore, extrinsic factors (such as neurotransmitters, hormones, and growth factors) and intrinsic factors (transcription factors, DNA rearrangement, RNA editing and epigenetic modifications of proteins), instructing these multipotent progenitors to generate neurons, astrocytes or oligodendrocytes, have been extensively investigated (49) [for a review of neuronal variability, see (50)].
The direction of cell fates into neurons or glia is also related with symmetric and asymmetric cell division, characterized by equal and unequal cytoplasmic distributions to progeny. Symmetric or proliferative cell division results in a rapid expansion of the progenitors. Neurogenesis starts by a change to asymmetric division (or neurogenic division), whereby stem cells produce another stem cell and a neuron or a neural progenitor. Asymmetric division is important for the generation of cellular diversity in the mammalian cortex (51, 52). Glial cells, in particular radial glial cells, exert special notable roles in the process of cell fate determination and have been more recently considered as a source of neuronal progenitors (53). According to Doetsch (54), radial glial cells comprise a widespread non-neuronal cell type in the developing CNS, appearing at the onset of neurogenesis and giving rise, directly or through progenitor formation, to most neurons of the cortex. In the embryonic telencephalon, radial glial cells reach the inner ventricular zone, extend radial processes at the pial surfaces of the neural tube and divide at the ventricular surface. Radial glia shows interkinetic nuclear migration during the cell cycle (55). The complexity of this mechanism and cell biology of neural differentiation was extensively reviewed by Gotz and Huttner (56).
In view of these recent findings, we are living one of the biggest revolution in the NSCs field as glia can act as NSCs (57). The traditional opinion was that neuroepithelial cells from the early neural tube generate two separate groups of glial and neuronal progenitors. The current accepted view of neurogenesis exhibits that neuroepithelial cells at the base of the lineage develop into radial glia and then into astrocytes. Neurons and glia are produced by asymmetric division of radial glial cells, which occurs directly or through an intermediate precursor. Radial glial cells disappear after birth in mammals, when they are thought to transform into astrocytes (58).
Even with recent data providing evidence that NSCs have some characteristics of glia, it is extremely improbable that all glial cells are NSC. The quest of high efficacy cell-specific markers and isolation methods together with investigations into the mechanisms of extrinsic factors contributing to neurogenesis, such as neurotransmitters, growth factors, and their receptors, will reveal how these cells “decide” their fates.
NEUROTRANSMITTERS AND NEUROPEPTIDES IN NEURAL DIFFERENTIATION
In this part of the review, we focus on the kallikrein-kinin system, and cholinergic and purinergic signaling and show the interrelationships among these three neurotransmission systems favoring the differentiation of neural cells and their roles in brain development.
Kallikreins and kinins form a complex system contributing to control of blood pressure, coagulation, pain, and development (59–61). The kallikrein-kinin system consists of a large number of proteins such as high- and low-molecular weight kininogen, enzymes as tissue and plasma kallikreins, and the biological active peptides bradykinin (BK), kallidin, des-Arg9-BK, and des-Arg9-kallidin. The peptides mediate their effects via kinin-B1 and -B2 receptors (B1BKR and B2BKR) (62).
The expression of tissue kallikrein mRNA was found in both prenatal and adult rat brains. During brain development, expression of kallikrein increases in different locations suggesting a possible role in neurogenesis (63). By using adenovirus as a vehicle to enhance kallikrein expression, Xia et al. (64) detected increased cerebral levels of nitric oxide, phosphorylated AKT, and Bcl-2, in addition to a decrease in neuronal and glial apoptosis from stroke induced by middle arterial cerebral occlusion. These neuroprotective effects were abolished after treatment with a specific antagonist of B2BKR, HOE-140 (64, 65). Moreover, tissue kallikrein specifically stimulates proliferation of NSC that can generate new neurons. This proliferation-inducing effect was not observed in non-neural HeLa cells and in PC12 and GH3 cells (66).
