Early Surgery

The observation that STN inactivation prevents nigral degeneration in rodent models of PD^18,19 led investigators to consider the possibility that, in addition to its symptomatic benefits, subthalamic DBS might be neuroprotective in humans. The neuroprotective mechanism was thought to be reduced excitotoxin-induced nigral degeneration, mediated by the glutamatergic projections from the STN to the nigra.^20,21 To examine this, Hilker and coworkers²² used ¹⁸F-fluorodopa (F-dopa) positron emission tomography (PET), which quantifies presynaptic dopaminergic function,^23–25 to study the progression of PD after high-frequency STN-DBS. Although the study did not include controls, the annual striatal reduction in F-dopa uptake in 30 STN-DBS–treated patients with advanced PD was roughly 10% to 12%, which is consistent with previously reported rates in PD.²⁶ This evidence, along with clinical experience, suggests that the neuroprotective effects seen in animal models cannot be readily confirmed in humans.

It remains uncertain whether a neuroprotective effect of high-frequency STN-DBS might be observed if patients were treated early in the course of their disease. Patients with advanced PD may no longer harbor the capacity to alter the course of degeneration, whereas the striatum of patients in the early stages of the disease might be more apt to gain a survival benefit from a reduction in STN-mediated glutamatergic input. In March 2006, a phase I trial of early surgery in PD began under the supervision of Dr. Charles at Vanderbilt University.²⁷ The hypothesis of this study is that STN-DBS slows disease progression in subjects with early PD. In this single-center study, which is expected to be completed in 2010, eligible patients who have received dopaminergic therapy for less than 4 years are randomized to surgery plus medical therapy versus medical therapy alone. Motor outcomes are assessed in a single-blind fashion. A multicenter phase III trial is planned if safety and tolerability criteria are met.

Another line of investigation suggests that early surgery in patients suffering from motor fluctuations or dyskinesias may promote the maintenance of quality of life and social functioning.²⁸ The mean time from PD diagnosis to surgery is more than 10 years; however, dopamine-related motor fluctuations and dyskinesias may arise much sooner and typically contribute to a progressive decline in social and occupational function.²⁹ Early reports suggest that patients with a disease duration of less than 10 years who have undergone operation can maintain a baseline involvement in their work and personal lives postoperatively.²⁸ Schüpbach and associates³⁰ performed a matched-pairs study of 20 PD patients with a mean symptom duration of 7 years, comparing functional outcomes in patients receiving DBS versus best medical therapy. Eighteen months postoperatively, the surgical patients experienced a 24% improvement in disease-specific quality of life versus 0% in the nonsurgical patients. The principal investigator of this study, Dr. Agid, is now leading a multicenter, randomized, open-label, phase III trial (EARLYSTIM Study) examining quality of life outcomes in patients with STN-DBS versus medical therapy, with 2-year follow-up data expected in 2010.²⁷ This promises to offer further insight into the potential benefits of early surgery.

A New Target: The Pedunculopontine Nucleus

Postural and gait dysfunction leading to falls is the leading cause of emergency room visits and is a significant cost generator in the care of PD patients.^31–33 In addition, the fear of falling results in depression and a loss of independence.³⁴ Postural and gait dysfunction is particularly resistant to current dopaminergic and surgical therapies, suggesting a greater involvement of nondopaminergic pathways and brain loci other than the pallidal and subthalamic nuclei in their pathophysiology.^{13,14,35–41}

The pedunculopontine nucleus (PPN) is part of a brainstem locomotive center that also processes sensory and behavioral information. Running roughly between the SNc and locus caeruleus, the PPN primarily connects with basal ganglia structures and the spinal cord.^42,43 Midfrequency electrical or glutamatergic stimulation of the PPN leads to increased activity of its neuronal population and produces controlled locomotion on a treadmill in decerebrate animals.^44,45 Other neurotransmitters, including γ-aminobutyric acid (GABA) and acetylcholine, decrease PPN activity and inhibit locomotion.⁴² It has been theorized that the PPN regulates tone and locomotion by modulating lower brainstem nuclei.^46,47 These nuclei, in turn, activate central pattern generators in the spinal cord, which generate the motor patterns governing locomotion and limb tone.^47–50 Given the evidence for the PPN’s role in the initiation and maintenance of gait and the stability of postural tone, as well as its pharmacologic and electrophysiologic properties that appear to be reliant on cholinergic and other nondopaminergic neurotransmitters, pathologic function related to this structure could theoretically play a significant role in the gait and postural manifestations of PD.⁴²

