Review: The future of cell therapies and brain repair: Parkinson's disease leads the way

Authors

  • G. H. Petit,

    Corresponding author
    1. Neuronal Survival Unit, Department of Experimental Medical Science, Wallenberg Neuroscience Center, Lund, Sweden
    • Correspondence: Géraldine H. Petit, Neuronal Survival Unit, Department of Experimental Medical Science, Wallenberg Neuroscience Center, BMC B11, 221 84 Lund, Sweden. Tel: +46 46 222 05 26; Fax: +46 46 222 05 58; E-mail: geraldine.petit@med.lu.se

    Search for more papers by this author
  • T. T. Olsson,

    1. Van Andel Research Institute, Center for Neurodegenerative Science, Grand Rapids, MI, USA
    Search for more papers by this author
  • P. Brundin

    1. Neuronal Survival Unit, Department of Experimental Medical Science, Wallenberg Neuroscience Center, Lund, Sweden
    2. Van Andel Research Institute, Center for Neurodegenerative Science, Grand Rapids, MI, USA
    Search for more papers by this author

Abstract

During the past 40 years brain tissue grafting techniques have been used both to study fundamental neurobiological questions and to treat neurological diseases. Motor symptoms of Parkinson's disease are largely due to degeneration of midbrain dopamine neurones. Because the nigrostriatal pathology is relatively focused anatomically, Parkinson's disease is considered the ideal candidate for brain repair by neural grafting and dopamine neurone transplantation for it has led the way in the neural transplantation research field. In this mini-review, we briefly highlight four important areas of development. First, we describe marked functional benefits up to 18 years after transplantation surgery in patients with Parkinson's disease. This is proof-of-principle that, using optimal techniques and patient selection, grafted dopamine neurones can work in humans and the duration of the benefit exceeds placebo effects associated with surgery. Second, we describe that eventually protein aggregates containing α-synuclein, identical to Lewy bodies, develop inside foetal dopamine neurones transplanted to patients with Parkinson's disease. This gives clues about pathogenetic mechanisms operating in Parkinson's disease, and also raises the question whether neural graft function will eventually decline as the result of the disease process. Third, we describe new emerging sources of transplantable dopamine neurones derived from pluripotent stem cells or reprogrammed adult somatic cells. Fourth, we highlight an important European Union-funded multicentre clinical trial involving transplantation of foetal dopamine neurones in Parkinson's disease. We describe the design of this ongoing trial and how it can impact on the overall future of cell therapy in Parkinson's disease.

Introduction

Parkinson's disease (PD) is the most common neurodegenerative movement disorder. It affects around 10 million people globally and the prevalence will rise with an ageing population [1]. Pathologically, PD is characterized by intraneuronal inclusions of aggregated proteins (called Lewy bodies and Lewy neurites) and degeneration of the dopaminergic neurones of the substantia nigra. The protein inclusions are found throughout the brain and even outside the central nervous system [2]. The cardinal symptoms of PD are resting tremor, rigidity, bradykinesia and postural instability [1,3]. These motor disturbances are primarily due to the loss of dopamine neurones that project from the substantia nigra to the striatum. Traditionally, PD was viewed as a pure movement disorder, but it is now recognized that patients also exhibit nonmotor signs and symptoms such as sleep disturbance, autonomic dysfunction, hyposmia, cognitive decline and depression [2,4]. Thus PD is clearly a complex disorder with multiple signs and symptoms and a neuropathology that affects widespread brain regions. Nonetheless, the relative prominence of the motor disturbances and their close relationship to the loss of nigral dopamine neurones has meant that PD has been considered amenable to cell replacement therapy.

In this mini-review we will briefly describe key developments in the short history of clinical neural transplantation in PD. The first documented clinical transplantation of foetal nigral tissue into the striatum was performed in 1987. In the 1990s, a series of open label trial suggested that grafted dopamine neurones effectively reinnervate the striatum in PD and ameliorate several of the key motor deficits. In 2001 and 2003, two double-blind controlled grafting trials in PD were published. The results were disappointing with the outcome being negative in both studies, and troubling side-effects receiving more attention than any possible benefit from the transplants. As a result, clinical trials came to a standstill. In 2008, histopathological studies of dopaminergic neurones grafted more than a decade before the patients died revealed Lewy pathology inside the implanted cells. These unexpected findings raised concerns that PD actually affects transplanted cells and impairs their function. The findings also triggered a new and vibrant research field suggesting that the pathogenesis of PD involves prion-like behaviour of α-synuclein.

