Until the mid-1990s, repair mechanisms in the adult central nervous system (CNS) were generally believed to be restricted to a postmitotic state, involving mechanisms such as sprouting of axon terminals and synaptic reorganization, despite the initial discovery of neurogenesis in postnatal rat brain approximately 40 years ago [1, 2]. Now, however, the demonstration of functional neurogenesis and the isolation of neural precursor cells from the adult mammalian, including human, CNS has offered perspectives for the development of new strategies for the treatment of CNS disorders. Stem cells are characterized by their self-renewal capacity and the ability to give rise to multiple differentiated phenotypes. Neural stem cells differentiate into neuroectodermal progeny (i.e., neurons, astroglia, and oligodendroglia) in vitro, as well as in vivo . Throughout the life span of rodents, adult neural stem cells reside in two neurogenic brain regions, the subventricular zone (SVZ) of the lateral ventricles and the subgranular layer (SGL) of the dentate gyrus . The SVZ along the lateral wall of the lateral ventricles harbors the largest pool of proliferating cells in the adult brain of almost all mammals, including humans. According to the current, widely accepted model, it contains four main cell types . Multiciliated ependymal cells (type E) separate the lumen of the ventricle from the underlying brain parenchyma and induce circulation of the cerebrospinal fluid. Slowly proliferating astroglial-like (glial fibrillary acidic protein [GFAP]-positive) type B cells generate actively proliferating type C cells, also called transit-amplifying progenitors, which in turn generate migratory neuroblasts (type A cells). Type A cells migrate tangentially in chains via a pathway called the rostral migratory stream (RMS) to the olfactory bulb (OB). The processes of astrocyte-like cells form a scaffold that unsheathes the migratory chains. Type C cells form clusters between migratory cells throughout the SVZ. After arrival in the center of the OB, neuroblasts detach from their chains and migrate in the overlying granular and periglomerular layers . Seventy-five percent to 99% of the cells differentiate into GABA-ergic granular neurons, whereas only 1%–25% differentiate into periglomerular neurons expressing GABA or tyrosine hydroxylase (TH). When and how the neuronal lineage diverges to generate the distinct types of neurons in the OB is not known. A recent study suggested not only that both types of neurons are generated in different areas (periglomerular neurons mainly in the RMS and granular cells in the SVZ) but also that their differentiation is under the control of different transcription factors . Further characterization of the organization of the SVZ and RMS will shed light on the genes involved and the cellular regulatory mechanisms in healthy brain and will eventually help in analyzing the alterations in disease.
The organization of the adult SVZ in humans is different from that in other mammalian species. The lateral ventricular wall consists of four layers with various thickness and cell densities: a monolayer of ependymal cells (layer I), a hypocellular gap containing astrocytic processes (layer II), a ribbon of cells composed of astrocytes (layer III), and a transitional zone into the brain parenchyma (layer IV) . Astrocytes proliferate in vivo and behave as multipotent progenitors in vitro, but no chain migration was observed in the human SVZ . However, newborn cells that express cell cycle proteins (Ki-67 and proliferating cell nuclear antigen [PCNA]) were detected in the granular and glomerular layers of the human OB , but no clear evidence of the presence of a migratory pathway from the SVZ has been demonstrated. Therefore, it was suggested that individual cells might migrate separately to the OB . These results indicate that in comparison with rodents, precursor cells in the human OB are rare but not completely absent.
A second region of continuous neurogenesis identified in rodents, monkeys, and humans is located in the SGL of the dentate gyrus from the hippocampus. Neural progenitor cells migrate a short distance into the molecular layer and give rise to mature granular cells, sending axons to the CA3 region and projecting dendrites to the outer molecular layer. They develop electrophysiological properties identical to those of mature granular neurons. To what extent these newly generated neurons contribute to the function of both the hippocampus and olfactory bulb is not completely known, but studies show their importance in the formation of new (olfactory) memories [10, 11].