Regulation of BK secretion and B2BKR expression during differentiation has been observed in various in vitro models. Expression of B2BKR and BK secretion remained undetectable in undifferentiated P19 embryonal carcinoma cells, a stem cell model, while receptor expression and activity as well as generation of BK rose to maximal levels after 7 days of differentiation. B2BKR activity also participated in differentiation of P19 cells to neurons with a cholinergic phenotype, as carbamoylcholine (carbachol)-induced [Ca2+]i transients and gene expression of M1, M2 and M3 muscarinic receptors were suppressed following chronic treatment of differentiating cells with HOE-140 (67). Different from results obtained with P19 cells, secretion of BK and B2BKR expression decreased during ongoing neurosphere differentiation to neurons and glial cells (41). Upregulation of B2BKR has not only been observed in neurogenesis but also in other developmental processes. B2BKR expression increased during ES cell differentiation into an epithelial lineage (68).
Acetylcholine (ACh) exerts its effects on the CNS and peripheral nervous system (PNS) throughout two distinct subtypes of receptors: Muscarinic AChRs (mAChRs) and muscle- and neuronal-type nicotinic AChR (nAChR). Neuronal nAChR subtypes are located in pre- and postsynaptic membranes in autonomic ganglia and in cholinergic neurons throughout the CNS, including mediobasal forebrain, brain stem, cerebral cortex, and hippocampus (69, 70). It is well accepted that brain nAChR contribute to complex functions, such as attention, memory, and cognition. Moreover, clinical data suggests their involvement in the pathogenesis of several disorders such as Alzheimer's disease (71), schizophrenia (72), Parkinson's disease (73), autoimmune autonomic neuropathy (74), and hereditary epilepsies (75). In this context, neuroprotective actions of nAChRs have been very well characterized (76). However, nicotine-exerted functions on developing neurons have not yet been thoroughly investigated.
Nicotinic receptors subunits are among the first membrane proteins to appear during CNS development. Gene expression of several nAChR subunits and their functional properties change during neural differentiation and maturation. In the rat cortex, the α3 subunit has been detected in E12, while α4 prevails after E18 (77). The α4 and α7 subunit proteins have been observed in the ventricular zone as early as E14. However, since active neuronal nAChR are assembled by five subunits (α2-α10 and β2-β4) as homomeric or heteromeric receptors, detection of subunit expression alone is not sufficient evidence for receptor activity. In this context, Atluri et al. (78) provided proof for functional nicotinic ACh receptors in fetal mouse cerebral cortex as early as on E10, as revealed by patch-clamp measurements. The obtained results revealed that nicotine and ACh evoked sizable inward currents characteristic for nicotinic receptors, as these currents were blocked by nAChR antagonists. The Ca2+ response to nicotinic agonists was markedly prolonged in cells from early embryonic stages when compared with cells from later stages of development.
The effects of nicotine on developing neurons have been studied in various systems. The conclusion of these studies was that nicotine-induced effects depend on the site of action. For instance, in hippocampal formation with high expression of α7 nAChR, characterized by the high Ca2+ permeability of formed channels, Ca2+ fluxes exerted by high nicotine concentrations may be cytotoxic and therefore reduce neurogenesis. In agreement, chronic treatment of rats with increasing concentrations of nicotine in doses of 0.1–1 mg/kg by intraperitoneal injection for 2 weeks revealed dose-dependent reduction in expression of the neural cell adhesion molecule (PSA-NCAM) and NeuN expression in hippocampal formation, indicating a decrease in hippocampal neurogenesis (79). However in postnatal rats (P0–P7) chronic nicotine stimulation of α4β2 nAChR resulted in reduction of apoptotic cells in CA3 strata oriens and radiatum and cerebellar granula (80). Thus, the demonstrated examples suggest that chronic nicotine application may exert opposite effects in different brain regions, depending on respective nAChR subtype expression in those regions.
Key roles in cell–cell communication have been suggested for muscarinic receptors during migration and differentiation of neural precursor cells in vitro and in vivo. The muscarinic M1 receptor subtype is expressed in the neuroepithelium of the rat forebrain, where it was found in progenitor cells and recently differentiated neurons. This data supports the idea that early expression of M1 subtypes participates in neural differentiation prior to synaptogenesis (81). The M2 subtype participates in proliferation and differentiation of neural progenitor cells (82) and inhibition of cell cycle progression in Schwann cells (83), while M3 receptors contribute to oligodendrocyte development (84), and muscarinic M4 subtype activation enhanced NGF-induced prosurvival via Akt signaling in PC12 cells. Muscarinic receptor mediated signal transduction was pertussis toxin-sensitive, suggesting that M4 mAChR, but not M1 and M5 mAChR, were associated with this synergistic Akt activation (85).