Although direct evidence linking PPN dysfunction to PD-related postural and gait disturbances is lacking, there are several hints at a pathophysiologic connection. The PPN in anesthetized rats with hydroxydopamine (6-OHDA) nigral lesions is characterized by abnormally enhanced activity compared with controls.⁵¹ Furthermore, some of the electrophysiologic changes recorded in the basal ganglia structures of these PD-model animals could be reversed by PPN lesions.^52,53 Matsumura and Kojima⁵⁴ observed a smaller degree of nigral degeneration and fewer parkinsonian symptoms if monkeys with 1-methyl, 4-phenyl, 1,2,3,6-tetrahydropyridine (MPTP) lesions were given PPN lesions. Similarly, Nandi and colleagues⁵⁵ showed that microinjections of the GABAergic antagonist bicuculline into the PPN improve parkinsonian symptoms in MPTP primates. Electrophysiologic changes have been matched to metabolic alterations in the PPN of parkinsonian rodents and nonhuman primates.^56,57 More important, studies of human pathology demonstrate that a significant proportion of the large cholinergic PPN neurons degenerate in PD.^58–60 These individual lines of evidence collectively invoke the possibility that PD-related degeneration of nondopaminergic PPN neurons leads to electrophysiologic and metabolic disturbances that culminate in the dysregulation of PPN-mediated gait and postural functions. Therefore, medical or surgical interventions that modulate PPN activity in PD patients might normalize gait initiation and maintenance and postural stability.

Following this hypothesis, DBS of the PPN is being evaluated as a therapeutic modality in PD patients with significant gait disturbance (Fig. 84-2).^61–63 Initial subjects have included PD patients with refractory and functionally significant impairment of gait initiation (debilitating freezing) and postural stability (associated with an unacceptable risk of falling). Those suffering from comorbidities that might adversely affect gait and posture, including dementia, primary affective disorders, and marked sensory impairments, are considered poor candidates for PPN stimulation.^64–67

FIGURE 84-2 Postoperative axial T1-weighted magnetic resonance image of a patient with Parkinson’s disease demonstrating a DBS electrode in the region of the left pedunculopontine nucleus (arrow).

Initial reports of PPN-DBS in PD patients documented the technical aspects and electrophysiologic findings of the procedure, as well as short-term indicators of clinical outcome.^61,62 A more recent study provided greater insight into the clinical outcome of six patients who underwent simultaneous bilateral PPN- and STN-DBS.⁶³ Six months after surgery, Unified Parkinson’s Disease Rating Scale (UPDRS) motor scores in the “off-medication–on-stimulation” condition were improved by an average of 33% with PPN-DBS alone. This was less than the improvement obtained from STN-DBS alone (54%) or simultaneous STN- and PPN-DBS (56%) in these same patients. Improvements specifically in axial symptoms (e.g., rising from a chair, posture, gait, postural stability) were similar with PPN-only, STN-only, and simultaneous PPN- and STN-DBS (60% to 70%). In patients taking medication, Stefani and colleagues⁶³ found a greater improvement in UPDRS motor scores with combined STN and PPN stimulation compared with stimulation at either target alone. Moreover, STN-only stimulation did not improve axial symptoms in patients who were taking medication. If only axial symptoms were taken into account in the on-medication condition, both PPN-only DBS and simultaneous PPN- and STN-DBS led to a 50% to 60% improvement versus no stimulation. Importantly, no major complications were described. The only minor side effect reported was paresthesias associated with high-frequency PPN stimulation, likely resulting from the target’s proximity to the medial lemniscus.⁶³ In summary, reports suggest the possibility of a therapeutic response to low-frequency PPN stimulation, but larger controlled studies are needed to verify these results and define its potential role in the treatment of PD.