During the past few years, there has been renewed interest in cell replacement therapy in PD. We will describe that advances in stem cell research and the application of cellular reprogramming to human cells have meant that novel sources of dopamine neurones suitable for clinical transplantation will soon be available. Furthermore, the long-term outcomes of two patients who were part of the first open label transplantation trials have now been analysed revealing, marked and permanent motor improvement, combined with robust evidence for graft survival monitored using brain imaging. We will also discuss a European Union-funded multicentre trial (TRANSEURO) that will revisit foetal dopamine neurone transplantation in PD, using the best available cell preparation and patient evaluation protocols.

Cell replacement therapy, results so far

In the latter half of the 1980s and during the 1990s, several open label uncontrolled trials using human foetal dopamine neurones were performed in numerous centres worldwide. It has been estimated that around 400 PD patients were operated during this period [5]. A minority of these operated patients was reported in the scientific literature. In short, the outcomes were frequently described as positive, with moderate to marked gradual improvements in motor function being reported [5]. The parallel positron emission tomography studies showed that 18F-FDOPA uptake increased significantly in the grafted striatum, starting about 6 months after surgery and staying elevated for the full duration of the studies, that is, for at least a decade [5,6]. The positive indications obtained from these uncontrolled studies, suggesting that neural transplants might be effective in PD, stimulated the National Institutes of Health (NIH) to sponsor two major randomized double-blind sham surgery placebo controlled trials with neural grafts in PD [7,8]. The outcomes of both these trials were very disappointing to PD researchers and clinicians, as well as the patient community that had been hoping for a new therapy, which would have major disease-modifying effects. In both cases, the trials failed to meet the primary end-points, and several researchers advocated that this was definitive proof that neural transplantation simply does not work in PD. The positive signs in the earlier open label studies were deemed to have been the results of placebo effects and/or observer bias, and we discuss this in greater detail later in this section.

In addition to the failure to demonstrate benefit, these two NIH-sponsored trials highlighted that some grafted PD patients develop graft-induced dyskinesias (GIDs) following the operation. Although these affected only 15–50% of operated patients and were mild in most cases, there were also cases severe enough to require deep brain stimulation surgery [9]. Interestingly a retrospective study of patients included in open label studies showed that some of them had also developed GIDs [10], but typically this had not been a focus of the evaluation in the post-operative follow-up period. Alternatively, the involuntary movements had actually been noted, but the investigators and patients had not considered that the grafts, as opposed to anti-parkinsonian medication, might be triggering them. One has to remember that all of these patients had displayed L-dopa-induced dyskinesias (LID) prior to surgery. Therefore, in some cases it was not immediately apparent that the involuntary movements actually remained after temporary cessation of L-dopa medication and thereby were GIDs by definition. Following a decade of research in animal models of GIDs we now have a better understanding of when and why GIDs develop [11]. As described in more detail in the section below devoted to TRANSEURO, researchers today believe that GIDs can be minimized or avoided in future trials.

