Journal of Neurology

, 258:346

Stem cell-based therapies in Parkinson’s disease: future hope or current treatment option?


  • Kai Loewenbrück
    • Department of NeurologyDresden University of Technology
    • Department of NeurologyDresden University of Technology
    • CRTD, Center for Regenerative Therapies Dresden, Dresden University of Technology
    • DZNE, German Center for Neurodegenerative Diseases, Research Site Dresden

DOI: 10.1007/s00415-011-5974-4

Cite this article as:
Loewenbrück, K. & Storch, A. J Neurol (2011) 258: 346. doi:10.1007/s00415-011-5974-4


Parkinson’s disease (PD) is one of the most frequent neurodegenerative diseases and represents a major therapeutic challenge because of the so far missing therapeutic means to influence the ongoing loss of dopaminergic innervation to the striatum. Cell replacement has raised hope to offer the first restorative treatment option. Clinical trials have provided “proof of principle” that transplantation of dopamine-producing neurons into the striatum of PD patients can achieve symptomatic relief given that the striatum is sufficiently re-innervated. Various cell sources have been tested, including fetal ventral midbrain tissue, embryonic stem cells, fetal and adult neural stem cells and, after a ground-breaking discovery, induced pluripotent stem cells. Although embryonic and induced pluripotent stem cells have emerged as the most promising candidates to overcome most of the obstacles to clinical successful cell replacement, each cell source has its unique drawbacks. This review does not only provide a comprehensive overview of the different cellular candidates, including their assets and drawbacks, but also of the various additional issues that need to be addressed in order to convert cellular replacement therapies from an experimental to a clinically relevant therapeutic alternative.


Embryonic stem cellsInduced pluripotent stem cellsNeural stem cellsParkinson’s diseaseNeurotransplantationNeuroregeneration


There is an urgent need for additional therapies in Parkinson’s disease (PD), one of the most frequent neurodegenerative disorders, that not only offers symptomatic relief, but also slows down or replaces the loss of dopaminergic neurons, mainly but not exclusively in the substantia nigra pars compacta (A9 region, SNpc). Cell replacement strategies (neurotransplantation) have been suggested to have a great potential for restorative therapy in PD [14, 18, 38]. This hope has fueled research on the appropriate cell source and transplantation strategy for the past 30 years. There are several reasons why PD represents a good candidate for successful cell replacement in comparison to other neurodegenerative diseases: Firstly, even though the hard to treat non-motor symptoms seem not to be related to the striatal dopamine deficit, the cardinal and clinically most disabling motor symptoms are the consequence of the degeneration of a highly specialized cell population in a confined neuroanatomical structure, the SNpc. Secondly, although functional network integration and complex innervation of striatal target structures has been shown to increase the efficacy of dopaminergic cell transplants, the simple provision of dopamine by clinical pharmaceutical treatment with levodopa or in rodent transplantation studies, even without signs of a high level of network integration, are clinically effective. In principle, cell replacement would have to be safe and offer additional benefits over the available pharmaceutical or neurosurgical treatment options regarding symptomatic relief or slowing of disease progression. Both animal research and two double-blind placebo-controlled clinical studies using fetal ventral midbrain tissue as transplants give positive signals that it should be possible to achieve these goals [14, 18, 38], but so far numerous unresolved issues remain to be addressed.

The outcome of the cited clinical trials is exemplary for many of these unresolved issues: (1) Even though fetal mesencephalic dopaminergic neurons represent to date the most promising cell type for cellular replacement of midbrain-specific dopaminergic neurons, the use of up to eight aborted fetuses per transplanted patient as a cell source prevents a large-scale clinical application per se [18]. (2) Individual patients benefited largely from the transplantation, but a statistically significant improvement could only be observed in a subpopulation of patients of younger (<60 years) age or with lower (<50) UPDRS baseline motor scores. (3) A so far unreported complication that had not been observed in rodents before the clinical trials occurred in the form of graft-induced off-medication dyskinesias (GID) in up to 56% of transplanted patients [14, 19, 38]. The pathophysiological basis of this complication so far has not been completely understood, but might involve other than dopaminergic neuron types such as the co-transplantation of serotoninergic nerve cells. (4) One patient transplanted in another study died of transplantation-related complications in the form of fourth ventricle obstruction and brain stem compression caused by intraventricular transplant-derived teratoma formation [18].

