A neuroprotective therapy that slows, stops, or reverses disease progression is the most important need in PD therapeutics. To date, no drug has been established to have neuroprotective or restorative effects in PD. MAO-B inhibitors were studied based on their capacity to block MPTP toxicity, and to provide antioxidant and antiapoptotic effects. The classic DATATOP study demonstrated that patients randomized to selegiline experienced a delay in developing disability necessitating the introduction of L-dopa. However, it could not be established whether this benefit resulted from a neuroprotective effect that slowed progression, or a confounding symptomatic effect that masked progression. The delayed-start design is a 2-period study that was developed to try to resolve this dilemma. In the first period, patients are randomized to an active intervention or placebo. Differences between groups at the end of this stage could be due to symptomatic and/or protective effects. During the second period, patients in both groups receive the same active study intervention. If at the end of the second period the 2 groups have comparable change from baseline, it is likely that the drug had only symptomatic effects. However, if the early start group shows less deterioration from baseline than the delayed start group despite both groups being on the same intervention, and this difference endures at multiple visits with no sign that the groups are converging, then the benefit is consistent with a disease-modifying effect. A limitation of this design is that it will not detect the neuroprotective effect of an intervention whose onset of action is beyond the duration of the first period, or which has powerful symptomatic effects that mask the difference between the early and delayed start groups.
Rasagiline is an MAO-B inhibitor with antiapoptotic properties that was tested in a delayed start study. Rasagiline 1mg met 3 predefined primary endpoints, consistent with a disease-modifying effect. However, rasagiline 2mg failed to meet all primary endpoints. Dopamine agonists have been shown to have protective effects in the laboratory, and to induce a reduced rate of decline in an imaging biomarker of dopaminergic function relative to L-dopa,[33, 34] but these studies may have been confounded by pharmacologic differences in the 2 interventions, and pramipexole failed in a delayed start study. Therefore, it remains uncertain whether any currently available drug has a disease-modifying effect, although this possibility has not been excluded. Importantly, it is now generally accepted that delayed start studies can help to differentiate symptomatic from disease-modifying effects, and will likely prove useful in evaluating putative neuroprotective therapies in the future.
Trophic factors have the capacity to protect and restore function to dopamine neurons in preclinical studies. Unfortunately, it has proven difficult to translate these laboratory findings into benefits for PD patients. Double-blind trials of glial-derived neurotrophic factor (GDNF) administered into the ventricle or directly into the putamen failed, possibly because they did not adequately distribute the protein throughout the target region.[37, 38] Gene delivery using a viral vector offers the potential to provide widespread and long-lasting delivery of a trophic factor with a single procedure. However, a double-blind, sham surgery-controlled study testing AAV2 change in delivery of the trophic factor neurturin to the putamen failed to meet its primary endpoint (Unified Parkinson Disease Rating Scale score in the practically defined off state at 12 months). Nonetheless, benefits were seen with respect to several secondary and exploratory endpoints, including the primary endpoint for the subset of patients who were followed for 18 months. Postmortem studies demonstrated neurturin expression in the putamen with very little staining in neurons of the SNc, where it is required to induce upregulation of repair genes necessary for a beneficial response. These observations suggest that impaired axonal transport in PD could have accounted for the limited and delayed benefits that were observed. However, a recently completed double-blind trial testing AAV2-neurturin delivered directly into both the SNc and the putamen also failed to provide benefits in comparison to placebo (C. W. Olanow, personal communication).
A limitation of trophic factors is the need to administer them directly into the target region, as they are large proteins that do not cross the blood–brain barrier. Consequently, even if current approaches prove successful, benefit would theoretically be limited to the nigrostriatal dopamine system, and might not benefit the nondopamine features of PD. PYM50028 (Cogane) is a small molecule that crosses the blood–brain barrier and is thought to induce upregulation of trophic factors such as brain-derived neurotrophic factor and GDNF in regions of damage throughout the nervous system. Unfortunately, a recently completed double-blind trial testing multiple doses of the drug failed to show benefit in comparison with placebo (C. W. Olanow, personal communication). Failure to see benefit with trophic factors might relate to recent observations suggesting that alpha synuclein downregulates RET, the receptor for trophic factors that have been tested to date.