The participation of nicotinic and muscarinic receptors in neurogenesis has also been studied in simplified systems. P19 embryonal carcinoma cells, widely used as in vitro model for early neurogenesis, synthesize ACh (86) and express nicotinic and muscarinic receptor subtypes change when cells have been induced to neuronal differentiation in the presence of retinoic acid (RA;87). Our laboratory has shown that nicotinic receptor subunits are already present in undifferentiated cells, and these cells reveal nicotinic receptor-mediated Ca2+-fluxes with possible participation of α7 nAChR subtypes. All muscarinic receptor subtypes with the exception of the M5 subtype were expressed in P19 cells and their activities increased along differentiation (87). While nicotinic compounds promoted proliferation and differentiation of neural progenitors into neurons to a small extent, much larger effects in proliferation and differentiation stimulation were mediated by mAChR. Muscarine-induced proliferation or differentiation of neural progenitors was reversed in the presence of the Gq/11 protein inhibitors YM-254890 and pertussis-toxin blocking Go/i-protein activity (88). The obtained results led to conclude that Gq/11-coupled M1, M3, and M5 subtypes contribute to cholinergic receptor-mediated proliferation, while M2 receptor activity contributes to triggering neuronal differentiation.
The first pharmacological evidence for the potent actions of adenine compounds was published by Drury and Szent-Gyorgyi (89). Only in 1970s, Burnstock suggested that ATP or other related nucleotides are neurotransmitters or co-transmitters in synapses of the PNS (90). Implicit in the concept of purinergic neurotransmission was the existence of post-junctional ATP receptors.
Pharmacological studies comparing the respective responses in different tissues had led to a first division of these receptors in P1 (adenosine receptors) and P2 (purinergic receptors) (91) with the last ones being subdivided in ionotropic P2X(1–7) and metabotropic P2Y(1–14) subtypes (92). The concept of the purinergic signaling expanded throughout the years, including the cotransmission with other neurotransmitters in different types of peripheral nerves and the CNS. Thus, much attention has been focused on the function of purinergic signaling and the measurement of changes in cellular activity of short duration. However, mounting evidence indicates that purinergic receptors participate in cellular communication of long duration, including cellular proliferation, apoptosis, and differentiation [reviewed in (93)].
The first evidence for the participation and importance of purinergic receptors in development came from experiments conducted by Laasberg (94), demonstrating that these receptors belong to the first functional ones during gastrulation of the chicken embryo. Recently, Cheung et al. (95) described the expression patterns of all P2X receptors during rat brain development and their inhibitory roles on motor axon outgrowth in the neural tube. Except for P2X1, all P2X subunits are highly expressed during encephalon development. Moreover, ATP inhibited motor axon outgrowth in neural tube cultures during early embryonic neurogenesis, most likely mediated via the P2X3 receptor. The expression of P2X4 and P2X5 subunits occurs during postnatal neurogenesis, and P2X6 receptor expression increases at the same time (95, 96). The P2X7 receptor is suggested to be involved in programmed cell death during embryogenesis. However, this receptor has also other functions in glia–neuron communication through Ca2+ waves and ATP release. As support for this hypothesis, progenitor cells of hippocampus express the functional P2X7 receptor prior to any other membrane receptors, as soon as cells start to differentiate (97).
It is now clear that NTDPase, degrading bioactive ATP and producing ADP, and P2Y receptors participate in neural proliferation. The most important contributions for understanding of the roles of NTDPases in neural signaling and the developing nervous system have been already reviewed (98). Lin et al. proposed that the presence of extracellular purines is critical for progenitor cell maintenance, driving the expansion of ventricular zone neural stem and progenitor cells by activation of P2Y receptors. Consequently, P2Y receptor antagonists suppressed proliferation and permitted differentiation into neurons and glia in vitro, while subsequent removal of these antagonists restored progenitor cell expansion. In this case, P2Y receptors acted as negative regulators of terminal neural differentiation (99).