This review will give an overview of potential stem cell sources for dopaminergic neurons currently being investigated. Especially the discovery of induced pluripotent (iPS) cells in 2006 has provided a new and very promising source of dopaminergic neurons [56]. Embryonic stem cells [32] and fetal or adult neural stem cells (NSCs) represent relevant alternatives. At this point, however, it remains an open race which cell type will be the best answer to the above mentioned issues of availability, clinical efficacy, and safety. Assets and drawbacks of the different cell types, methods to increase the yield of dopaminergic neurons, as well as novel implantation strategies will be covered by this review.

Stem cell sources

Stem cells are the most promising candidate to overcome the obstacles to cell-based regenerative therapy. Minimal defining characteristics for stemness are the ability to self-replicate and to be able to differentiate into cells of different germ layers [for e.g., pluripotent embryonic stem (ES) cells] or into the different cell types of one germ layer [for e.g., tissue-specific, multipotent fetal or adult neural stem cells (fNSCs or aNSCs)]. The discovery of the possibility to reprogram differentiated somatic cells by temporary over-expression of stemness-inducing transcription factors into a self-sustaining pluripotent state almost identical to ES cells [induced pluripotent stem (iPS) cells] has opened up the exciting possibility, theoretically, to create individualized differentiated cells of any type for autologous transplantation strategies [56].

Whereas the ability of prolonged self-replication by symmetric cell division is a major prerequisite to overcome limitations in the number of desired cells, it is accompanied by frequent teratoma formation, which has so far prevented any clinical study based on ESCs or ESC-derived differentiated cell preparations [5]. Unfortunately, the undesired teratoma formation is shared by iPS cells so that strategies to prevent tumor formation are an irreplaceable presupposition for usage of undifferentiated, pluripotent cells for transplantation [58]. In contrast, with multipotent germ-layer restricted NSCs, tumor formation has not been observed. However, with NSCs availability becomes a major concern again, because they are either derived from fetuses for allogenic transplantation approaches or would have to be harvested stereotactically from patients’ neurogenic niches in the case of autologous transplantation strategies. In addition, differentiation of NSCs into the desired dopaminergic midbrain phenotype has been shown to sometimes only result in disappointingly low yields and also to vary strongly between different species.

Embryonic stem cells

ES cells (ESCs) are derived from the inner cell mass of early post-fertilization blastocysts [20]. Because of their ability to generate an entire organism, they hold a high proliferate potential that can be used to culture them over extended time periods and to derive large numbers that later on can be differentiated into the desired phenotype. The major limitation is their embryonic origin and the need to destroy embryos for their cultivation. When it comes to ES cells as a cell source for cell replacement, two basic strategies can be employed: On the one hand, they can be used without any prior in vitro differentiation based on the hypothesis that regional, host-tissue derived inductive cues are the best setting to obtain the required cell type. An alternative strategy would be to partly or totally differentiate ES cells into the desired cells based on the hypothesis that the host tissue-derived inductive cues are not sufficient to achieve this and to avoid tumor formation at the same time.

