Cell-Based Therapies for Parkinson’s Disease: Preclinical and Clinical Perspectives

Parkinson’s Disease (PD) is a highly prevalent neurodegenerative disease that affects millions of people globally and remains without definitive treatment. There have been many recent advances in cell-based therapy to replace lost neural circuitry and provide chronic biological sources of therapeutic agents to disease-affected brain regions. Early neural transplantation studies highlighted the challenges of immune rejection, graft integration, and the need for renewable, autologous graft sources. Neurotrophic factors (NTFs) offer a potential class of cytoprotective agents that may complement dopa‐ mine (DA) replacement and cell-based therapies in PD. In fact, chronic NTF delivery may be an integral goal of cell transplantation in PD, with ideal grafts consisting of autologous drug (e.g., DA, NTF)-producing cells capable of integration and function in the host brain. This chapter outlines the past and recent preclinical and clinical advances in cell-based and NTF therapies as promising and integrated approaches for the treatment of PD. PD patients. NTFs offer a potential class of cytoprotective agents that complement DA replacement and cell-based therapies in PD, with ideal grafts consisting of immunologically inert cells that continuously produce and release these agents in the host brain. Further development and refinement of these potential therapies is essential to develop personalized care for PD patients.


Introduction
Parkinson's disease (PD) is the second most common neurodegenerative disorder, following Alzheimer's disease. In the developed world, the prevalence of PD is approximately 0.3% of the population and 1% of those over 60 years of age [1]. Hallmarks of PD include degenera-and tremor, respectively [12,13]. Pallidotomy has been demonstrated to provide sustained improvement for tremor, rigidity, bradykinesia, and drug-induced dyskinesias, compared with medical therapy [14]. In the past decades, deep brain stimulation (DBS) has become the standard of surgical care for PD owing to the versatility of stimulator programming, and the avoidance of creating a permanent surgical lesion [15]. The two primary targets of DBS are the internal globus pallidus (GPi) and subthalamic nucleus (STN). DBS improves motor symptoms and often permits a reduction of medication dose and associated side effects, but does not slow or halt progression of the disease [16].

Cell-based therapies
As PD is characterized by the loss of dopaminergic nigrostriatal neurons, cell-based therapies initially focused on the potential to replace these neurons and replenish DA supply in the striatum using fetal mesencephalic neural grafts. More recently, studies have included the transplantation of induced pluripotent stem cells (iPSCs), reprogrammed somatic cells or induced neural progenitor cells (iNPCs).

Preclinical studies
Early PD transplantation studies involved grafting of fetal ventral mesencephalic (fVM) tissue into the anterior chamber of the rat eye [17]. These studies identified the optimal developmental stage of the neural tissue to be used to promote DA neuron survival and outgrowth [17]. The first graft transplantation studies in a unilaterally lesioned 6-hydroxydopamine (6-OHDA) PD rat model examined the effects of solid grafts of fetal adrenal medullary or fVM tissue implanted into the lateral ventricle or preformed cavities adjacent to the striatum and reported reduced amphetamine-induced rotation behavior [18]. Subsequent studies showed that grafting cell suspensions of fVM tissue from 14-to 15-day-old rat fetuses into the striatum of 6-OHDA rats also reduced amphetamine-induced rotation behavior [19]. Follow-up studies used fVM tissue from 9-to 19-week-old human fetuses. These implants reduced and even reversed motor asymmetry in unilaterally lesioned 6-OHDA rats [20].