In P19 embryonal carcinoma cells as in vitro model for early neurogenesis, we have shown that ATP and many other purinergic receptor agonists promote proliferation of undifferentiated and neural progenitor cells as well as differentiation of progenitor cells into neurons (100). Pharmacological profiling in the presence of various purinergic receptor agonists and antagonists led us to conclude that ATP-induced proliferation stimulation was mediated by P2Y1 and P2Y2 receptors. Acceleration of neuronal differentiation determined as enhanced expression of marker proteins for neural progenitors and neurons was also due to P2Y1 and P2Y2 receptor activity. Moreover, purinergic receptor activity also participated in P19 cell differentiation into neurons expressing functional cholinergic and N-methyl-D-aspartate (NMDA)-glutamate receptors, as carbamoylcholine- and NMDA-induced Ca2+ fluxes were inhibited in neuronal-differentiating cells in the presence of purinergic receptor antagonists (101). The herein discussed data indicate that there is an intrinsic regulation (cross-talk) between purinergic and cholinergic receptor expression and activity during neuronal differentiation of P19 cells, as already shown for kinin-B2 and cholinergic receptors (67). Yet unpublished data of our laboratory also suggest such cross-talk between kinin-B2 and purinergic receptors.
The finding described in this part corroborate that extrinsic factors regulate proliferation and brain development. In addition, variations of [Ca2+]i induced by a network of receptors and ion channels result in the overall control of differentiation progress (Fig. 3). The time course and pattern of calcium signaling during neuronal development in vivo has been shown to contribute to acquisition of the exact phenotype, including the formation of excitatory and inhibitory synapses, the liberation of neurotransmitters, and glia–neuron communication (102). Time points, frequencies, and peak heights of [Ca2+]i transients during in vivo as well as in vitro differentiation are controlled by regulated expression and activities of cell surface receptors and ion channels. These [Ca2+]i transients regulate metabolic functions and participate in the control of neurotransmitter liberation and synaptic activity. Recent studies have brought evidence for receptor-induced [Ca2+]i transients in the regulation of trophic processes, including neuronal development and cellular migration, changes in cell architecture, and organogenesis (103).
NSC ENRICHMENT: CONTROLLING CELL FATE
Stem cell research has advanced during the past years, and NSCs may be ideal candidates for neural transplantation in a wide range of neurological disorders. However, little is known about the mechanisms determining cell fate. Manipulation of cell environment for directing differentiation into specific phenotypes has turned into an important tool for understanding the mechanisms of neural and glial development. For future cell therapies using NSC, scientists have to learn how to control cell fate and develop ideal culture media for NSC expansion in vitro, without altering their plasticity. One can assume from studies of neural development, that on and off periods of growth factors might provide specific patterns of cell proliferation, migration and differentiation. Studies on the isolation and propagation of multipotent NSC as neurospheres suggest their potential use in the reconstitution of neurons and oligodendrocytes in neurodegenerative diseases. In order to ensure that an adequate number of functionally relevant cells are present after transplantation, in vitro manipulation of cell fate before transplantation may be necessary to control the terminal phenotype of these cells.
The adult rat hippocampus contains stem cells that are responsive to fibroblast growth factor 2 (FGF-2). These cells have been used as models to study the effects of retinoic acid (RA) on neuronal differentiation. RA has been shown to induce cell cycle exit by increasing the expression of specific proteins that block cell cycle progression, such as p21 and the transcription factor neurogenic differentiation (NeuroD). Since decrease in cell proliferation increases differentiation, it was also found that RA promoted a three-fold increase in the number of immature neurons. Interestingly, in the absence of RA, only a few cells were responsive to neurotrophins (NT), while RA treatment followed by addition of brain-derived neurotrophic factor or NT-3 significantly increased the number of mature neurons (positive for GABA, acetylcholinesterase, TH, and calbindin) (108). In the same way, forskolin cooperates with growth factors to induce differentiation into dopaminergic neurons. When treated with FGF-8 alone, human mesencephalic NSC gave rise to a slight expression of dopaminergic phenotypes. The simultaneous addition of forskolin and FGF-8 led to an increase in the number of dopaminergic neurons, expressing characteristic genes for dopaminergic differentiation, including TH, nuclear receptor-related factor 1 (Nurr 1) and D2 receptor. Furthermore, these new cells produced the neurotransmitter dopamine (109).