Epidural and Subdural Electrodes

Internalized epidural and subdural electrodes provide titratable systems with relatively precise target specificity, allowing focal stimulation while avoiding the greater risk of intraparenchymal placement. Since 2000, five reports have described 19 patients with advanced PD who were treated with epidural or subdural cMCS (Table 84-1).^99–103 Although these publications collectively suggest a trend toward varying degrees of improvement across all clinically relevant motor domains, the modest benefits reported remain unconvincing. One poorly understood finding is the relatively consistent occurrence of bilateral improvement in symptoms despite unilateral cMCS, a phenomenon that is generally not observed with DBS. Further evidence indicates that cMCS in PD patients produces no alteration in movement-related regional cerebral blood flow as measured by PET scans, casting further doubt on its potential utility in PD.¹⁰⁴ An additional report investigated the use of subdural cMCS for multiple system atrophy with predominant parkinsonism, a rapidly progressive neurodegenerative disease that shares some motor features with PD but for which treatments are generally ineffective.¹⁰⁵ In the five patients studied, neither low- nor high-frequency stimulation had any benefit. Larger studies in appropriate patients may better define the efficacy and potential indications for epidural and subdural cMCS in PD. At the time of this writing, additional small phase I-II randomized, controlled, and crossover clinical trials investigating cMCS in PD are ongoing in France and Italy, with expected completion in the next couple of years.²⁷

Transcranial Magnetic Stimulation

TMS is a noninvasive technique that induces changes in cortical activity through electromagnetic induction (see Fig. 84-3). Repetitive TMS (rTMS) delivered to the motor cortex at 5 to 25 Hz improves motor performance in PD patients.^106–111 A recent systematic review of noninvasive neurostimulation in PD pooled the results from eight double-blind, sham-controlled studies of rTMS involving patients with a mean age and disease duration of 63 and 7 years, respectively. The standardized mean difference in pre- and posttreatment UPDRS scores was a modest 0.60 (95% confidence interval, 0.24 to 0.96), with no significant heterogeneity. Moreover, the longevity of this modest effect is uncertain, given the less than 2 months of follow-up.¹¹² Frequencies of approximately 1 Hz applied to the supplementary motor area worsened motor performance but improved levodopa-induced dyskinesias.^113,114 One interesting study demonstrated a concomitant increase in motor performance and relief of depression with rTMS applied to the dorsolateral prefrontal cortex, raising the possibility that modulating cortical activity might address the often debilitating comorbid depression that is often observed in patients with advancing PD.¹¹⁵ Other studies have failed to demonstrate any motor benefits with stimulation at the dorsolateral prefrontal cortex.^116–118 Despite the appeal of its minimal invasiveness, the modest demonstrated efficacy of TMS and the practical issues involved in using it over a long period may ultimately limit its clinical utility.

Augmenting Striatal Dopamine Production

This strategy aims to increase dopamine production in neurons and glia within the striatum by ectopically expressing the enzyme aromatic L-amino-acid decarboxylase (AADC), which catalyzes the conversion of levodopa to dopamine (Fig. 84-4). In essence, this is an extension of current pharmacologic interventions with dopaminergic drugs. Bankiewicz and associates¹³⁸ theorized that declining levels of AADC not only contribute to the disease by reducing dopamine production but also lead to a backlog of exogenous levodopa, which may promote the development of levodopa-induced dyskinesias.

FIGURE 84-4 Dopamine synthesis pathway.

Convection-enhanced delivery of AAV-AADC to the striatum increases that nucleus’s capacity to decarboxylate exogenous levodopa in animal models of PD. Efficient and long-term expression of the transgene in the striatum restores local dopamine production and promotes behavioral recovery in 6-OHDA rats and MPTP monkeys treated with levodopa.^138–142 PET data suggest that these same monkeys exhibit persistent AADC activity for at least 6 years after transfection.^138,140 Similar studies using the simultaneous transduction of tyrosine hydroxylase and AADC showed a significant increase in intracellular dopamine in the striatum, producing better behavioral recovery than with the transduction of either enzyme alone; however, this multipronged approach adds complexity to the delivery.^143–145

A human phase I study of AAV-AADC gene therapy for idiopathic PD is now under way.¹⁴⁶ A noteworthy inclusion criterion is the presence of intractable motor fluctuations. Patients receive three escalating doses through bilateral stereotactic infusions into the postcommissural putamen, delivered in two 50-µL volumes spaced approximately 6 mm apart. The key outcome measures are an improvement in ¹⁸F-fluorometatyrosine PET signal at 1 and 6 months, indicating the presence of the AADC transgene, and an improvement in standard clinical rating scales. Preliminary data in five patients who received the lowest dose demonstrated increases in PET signal, suggesting successful AADC transgene expression.¹⁴⁶ Future results, particularly from the higher dose groups, should shed light on the potential safety and efficacy of this approach.