Clinical studies
In 1987, solid graft adrenal medullary transplants were implanted in the head of the caudate in two patients and produced significant clinical improvement, including reduced tremor [21]. Unfortunately, follow-up studies showed only modest clinical effect, with concerns regarding efficacy and safety of this technique. Many patients suffered major postsurgical complications and psychiatric problems, thus this transplant approach was abandoned. Subsequent open label studies in six human patients utilized human fVM tissue from 6-to 8-week-old fetuses grafted into the caudate and putamen, demonstrating overall clinical improvement and normal DA signaling seen by 18 F-Fluorodopa ( 18 F-FDOPA) uptake in Positron Emission Tomography (PET) imaging [22][23][24]. Two patients in this study continued to demonstrate clinical improvement 20 years later [25]. In a subsequent open label study nigral grafts from 6-to 7-week-old embryos were implanted into the caudate and putamen of seven PD patients [26,27]. Significant improvement in the activities of daily living was noted after 12 months, in both "on" and "off" states. The dose of levodopa could be reduced by an average of 39%. Four patients reported an "important difference in their daily lives," two patients reported improvement in "some respects," and one patient did not improve. Other open label studies with a small number of patients also showed mostly beneficial effects ( [28][29][30][31], reviewed in [32]).
A randomized double-blind controlled trial (RDBCT) enlisted 34 patients that underwent transplantation of fetal mesencephalic tissue into the putamen or sham surgery. The patients showed limited clinical improvement, despite graft survival and significant reinnervation of the striatum as confirmed by PET and at autopsy. Interestingly, patients with less severe motor dysfunction showed significant clinical improvement, suggesting this technique may have produced some degree of neuroprotection. Furthermore, graft-induced dyskinesias were observed in over half of the patients [33]. Interestingly, these patients underwent a 2-and 4year follow-up RDBCT that demonstrated clinical improvement regardless of the age of the patient, which was accompanied by significantly increased 18 F-FDOPA uptake in the putamen [34]. Another RDBCT had 40 patients with advanced PD undergo transplantation. When results were normalized according to age, patients under the age of 60 showed significant clinical improvement, as measured with the Unified Parkinson's Disease Rating Scale (UPDRS) and Schwab and England score, while those over the age of 60 did not [35]. Clinical improvement was correlated with increased 18 F-FDOPA uptake in 85% of patients with a transplant and postmortem examination confirmed dopaminergic cell survival and fibre outgrowth; however, 15% of patients developed graft-induced dyskinesias or dystonia [35].
To more accurately assess the potential of fetal grafts, a new European study has been designed to optimize and control for patient selection, tissue composition, tissue placement, and trial design. TRANSEURO is an open label multicenter trial to define the feasibility and efficacy of human fetal ventral mesencephalic grafts in patients with PD (https://clinicaltrials.gov/ct2/ show/NCT01898390). The primary outcome measure of this study is the change in motor UPDRS scores in the absence of PD medications at 3 years posttransplantation. It is hoped that this new trial will shed light on the true potential of dopaminergic allografts for PD treatment.
The use of human fVM tissue, however, is complicated by ethical issues and difficulty in obtaining human tissue. Strategies are being developed that involve expansion of fVM tissue and its dopaminergic neuroblasts [36], and other cell sources are also being investigated for cell-based treatment of PD.

Preclinical studies
Human embryonic stem cells (hESCs) were first isolated from the inner cell mass of blastocysts. These cells demonstrate pluripotency and have been shown to differentiate into neural cells, including neurons, astrocytes, and oligodendrocytes [37,38]. hESCs may prove useful to avoid the technical concerns associated with the use of fetal tissue. hESCs have been shown to differentiate into midbrain DA neurons, and injection of these cells in 6-OHDA lesioned rats [39] and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-lesioned monkeys [40] leads to significantly improved motor function in both models. However, the development of clinical applications using these cells has been slowed by various biological and social factors, including the potential for immune rejection and tumor formation, as well as ethical and political opposition [41].
Alternative stem cell sources have been investigated for differentiation into a neural lineage, in particular mesenchymal stem cells (MSCs) from bone marrow, umbilical cord blood, dental pulp, and adipose tissue [42]. Autologous MSCs are favorable due to their availability, potential for differentiation, and the absence of ethical issues associated with hESCs. In addition, MSCs have been demonstrated to exert regenerative and neuroprotective effects in a number of animal PD models potentially, due to endogenous NTF expression. MSCs have been shown to differentiate into dopaminergic cells that express tyrosine hydroxylase [43]. Implantation of these differentiated MSCs into the striatum of 6-OHDA mice led to significant behavioral improvement, with striatal graft cells confirmed at postmortem analysis [43].