In addition, precursor cells were isolated from several so-called non-neurogenic regions (e.g., retina, optic nerve, hypothalamus, and spinal cord), but although they retain the capacity to differentiate into neurons and glia, these cells showed limited self-renewal potential in cell culture [12, , , , , , –19]. Evidence for neurogenesis in vivo outside the neurogenic regions is more controversial. However, regions with at most extremely limited adult neurogenesis seem to acquire the ability to induce neurogenesis under specific pathological conditions. Indeed, next to increased neurogenesis in the SVZ and SGL in response to several brain injuries , also regions such as the striatum [20, 21], the neocortex , the corticospinal motor neurons , and the CA (corno ammonis) layer of the hippocampus  induce modest levels of neurogenesis in response to selective neuronal death or degeneration. This suggests that either changes in the neurogenic permissiveness activate quiescent precursor cells or that recruitment occurs from neurogenic brain regions. Since this mechanism does not reach full cellular and functional recovery, intelligent manipulation of endogenous stem cell populations may represent an interesting target for future therapeutic applications.
A thorough understanding of the mechanisms that regulate neurogenesis in normal and diseased brain is essential prior to the potential use of endogenous neural stem cells for brain repair. Alternatively, stem cells isolated from (postmortem) brain tissue  or other sources may be directed toward neuronal differentiation in culture before transplantation. Those other sources potentially include embryonic stem cells  or adult stem cells isolated from readily available tissues, such as bone marrow-derived  or adipose-derived  adult stem cells. Progress will depend on rigorously defining all developmental potentials of these cells and finding ways to manipulate the production of specific types of progeny and clearly demonstrate their functionality. In this regard, genetic manipulation of cells may be necessary. Retroviral and lentiviral vector systems are reported to be efficient vehicles for gene transfer into neural precursor cells [29, –31]. Since we do not know the best stem cell source at this moment, research should proceed in parallel on various cell types.
The extent of anatomical repair required to achieve significant functional recovery is different in various pathological conditions, depending on underlying factors: the nature of the disease, the identity of neuronal systems involved, and the complexity of networks affected. Demyelinating diseases require specific phenotypic replacement (e.g., oligodendrocytes) and restoration of myelin through cellular interactions. Brain injury following trauma or stroke requires replacement of multiple phenotypes, restorations of local circuits, and long-distance projections. Progressive neurodegenerative diseases, such as Parkinson disease and Huntington disease, require restoration of a specific neuronal phenotype and its afferent-efferent connections. Probably, not all neurological disorders are good candidates for cell-based therapies. Despite the complexity of the human brain structure and function, there is hope that one day cell-based treatments may restore damaged brain functions. Whether transplanted or endogenous stem cells in the brain will be used depends on a rigorous understanding of how to control neurogenesis and integration into neural networks.
Regulation of Embryonic Dopaminergic Differentiation
Successful stem cell therapies, whether based on endogenous stem cells or transplantation, require understanding of the normal pathways guiding neuronal differentiation. During embryonic development, almost all central nervous system neurons are generated under control of local inductive signals. Ventral midbrain dopaminergic neurons are born at mouse embryonic day (E) 10–11.5 by precise concentration gradients of both sonic hedgehog (Shh) and fibroblast growth factor 8 (FGF-8) . Initially, the isthmus, a neuroepithelial signaling center located between the future midbrain and hindbrain, expresses FGF-8, which acts as an anterior-posterior morphogen. In addition, dorsoventral patterning occurs through Shh, which is secreted by ventral midbrain ventricular neuroepithelial cells at the floor plate. In proliferating ventral midline Nkx6.1 neuroepithelial cells at E9, Shh induces expression of Lmx1a, a homeodomain transcriptional activator that determines generation of midbrain dopaminergic neurons . At E9.5, Lmx1a expression upregulates Msx1 expression, which is a Groucho/TLE-dependent homeodomain repressor. At the same time, Nkx6.1 expression in floor-plate cells is downregulated, probably via Msx1. Thus, Msx1 expression triggers a glial-to-neuronal switch that is consistent with the expression of the proneural helix-loop-helix protein Ngn2, thereby setting the timing of dopaminergic (DA) neuron generation at the ventral midline. At E10, precursors migrate to the pial surface, guided by radial glial processes, and start expressing the orphan nuclear related receptor 1 (Nurr-1), thereby withdrawing the cells of the cell cycle and making them postmitotic. In addition, postmitotic DA neurons express engrailed 1, Lmx1b, Pitx3, and TH, the enzyme that catalyzes the conversion of l-tyrosine to l-dopamine [36, , , , –41]. By E13, the DA neurons take their prospective positions corresponding to the future substantia nigra pars compacta (A9 nuclei), ventral tegmental area (A10 nuclei), retrorubral field (A8 nuclei), and the interfascicular nuclei. Until birth, DA neurons undergo axonal outgrowth in a rostral direction and establish contacts with their target cells in the striatum and cortex. A better insight into DA neuron development during embryogenesis, as well as during regeneration of the dopaminergic system in amphibians (frogs and salamanders) [42, 43], will help us to understand dopaminergic regeneration during adulthood, facilitating the development of new regenerative therapies for PD.