Altering Pathologic Subthalamic Nucleus Output

The second approach to gene therapy for PD under active investigation aims to modify STN output through ectopic expression of glutamic acid decarboxylase (GAD), which catalyzes the production of GABA. The theory behind the therapy is that the local production of GAD transforms the STN projection neurons, which are naturally glutamatergic and excitatory, into inhibitory GABAergic neurons (Fig. 84-5), thereby reducing the excessive stimulatory effects of the STN on the GPi, which are thought to contribute to many parkinsonian symptoms. This follows the theory that STN-DBS generates its positive effects in PD by similarly reducing STN excitatory output.¹⁴⁷

FIGURE 84-5 Synthesis of γ-aminobutyric acid (GABA) from glutamate.

Before considering a clinical application of this concept, AAV-GAD was injected into the STN of 6-OHDA rats.¹⁴⁸ These animals demonstrated activity-dependent GABA release from STN nerve terminals, increased survival of dopaminergic neurons, and marked behavioral recovery. Moreover, GAD transduction of the STN appeared to protect SNc dopaminergic neurons by preventing the lesion induced by 6-OHDA. A recently published study in MPTP monkeys given AAV-GAD injections found more modest improvements (about 20%) in motor behavior, although statistically significant gross motor skill and tremor improvements were noted at 1 year.¹⁴⁹

A human phase I clinical trial of AAV-GAD gene therapy is well under way, with interim data already published.¹⁵⁰ Twelve PD patients with reduced efficacy or tolerance of levodopa underwent unilateral stereotactic STN injections of one of three vector doses based on magnetic resonance imaging targeting. Outcome measures include changes in brain metabolism on PET scans and standard clinical rating scales. No study-related significant adverse events or withdrawals occurred, and no new antivector or anti-GAD peripheral antibodies were detected. At 1 year of follow-up, modest but significant improvements were noted in UPDRS scores, along with a trend toward superior activities of daily living scores. Further, fluorodeoxyglucose (FDG) PET demonstrated reduced thalamic metabolism in the operated hemisphere and increased local metabolism in the ipsilateral supplementary motor area, the latter being highly correlated with UPDRS score improvement.^150,151 These promising clinical findings and supportive imaging studies, along with a favorable safety profile, suggest that AAV-GAD gene therapy may be beneficial in PD. Although future data from this trial will help confirm this early impression, a subsequent phase II trial will be required to judge whether the magnitude of effect is clinically meaningful.

Growth Factor Gene Therapy

Unlike the other two strategies, which are directed at controlling PD symptoms, the goal of this therapy is to prevent ongoing degeneration of dopamine neurons by ectopically expressing the growth factor neurturin throughout the striatum. Growth factors (particularly GDNF) attenuate the loss of nigrostriatal dopaminergic neurons and restore function of the nigrostriatal pathway in animal models of PD.^119,152 Accordingly, targeted delivery of a gene that promotes growth factor production in the human SNc might achieve clinically meaningful neuroprotection or neurorestoration in PD patients. Owing to intellectual property issues, clinical gene therapy with GDNF is impossible at this time; however, work is proceeding with a related neurotrophin, neurturin. Neurturin is a naturally occurring structural and functional GDNF analogue and has already been used in patients.^153–155

Lentivirus- and AAV-mediated GDNF gene therapy and striatal neurturin infusion have shown protective effects against the loss of dopaminergic nigral neurons and the motor consequences of MPTP intoxication in primates and 6-OHDA lesioning in rats.^122,156,157 However, it was uncertain whether benefits would also be realized when nigral degeneration was quite advanced at the time of disease diagnosis. Kordower and colleagues¹⁵⁸ performed AAV-neurturin gene therapy in MPTP monkeys and showed that neurturin expression is associated with substantial histologic and clinical neuroprotection. Importantly, these monkeys received the vector infusion following MPTP administration at a time when nigral destruction was under way. AAV-neurturin also improved motor function in aged monkeys, restoring failing nigrostrial function.¹⁵⁹ As further confirmation of its biologic effect, AAV-neurturin provided efficient, persistent, and behaviorally relevant neuroprotection in the 6-OHDA rat model.¹⁶⁰ These data suggest that neurturin gene therapy might benefit patients with preexisting PD.