Because of the limited success of early in vitro differentiation protocols, undifferentiated ES cell transplantation was initially studied. Results showed that local inductive cues are indeed sufficient to induce dopaminergic differentiation and concomitant clinical improvements in parkinsonian rats [5]. Whereas some neurons expressed an A9-specific marker combination, also A10-specific dopaminergic neurons were found and a high proportion of serotoninergic neurons, leading to a relation of 2:1 of dopaminergic to serotoninergic neurons. Taking into account that serotoninergic neurons in grafts have been shown to be responsible for levodopa-induced dyskinesias in clinical transplantation studies [8], it is rather unlikely that the resulting cell composition is well suited for successful transplantation. A more limiting problem was that 20% of the grafted animals had to be sacrificed before the defined study endpoint because of teratoma formation. Even if several strategies have been successfully employed to reduce the risk of tumor formation [12, 31, 48], the use of undifferentiated ESCs remains an unsafe strategy.

Various culture strategies, including the usage of different growth and differentiation factors [30], co-culture systems with different feeder cells [25] and forced over-expression of transcription factors have been developed to differentiate ES cells into dopaminergic phenotype of midbrain specificity before transplantation [11, 13, 27]. However, the major problem has been the variability and limitation of success rates of dopaminergic differentiation, dopaminergic sub-specification into A9-specific cells and phenotype stability [25, 30, 43, 61]. During recent years, however, the increasing knowledge about growth and differentiation factors, as well as transcription factors that govern embryonic midbrain dopaminergic differentiation has allowed increasing the yield of dopaminergic neurons by combination of forced over-expression of transcription factors, such as Nurr1 and Pitx3, with the above mentioned culture conditions [1, 11, 13, 27].

Although effective approaches are available to guide ES cells into midbrain dopaminergic phenotype, the transplantation of such cells into parkinsonian animal models so far shows limited long-term behavioral efficacy. Whereas most studies using mouse ES cells were able to show significant improvements of parkinsonian rats in the rotational test [25, 27], grafts with human ESCs showed only partial [45, 59] or even no [7] improvement. In addition to this sometimes limited clinical effectiveness, accompanying histological analysis of transplants so far render any clinical application impossible: Even if in vitro a fully functional dopaminergic phenotype can be effectively predetermined, in vivo conditions seem to be unfavorable for phenotypic stability and survival [13, 45, 59]. Together, prolonged in vitro differentiation seems to help prevention of tumor formation, but at the same time strongly compromises the ability of the transplant to integrate into the host tissue and to exert a clinically relevant effect. However, encouraging results were achieved in a study in using a differentiation protocol using Wnt5a showing increased phenotypic stability and behavioral efficacy [46].

Neural stem cells

Neural stem cells, already neuroectodermally restricted but still able to self-propagate, could represent an ideal cell source to achieve, at the same time, sufficient cell numbers and midbrain dopaminergic differentiation. In addition, because of their germ layer restriction, problems with contamination by unwanted cell types and teratoma formation could be expected to be less pronounced than with pluripotent stem cell types. At the same time, dopaminergic differentiation potential of NSCs seems to be more limited than in pluripotent stem cells.

Fetal neural stem cells

Even though the proliferate potential of fetal NSCs, especially in the case of midbrain-derived NSCs, is more limited than in ESCs, they can be sufficiently propagated using FGF-2 and EGF as growth factors [17, 39, 40]. However, even though not easily explained, the neural restriction of NSCs seems not to be of advantage in comparison to unrestricted ESCs when it comes to the generation of midbrain dopaminergic neurons. Although fetal mesencephalic NSCs readily differentiate into neurons upon growth factor withdrawal and exposure to specific culture conditions, this resulted in rats and humans in relatively low yields of dopaminergic neurons (approx. 1–5% [24, 5355]), whereas in mice fetal mesencephalic NSCs might have a more restricted potential to differentiate into dopaminergic neurons. As with ESCs, the addition of different cytokines or growth factors, such as leukemia inhibiting factor (LIF), IL-1β, IL-11 and glial cell line-derived neurotrophic factor (GDNF) have been found to increase the yield in dopaminergic neurons to 20–25%, allowing for clinical successful transplantation in 6OHDA-lesioned rats [9, 49, 54]. Others have successfully used various factors including forskolin, brain-derived growth factor (BNDF), dopamine or retinoic acid [44, 60].