Cell Transplantation Therapy for Parkinson Disease
Parkinson disease is the second most common neurodegenerative disorder. The lifetime risk of developing Parkinson disease is 2% for men and 1.3% for women. Although the cause of the disease remains unknown, combinations of genetic and environmental factors are believed to be involved in the pathogenesis of the disease. Idiopathic PD is pathologically hallmarked by the presence of intraneuronal Lewy bodies and a progressive neurodegeneration of dopaminergic neurons in the substantia nigra, resulting in depletion of striatal dopamine. This is clinically manifested in motor dysfunctions such as bradykinesia, hypokinesia, and rigidity, sometimes combined with rest tremor and postural changes [44, 45]. Degeneration can also occur in other brain regions, such as the locus coeruleus, the nucleus basalis of Meynert, the dorsal motor nucleus of the vagus, and the pedunculopontine nucleus. Symptoms do not appear until the striatal dopamine levels have decreased by more than 70% , implying that compensatory mechanisms, involving many different cell types and inflammatory or trophic molecules, are present. In the early stages of the disease, dopamine replacement therapy, using the dopamine precursor levodopa, is effective, but the dose response decreases with disease progression, and motor complications (dyskinesias) and other side effects (e.g., mood disorders and sleep disturbances) arise after chronic treatment. These complications may be due to either the advanced stage of the disease (in which degenerating dopaminergic neurons cannot buffer the fluctuating plasma levels of levodopa, resulting in pulsatile stimulation of the dopamine receptors) or the further degeneration in nondopaminergic regions [47, 48]. Since the underlying mechanisms of neuronal loss in patients are not known, current therapies are mainly symptomatic and not curative or preventive .
Early proof-of-principle of cell-based therapies for neurodegeneration in PD patients was provided by transplantation of human fetal mesencephalic tissue from aborted fetuses, rich in primary dopaminergic neurons, in the putamen or caudate nucleus . The clinical outcome in early open-label clinical trials was encouraging in some patients, where grafted neurons reinnervated the striatum up to 10 years after transplantation, despite an ongoing disease process. However, two double-blinded, randomized, controlled trials set back the initial positivism. In these trials, transplantation did not prove to be beneficial, perhaps because of the disease state, immunological problems, or low graft survival . In addition, a subset of patients developed graft-induced dyskinesias. The underlying mechanism of these dyskinesias remains obscure, but it might be related to the composition of the graft (proportion of nondopaminergic neurons and other cell types or different subtypes of dopaminergic neurons) and partial, incomplete reinnervation of the striatum . Therefore, few if any cell transplantations have been carried out in PD patients in the last few years, since cell transplantation has turned out to be less effective than deep brain stimulation [50, 51]. For the moment, further developments are needed for cell-based approaches to become clinically competitive treatments, such as (a) evaluating other cell sources for dopaminergic neurons (embryonic stem cells; neural stem cells from embryonic, fetal, or adult CNS or peripheral nervous system; and bone marrow or umbilical cord stem cells), (b) improvement of purification, differentiation, and implantation protocols, and (c) evaluation of selection criteria for patients . Nevertheless, these transplantation studies were important as the first experimental cell-based treatment strategies and provided a revolutionary treatment strategy for neurodegeneration: they proved the usefulness of pursuing cell-based approaches in the treatment of neurodegenerative disorders.