These discouraging results have undermined the prospects for trophic factors as a therapy for PD, although it remains possible that better results might be attained with administration to patients in an earlier state of the disease different trophic factors, different doses, different targets, and different delivery techniques.
Therapies Targeting Pathways Implicated by Pathological Studies
Most studies of putative neuroprotective agents performed to date have been directed at targets implicated in postmortem calcium mediated and biochemical studies. These include oxidative stress, excitotoxicity, inflammation, mitochondrial dysfunction, and signal-mediated apoptosis. However, clinical trials to date have failed to define a neuroprotective agent aimed at these targets, possibly because we do not know which of these factors is primary, whether it is the same in all individuals, whether multiple factors contribute to a pathogenic cascade such that a cocktail of agents is required, or whether the factors identified to date are epiphenomena and the true pathogenic mechanism remains to be discovered.
Recent interest has focused on calcium channel blockers based on preclinical studies demonstrating that, with aging, dopamine neurons increasingly rely on L-type Ca(v).1.3 calcium channels to drive rhythmic pacemaking, thereby increasing the risk of cytotoxicity in these high-energy vulnerable nerve cells. It has been proposed that agents that block calcium channels might be neuroprotective by promoting pacing through sodium channels that remain latent in adulthood. Preliminary studies have assessed the safety and tolerability of the calcium channel blocker isradipine in PD, but clinical benefits remain to be demonstrated, and testing with more specific Ca(v).1.3 calcium channel blockers is awaited.
Neuroinflammation with activated microglia and toxic cytokines has long been recognized in areas of neurodegeneration in PD, and is considered a possible target for a neuroprotective therapy. Interest in this target has increased with evidence suggesting that activated microglia with release of toxic cytokines and oxidizing agents play a role in α-synuclein-mediated toxicity. Experimental studies demonstrate that inflammation associated with lipopolysaccharide injected into the SNc is associated with α-synuclein accumulation and dopamine neuronal degeneration in transgenic mice that overexpress wild-type or conversely, mutant α-synuclein, but not in α-synuclein knockout mice. Significant neuroprotection was observed with abatement of microglia-derived nitric oxide and superoxide. It has been proposed that extracellular α-synuclein is taken up by and activates microglia, which in turn release toxic cytokines that contribute to damage in affected cells, and promote misfolding of wild-type α-synuclein, leading to aggregate formation and damage in unaffected neurons. Such a mechanism could account for the development of the Lewy pathology that has been observed in transplanted embryonic dopamine neurons.[48, 49] A prominent inflammatory response with activated microglia was observed in transplanted regions many years prior to the emergence of α-synuclein aggregates in these cells, consistent with the possibility that inflammation is an initiating or important contributing event to the development of Lewy pathology in these cells. Anti-inflammatory therapies have been shown however, to provide protective effects in a variety of PD models. Minocycline, the only anti-inflammatory agent that has been tested in a double-blind clinical trial in PD, was rejected in a futility study. Further testing of agents that suppress microglia activation and T-cell activity remain of interest.
The nuclear receptor related 1 protein (Nurr1) has also attracted considerable interest as a putative neuroprotective target because it is a transcription factor that plays a key role in regulating dopamine neuronal development and survival, and is markedly downregulated in the mesencephalon in PD. Reduced levels of Nurr1 increase vulnerability of dopamine cells to MPTP-induced neurodegenerataion, and gene delivery of α-synuclein induces downregulation of Nurr1 and neurodegeneration that is resistant to trophic factor protection. By contrast, overexpression of Nurr1 is protective, and enables cells to respond to trophic factors and resist toxicity induced by α-synuclein. It is thus possible that restoring Nurr1 levels in PD patients will have protective and/or restorative effects. Several groups are now actively investigating agents that activate Nurr1.