Other strategies to optimize dopaminergic differentiation include the over-expression of dopaminergic key-fate-determining transcription factors. Intrinsically, fetal NSCs from different brain regions already show regional specification, as illustrated by the finding that midbrain dopaminergic neurons can solely be derived from midbrain-derived fetal NSCs [39]. By over-expression of such a transcription factor, namely Nurr1, fetal forebrain-derived NSCs could be driven into a midbrain-specific dopaminergic phenotype, which did not showed any efficiency after transplantation into parkinsonian rats [42]. Over-expression of Pitx3 in midbrain-derived NSCs also results in an improvement of dopaminergic stability and graft survival and in significant clinical improvements of lesioned rats [37]. Combinations of key fate-determining transcription factors might enhance phenotype differentiation and stability [3].

In summary, even though fetal NSCs appear as an attractive source to attain midbrain dopaminergic neurons, the low growth rate leading to limited availability and the restricted dopaminergic differentiation potential of fetal NSCs limits their preclinical and clinical application. The lack of evidence for tumor formation appears as a big advantage, but so far there is no evidence that human fetal NSCs can be grafted with the level of functional integration and phenotypic stability that is the basic prerequisite for clinical effectiveness. In addition, even if the technical limitations might be overcome, their fetal origin still represents a major ethical obstacle to routine clinical application.

Adult neural stem cells

Adult NSCs represent a very attractive cell source, because an autologous transplantation protocol would also avoid the danger of immunological graft rejection and the ethical concerns regarding cell exploitation from fetuses or embryos. The finding of adult neurogenesis and its contribution to the functional integrity in some brain regions in mammals was epochal and basically allows for two strategies in cell replacement therapies in PD [2]: Firstly, in vivo recruitment of NSCs and their differentiation into dopaminergic neurons could be the basis for an intrinsic cellular replacement. Secondly, adult NSCs could be stereotactically harvested, cultured and pre-differentiated in vitro, and then retransplanted.

Unfortunately, for the first elegant strategy so far no unequivocal evidence is available [16, 23]. Most studies that claim that there is an intrinsic midbrain- or even substantia nigra-derived neurogenesis are affected by relevant methodological concerns [50, 62]. Interestingly, it has been shown that in spite of the lack of evidence for any intrinsic activity, adult NSCs from mouse midbrain can be harvested, propagated and then be driven into a midbrain dopaminergic phenotype [22, 23]. Another study was able to recruit mouse midbrain progenitor cells by intraventricular growth factor infusion in vivo [33]. Together, adult NSCs seems to have the theoretical potential to contribute to dopaminergic cell replacement within the midbrain, but that possibly environmental or cell-intrinsic factors prevent them from taking action in vivo.

Evidence for the second strategy—explanting adults NSCs and transplanting them after an in vitro expansion and differentiation period—is also very limited. Whereas the problem of tumor formation has not been reported for adult NSCs, their in vitro proliferation potential and stability is low and could compromise the achievement of sufficient cell numbers needed for successful transplantation [23]. An important study, therefore, tried to drive adult NSCs to the midbrain dopaminergic phenotype by over-expressing the above mentioned transcription factor Nurr1 [52]. A study comparing adult mouse NSCs and ES cells found very poor survival of adult NSCs in the final differentiation stage and could find depolarization-induced dopamine release (a hallmark of functional dopaminergic neurons) only in ESC-derived neurons [41]. So far, no study has been performed to show clinical effectiveness of this approach in PD animal models. In spite of this very limited information, there is one incomplete report of an autologous transplantation attempt in a patient. Neural progenitors were harvested from a cortical autopsy and a 15% dopaminergic differentiation rate was achieved after an extended in vitro culture period. Upon transplantation, there was an increase in fluorodopa uptake by 55% accompanied by partial clinical improvement [47].