Evidence for Neurogenesis in the Striatum and the Subventricular Zone in PD
It has recently become clear that the existence of endogenous neurogenesis is opening possibilities for a second cell-based approach in the treatment of neurodegeneration. However, a thorough understanding of the regulation of adult neurogenesis is necessary before attempting to modulate and recruit endogenously produced neural precursors for cell replacement. To study the pathophysiology of PD, neurotoxin models (based on methyl-4-phenyl-1,2,3,6-tetrahydropyridine [MPTP] or 6-hydroxydopamine [6-OHDA]) , as well as transgenic animal models (e.g., α-synuclein transgenic mice)  and viral vector-based models [55, 56] have been developed. However, we should keep in mind that these models do not fully reconstitute the pathology and disease progression seen in PD patients.
Before the emergence of the stem cell field, there was already evidence that the brains of PD patients and of PD animal models were subject to changes associated with cell plasticity. First, neuronal death in neurodegenerative diseases , as well as in cases of induced brain insult, was associated with a massive astrogliosis . Second, MPTP selective lesioning of nigral dopaminergic neurons was followed by a restoration of dopamine in the dorsal striatum to nearly normal levels 30 days post-lesion . This is reportedly mediated by collateral sprouting from uninjured dopaminergic neurons in the ventral tegmental area or in the substantia nigra. Third, despite the classic view that the substantia nigra pars compacta harbors the majority of dopaminergic neurons, cells expressing TH, the rate-limiting enzyme in the biosynthesis of catecholamines, have also been described in the striatum of rodents, monkeys, and even humans [60, , –63]. Interestingly, in monkeys, nigrostriatal degeneration by MPTP intoxication causes a 3.5-fold increase in TH-immunoreactive cells in the dorsal striatum [64, 65], whereas similar results were observed in rats and mice after 6-OHDA and MPTP lesioning, respectively [62, 66, –68]. In the striatum of parkinsonian patients, the number of TH-positive cells was also reported to be markedly increased . The increase in TH-positive cells is believed to be a compensatory response, where a phenotypic switch takes place in striatal neurons, to restore dopamine levels because of the absence of dopaminergic inputs. However, with the new concept of neurogenesis in mind, these results may have to be reconsidered. It is possible that both TH-positive and astroglial cells are either newly generated from precursor cells in the striatum or have migrated from the SVZ. Finding a way to influence cell proliferation, migration, and differentiation (e.g., promoting neuronal differentiation and suppressing astroglial) might help in the development of treatments for PD and other brain disorders.
Early after MPTP lesioning, 5-bromo-2′-deoxyuridine (BrdU) labeling of proliferating cells demonstrated a significant increase of labeled cells in the striatum of adult mice. These cells mainly differentiated into mature astroglial cells, participating in the injury-induced glial reaction [68, 70]. It is not clear whether these newly generated cells are derived from local or SVZ precursors. However, it is more likely that they are locally derived, since they appear in the striatum early after lesioning and are often found in clusters. No colabeling of BrdU-positive and TH-positive cells was observed in adult mice or, in an analogous study, in aged macaques . This suggests that new TH-positive cells in the striatum of experimental animals after lesioning or in PD patients most likely represent striatal cells undergoing a phenotypic switch to compensate early dopamine loss. Since dopamine is an important regulator for early ontogenetic neurogenesis, the role of this neurotransmitter in PD patients and animal models was investigated. Experimental depletion of dopamine in MPTP-intoxicated mice and 6-OHDA-lesioned rats and mice resulted in a decreased proliferation of progenitors both in the SVZ and SGZ [72, 73]. Indeed, dopaminergic