The discussion on why the NIH-sponsored trials had failed when the open labelled trials suggested that neural transplants could be effective almost split the field into two camps. While some maintained that the open label trials had been misleading, due to the lack of a placebo control and the fact that the investigators had not been blinded, others argued that the design of the NIH-sponsored trials had not been optimal. Several earlier reviews [5,12] have described in detail the shortcomings with the NIH trials. It has to be emphasized that neural transplantation trials are complex with multiple parameters that can be varied. Failing to make the right choice for just one of these critical parameters can jeopardize the outcome of the whole trial, resulting in a misleading conclusion on the overall value of neural transplantation in PD. Because the two NIH-sponsored trials differed significantly from the open-label trials regarding techniques and trial design, it is difficult to briefly summarize the potential problems with the trials. In short, the most important concerns focused on the preparation and storage of the donor tissue; the choice of potentially inadequate immunosuppression (or no immunosuppression at all); the selection of relatively short follow-up periods with outcome parameters that are exceptionally sensitive to placebo effects. For example, in one of the NIH-sponsored trials the donor tissue had been stored for up to 4 weeks prior to implantation, without this storage method being well validated [7]. In the same trial no immunosuppression at all was given (despite animal experiments showing that the ‘immune privilege’ of the brain is incomplete), and in the other NIH-sponsored trial only a single immunosuppressant (cyclosporin) was given for just 6 months [8]. Finally, despite open label trials showing that it takes several years for the effects to be fully developed [6,13,14], one trial had the final evaluation at 12 months [7] and the other at 24 months [8]. In the Freed/Fahn study, the primary outcome parameter was a subjective global rating of the change in the severity of disease after 1 year, scored by the patients [7]. Intuitively, this type of outcome parameter seems likely to be exceptionally sensitive to placebo effects. Interestingly, the same team of investigators reported 9 years later that when a subgroup of the Freed/Fahn cohort was followed up for 2–4 years after surgery, in an unblinded phase, the grafts were effective at reducing the PD motor symptoms [15]. At 2 years, clinical improvement was almost twice the level observed at 1 year and this was sustained at 4 years. Likewise, the increase of 18F-FDOPA uptake was evident at 2 and 4 years with significant clinical-PET correlations. This strongly suggests that the follow-up of the initial study was too short, and the chances that the neural grafts could meet the primary end-point after 1 year were almost zero.

A recent paper reporting on long-term follow-up of two grafted patients clearly showed that this therapeutic strategy needed years to reach its full effect. Two patients from the Lund open-label transplantation series were followed up to 15 (patient 15) and 18 (patient 7) years after surgery [13]. Patient 7 gradually improved his motor performance over the first 4 years post-transplantation. In patient 15 the transplants did not exert any effect on UPDRS motor scores during the first 2 years, indeed 4 years post-graft were necessary for motor improvement to become evident. Symptoms of LID were reduced for both patients following grafting. Both patients started to develop mild to moderate GID. However, the GIDs did not cause distress and disability and were experienced to be less of an impairment than LID. Most importantly, these two patients could stop their dopaminergic medication around 5 years post-graft and motor benefits are still constant at their last assessment, 15–18 years post-transplantation.

What is the importance of Parkinson-like neuropathology in the grafts?

A defining histopathological feature of PD is intraneuronal protein inclusions consisting mainly of α-synuclein; these are called Lewy bodies or neurites depending on localization.

In 2008, Lewy bodies and Lewy neurites were found in brain tissue grafted to PD patients more than a decade earlier [16,17]. One patient received grafts into putamen on both sides 4 years apart, 12 and 16 years before death. In the 12-year-old graft, 1.9% of the neuromelanin-positive grafted neurones exhibited Lewy body pathology and in the 16-year-old that number was 5% [18]. The findings of Lewy pathology were surprising because when the patients died the grafted neurones were of an age when one typically does not develop PD. In normal young individuals (<20 years) α-synuclein staining is not even detectable in the cytoplasm of nigral dopamine neurones [19]. One speculative explanation for the Lewy pathology in the grafted cells could be that they expressed increased endogenous levels of α-synuclein and that this increased the likelihood for aggregates forming and escaping the normal clearance mechanisms. One could imagine that the increased α-synuclein levels and aggregation would be promoted by the disease environment of the host brain which could feature pathological inflammation, oxidative stress and excitotoxicity, all of which are known to promote misfolding of α-synuclein [20]. However, the above pathological conditions are not unique to the PD brain, and grafted neurones implanted into the putamen of Huntington's disease patients (another neurodegenerative disorder associated with neuroinflammation) do not develop Lewy body pathology. In addition, the level of neuroinflammation varied between the grafted PD patients and was not correlated to the extent of Lewy body pathology in the transplant [21].

Another hypothesis, which could explain the presence of Lewy pathology in grafted neurones, is gaining increasing support. It states that misfolded α-synuclein transfers from host neurones to the grafted neurones and once inside the new cells triggers the formation of α-synuclein aggregates in a prion-like manner [20]. According to the prion-like hypothesis, misfolded α-synuclein transfers between cells. Once in a new cell, it can act as a template for further misfolding of α-synuclein, eventually leading to Lewy body formation, production of toxic species of α-synuclein and further release of misfolded α-synuclein into the extracellular space.