Clinical studies
There is currently very limited clinical data on the efficacy of hESCs in PD. An initial clinical study in seven PD patients demonstrated that transplantation of autologous bone-marrowderived MSCs was safe and feasible with no serious adverse side effects. Unfortunately, no clinical efficacy was observed, potentially due to the small number of patients and uncontrolled nature of the trial [44].

Preclinical studies
With the discovery that somatic cells, such as fibroblasts, can be reprogrammed to a pluripotent state by viral delivery of four transcription factors, Oct4, Sox2, Klf4, and cMyc [45,46], studies have focused on the potential of these induced pluripotent stem cells (iPSCs) to improve current cell-based therapies for treatment of many degenerative medical conditions, including PD. Compared with fetal grafting and hESC cells, iPSCs provide increased accessibility as well as ethical advantages. These cells can differentiate into many different cell types including cardiomyocytes [47], hepatocytes [48], oligodendrocytes [49], glia, and neuronal subtypes [50]. Murine iPSCs have also been shown to be reprogrammed into dopaminergic neurons that express the transcription factors Nurr1, Pitx3 and tyrosine hydroxylase and demonstrate electrophysiological properties of DA neurons [51]. Subsequent studies demonstrated successful differentiation of dopaminergic neurons from both established human iPSC lines and patient-derived somatic cells [52,53].
In all of these studies, iPSC-derived DA cells demonstrated expression of key dopaminergic markers and electrophysiological properties of DA neurons. Furthermore, these DA cells were successfully incorporated into a 6-OHDA rat model of PD, leading to significantly reduced motor asymmetry [52]. Most recently, primate-derived iPSCs were successfully transplanted back into the putamen of MPTP lesioned monkeys; these autografts led to significant motor improvements, and postmortem analysis showed graft survival and outgrowth into the transplanted putamen [54]. Despite promising results in preclinical studies, several factors have provided road blocks to utilization of these cells in humans, including use of viruses to modify cells and risk of tumorigenicity [55].

Clinical studies
The first pilot study in humans was performed in Japan in 2014, and utilized iPSC-derived autologous pigmented retinal epithelial cells for treatment of macular degeneration. Transplantation in the first patient was completed without adverse effects; however, long-term follow-up is necessary [56]. Unfortunately, iPSCs derived from fibroblasts of a second patient were discovered to have genetic mutations, including three single-nucleotide variations and three copy-number variants, prompting suspension of the trial [57]. Studies are now focusing on use of allogenic partially matched donor cells from the Center for iPS Cell Research and Application, an iPSC bank, for treatment of macular degeneration [57]. There is also potential to transform fibroblasts directly into neurons (iN cells) or dopamine cells (iDA cells) using specific transcription factors: Ascl1, Brn2, Mrt1l, without or with Lmx1a and FoxA2, respectively [58][59][60]; however, this technology still needs to be tested in preclinical models.

Autologous brain-derived progenitor cells (BDPCs)
Recently, the safety and feasibility of performing small volume brain biopsies has been demonstrated in PD patients undergoing DBS surgery [61]. These tissue specimens yield an expandable cell population that expresses several NTFs known to be highly protective against PD neurodegeneration, including glial-derived neurotrophic factor (GDNF), brain-derived neurotrophic factor (BDNF), and cerebral dopamine neurotrophic factor (CDNF) [61]. The cultures yield large numbers (i.e. 10 7 ) of cells with limited capacity for self-renewal. These cells, called BDPCs, also show expression of progenitor and neural markers including nestin, Olig1, and GalC. Colocalization of neural and oligodendroglial markers suggests these BDPCs may be grafted into the host brain to integrate as autologous glia [61]. A patient-derived cellular source of neuroprotective agents, reimplanted into the host brain, may confer long-lasting therapeutic benefit in PD patients. Preclinical studies on the potential for these BDPCs to be used as an autologous cell-based therapy for PD are currently underway.