nerve terminals were detected both in the SVZ in close contact with epidermal growth factor receptor (EGFR)-expressing cells (i.e., type C cells and a small percentage of type B cells) and in the SGL in close contact with BrdU-labeled cells . Detailed analysis of dopamine depletion in MPTP mice demonstrated a direct decrease in the proliferation of type C cells in the SVZ, resulting in a reduced migration and neuronal differentiation in the OB and a decreased proliferation (PCNA+ cells) in the SGL . A recent study in macaques revealed a topographically organized projection from the substantia nigra pars compacta to the SVZ  and a significant decrease in the number of PCNA+ cells and polysialated-neural cell-adhesion molecule (PSA-NCAM+) neuroblasts in the SVZ after MPTP lesioning. Importantly, close contact between dopaminergic fibers and EGFR+ cells is conserved in the SVZ in humans . Moreover, endogenous cell proliferation was also impaired in the SVZ, OB, and SGZ of parkinsonian patients, suggesting a conserved role of dopamine throughout different species not only during embryonic development but also in the adult brain. On the contrary, levodopa significantly restored the cell proliferation (PCNA+ cells) in the SVZ of 6-OHDA-lesioned rats, again pointing to an important role of dopamine in increasing adult neurogenesis . Studies with dopamine receptor agonists and antagonists suggest that dopamine directly interacts with the neural precursor pool. The dopamine receptor D1L was detected in type A and C cells, but D2L receptors were more abundantly present in the cytoplasm and plasma membrane of type C cells. Administration of the selective dopamine D2 receptor (D2L) agonist ropinirole increased the number of PCNA-expressing cells at both the lesioned and nonlesioned sides . However, this effect can be due to changes in the proliferation or the survival. Next to D2 receptors, D3 receptors are also reported to be expressed in the SVZ [75, 76]. Intraventricular infusion of a D3 receptor agonist (7-hydroxy-N,N-di-n-propyl-2-aminotetralin [7-OH-DPAT]) resulted in more BrdU-labeled cells in SVZ and RMS of rats . This effect was blocked by co-administration of a selective dopamine D3 receptor antagonist. Interestingly, D3 receptor stimulation also resulted in newly generated BrdU-labeled cells in the striatum, which adopted a neuronal phenotype and sometimes expressed TH. Since BrdU-labeled cells in the striatum had already appeared 3 days after treatment, these cells probably originated from resident precursors. However, to conclude that there is a direct effect on either survival or proliferation of these cells or subtypes of them, D3 receptor expression in the exact SVZ precursor cell type in vivo still has to be demonstrated clearly. In mice, the same D3 receptor agonist failed to induce proliferation, pointing to species difference in dopamine receptor expression .
The effect of the dopamine D2 receptor blocker haloperidol on neurogenesis is more contradictory. Studies report a decreased cell proliferation in the SVZ of early postnatal rats , no effect , and an increase in neural progenitor cells after chronic treatment . The latter study suggested that endogenous dopamine inhibits proliferation of NSCs. The discrepancy in the results may be due to the many pharmacological properties of haloperidol. Next to dopamine receptor blockade, it also activates σ opiate receptors, antagonizes α1-adrenergic and 5HT2-serotoninergic receptors. Therefore, haloperidol may affect more than just dopaminergic transmission. In addition, PD animal models induce alterations in several neurotrophic factors  and synaptic reorganization in the striatum . Hence, the effect of dopamine on SVZ precursors may depend on several factors, including a direct effect through interaction with dopamine receptors and an indirect effect through growth factor release. However, the dopamine receptor subtype expressed by NSCs or progenitor cells may change after brain insults. Together, these studies imply the dopaminergic control of the adult SVZ through the nigrostriatal pathway.