At the cellular and molecular levels, there is now a great deal of support for this prion-like hypothesis. Thus, cell culture studies have shown that (i) α-synuclein can spread between cells in a partially endocytosis-dependent manner [22–25]; (ii) misfolded exogenous α-synuclein can recruit endogenous α-synuclein thus acting as a ‘seed’ [23,26]; and (iii) α-synuclein can be axonally transported [27]. Axonal transport could be the means by which misfolded α-synuclein spreads between brain regions, or, in this context, specifically from the host brain into the grafted tissue. Animal studies also indicated transfer of host α-synuclein into neurones grafted into hippocampus and striatum of mice and rats overexpressing human α-synuclein, mimicking the results seen in humans receiving grafts [22,28–30].

The idea that α-synuclein aggregates form in grafted neurones following a prion-like process has also stimulated researchers to consider that this might also be how α-synuclein pathology spreads throughout the brain in PD. Cerebrospinal fluid from PD patients show higher levels of oligomeric α-synuclein [31] as well as lower levels of total α-synuclein, suggestive of an increase of insoluble α-synuclein [32]. Transfer of α-synuclein between neurones is also in line with the suggestion of Braak and colleagues that the neuropathology of idiopathic PD follows an anatomically stereotypic pattern [33]. Braak and collaborators have proposed a neuroanatomical staging scheme of PD; Lewy pathology is first seen in the olfactory and vagal nuclei, proceeds up the brain stem and the midbrain and finally it reaches cortical areas [2]. Braak and co-workers never suggested that misfolded α-synuclein was the culprit and actually hinted that a neurotropic virus might be the agent that transferred from one cell to another in [34]. However, the recent demonstrations that α-synuclein can act in a prion-like manner might now provide an explanation for Braak's neuropathological stages of PD. Considering that the olfactory system and motor vagal nerve have been proposed to display α-synuclein aggregates first in PD, it has been suggested that this is where the first α-synuclein misfolding events take place [35].

The mere presence of Lewy bodies in a small proportion of the grafted cells might not affect the graft function. Lewy bodies are found in almost all PD brains, but they are also found in 5–20% of nonsymptomatic individuals older than 60 years [36]. However, a significant proportion of grafted cells also displayed decreased levels of dopamine transporter (DAT) and tyrosine hydroxylase (TH). This suggests impaired function in a subset of grafted dopamine neurones. Furthermore, cytoplasmic α-synuclein staining appeared between 1.5 and 4 years after transplantation, increased with graft age (80% of cells at 16 years) and was inversely correlated with DAT levels [18,19]. Taken together, these histological findings support the idea that the grafts are actually affected by PD pathology.

With PD spreading to the graft, is cell replacement still a valid therapeutic approach? For most patients the answer is yes, as we have seen, some patients show motor improvements for up to 15–18 years [13]. The subset of cells exhibiting Lewy bodies is small and even in a 22-year-old graft, the oldest graft examined so far, a majority of the cells containing neuromelanin still expressed TH [37]. Even if at some time point, several decades later, the spread of Lewy pathology causes the graft to fail one can argue that for many patients the benefits outweigh the costs.

In summary, grafted cells develop PD pathology similar to the one of the host nigral neurones and the prion-like spread of α-synuclein could be the underlying mechanism. This probably leads to cellular dysfunction that reduces graft function. Considering the recent report on highly functional grafts 15–18 years after surgery in two patients (described above) [13], α-synuclein pathology in the grafts does not invalidate neural grafting as a potential clinical therapy. With further advances in stem cell technology, which will be discussed in the next section, it might be possible to reduce the uptake of α-synuclein, prevent its misfolding or increase its degradation to better maintain protein homeostasis and grafted cell function.

New sources of cells for grafting in PD

The use of foetal neural cells for transplantation in PD comes with major logistical and ethical issues. Even though the clinical trials with transplantation of foetal ventral mesencephalic neurones showed encouraging results, grafting foetal tissue will never be a widespread therapy as tissue availability is limited and the ethical issues that the use of foetal tissue raises in many cultures. New cell sources for transplant in PD are clearly necessary for this to become a widespread therapy. Currently, there are four main candidates: (1) human embryonic stem cells (hESC); (2) induced pluripotent stem cells (iPSC); (3) induced neuronal cells (iN cells); and (4) induced dopaminergic cells (iDA cells) (Figure 1). In this section, we will briefly describe these different cells types and their pros and cons for cell replacement therapy.