Neurotrophic factors
Neurotrophic factors (NTFs) are being intensively evaluated as therapeutic agents for PD owing to their known roles in neuronal survival, differentiation, and plasticity. Additionally, NTF deficiency has been associated with PD and replacement or enhancement of NTF signaling confers neuronal protection in both in vitro and in vivo preclinical PD models [62]. These secreted proteins regulate vital biological programs in the developing and adult nervous systems and are currently the most potent cytoprotective agents known against PD-related degeneration in the brain [63]. The NTF families include (1) glial-derived neurotrophic factor (GDNF) family of ligands (GFL), (2) neurotrophins, (3) neuropoietic cytokines (neurokines), and (4) cerebral DA neurotrophic factor (CDNF)/mesencephalic astrocyte-derived neurotrophic factor family (MANF). To date, GDNF and other GFL members have received the most attention in development of potential new clinical therapies for PD [64].

GDNF family of ligands
GDNF was the first member of the GDNF family of ligands (GFL) to be discovered. Other members include neurturin (NRTN), persephin (PSPN), and artemin (ARTN). GFLs are important for cell survival, neurite outgrowth, cell differentiation, and cell migration [65].

Preclinical studies
Application of GDNF to rat ventral mesencephalic cultures increased survival, neurite length, and differentiation of DA neurons [66]. GDNF also reduced apoptosis and enhanced cell survival in cultures derived from monkey, porcine, and human mesencephalic tissues [67]. These effects extend to promote differentiation and protection of dopaminergic neurons against 6-OHDA and MPTP neurotoxins [65]. Numerous in vivo studies have demonstrated therapeutic effects of GDNF in the 6-OHDA rat model of PD; the infusion of recombinant GDNF significantly increased DA neuron survival in both the SN and ventral tegmental area and improved parkinsonian symptomology, including motor impairments and amphetamine-induced rotational behavior [68][69][70][71]. GDNF has also been shown to be neuroprotective in the MPTP mouse model of PD [72,73]. Studies in nonhuman primates have demonstrated that intracerebral administration of GDNF in MPTP-treated rhesus monkeys results in significant improvements in bradykinesia, rigidity, and postural instability, as well as increased DA levels in the midbrain, globus pallidus, and SN [74].

Clinical studies
Based on the success of using GDNF in preclinical models of PD, GDNF has now been studied in four clinical trials via infusion into the ventricular system or putamen [75]. The first RDBCT compared effects of intracerebroventricular administration of recombinant methionyl human GDNF (r-metHuGDNF, Liatermin®; Amgen) in escalating doses or placebo in 50 PD patients over a period of 8 months. No significant improvement in "on" and "off" total and motor UPDRS was seen in patients treated with GDNF. Adverse effects included paresthesias, nausea, and vomiting. A follow-up open label study in 16 of these patients for 20 months showed no additional improvement in PD symptomology. It was felt that the adverse effects resulted from off-target GDNF influence and the lack of therapeutic benefit from an inability of GDNF to diffuse into the parenchyma from the ventricular source [76].
A subsequent open label study that enrolled 5 PD patients investigated the effects of intraparenchymal delivery of GDNF via implanted catheters in the dorsal putamen (unilateral in one patient; bilaterally in four patients) and connected to an extracranial pump system [77].
After one year, there were no serious clinical side effects, a 39% improvement in the offmedication UPDRS motor scores and a 61% improvement in the activities of daily living (ADL) subscore. Medication-induced dyskinesias were considerably reduced and (PET) scans of 18 F-FDOPA uptake showed a significant 28% increase in putamen DA storage after 18 months [78]. In a follow-up report, the group described one of the patients with bilateral GDNF infusions who had received treatment for 39 months, then was followed clinically and with PET for another 36 months. The UPDRS motor and ADL scores "off" medication remained improved by 74% and 76%, respectively, levodopa usage ceased after a year, and at 36 months post-GDNF cessation, the 18 F-FDOPA uptake remained 29% higher in the posterior putamen [79]. Another group led a second open label study that enrolled 10 patients treated unilaterally with intraputamenal GDNF [80]. A significant increase in total and motor UPDRS scores was observed after 24 weeks, but benefit was lost with cessation of treatment. These positive outcomes spurred a second multicenter, placebo-controlled trial in which 34 PD patients were randomized to receive bilateral intraputamenal GDNF (15 μg/putamen/day; a dose lower than that of the previous studies) or placebo via continuous infusion. At 6 months, there was no significant treatment benefit reflected in the "off" UPDRS motor scores; however, a 32.5% increase in putamenal 18 F-FDOPA uptake was observed in the GDNF-treated cohort [81]. The disparate outcomes of these studies may reflect differences in study design, cohort size, drug dosage, and/or delivery systems. The r-metHuGDNF manufacturing company subsequently withdrew the agent on the grounds of safety concerns regarding production of neutralizing antibodies in several patients and related cerebellar injuries in animal studies, although no such injuries were reported in human trials. Efforts are now underway to evaluate adenoassociated virus (AAV)-mediated GDNF in an open label phase I for patients with advanced PD (https://www.clinicaltrials.gov/ct2/show/NCT01621581).