Another interesting fact is that a single ligand can regulate cell fate by activating distinct cytoplasmic signals. This is the case of the bone morphogenetic protein (BMP-4), which can induce differentiation into a wide range of dorsal CNS and neural crest cell types. Glial differentiation can be induced by BMP-4 through a novel pathway mediated by activation of FKBP12/rapamycin-associated protein (FRAP) and signal transducers and activators of transcription (STAT) proteins. The induction of differentiation occurs at high local densities in response to BMP-4 and is specifically blocked by a dominant-negative mutant stat-3. BMP-4 treatment causes FRAP-induced STAT activation, promoting glial differentiation (110). It was reported that conditioned medium from rat neuroblastoma B104 cells (B104 CM) contained some mitogens promoting proliferation and differentiation of oligodendrocyte/type2 astrocyte (O-2A) progenitor cells. Later studies showed that B104 cells produced and secreted platelet-derived growth factor alpha (PDGF)-A homodimers, but not PDGF-B. Additional experiments revealed that neutralization of B104 CM with anti-PDGF-A antibody reduced proliferation of O-2A progenitor cells. B104 CM maintained a high expression of PDGF-A receptors in oligodendrocytes. On the other hand, PDGF-A addition was not equivalent to the effect of B104 CM, indicating the presence of other unidentified growth factors (111).
Neurospheres derived from embryonic rat brain were also cultured in B104 CM to enhance the generation of oligodendroglia. Long-term growth of the neurospheres in this medium markedly augmented the number of oligodendrocyte progenitor cells. Besides B104 CM, other growth factors known to participate in oligodendrocyte development, such as sonic hedgehog (Shh), PDGF, and FGF-2, were used to increase the number of oligodendroglia derived from E18 cortical neurospheres. Among the tested growth factors, PDGF was the most potent mitogen to induce the formation of oligodendrocyte progenitor cells, despite the other growth factors also being capable to generate oligodendroglia as well, but in minor proportions. Shh favored the formation of mature oligodendrocytes (112).
Another approach for recovery of the damage caused by insults to the CNS or by neurodegenerative diseases lies in the activation of the endogenous stem cells for cell regeneration. It is stated that some regions of the adult brain, such as the hippocampus, the SVZ, and the olfactory bulb, are sources of NSC. However, studies have so far failed to identify hippocampal stem cells capable of providing the renewable source of these neurons. Interestingly, it has been recently shown that depolarizing levels of KCl produce a threefold increase in the number of neurospheres generated from the adult mouse hippocampus. Moreover, depolarizing levels of KCl led to the emergence of a subpopulation of precursors generating large neurospheres (>250 μm in diameter). Conversely, the same conditions led to a 40% decrease in the number of neurospheres in the SVZ. Walker and co-workers have shown that latent hippocampal progenitor cells can be activated in vivo in response to prolonged neural activity. Understanding and defining the mechanisms underlying its activation may provide new insights for the combat of cognitive deficits associated with a decline in neurogenesis (113).
We have shown that NGF (20 ng/ml) treatment of rat NSC resulted in enrichment of the neuronal cell population after 7 days of differentiation, when qualitatively compared with the control experiment (Fig. 4). In the same manner, RA (1 μM) application increased the percentage of neuronal cells. However, it is still unclear whether neurogenesis was favored or glial cell proliferation decreased in the presence of NGF and RA.
EGF and FGF-2 promoted proliferation of rodent and human neurospheres, while retaining the cells in an undifferentiated state (114, 115). Removal of these mitogens during 10 days of neurosphere culture of three different human NSC lineages (M006CX, M007CX, and M009CX) induced differentiation into neuronal and glial cells in suspension (data not shown) and caused a decrease in cell proliferation together with an increase in apoptosis (Fig. 5). The decrease in proliferation favored cell differentiation in suspension cultures of neurospheres. In addition, these cells became primed to differentiate in a distinct way after plating. These observations gave rise to a new question: can growth factors removal influence the capability of migration, integration, and differentiation of NSC after cell transplantation? Priming the cells could be relevant for implantation in the course of regenerative strategies. Altering their culture conditions might be critical for defining the potential for survival and integration and enhance the ability of these cells to graft and provide functional recovery. Rather than providing optimal surviving conditions for cultures, we hypothesize that adjusting culture parameters (including growth factors withdrawal) might be essential for achieving success when grafting these cells in damaged systems.
CELLULAR THERAPY WITH NSC
Spinal Cord Injury
Spinal cord injury (SCI) affects about 2.5 million people, with more than 130,000 new injuries reported each year, and causes a significant impact on quality of life and economic burden, with costs associated to primary care and income loss. No efficient restorative strategies to SCI were discovered as yet. One can predict that combinations of different strategies, such as cell therapy and physical therapy treatment will lead to improvements in outcome after SCI. Moreover, it is important to provide a rational basis for tailoring specific combinations of clinical therapies to different types of SCI (116).