Figure 1.

Schematic of the possible sources of cells for transplantation in Parkinson's disease. Donor tissues including foetal mesencephalic tissue, blastocyst-derived hESC and patient-derived somatic cells could be used to generate dopaminergic cells.

Human embryonic stem cells

hESC are pluripotent stem cells derived from the inner cell mass of blastocysts. The first hESC line was reported 15 years ago [38]. They have unlimited self-renewal capacity and can be differentiated to dopaminergic neurones [39]. Two major concerns are that the hESC-derived dopamine neurones need to exhibit the correct ventral midbrain dopaminergic identity (as opposed to other types of dopamine neurones) and that they must survive, grow axons and function after transplantation. Through the activation of canonical Wnt signalling, hESC can exhibit rapid and efficient differentiation into neuronal progenitors. When such neural progenitors are transplanted into the immunosuppressed rat striatum, many develop a dopaminergic phenotype in terms of morphology, projection pattern and protein expression [40]. A similar protocol with, in addition, an activator of sonic hedgehog also demonstrated long-term survival in mice, rats and monkeys [41]. The clinical use of hESC still faces a major obstacle. Their high capacity of self-renewing and differentiation make them prone to form tumours, mostly teratomas or overgrowth of neural progenitors. Moreover, the immunogenicity, which is already a limitation of foetal nigral grafts and can lead to graft rejection, is still unresolved with hESC [42]. Finally, the origin of the hESC still raise ethical issues that may compromise its use for clinical therapy [43].

Induced pluripotent stem cells

iPSC are somatic cells, such as fibroblasts, that are reprogrammed into pluripotent cells, which can be differentiated to dopaminergic neurones [12,44]. A first study showed that the expression of only four factors (Oct4, Sox2, c-Myc and Klf4) are enough to reprogram mouse adult fibroblasts to pluripotent cells [45]. This seminal discovery led to Dr Shinya Yamanaka being awarded the Nobel Prize (shared with Dr John Gurdon) in Medicine or Physiology in 2012. Later, human fibroblasts were also reprogrammed into iPSC [46,47]. Neural progenitor cells derived from human iPSC (hiPSC) have been transplanted in rat models of PD and been reported to improve motor behaviour [48]. Similar cells were found to survive as long as 6 months in a primate model of PD [49]. This remarkable possibility to reprogramme differentiated cells to pluripotent cells has opened several new perspectives in cell replacement therapies. This approach solves ethical issue by avoiding embryonic or foetal cells. With hiPSC it also becomes conceivable to use patient-specific cells for cell replacement therapy. Thereby, the grafted cells will be genetically compatible and hence avoid immune rejection. One major issue with hiPSC-derived transplants is the use of viral vectors that will induce insertion of genes in multiple sites of the genome. This raises regulatory concerns and is not compatible with clinical therapy. However, this issue seems to be resolved in a recent study showing that iPSC could be reprogrammed from mouse somatic cells using only small molecules [50]. Although hiPSC cells are very promising as a source of transplantable neurones, at least two major hurdles still need to be passed before clinical use. First, it is still a challenge to get a high proportion of hiPSCs to differentiate into dopaminergic neurones. Second, the hiPSC have the same proliferative capacity as hESC and so they can still potentially induce tumour formation and neural overgrowth after transplantation.

Induced neuronal cells

iN cells are somatic cells directly transformed into neurones without going through any undifferentiated state. Using expression of three neural-lineage-specific transcription factors (Ascl1, Brn2 and Myt1l), mouse fibroblasts were directly converted to functional neurones generating action potentials and forming functional synapses [51]. This direct conversion was later observed using human fibroblasts [52]. This direct strategy, which avoids pluripotent cells, addresses concern regarding the proliferative capacity of hESC and iPSC, which was a major issue for clinical application. However, to be a good source of cells for cell replacement in PD, iN cells should be able to generate a higher proportion of dopaminergic cells [52].