Therapies Targeting Pathways Implicated by Genetic Studies
Perhaps the most promising opportunities for developing a neuroprotective therapy for PD are targets or pathways implicated by gene mutations that are associated with the disease. Although most cases of PD occur sporadically, it is anticipated that insight into the cause of neurodegeneration in these genetic cases will be relevant to sporadic PD as well. Among the mutations associated with PD, most interest has focused on pathways implicated by mutations in the LRRK2, parkin/PINK1, and alpha synuclein pathways.
LRRK2 mutations are the most common genetic cause of PD, and affect as many as 40% of cases in selected populations such as Ashkenazi Jews and North African Berbers. Furthermore, individual cases with typical clinical and pathological features of sporadic PD have been identified to have LRRK2 mutations. LRRK2 is a protein kinase with a GTP domain. Most mutations are associated with the kinase domain and are thought to involve alterations in autophosphorylation, or phosphorylation of target proteins, although all of the physiologic substrates of LRRK2 and the exact mechanism leading to neurodegeneration are not precisely known. Transgenic animals that express LRRK2 mutations demonstrate alterations in dopaminergic neuronal function and terminal degeneration, but do not precisely replicate the pathologic features of PD. In the laboratory, kinase inhibitors have been shown to protect against cell death induced by LRRK2 mutations. These observations suggest that LRRK2 might be an appropriate target for a neuroprotective treatment in PD, and many pharmaceutical and biotech companies are currently pursuing this line of research. It remains unclear, however, whether such therapies will be safe, as the physiologic role of LRRK2 is not known, and it is uncertain whether therapies targeted specifically at LRRK2 will be effective in patients with sporadic PD who lack this mutation.
Parkin and PINK1 mutations are rare causes of PD, which are associated with defects in mitochondrial function.[60, 61] This target is of particular interest, as sporadic PD is associated with a defect in complex I of the mitochondrial respiratory chain, and agents that interfere with complex I such as MPTP and rotenone induce dopamine neuronal degeneration and a model of PD. Parkin and PINK1 are now known to regulate mitochondrial morphology and to promote the autophagic clearance of damaged mitochondria (mitophagy). Parkin, a ubiquitin ligase, translocates from cytosol to mitochondrion in response to a fall in mitochondrial membrane potential and promotes mitophagy with the clearance of abnormal mitochondria. Overexpression of parkin protects against PINK1 mutations, but the converse is not true, suggesting that parkin is downstream of PINK1 in the mitophagy pathway.[60, 61, 64] These observations suggest that agents that upregulate parkin and promote mitophagy might be protective in PD. Rapamycin promotes autophagy and is protective in drosophila that express PINK1 or parkin mutations. Attempts have been made to provide neuroprotection in PD through the use of bioenergetics that enhance mitochondrial function. Coenzyme Q10 showed positive results in preliminary trials, but failed in large-scale double-blind studies. Mito-Q is a modified form of coenzyme Q10 with increased brain penetration that is currently being investigated. Creatine is a bioenergetic that was not rejected in a double-blind futility study and is currently being evaluated in a long-term simple study. An alternate approach to enhance mitochondrial function is the use of agents such as rosiglitazone, pioglitazone, and troglitazone that activate the peroxisome proliferator-activated receptor γ coactivator-1α (PGC-1α), which regulates mitochondrial activity and functions together with SIRT1 to influence mitochondrial biogenesis. These drugs are currently approved for the treatment of type II diabetes mellitus, and demonstrate protective effects in models of PD. Preliminary clinical trials with pioglitazone are underway in PD. There is also interest in resveratrol, a naturally occurring polyphenolic compound found in the skin of grapes, that increases SIRT1 and PGC-1α activities, promotes mitochondrial biogenesis, and has with this agent protective effects in animal models of PD. Clinical trials are anticipated in the near future.