NSCs have been broadly used for SCI therapies, and the results are still unclear. Neurospheres obtained from the rat brain, after transplantation in Sprague-Dawley rats subjected to spinal contusion at T8–T9 level, slightly improved locomotor function of the animals (117). Briefly, 1 × 105 cells were injected into the epidural space close to the lesion site 7 days after lesion induction. Then the animals were evaluated for ∼2 months. Stem-cell injected animals revealed a small improvement in locomotor function, evaluated by using the BBB score (118), when compared with animals that had received only saline solution. These results suggested that transplanted neurospheres did not play an important role in motor neuron recovery. Moreover, there was a significant increase in allodynia after cell transplantation. Allodynia is a painful response to a usually nonpainful stimulus. Forelimbs were not affected by the lesion in SCI animals, but became hyper-sensitive and painful after cell transplantation. The thermal hot plate stimulus was performed to evaluate the allodynia response in transplanted animals. Animals that had received neurospheres after SCI were more sensitive to the hot stimulus than those injected with saline solution only.
Macias and coworkers used the same approach to study the effect of neurospheres transplantation following SCI induction. Rats were injected with neurospheres 8 days after SCI at the T8 level. In agreement with the results of Basso and co-workers, the gain in motor function was very slight, but allodynia significantly increased after cell transplantation (119). It is noteworthy that in these experiments genetically modified neurospheres were used in addition to naive neurospheres. The Hofstetter group modified neurospheres to produce Neurogenin 2 (Ngn2), a transcription factor involved in the determination and differentiation of multiple neural lineages during development, while Macias and colleagues genetically engineered neurospheres to produce glial cell line-derived neurotrophic factor (GDNF), a trophic factor with proliferative and neuroprotective effects. As promising results, both Ngn2 and GDNF protected against the allodynia side-effect. Moreover, motor recovery was improved following transplantation of genetically modified neurospheres. As conclusion, NSCs with genetic modifications can be transplanted to improve motor recovery without causing undesired pain.
In 2006, scientists from Johns Hopkins University demonstrated a significant recovery of paralysis in rats infected with Sindbis virus that specifically kills motor neurons in rodents (120). They predifferentiated ES cells into neurons in the presence of RA and Shh, and then transplanted those cells into the affected animals. In addition, dcAMP and Rolipram (a phosphodiesterase inhibitor) were used to overcome the inhibition of neural outgrowth. NCS were injected at the same time to secrete trophic factors supporting neuronal survival and muscle targeting for neuromuscular junction formation.
Parkinson's disease is a neurodegenerative illness that causes shaking, muscle rigidity, loss of motor coordination, and equilibrium as well as alterations of speech and writing. Despite scientific advances, it remains incurable, is progressive, and its cause is still not known. In Parkinson's disease, cells from the substantia nigra, responsible for dopamine production, degenerate, and die. Dopamine is a neurotransmitter that among other functions is involved in controlling body movements, and the decrease of its levels leads to above-mentioned symptoms.
In 2003, a group of scientists and physicians from United Kingdom started the first clinical trial using a trophic factor to treat five patients with advanced Parkinson's disease. GDNF was chosen because of its described beneficial effects in rodent and primate models for Parkinson's disease. GDNF was directly delivered into the patient brain using an implanted pump. Two years later obtained results were very encouraging. Severe clinical side effects were not detected, and there was a 39% improvement in the Unified Parkinson's Disease Rating Scale subscore in patients out of L-dopa medication. Moreover, a 69% improvement in daily activities was observed, and a 64% reduction in diskynesias caused by L-dopa was detected. Scientists also reported a 28% increase in dopamine storage after 18 months (121, 122). However, in February 2005 Amgen Inc., who sponsored clinical trials for Parkinson disease using GDNF, decided to halt those clinical trials. Moreover, Amgen announced that it will not provide GDNF to the patients of the clinical trials terminated in 2004. Leading investigators of the clinical trials protested the company's decision to stop the study. The company had three good arguments for their decision to stop the providing of GDNF. First, the results of one clinical trial did not let to conclude anything about the benefits of GDNF for the patients, since the statistical analysis of obtained data did not reveal any significant difference. Second, some patients presented antibodies against GDNF in their blood. And finally, toxicology studies with rhesus monkeys that had received high doses of GDNF showed focal cerebellar lesions. Part of the investigators sided with Amgen, abandoning the clinical trials. Other investigators, however, decided not to turn off the pumps that have been implanted in 32 patients, opposing the company's decision (123). In response to the Amgen announcement, Dr. Michael Hutchinson, MD, PhD, Associate Professor of Neurology at New York University School of Medicine, sent a concerned scientists' statement synthesizing the objections of investigators against Amgen's decision. He said it was an unusual situation, and the majority of the investigators were in complete disagreement with the company's decision. An alternative method for GDNF delivery using mechanical pumps can be the use of biological minipumps, such as genetically modified neurospheres. Neurospheres modified to produce GDNF can be transplanted into the brain for delivery of the trophic factor. Using an inducible viral system, such as that described by Capowski et al. (124), it is possible to control the amounts of GDNF delivery as well as to stop the therapy at any time. This type of cell therapy associated with gene therapy is still under evaluation in animal models.