Following the success of the phase I trial, a phase II sham-surgery controlled efficacy study enrolling 58 patients began in 2007.¹⁶¹ Disappointingly, despite being safe and well tolerated, there was no appreciable difference between experimental and control groups.¹⁶¹ Both showed an approximate 7-point improvement in UPDRS motor scores at 12 months, relative to a baseline mean of approximately 39 points.¹⁶¹ Recently released 18-month follow-up data from 30 patients found a clinically modest but statistically significant treatment effect.¹⁶¹ Further details have not been released, other than postmortem results from 2 patients indicating neurturin expression in the putamen that was not transported to the substantia nigra. These findings leave the status of further efforts in doubt.

Fetal Tissue Transplantation

Neural transplantation strategies seek to reverse some of the symptoms of PD by replacing the degenerating dopaminergic cells. Although it is increasingly recognized that PD involves neurodegeneration in both dopaminergic and nondopaminergic systems,^162,163 attention has focused on replacing the dopaminergic neurons of the SNc, whose degeneration is PD’s hallmark. Thus far, fetal dopaminergic cells have been transplanted not into their ontogenic site in the SNc but ectopically into the putamen, because cells transplanted into the SNc failed to extend their processes to their target nuclei—the putamen and caudate nuclei.¹⁶⁴ When assessing the potential impact of transplantation, it is important to note that an exclusive focus on repleting lost dopaminergic cells may limit the ability of these technologies to fully address the disabilities caused by PD.

Experimental neural transplantation therapies for PD are based on almost 40 years of preclinical work using animal models in which nigrostriatal dopaminergic neurons were selectively lesioned with neurotoxins. Initial studies performed in the mid-1970s were concerned mainly with histologic aspects of graft survival.^165,166 Later, the functional effects of transplanted cells were assessed.^164,167,168 Transplanted dopaminergic neurons have been shown to survive after transplantation,¹⁶⁹ produce and secrete dopamine,¹⁷⁰ form synaptic connections with host neurons,¹⁷¹ and attenuate both simple motor behaviors, such as asymmetric rotational behaviors induced by the administration of dopamine agonists, and more complex sensorimotor deficits, such as akinesia and motor forelimb function.¹⁷²

These encouraging results provided experimental evidence that cell restoration strategies may be beneficial in the treatment of PD. Results of open-label clinical studies in which ventral mesencephalic tissue procured from 6- to 8-week fetuses was transplanted into the putamen of PD patients resemble those obtained from animal experiments. The first clinical trials of transplanted fetal dopaminergic cell therapy began in 1987.^173,174 Since then more than 350 patients have received fetal dopaminergic cell transplants, although the number of centers performing the procedure is limited. Initially, solid tissue grafts were implanted via an open microsurgical technique on the ventricular surface of the caudate nucleus.^175,176 Stereotactic injection of grafted tissue gradually became the preferred approach, with tissue implanted either as solid strands¹⁷⁷ or as a dissociated single-cell suspension.^178–180 Cell suspension grafts appear to develop greater vascularization,¹⁸¹ integration with host structures,¹⁸⁰ and axonal reinnervation¹⁸² than their solid tissue counterparts. Autopsy studies demonstrate that transplanted cells survive (Fig. 84-6)—in some cases, for as long as a decade^180,183—although the implanted cells may undergo early senescence with Lewy body inclusions.¹⁸⁴ Graft survival and function are further supported by corresponding increases in F-dopa signal and evidence of enhanced dopamine release on PET.¹⁸⁵ Objective measures of improved dopaminergic function are mirrored by some patients who have undergone transplantation and who exhibit significant improvements in motor performance,^186–189 including reductions in bradykinesia and rigidity, improved “on” times (defined as periods of adequate disease control while on medication), and decreased daily doses of levodopa, with some patients not requiring any medication several years after transplantation.^186,189,190

FIGURE 84-6 A and B, Transplanted fetal dopaminergic neurons, visualized here in a coronal brain section with immunohistochemistry for tyrosine hydroxylase (the rate-limiting enzyme involved in the production of dopamine), survive several years after transplantation into the putamen of a patient with Parkinson’s disease. Scale bars = 500 µm (A), 100 µm (B).