Perhaps the therapeutic target that has generated the most interest is α-synuclein. In 1997, it was discovered that mutations in the α-synuclein gene are associated with rare familial cases of PD, that α-synuclein is a risk factor for sporadic PD, and that α-synuclein is a principal component of Lewy pathology. It is also now appreciated that duplication/triplication of the wild-type α-synuclein gene cause a form of PD,[74, 75] indicating that overexpression of the native protein is itself sufficient to cause the disease. The discovery of α-synuclein pathology in healthy embryonic dopaminergic neurons that had been implanted into the striatum of PD patients[43, 44] further suggests that α-synuclein aggregates can transfer from affected to unaffected neurons.
These observations raise the remarkable possibility that α-synuclein is a prion, and PD a prion disorder. Similar to the PrPC prion protein, α-synuclein assumes an α-helical configuration when bound to membranes, misfolds to form beta-rich sheets when present in a mutant form or in excess concentration, generates toxic oligomers and aggregates, polymerizes to form amyloid plaques, induces neurodegeneration, and can transfer to involve unaffected neurons to extend the degenerative process. Laboratory studies confirm that injection of α-synuclein fibrils can lead to aggregate formation in healthy neurons, with subsequent behavioral abnormalities, neurodegeneration, and transfer to anatomically related neurons.[77, 78] Importantly, aggregates do not form and neurodegeneration does not occur in α-synuclein null cells, implicating a prion conformer reaction whereby the misfolded protein acts as a template to promote misfolding of the wild-type protein in the neurodegenerative process. Further support for the concept that α-synuclein is a prion comes from studies demonstrating that inoculates derived from the brains of elderly transgenic animals that express mutant α-synuclein accelerate neurodegeneration and death when injected into the brains of young and clinically unaffected transgenic animals. A proposed schema for how α-synuclein might induce neurodegeneration is provided in the Figure.
Figure 1. Proposed model illustrating a possible sequence of events leading to α-synuclein misfolding with neurodegeneration, and possible targets for neuroprotective therapies. The proposed schema accounting for α-synuclein misfolding and neurodegeneration includes the following elements: (1) native α-synuclein misfolds to form beta-rich sheets as a result of a mutation, increased level of the protein (due to gene duplication/triplication or impaired clearance), or stochastically; (2) increased levels of toxic α-synuclein are normally cleared by the ubiquitin–proteasome system (UPS) or the autophagy/lysosomal system; (3) failure to adequately clear α-synuclein due to increased production or impaired clearance leads to the accumulation of the misfolded protein, which serves as a template to promote misfolding of the wild-type protein; (4) misfolded proteins combine to form oligomers, fibrils, and aggregates; additionally, (5) aggregates interfere with UPS and autophagy/lysosomal function, which further impairs clearance mechanisms; (6) aggregates polymerize to form amyloid structures (Lewy bodies), whose role is not known but may be protective; (7) increased levels of toxic α-synuclein species and other species (presumably oligomers) interfere with vital cell functions, leading to neurodegeneration; and (8) α-synuclein aggregates are released from affected cells and taken up by microglia and/or by healthy unaffected cells to promote and extend the neurodegenerative process. This schema offers several novel opportunities to provide a neuroprotective therapy for Parkinson disease including: (1) agents that prevent misfolding or promote refolding of proteins; (2) agents that promote clearance of misfolded α-synuclein via the UPS or lysosomal system; (3) immunotherapies that remove toxic oligomers/aggregates; (4) agents that inhibit transfer of α-synuclein to microglia or to unaffected cells; (5) agents that interfere with microglia activation and/or block cytokine-mediated toxicity; (6) agents that block the prion conformer reaction, whereby misfolded protein promotes misfolding of wild-type protein; and (7) agents that removal native α-synuclein to eliminate the substrate for the prion reaction. [Color figure can be viewed in the online issue, which is available at www.annalsofneurology.org.]