Induced dopaminergic cells

Particularly relevant to transplantation in PD, recent studies have demonstrated that it is possible to directly reprogramme human fibroblasts into dopaminergic neurones, that is, create induced dopaminergic cells (iDA cells). One approach to achieve this is to use the three neural-lineage-specific transcription factors (Ascl1, Brn2 and Myt1l) in addition to the expression of two transcription factors involved in dopaminergic neurone differentiation (Lmx1a and FoxA2) [53]. Another approach identified a set of only three transcription factors (Nurr1, Mash1 and Lmx1a) which supported the generation of dopaminergic neurones directly from mouse and human fibroblasts without reverting to undifferentiated cells [54]. Interestingly these iDA cells exhibited properties such as dopamine release and specific spontaneous electrical activity. As for iN cells, the conversion rates from fibroblast to iDA cells is still relatively low. Furthermore, because they are patient-derived cells and could potentially express PD susceptibility genes, the use of iDA cells for clinical trials would require further investigation to determine whether they will be safe or might degenerate after grafting.

As mentioned above, foetal cell transplantation can only be a transitory step in cell therapy for PD due to limitations of tissue availability and the ethical issues involved. Therefore, cell replacement therapy is likely transitioning into an era with new emerging sources of transplantable dopamine neurones derived from pluripotent stem cells or reprogrammed adult somatic cells. However, even though this field is progressing quickly and will likely lead to future stem cell therapies in PD, several hurdles must be overcome in the next 5–10 years including the safety and regulation of transplanted cells. For instance, a major hurdle that must be solved is the risk of cell proliferation and tumour formation of transplanted cells. The future production of dopamine cells for transplantation must be completely free of proliferative cells [55]. Another important obstacle for reprogramming adult somatic cells is the percentage of dopaminergic cells derived, which must be improved for clinical relevance.

A new European multicentre trial: TRANSEURO

From the clinical trials described in the first section we learnt that ventral midbrain dopaminergic neurones can be safely transplanted into the striatum of PD patients. In open-label trials, beneficial effects were more evident after long (2–4 years) follow-up and for some patients the graft enabled them to discontinue their medication. Despite these encouraging results, limitations of these first clinical trials appears to be the high variability of the outcomes as well as, for some patients, the development of GID. As alluded to above, variability in clinical procedures ranging from patient selection, to cell preparation, surgical method and trial design might contribute to the differences in outcome [12]. In this context, a new European multicentre project called TRANSEURO, led by Dr Roger Barker, Cambridge, UK, initiated in 2010 (http://www.transeuro.org.uk). The objective of TRANSEURO is to optimize four critical issues: patient selection, tissue composition, tissue placement and, finally, trial design/end-points. These issues will be addressed in a small open label study of foetal ventral midbrain transplants to patients with early PD.

Patient selection

Cohorts of PD patients have identified subgroups of patients that might not respond to the same extent to dopaminergic cell therapy [56–58]. Some young patients, who may have a degree of frontostriatal cognitive impairments, seem to have a more benign disease course with a restricted nigral pathology and would be the ideal target group for cell transplantation therapy. In contrast, another profile of patients, older with MCI (Mild Cognitive Impairment), displays a more posterior cortical impairment and widespread pathology and will probably not be the most responsive patients to dopaminergic cell therapy focusing on the nigrostriatal pathology.

Disease severity should also be taken in consideration in the patient selection criteria. Interestingly, the double-blind trial performed in Olanow/Freeman trial showed a significant effect of cell transplantation in the less severe disease cases [8]. Moreover, the development of LID prior to cell transplantation in a rat model of PD, is correlated with the severity of GID [59,60]. So to improve the efficiency of cell replacement therapy and to reduce the risk for GID, the TRANSEURO trial will select patients with no cognitive impairment, who are younger than 65 years old, have less than 10 years of disease duration and do not show significant LID.

Tissue composition

Cell composition may also be an important factor in the graft efficiency and on the risk of developing GID. Some data suggest that the presence of serotonergic neurones might play a role in the development of GID. Two transplanted patients that developed GID, had their dyskinesias suppressed after the administration of a serotonin 5-HT1A receptor agonist (buspirone) that inhibits the activity of serotonergic neurones [61]. In a rat model of GID, induced by low dose of amphetamine administration, the role of serotonin neurones in the graft is less clear. Endogenous serotonin neurones (projecting from the raphe to the striatum) seems to play a modulatory role on expression of GID [62] and serotoninergic neurones in the graft might be involved in the development of GID, but given sufficient dopaminergic cells, their impact might be minor [63]. However, to minimize the potential side-effect of serotoninergic neurones in the graft, the dissection will be standardized and optimized to get a high ratio of dopaminergic to serotoninergic neurones.