Dopamine has also been suggested to be an important regulator of adult neurogenesis in both the OB and hippocampus. Results from animal experiments indicate that olfactory neurogenesis is important for odor memory and hippocampal neurogenesis for mood and mnestic functions [10, 11]. In MPTP and 6-OHDA models, the number of granular GABA-ergic neurons decreases after lesioning, whereas the number of dopaminergic neurons in the glomerular layer of the OB increases [84, 85]. Since both granular and glomerular neurons are derived from the SVZ and SVZ proliferation is reduced after lesioning, the results suggest a shift in the ratio of newly generated interneurons induced by changes in the neurotransmitter dopamine. Similarly, in PD patients, the number of dopaminergic cells in the OB is doubled , whereas the number of nestin+ precursors is reduced . Intriguingly, depression, impaired spatial memory, and olfactory dysfunction often occur in PD patients before the onset of motor symptoms, and olfactory dysfunction coincides with subclinical striatal depletion of dopamine [87, –89]. Because dopamine inhibits olfactory transmission in the glomeruli, the increase in dopaminergic neurons, together with the decrease in GABA-ergic neurons, may be responsible for the hyposmia in PD patients. Moreover, the higher dopamine levels in the OB may explain why olfaction does not improve upon levodopa therapy. The results described suggest that next to the neuropathological changes observed in the olfactory bulb of PD patients [90, 91], impaired neurogenesis may also be important in the development of hyposmia. This raises the important question of whether reduced neurogenesis might contribute to the development of Parkinson disease symptoms.
In an α-synuclein transgenic mouse model for PD, where no dopaminergic cell death occurs, endogenous neurogenesis in the SVZ and SGZ was reported to be impaired because of reduced survival of neuronal precursors . Together, these results indicate that not only dopamine but also other disease-related mechanisms may contribute in decreased neurogenesis in Parkinson disease.
Evidence for Endogenous Dopaminergic Neurogenesis in the Substantia Nigra in PD
Whether dopaminergic neurogenesis occurs in the adult substantia nigra in normal brain or in PD animal models is still a matter of debate. The controversy may be due to methodological differences between the studies reported (BrdU dose, section thickness, etc.). On the one hand, it has been postulated that dopaminergic differentiation occurs at a very low level in the substantia nigra (SN) of healthy mice and that this process increases after MPTP lesioning . Basal levels of neurogenesis and increased proliferation and dopaminergic differentiation after MPTP administration were also demonstrated in nestin-LacZ transgenic mice , although the levels of dopaminergic differentiation were very low. Approximately two newly generated DA cells were detected in control mice, compared with nine DA cells after MPTP treatment, of which only four colabeled with BrdU . On the other hand, other groups could not confirm this independently in healthy mice or in 6-OHDA-lesioned rats [15, 95, 96]. Results from several laboratories demonstrated that 6-OHDA lesioning in rats or MPTP lesioning in mice resulted in cell proliferation in the SN without dopaminergic differentiation [15, 68, 70, 96]. Interestingly, when these proliferating cells were isolated and implanted in a neurogenic environment, such as the dentate gyrus, 20% of the cells differentiated into mature neurons . Transplantation into the SN of healthy rats resulted only in glial differentiation. Thus, regional environmental factors restrict differentiation to distinct neural lineages in situ. In addition, part of the proliferating cells in the SN expressed GFAP or NG2. NG2 is considered to be a marker of a population of oligodendrocytic precursors, but NG2-expressing cells were also reported to have the capacity to generate neuronal cells. A large part of the proliferating cells could not be characterized with known neuronal and glial markers , suggesting that they remain uncommitted progenitors [70, 94]. Hermann et al.  isolated NSCs from the tegmental tissue of the midbrain and hindbrain, including the ependymal zone of the aqueduct and the fourth ventricle. These cells displayed the same characteristics in cell culture as NSCs isolated from fetal brain and from the principal neurogenic regions of the adult brain. More importantly, they reported for the first time on functional dopaminergic differentiation in cell culture from adult tegmental NSCs. Both in vitro and in vivo studies suggest that cells with dopaminergic potential are present in the midbrain, but although cell proliferation occurs after lesioning and although dopaminergic neurons are generated in cell culture, the low level of dopaminergic differentiation (if occurring) in situ suggests that a restrictive environment impairs dopaminergic neurogenesis in the SN. Moreover, there is still some debate about the origin of the proliferating cells in the SN. They may be derived from precursors residing in the brain parenchyma or from precursors lining the cerebroventricular system. In conclusion, there is a need for specific markers and methods to undisputedly illustrate ongoing neurogenesis in the SN of both PD patients and animal models of PD. Even if dopaminergic neurogenesis takes place in the substantia nigra, it will only become therapeutically relevant if the levels can be boosted considerably. Recently, Van Kampen et al. showed that pharmacological activation of the dopamine D3 receptor by 7-OH-DPAT triggers neurogenesis in the SVZ and the substantia nigra of the healthy adult rat brain [77, 98]. Dopamine D3 receptor stimulation by the same agonist also increased cell proliferation and dopaminergic differentiation in the substantia nigra of 6-OHDA-lesioned rats . Moreover, striatal innervation from these cells and long-term improvement in locomotor function demonstrate that pharmacological intervention by changing environmental signals may have important implications in the design of novel treatment strategies for PD.