(Courtesy of I. Mendez and K. Mukhida, Dalhousie University, Halifax, Nova Scotia, Canada.)

Results from two National Institutes of Health–sponsored double-blind, randomized, controlled trials undertaken to assess the effects of neural transplantation in PD were less promising than those of the open-label studies.^177,191 Both were controversial, in that they included groups of patients who received placebo surgery consisting of incision and bur holes with no implantation of cells or brain penetration. In the first study, a transfrontal approach was used to stereotactically deliver strands of solid ventral mesencephalic tissue to the putamen bilaterally.¹⁷⁷ By 1 year after transplantation, patients who received tissue grafts demonstrated an insignificant 18% improvement in the total UPDRS score in the “off” phase (defined as the clinical state while off levodopa for at least 12 hours), although a subset of patients younger than 60 years exhibited a significant 34% improvement in UPDRS motor subscores. Patients demonstrated relatively consistent graft viability as assessed by F-dopa PET imaging. In the second trial, patients were randomized to either placebo surgery or transplantation with solid ventral mesencephalic grafts derived from either one or four fetuses (i.e., a three-armed study).¹⁹¹ As in the first clinical trial, no significant improvement in UPDRS motor scores was observed among the groups of patients. Patients who received grafts derived from four fetuses demonstrated clinical improvement up to 9 months after transplantation, but their UPDRS scores returned to preoperative levels by the end of the study period (24 months). Aside from the lack of demonstrated benefit of fetal transplantation, the trials were also disappointing in that 15%¹⁷⁷ and 56%¹⁹¹ of patients developed severe off-phase dyskinesias as a result of the transplants. Critics of these trials proposed that the suboptimal results were due to flaws in the study design and methodology, including the absence¹⁷⁷ or only short-term use¹⁹¹ of immunosuppression, the culturing of fetal tissue for extended periods (up to 1 month),¹⁷⁷ poor transplanted cell survival,¹⁷⁷ and the inclusion of patients with more severe disease than were included in open-label trials.^177,191

These disappointing results reinforce the need to optimize cell restoration strategies in experimental settings before attempting their clinical use. Optimization should include standardization of transplantation procedures, including patient selection; procurement, storage, and preparation of fetal cells; and transplantation technique.¹⁹² Absent optimization, it may be premature to dismiss transplantation entirely without due consideration of the variables that may be critical to improve its efficacy. These include transplantation targets, graft-induced dyskinesias, and cell sources.

Neural Stem Cell Transplantation

Neural stem cells are attractive because of their ability to self-renew; this allows their growth in culture, which can generate the large numbers required for widespread clinical transplantation.^205,206 Moreover, their multipotential ability to differentiate into enriched populations of neurons and glia may one day allow them to be used in numerous transplantation strategies (Fig. 84-7).²⁰⁷

FIGURE 84-7 Neural precursor cells obtained from the human telencephalon grow as neurospheres (spherical aggregates of cells) and display immunoreactivity for the stem cell marker nestin (A). These cells have the potential to differentiate into all cell types found in the central nervous system, including astrocytes (B), visualized using immunohistochemistry for glial fibrillary acidic protein, and neurons (C), visualized using immunohistochemistry for the immature neuronal marker βIII-tubulin. Scale bars = 200 µm (A), 20 µm (B and C).

(Courtesy of K. Mukhida and I. Mendez.)

Recent studies have demonstrated the feasibility of generating dopaminergic cells from neural stem cells. Embryonic stem cells transplanted into the striatum of a parkinsonian rodent model spontaneously differentiated into dopaminergic cells, albeit in low numbers.²⁰⁸ The efficiency of dopaminergic differentiation has been improved using a variety of techniques, including modification of the culture environment with mitogens (e.g., sonic hedgehog),²⁰⁹ neurotrophic factors (e.g., fibroblast growth factor 8, GDNF, brain-derived neurotrophic factor),²¹⁰ low oxygen,²¹¹ and cytokines.²¹² Additionally, genetic manipulation to overexpress regulatory genes of dopaminergic differentiation, such as Nurr1,²⁰⁹ Lmx1b,²¹³ and Wnt5a,²¹⁴ has been used to enhance dopamine cell production. Neural stem cells modified under these conditions can improve cortical activation²⁰⁸ and functional outcomes in parkinsonian animals, such as dopamine agonist–induced rotational asymmetry in rodents.^208,209,212