The number of dopaminergic neurones in the graft is crucial to the reduction of motor symptoms. To collect enough foetal tissues to transplant, and to allow for informed consent to be obtained from the woman undergoing abortion, some tissue in some trials had been stored up to 1 month [7]. However, with prolonged time in storage, dopaminergic cell survival decreases drastically [64] therefore an optimization of the hibernation protocol to limit dopaminergic neurone death is essential. Tirilazad mesylate, a lipid peroxidation inhibitor, appears to be a good candidate [65]. PD patients grafted with less mesencephalic tissue but prepared with tirilazad mesylate showed similar motor improvement and graft survival as patients transplanted with 42–50% more tissue but not treated with tirilazad mesylate [66]. To optimize the amount of tissue available for transplantations in TRANSEURO trial, the tissue collection will be performed in several centres with possible transfer of tissue between centres, even across national borders, and stored up to a maximum 4 days in a hibernation medium with tirilazad mesylate.

Tissue placement

Surgical technique, in particular tissue placement, is another variable factor observed between clinical trials. An uneven increase in dopaminergic reinnervation (creating so-called ‘hot spots’) in the striatum has been suggested to promote the development of GID in PD patients [67]. Hence proper distribution of the grafted material across the striatum is essential. To minimize GID and to ensure the same operating procedure in the different centres, the TRANSEURO clinical trial will use a standardized instrument for grafting and will deliver the tissue through five to seven tracts to the posterior putamen.

Trial design and end-points

Trial design is a real challenge in cell transplantation especially regarding the question of appropriate controls and placebo effects [68]. In acute studies involving pharmacotherapies, the patients' expectations regarding treatment benefit can potentially induce dopamine release interfering with the proper treatment effect [69]. Nevertheless, an open-label trial should provide conclusive findings if the duration of the follow-up is long enough (minimizes placebo effect and optimizes effect of the graft) and if the assessment is done in an unbiased way [5]. The TRANSEURO design includes an observational study on 150 patients. Out of these 150 patients, 40 that fulfil inclusion/exclusion criteria will be randomized, to go through positron emission tomography and be assigned to either the transplant group or the control group. All motor assessments done in all 150 patients will be rated blindly. As we previously mentioned, the duration of the follow up has been shown to be crucial in the interpretation of the outcomes. In the new TRANSEURO trial, the primary end-point will be 3 years post-transplantation. As there is uncertainty from the comparisons between the open label and NIH-sponsored trials whether immunosuppression is important, TRANSEURO has adopted a cautious approach and will immunosuppress the patients for a year to minimize the risk for immune rejection.

Conclusion

The first open label clinical trials on foetal dopaminergic cell replacement demonstrated the potential of this therapy for a subset of patients. The subsequent NIH-sponsored trials failed to meet their primary end-points and were tainted by unwanted side-effects of the grafts. However, it is fair to suggest that inadequate transplant protocols (tissue storage, poor immunosuppression) and suboptimal trial design (short follow-up, subjective outcome parameters, etc.) contributed to their failures. It is reasonable to suggest that with an improved transplant protocol and more careful patient selection, future neural transplantation studies in PD, like TRANSEURO, stand a good chance of generating outcomes with fewer side-effects and better symptomatic relief. Studies of neural grafts have yielded important insights into the pathogenesis of PD and significantly contributed to the development of the prion-like hypothesis regarding the spread of α-synuclein. Finally, new cell technologies based on pluripotent stem cells or reprogrammed adult somatic cells could give us an abundance of safe dopaminergic neurones for transplantation in the future.

Acknowledgements

GHP and PB acknowledge the support of the EU-FP7 for funding TRANSEURO. PB has a European Research Council Advanced Award (PRISTINE-PD). TTO is supported by a travel grant from Svenska Läkaresällskapet (The Swedish Society of Medicine). All authors are active in the Strong Research Environment Multipark (Multidisciplinary research in Parkinson's disease at Lund University).

Conflict of interest

GHP and TTO have no conflict of interest. Of relevance to this review, PB is a cofounder and shareholder for the biotechnology company ParkCell AB.

Ancillary