In the substantia nigra of PD patients, cells expressing PSA, a marker for immature neurons, were detected, which sometimes colabeled with TH . However, this finding cannot be considered solid evidence for ongoing dopaminergic neurogenesis, since PSA expression can be upregulated in pathological conditions [101, 102].
Therapeutic Strategies Based on Endogenous Neurogenesis
The response of endogenous neural precursors in PD animal models is complex. Because of partial dopaminergic control of the SVZ progenitors in the vicinity of the degenerating axon terminals, proliferation in the SVZ decreases, whereas new astroglia are generated in the striatum. Progenitors in the substantia nigra, in the vicinity of degenerating dopaminergic neurons, respond by increasing proliferation (Fig. 1) with minor dopaminergic differentiation. This implies that different regulatory mechanisms mediated by the changes in the local microenvironment result in different responses in both brain regions. Hence, different therapeutic strategies may be necessary for both regions.
Figure Figure 1.. Overview of responsive regions of cell proliferation in rat and mouse models of Parkinson disease. Areas in red correspond to regions of decreased cell proliferation, whereas areas in green correspond to regions of increased cell proliferation. Neurogenesis in the subventricular zone of the lateral ventricle and the DG of the hippocampus is decreased. Less proliferative cells were detected in the Gr. However, the amount of dopaminergic neurons in the Gl is increased. Neurotoxic lesioning also results in glial cell proliferation in the striatum and increased cell proliferation in the SN. Whether neurogenesis is occurring in the SN is still a matter of debate. Abbreviations: cc, corpus callosum; CPu, caudate putamen (striatum); DG, dentate gyrus; EPl, external plexiform layer of olfactory bulb; Gl, glomerular layer of olfactory bulb; Gr, granular cell layer of olfactory bulb; IPl, internal plexiform layer of olfactory bulb; LV, lateral ventricle; LR4V, lateral recess of the 4th ventricle; Mi, mitral cell layer of olfactory bulb; SN, substantia nigra.
Download figure to PowerPoint
As long as no causal therapy for PD can be designed, it will remain important to increase dopamine levels in the striatum. Based on the fetal transplantation studies, it has been estimated that only 2%–4% of the normal number of dopaminergic neurons that innervate the striatum should be restored for the first signs of functional recovery to occur in rodents with nigrostriatal lesions . In addition, in autopsy studies performed on patients who received fetal transplants, it was demonstrated that a good clinical response was obtained after reinnervation of approximately one third to one half of the putaminal volume .
Successful endogenous stem cell-based therapy has to result in efficient progenitor cell proliferation, dopaminergic differentiation, and survival of newly generated cells. Therefore, different strategies can be pursued: It may be feasible to generate dopaminergic neurons in the striatum by either recruitment of endogenous progenitors from the SVZ or stimulation of resident cells in the striatum. Next, since the SN harbors proliferating cells, it may be feasible to stimulate differentiation into dopaminergic neurons. However, restoration of nigrostriatal projections may be the major challenge here. To date, the optimal approach for endogenous stem cell therapy is not known. Stimulation of cell proliferation or induction of dopaminergic differentiation may be mediated by viral vector-mediated local overexpression of either growth or transcription factors or by pharmacological intervention. Alternatively, alteration of the local microenvironment by overexpression of growth factors may increase cell survival and help in dopaminergic differentiation. Since it is difficult to predict whether intrinsic and extrinsic signals or a combination of both will be necessary, careful investigation in animal models of Parkinson disease is required to shed light on the feasibility of cell therapy in combination with gene or pharmacotherapy to induce dopaminergic differentiation.