Preclinical studies
Neurturin (NTRN) shares 40% sequence homology with GDNF [82] and has been shown to promote survival of DA neurons in the nigrostriatal system [82][83][84]. In vitro, NTRN leads to neurite outgrowth in cultured spinal motor neurons and protects against glutamate toxicity. Early studies infusing NTRN directly into the SN was shown to be neuroprotective against 6-OHDA toxicity, while striatal infusion improved behavioral parameters of DA neuronal function in rats [83,85]. In MPTP-lesioned monkeys, intraputamenal infusion of NTRN led to significant improvement in parkinsonian deficits as well as increased DA metabolite levels in the globus pallidus [86]. CERE-120, an (AAV) vector expressing NTRN, has also shown potential therapeutic benefit in preclinical studies [87]. When MPTP-lesioned monkeys were given CERE-120 into the striatum, motor symptoms were improved and loss of DA neurons was reduced [88]. After one year follow-up, no toxic adverse effects were observed [89].

Clinical studies
A Phase 1 open-label clinical trial demonstrated safety, tolerability, and potential therapeutic benefit in PD patients after one year [90]. A subsequent RDBCT enrolled 58 patients to receive AAV2-NTRN bilaterally into the putamen or sham surgery. The primary endpoint was change from baseline to 12 months in the UPDRS motor score in the off state, and no significant difference was found between patients treated with AAV2-NTRN compared with control individuals. Three of 38 patients in the AAV2-NTRN group and two of 20 in the sham surgery group developed tumors, with uncertain relations to the actual treatment [91]. Postmortem analysis of two patients revealed that, unlike the animal studies, putamenal AAV-NTRN injections did not confer adequate retrograde labeling of neurons in the SN [92]. This deficiency in axonal transport of AAV-NTRN to the SN was addressed in a phase 1 safety study that enrolled six patients who received bilateral dual injections into the putamen and SN [93]. Twoyear follow-up suggested that the procedures were well-tolerated and no serious adverse effects were reported. A second phase 2 RDBCT was then conducted, enrolling 51 patients to receive bilateral putamen and SN AAV-NTRN (https://clinicaltrials.gov/ct2/show/ NCT00985517). In 2013, it was announced that the trial did not demonstrate statistically significant improvement in patient UPDRS scores after 15-24 months of follow-up. However, a more robust response to CERE-120 was observed in PD patients treated within 5 years of diagnosis, and no safety concerns were raised. There was a marked placebo effect as the control patients and the CERE-120 treated patients both improved significantly following surgery. Long-term observational studies of the participants are planned to assess delayed clinical effect (http://www.prnewswire.com/news-releases/ceregene-reports-data-from-parkinsonsdisease-phase-2b-study-203803541.html).