Wernig and coworkers showed that iPS cells can give rise to neuronal and glial cell types in vitro and in vivo. Electrophysiological recordings and morphological analysis demonstrated that, after transplantation of mouse iPS-cells, the grafted neurons were functionally integrated in the host brain. The iPS cells differentiated into dopaminergic neurons and improved the behavior in a rat model of Parkinson's disease. The iPS-cells were purified by FACS, to exclude pluripotent cells, avoiding tumor formation (125).
Amyotrophic Lateral Sclerosis
Amyotrophic lateral sclerosis (ALS) is a fatal progressive neurodegenerative disease where motor neurons in the spinal cord and brain stem die and causes loss of muscle control and paralysis (126, 127). ALS is mainly divided in two types: sporadic ALS and familial ALS. The cause of sporadic ALS is unclear, but the disease is clearly related to point mutations in the gene coding for superoxide dismutase 1 (SOD1) in some familial ALS patients (128). The most useful models to study ALS have been transgenic mice and rats over-expressing mutant sod1 (SOD1G93A). These models reveal many of the characteristics of ALS and can be helpful to understand the mechanisms underlying this neurodegenerative disorder (129, 130). Recently, these models contributed to elucidate the role of glial cells in excitotoxicity-caused degeneration of motor neurons. The obtained results suggest a new perspective of therapy, combining cell and gene therapies, to produce growth factors and protect neurons from death. The potential to maintain dying motor neurons and to prevent the death of the remaining healthy neurons by delivering GDNF using NSC represents a powerful strategy for ALS.
As already mentioned, NSC can be genetically modified to release growth factors and thus have been shown to work as long-term minipumps in the rodent and primate brain. Suzuki and co-workers provided evidence that genetically modified human NSCs releasing GDNF (hNSC-GDNF) can be transplanted into the spinal cord of SOD1G93A rats. Following unilateral transplantation of hNSC-GDNF into the spinal cord, robust cellular migration into degenerating areas, efficient delivery of GDNF and remarkable preservation of motor neurons were noticed at early and end stages of the disease. However, motor neuron survival was not accompanied by innervations of muscle end plates, resulting in no improvement in ipsilateral limb use. These results suggest that additional approaches may be required for functional recovery, but without any doubts the obtained results are the first step of a therapeutic advance for this incurable disease (131).
The systemic or intracerebral administration of high doses of a potent mAChR agonist, pilocarpine hydrochloride, in rats and mice promotes seizures and sequential behavioral and electrographic changes. These changes can be divided in acute, silent, and chronic periods. Therefore, this novel experimental approach serves as a model of temporal lobe epilepsy (TLE) mimicking the human condition (132). However, the potential contribution of cell damage—and not only cell loss—to the epileptogenesis is still unclear in both, experimental models and humans. Whether neuronal injury and loss in TLE is the cause or consequence of seizures has been the subject of recent debates (133). Chu et al. (134) examined the effects of grafting NSC on spontaneous recurrent motor seizures (SRMS) in rats 1 day after status epilepticus (SE) induction. Approximately 1 month after onset of SE, 87% of animals, which had not received NSC, showed spontaneous recurrence of seizure activity. On the other hand, only 13% of animals treated with NSC displayed SRMS as compared to those of untreated control animals. SRMS in NSC-treated animals also revealed milder intensities. Phenotypic investigation showed that only a small number of injected NSC expressed markers of mature neurons, and the majority of these cells did not reveal morphologies of hippocampal neurons. However, some transplanted cells co-expressed phenotypic markers of interneurons (GABA, parvalbumin), suggesting that NSC can differentiate into GABA-synthesizing cells after transplantation into the lesioned hippocampus. As mechanism, these new GABA-synthesizing cells may decrease neuronal excitability and suppress SRMS. However, whether these cells can integrate into the brain circuitry remains unclear. It is also possible that these new cells do not differentiate into new interneurons, and therefore the presence of GABA-synthesizing cells may be due to a fusion process between injected NSC and host interneurons.