To conclude, even though an autologous transplantation strategy circumvents some limitations of allogenic transplantation, evidence so far is too limited to postulate that this strategy will lead to a clinical relevant therapeutic option. The lack of evidence for tumor formation and the possibility to drive forebrain adult NSCs into a midbrain dopaminergic phenotype is encouraging [52].

Induced pluripotent stem cells

Induced pluripotent stem (iPS) cells very likely represent the key to autologous cell replacement therapies, including in PD. iPS cells can be theoretically generated from any differentiated somatic cell by the temporary forced over-expression of key pluripotency-inducing transcription factors, whereas also less than the initially used four factors c-Myc, Oct-4, Klf4 and Sox2 seem to do the job [20]. Upon temporary artificial expression of all or some of these factors a self-sustaining autoregulatory transcriptional loop is initiated that leads to a stable pluripotent phenotype, mimicking most of the ES cell-derived criteria for pluripotency, such as self-propagation, expression of key pluripotency markers, differentiation into differentiated cells of all germ layers, teratoma and chimera formation. However, being an almost ES cell-identical cell type means that iPS cells as a source for cellular replacement suffer from the same disadvantages as ES cells, in addition to concerns related to the necessity of genetic modifications that harbor the risk to even increase tumor formation, because some of the reprogramming factors represents potent oncogenes. Therefore, alternative strategies of achieving induced pluripotency have been successfully tested, including a reduction of the factors induced (usage of only Klf-4 and Oct4) [28, 29, 51], transient transfection by plasmids and transfection with loxp-flanked gene constructs, allowing for a Cre recombinase-induced excision upon achieved reprogramming. The most convincing approaches so far for a clinical application totally passed on DNA modification and induced pluripotency efficiently by the administration of modified synthetic mRNA or recombinant proteins [26, 57].

iPS cells were successfully differentiated into the dopaminergic phenotype by one of the protocols formerly successfully employed in ESCs differentiation [30] and transplanted into parkinsonian rats showing a clinically relevant striatal re-innervation [58]. These animals showed clinical improvements comparable to those achieved with ESC-derived transplants, in addition to an efficient striatal re-innervation and complex dendritic arborization. The expected complication of continuous proliferation and teratoma formation could be prevented by pre-implantation FACS-based negative selection of residual undifferentiated cells. In addition, clinical successful transplantation was achieved in parkinsonian rats upon the transplantation of iPS cell-derived dopaminergic neurons from PD patients [21]. However, axonal outgrowth and striatal innervation pattern was poor compared to the transplantation of rat iPS cell-derived transplants. These studies on the clinical effectiveness of human iPS cells are important, because significant differences are known to exist between human and rodent iPS cells that could impede the translation of results from rodent to human iPS cells [20].

In spite of the many issues to be addressed before a clinical trial with iPS cell-derived dopaminergic neurons can be envisaged, so far achieved milestones make it appear possible to attain iPS cell-derived cell preparations in future that are safe for usage in clinical trials: iPS cells can already be derived without the necessity of DNA modification and the risk of tumor formation can be lowered by negative selection of remaining undifferentiated cells. This progress regarding safety concerns is accompanied by convincing evidence for iPS cells—in contrast to other autologous cell sources—to provide clinically effective striatal re-innervation.

Transplantation strategies

Besides the tremendous efforts invested in finding the right cell source and differentiation protocol, there remain numerous additional questions that require an answer in order to develop cell replacement therapy in PD into a relevant therapeutic alternative.