What are the results described so far? In animals, SVZ progenitors can be recruited to the striatum after administration of several growth factors (e.g., transforming growth factor-α [TGF-α] and brain-derived neurotrophic factor [BDNF]) [104, 105]. Intrastriatal TGF-α infusion in 6-OHDA-lesioned rats resulted in differentiation of SVZ precursors to dopaminergic neurons in the striatum  in combination with an improved behavioral response to the dopamine agonist apomorphine, but this could not be independently confirmed . Although this study suggests that recovery may be stimulated by dopaminergic differentiation from endogenous precursor cells, it should be noted that no clear correlation between newly born dopaminergic neurons and functional recovery was demonstrated. In addition, platelet-derived growth factor and BDNF were also able to recruit new cells to the striatum and the substantia nigra in 6-OHDA-lesioned rats . In this study, newly generated cells in the striatum expressed markers for immature (Dcx), early striatal (Pbx) and mature neurons (NeuN), but without adopting the phenotype of the major resident cell type (DARPP-32) and without indications of dopaminergic differentiation. In the SN, no newborn neurons were detected. Lentiviral vector-mediated overexpression of glial cell line-derived neurotrophic factor (GDNF) resulted in an eightfold increase in the number of striatal TH-positive cells . Although BrdU birthdating studies indicated a GDNF-mediated increase in cell proliferation, they failed to demonstrate that these TH-positive cells were newly generated , providing further evidence for the restorative actions of this growth factor. These initial studies have thus demonstrated that cell recruitment to the striatum is possible in animal models of PD. However, the next challenge lies in stimulating the proliferated cells to differentiate into dopamine-secreting cells and demonstrating a clear correlation between dopaminergic differentiation and functional recovery.
In addition to growth factors, neurotransmitters such as dopamine are reported to play a role in the control of embryonic and adult neurogenesis. As mentioned before, dopamine receptor stimulation may be an interesting strategy to induce cell proliferation. Stimulation of the dopamine D3 receptor by a preferential agonist, 7-OH-DPAT, not only increased cell proliferation in the SVZ and striatum  but also significantly increased proliferation in the lining of the third ventricle and the SN. A proportion of BrdU-labeled cells adopted a neuronal phenotype (NeuN-positive) or differentiated into TH-positive cells . Since D3 receptors are almost exclusively expressed in the nervous system, pharmacotherapy with 7-OH-DAPT or its analogs may represent an important treatment strategy. Proof of concept in an animal model of PD demonstrated an improvement in locomotor function, as explained before . Further testing in different animal models of PD will be of much importance in shedding light on the pharmacologic intervention with progenitor-specific agonists.
Several characteristic transcription factors have been described to be involved in the maturation of dopaminergic neurons, such as Nurr-1, Pitx3, Engrailed, and Lmx1b [36, , , , –41]. An important step forward in generating dopaminergic neurons was accomplished by defining Lmx1a and Msx1 as intrinsic determinants for midbrain dopaminergic neurons . Overexpressing both factors in embryonic stem cells resulted in efficient DA differentiation. Functional integration and improvement in locomotor function after transplanting these cells in 6-OHDA-lesioned rats still have to be demonstrated. Although DA differentiation from embryonic stem cells was highly efficient, there is still the risk of teratoma formation, even if only a few nondifferentiated cells are transplanted. Therefore, it will be interesting to study the effects of overexpressing Lmx1 and Msx1 in alternative sources of stem cells with no known risk of teratoma formation, such as multipotent adult progenitor cells or adult neural stem cells . Alternatively, overexpression in precursors in the substantia nigra might increase dopaminergic differentiation. It will be of interest to study the role of these differentiation factors with or without additional growth factors after dopaminergic denervation in animal models of PD.