We injected mouse GFP-NSC into the ventricle of mice systemically treated with pilocarpine and observed 21 days after transplantation, that GFP-NSC were yet present at the injection site, but also migrated into the region of the pilocarpine lesion (Fig. 6). As beneficial effects of anticonvulsant delivery, such as Neuropeptide Y (NPY) and GDNF, in animal models of TLE have been described (135, 136), we can predict that cell therapy combined with gene delivery will provide a useful therapeutic approach for chronic epilepsy (137).
The results of clinical trials are summarized in Table 2.
Table 2. Clinical trials for neurological diseases using stem cells for therapy
TROPHIC FACTORS CELLS
Spinal cord injury
SD rat subjected to spinal contusion at T8–T9 levels
Cell injection in the epidural space close to the lesion
Expression of markers for mature neurons only by a small number of injected NSC.
NSC can be obtained from the fetal or adult CNS. As they can be propagated in culture, they are an unlimited source for studying mechanisms of neural differentiation, as proliferation, migration, and differentiation of NSC reflect their behavior during in vivo embryonic and adult neurogenesis. In vitro studies with NSC will contribute to understand the participation of extrinsic and intrinsic factors in neurogenesis, thereby optimizing the yield of the obtained neural phenotypes. Replacement therapies for neurodevelopmental and neurodegenerative diseases have revealed success when stem cells were placed into a microenvironment favoring differentiation and survival of newborn cells. Such environmental conditions can be generated by transplantation of genetically engineered NSC, which produce and secrete growth factors and neurotransmitters with neuroprotective and trophic actions. Moreover, the time-specific up and down-regulation of key proteins expression to promote differentiation in a unique determined phenotype, i.e., dopaminergic neurons, can be achieved by using NSC transfected with inducible viral expression systems. Besides triggering differentiation and survival of transplanted cells by autocrine and paracrine mechanisms, genetically modified NSC secreting trophic and differentiation-inducing factors can also recruit endogenous NSC and neural progenitors for neural tissue repair.
In the past, the importance of ion channels and metabotropic receptors has been mostly associated with neurotransmission and regulation of neuronal activity. However, recent research revealing that stem cells already express ion channels has changed this vision and led to the conclusion that these channels mediate excitability and trigger Ca2+ waves, defining the progress of differentiation into a distinct phenotype. In view of that, studies carried out by our laboratory showing loss of purinergic and cholinergic receptor activity; when kinin-B2 receptor activity was inhibited along differentiation, indicate that the phenotypic outcome of differentiation may be directed by modulation of receptor-induced Ca2+ fluxes. This hypothesis is in agreement with unpublished data from our laboratory revealing changes in cholinergic and purinergic receptor expression when in vitro neuronal differentiation was induced under depolarizing or Ca2+-deficient extracellular conditions. Therefore, parallel to naive NSC that already express well-known factors with trophic and differentiation-inducing properties, i.e., Shh and GDNF, modified NSC can be engineered displaying up- or down-regulated gene expression for neurotransmitter receptors and neurotransmitter production. The effects of these genetic modifications can be tested following transplantation to evaluate NSC survival and capacity of integration into neuronal networks in damaged neural areas.
In conclusion, future efforts will focus on the elucidation of molecular bases of NSC proliferation and directed differentiation to distinct neural phenotypes for regenerative therapy. The success of therapeutic intervention in SCI, TLE and neurodegenerative diseases will mostly depend on genetically modified NSC producing essential extrinsic factors at the site of the damaged CNS.
T.T.S. thanks Prof. Clive. N. Svendsen, Waisman Center, University of Wisconsin, Madison, for having performed in his laboratory some of the experiments with hNSC lineages.