Although neural stem cell–derived dopaminergic cells are a promising source of cells for neural transplantation, clinical applicability is limited by a number of factors. First, differentiation of stem cells exclusively into non-neoplastic neurons has not been consistent,²⁰⁸ and long-term studies of their teratogenic potential are lacking.²¹⁵ Second, dopaminergic stem cells exhibit poor survival following transplantation, with some reports estimating that as few as 1% of the transplanted cells survive beyond 6 weeks.^211,212 Last, although a few studies have shown that transplanted dopaminergic neural stem cells have some of the electrophysiologic properties of midbrain dopaminergic neurons,^209,214,216 stem cell–derived dopaminergic cells need to demonstrate functional reinnervation of the basal ganglia, in addition to the molecular and morphologic properties of SN dopaminergic neurons, if they are ultimately to be used to replace them.²¹⁷

In addition to being a source of replacement cells, stem cells may be useful for their ability to secrete factors that affect the fate of neighboring cells. These factors can support the survival of endogenous or grafted dopaminergic neurons (e.g., GDNF, sonic hedgehog).^218,219 The production of an array of trophic factors, especially at physiologically relevant levels, may be important for transplantation strategies in which transplanted fetal dopaminergic neurons require exposure to a variety of neurotrophic factors than cannot be provided by cells genetically programmed to produce only one or two of them.²²⁰ Neural stem cell–produced factors are neuroprotective in parkinsonian rodent models, protecting endogenous dopaminergic neurons in the SN from MPTP-induced²¹⁸ or 6-OHDA–induced²²¹ neurotoxicity, restoring the size of caudate nucleus type 1 dopaminergic cells to control levels, and preventing the upregulation of dopaminergic cell numbers in the caudate and putamen in primates with MPTP lesions.²²² Cotransplantation with murine subventricular zone neural stem cells approximately doubled the number of surviving fetal dopaminergic cells.²¹⁹ All these measures of neuroprotection correlated with significant attenuation of parkinsonian motor behaviors. Therefore, the utility of neural stem cells may be as adjuncts to fetal tissue transplantation or to deliver neuroprotective agents to preserve endogenous dopaminergic nigrostriatal function.

Repair by Endogenous Stem Cells

Rather than implanting exogenous, genetically altered embryonic stem cells in the brains of PD patients, one can harness endogenous stem cells, located mainly in the subventricular zone and subgranular zone of the dentate gyrus of the adult mammalian brain,²²³ to repair the dopaminergic nigrostriatal system. Normally, few dopaminergic cells are present within the putamen and caudate nuclei^224,225; however their numbers increase significantly in these locations in PD, suggesting a compensatory dopaminergic differentiation of endogenous neural stem cells that migrate out of the subventricular zone and into the striatum.^226,227 The number of subventricular zone stem cells that proliferate, migrate into the striatum, and then differentiate into neurons can be enhanced by the intraventricular injection of trophic factors such as brain-derived neurotrophic factor, platelet-derived growth factor, and transforming growth factor-α^228,229; in other studies, however, neuronal differentiation was not observed.²³⁰ Administration of agonists to the dopamine D₃ receptor can also increase stem cell proliferation in the subventricular zone, striatum, SN, and ependymal layer of the third ventricle and promote their differentiation into dopaminergic cells.^231,232 Whether endogenous dopamine neurogenesis occurs in the SN is still a matter of debate.^233,234

The prospect of using pharmacotherapies to induce endogenous neural stem cells to proliferate, migrate to areas of dopaminergic neurodegeneration, and differentiate into dopaminergic neurons is attractive because it would avoid many of the technical and logistical difficulties associated with fetal- or stem cell–derived transplantation strategies.²³⁵ Further studies exploring endogenous brain repair are needed, although the identification of neural stem cells throughout the neuraxis also means that such therapies need to be approached with caution²³⁵ because they could contribute to the formation of tumors.²³⁶