Download figure to PowerPoint
The concept that α-synuclein is a prion-like protein and PD a prion-like disease suggests novel targets for putative neuroprotective therapies in PD (see Fig). Some of these approaches are already being actively studied in the clinic. One approach is to promote the clearance of misfolded α-synuclein. α-Synuclein is cleared via the ubiquitin–proteasome and lysosomal systems, both of which are affected in PD.[81, 82] This could further lead to α-synuclein accumulation and aggregation, and in turn α-synuclein aggregates could interfere with normal lysosome and proteasome function.[83, 84] Experimentally, degeneration of dopamine neurons with inclusions that stain for α-synuclein can occur as a result of interference with proteasome or autophagy function or overexpression of α-synuclein.[85-87] One can thus envision a vicious cycle in which α-synuclein aggregates impair clearance, and impaired clearance leads to the accumulation of α-synuclein aggregates, ultimately resulting in neurodegeneration. Thus, therapies that enhance clearance of α-synuclein might be neuroprotective in PD. Rapamycin acts through TOR inhibition to promote autophagy with clearance of α-synuclein and provides neuroprotection in model systems.[88, 89] However, the drug is associated with toxicity primarily related to the immune system. A series of small molecules that enhance rapamycin (SMERs) have been developed that provide similar effects without toxicity and might be suitable for clinical trials. Similar observations have been noted with agents such as trehalose that promote autophagy in a TOR-independent manner. Latrepirdine (Dimebon) is an antihistamine that has been widely used in Russia and promotes autophagy and clearance of α-synuclein. Latrepirdine been tested in short-term clinical trials in Alzheimer disease (AD) and failed to meet the primary endpoint, but was well tolerated with no significant toxicity. Preclinical studies suggest that the drug may have a preferential effect on clearing α-synuclein, and it has been reported to protect dopamine neurons in cultured cells and mice, but these results were not observed in transgenic mice expressing human α-synuclein. Clinical trials in PD are being considered.
A second approach is to target potentially toxic α-synuclein species (oligomers/fibrils/aggregates) using immunotherapy. Current thinking suggests that the toxic species is probably an insoluble oligomer or fibril. Vaccination with human α-synuclein has been shown to reduce α-synuclein aggregates in cell bodies and synapses of α-synuclein transgenic mice . More recent studies have demonstrated a reduction in α-synuclein aggregates in animal models following either intracerebral injection of antibodies or systemic vaccination. These studies suggest that antibodies target extracellular α-synuclein, and facilitate their uptake and clearance by microglia. Based on these data, AFFiRiS AG has initiated an immunotherapy trial in PD patients using the Parkinson vaccine PDO1A to target α-synuclein. Caution is urged, however, as the precise species responsible for α-synuclein toxicity is not known, and theoretically it is possible that removal of protective species could accelerate the neurodegenerative process. Furthermore, antibodies could target unrelated epitopes and lead to unwanted side effects such as meningoencephalitis, as seen in AD patients immunized with Abeta42 (AN1792).
A third approach is to prevent the formation of misfolded α-synuclein by either interfering with the “prion conformer” reaction or by knocking down levels of the wild-type protein so that they cannot participate in the neurodegenerative process. In the prion conformer reaction, misfolded α-synuclein acts as a template to promote misfolding of the native protein. The precise molecular mechanism responsible for templating is not known, but understanding the molecular pathway responsible for misfolding of the wild-type protein would provide potentially important therapeutic targets that could stop the prion chain reaction. Knockdown of wild-type α-synuclein has attracted considerable attention, as native α-synuclein is essential for aggregate formation and neurodegeneration in healthy neurons. This is illustrated by studies showing that α-synuclein fibrils do not induce aggregate formation in α-synuclein null cells. Antisense oligonucleotides and siRNAs have the capacity to knock down alpha synuclein mRNA and protein and limit the availability of the protein to participate in the prion process. The Spanish company nLife has developed antisense oligonucleotides that specifically target aminergic neurons following intranasal delivery and can reduce α-synuclein mRNA and protein by 30% and 50%, respectively. The physiologic role of α-synuclein is not well understood, and it remains uncertain what the consequences of knockdown will be. However, transgenic animals with α-synuclein knockout appear to suffer only minor consequences,[100, 101] and partial/short-term knockout may be all that is required to reset the equilibrium for protein clearance and to restore neuronal function to begin. Clinical trials testing these approaches are anticipated in the near future.