Persephin/artemin
Persephin (PSPN) shows approximately 40% sequence homology to GDNF and NTRN [94]. PSPN promotes survival of cultured ventral midbrain dopaminergic neurons as well as motor neurons and prevents their degeneration after 6-OHDA toxicity [94]. PSPN-overexpressing neural stem cells grafted into the striatum prevented loss of DA neurons and led to behavioral improvements in 6-OHDA lesioned rats [95]. Artemin (ARTN) promotes survival of DA neurons in culture [96] and also protects against DA neuron degeneration in the striatum following neurotoxic doses of methamphetamine [97]. Although early preclinical studies have shown therapeutic benefit of both PSPN and ARTN, more studies are necessary before these NTFs can be tested in a clinical setting.

BDNF
Brain-derived neurotrophic factor (BDNF) is an essential regulator of neuronal differentiation and plasticity. It has been suggested that alterations in BDNF expression may be responsible for the development of neurodegenerative disorders [98]. Postmortem studies of PD patients have demonstrated that BDNF levels are reduced in the substantia nigra pars compacta as a result of decreased transcription of the BDNF gene [99]. Another study reported that only 10% of melanized neurons in the substantia nigra of PD patients were immunoreactive for BDNF expression, compared with 65% in healthy controls [100]. Serum BDNF levels have also been shown to correlate with a loss of striatal DA transporter binding in PD patients, suggesting an influence on striatal neurodegeneration [101]. Animal studies have demonstrated that BDNF antisense oligonucleotide infusion in rats produces a Parkinsonian phenotype [102], and BDNF knockout mice have reduced dopaminergic neurons in the substantia nigra [103]. BDNF promotes in vitro survival and differentiation of human and rat embryonic dopaminergic neurons, and it has protective effects against various toxins including 6-OHDA and MPTP [99]. In a nonhuman primate model of PD, intrathecal BDNF infusion resulted in milder PD symptoms and less neuronal cell loss in the substantia nigra [104]. To date, there are no clinical studies evaluating the efficacy of BDNF therapy in human PD, likely due to the logistical challenges of CNS drug delivery and dosing, as BDNF has poor blood-brain barrier penetration if administered parenterally, and intrathecal or intraventricular delivery results in poor penetration of the brain parenchyma [105].

CDNF/MANF
Mesencephalic astrocyte-derived neurotrophic factor (MANF) was first discovered in 2003 and was shown to be selectively neuroprotective for dopaminergic neurons [106]. Later, cerebral DA neurotrophic factor (CDNF) was discovered as a homologue of MANF with 59% sequence homology [107]. CDNF has been shown to be neuroprotective to DA neurons and intrastriatal injection of CDNF or AAV-CDNF reduces degeneration of DA neurons and parkinsonian behavior in rats and increases TH levels in the striatum and SN [107][108][109][110]. Interestingly, intranigral infusion of a combination of both CDNF and MANF via lentiviral mediated delivery reduced amphetamine-induced rotational behavior and increased striatal TH-fibers and TH-positive neurons in the substantia nigra [111]. Intranigral CDNF alone also improved behavior and increased TH fibers in the striatum, but both to a lesser extent than with CDNF/ MANF together, and did not protect against TH neuronal loss in the SN [111]. Intra-nigral MANF alone did not affect behavior or striatal TH fibers, but did protect against SN neuronal loss [111]. Results of these studies suggest that combined delivery of CDNF/MANF may be more effective than single NTFs and may be a more effective potential therapeutic treatment for PD, although neither NTF has been tested in a clinical setting.

Conclusions
There remains a critical need for new therapies to delay or prevent the progression of PD. As discussed in this chapter, cell-based therapies may provide a promising therapeutic benefit to PD patients. NTFs offer a potential class of cytoprotective agents that complement DA replacement and cell-based therapies in PD, with ideal grafts consisting of immunologically inert cells that continuously produce and release these agents in the host brain. Further development and refinement of these potential therapies is essential to develop personalized care for PD patients.