Cellular composition

Although the clinical effectiveness of fetal ventral midbrain tissue has been the most convincingly demonstrated so far, this tissue does not represent a primarily dopaminergic cell source, but a mixture of all neural cell types homing to the fetal midbrain, among which dopaminergic neurons are only one of the main constituents. To date, apart from few exceptions, it is unclear which of the other constituents are of positive or negative impact on the clinical effectiveness of the transplants. There is evidence that containing serotoninergic neurons might be responsible for dyskinetic side effects. No information is available on the effect of GABAergic or noradrenergic neurons also present in midbrain transplants and on potential pleiotrophic effects of astroglial cells on the grafts as well as host tissue in vivo. In vitro, it has been shown that fetal midbrain-derived astrocytes have beneficial effects on dopaminergic differentiation. Transplanting a mixture of dopaminergic precursors/neurons and midbrain-specific astrocytes might thus be more effective than a purified dopaminergic cell source not being able to take advantage of a supportive pleiotrophic environment.

Transplant location

Striatal transplantation studies have shown that the clinical success depends on the level of striatal dopaminergic re-innervation. Principally, ectopic transplantation directly into the striatum instead of orthotopic into the substantia nigra pars compacta has been shown to result into sometimes very efficient dopaminergic re-innervation in patients (normalization of fluorodopa uptake) accompanied by only a limited degree and limited time period of clinical effectiveness [14, 18, 38]. On the other hand, studies in postnatal rats showed that striatal re-innervation by orthotopic transplantation into the substantia nigra pars compacta can be achieved only up to postnatal day 10, providing a functionally more efficient re-innervation with better clinical improvements than intrastriatal grafts [4, 35, 36]. To overcome this limitation of intrinsic axonal outgrowth in adult animals, bridging transplantation strategies were successfully employed. This means providing additional guidance cues to the orthotopically transplanted cells into the substantia nigra to allow the cells to target the striatum with their outgrowing axons. Indeed, by directly injecting excitatory amino acids (kainic acid) or by additionally transplanting GDNF-secreting fetal kidney cells into the striatal target region, partial dopaminergic re-innervation of the striatum with significant improvement in the amphetamine-induced rotational test could be demonstrated [6, 10, 63]. Recently, novel biomaterials and engineered extracellular matrix molecules (ECM) have caught increasing attention as a strategy in bridging transplantation models. Such molecules have besides their controlled and well-defined molecular structure the advantage that they can be uploaded with bioactive reagents, such as GDNF or other neurotrophic or axon-guiding factors [15]. However, so far no applications of such biomaterials or artificial ECM in bridging transplantation strategies have been reported in parkinsonian models.

Further alternative transplantation locations apart from the striatum have been explored. Under consideration of the high clinical effectiveness of the blockage of subthalamic nucleus hyperactivity by deep brain stimulation, Mukhida et al. [34] investigated to co-transplantation of inhibitory GABAergic neurons into the subthalamic nucleus and the equally disinhibited substantia nigra pars reticularis together with fetal dopaminergic cells into the striatum in parkinsonian rats. In contrast to isolated striatal dopaminergic transplantation, this co-transplantation approach was able to induce a significantly better improvement in the cylinder test and in the step adjustment test compared to the standard approach.


At this point, it appears legitimate to state that routine clinical application of cellular replacement is in sight, but still out of reach. Concerning both their proliferate and differentiation potential, pluripotent stem cells are a very promising replacement candidate, with iPSCs now even offering the option of an autologous cell source. Besides the important focus on cell sources and differentiation protocols, studies addressing alternative stereotactic transplantation strategies and transplant compositions promise to add valuable knowledge to the long way to go. It will be one of the most interesting questions in the future treatment of PD, whether cellular replacement will be able to live up to the hope that it will become an option equivalent to current treatment or even superior by being the first having a beneficial effect on disease progression.

Conflict of interest

The research of the authors was supported by the Dresden Medical Faculty Research Program MeDDrive, the Bundesministerium für Bildung und Forschung, the Deutsche Forschungsgemeinschaft (DFG) through the Sonderforschungsbereich 655 “From cells to tissues” and the DFG-Research Center and Cluster of Excellence “Center for Regenerative Therapies Dresden (CRTD)”, the Thyssen-Stiftung, and the Landesstiftung Baden-Württemberg.

Copyright information

© Springer-Verlag 2011