Why Have We Failed In Our Attempts To Develop A Neuroprotective Therapy For Parkinson's? How Can We Improve This Situation?
We are living in an exciting time. Never before have we been so close to the realistic chance that it may be possible to slow the progression of Parkinson’s, rather than just treat the symptoms. Therapies under investigation include existing treatments developed for other conditions such as diabetes and either active (vaccine) or passive (monoclonal antibody) treatments designed to specifically target the abnormal deposition of α-synuclein in the brain. Additionally, treatments designed to either replace glucocerebrosidase (GBA) or prevent the accumulation of its substrate glycosylceramide may help reduce α-synuclein accumulation even in people with Parkinson’s who do not have GBA mutations. Clinical trials of these therapies are already underway, with more to come in the next few months. Although potential problems remain, the rational basis for these approaches is obvious.
As exciting as these developments are, it is also prudent to step back and consider the numerous other approaches that have failed to deliver the promised impact on disease progression over the past several years. While the purported mechanisms of benefit were different, it may be helpful to bear some of the lessons learned in mind in assessing the newer approaches.
1. Parkinson’s almost certainly represents more than one disease – numerous mechanisms contribute
While the deposition of aberrantly folded α-synuclein is common to most (but not all) forms of Parkinson’s, there may be many roads to this destination, and while there is a considerable body of evidence that it is indeed harmful in its own right, not everyone agrees that α-synuclein deposition is the problem, as opposed to a bystander or attempt to compensate for underlying pathology. Only a minority of patients with Parkinson’s have a genetic mutation leading to abnormal conformation or accumulation of α-synuclein, so in most people, it is assumed that some other factor(s) contribute, including abnormal protein trafficking and recycling machinery, defects in cellular energy, abnormal cell calcium signaling, brain inflammation and a host of environmental factors, some of which may only be relevant in the setting of permissive genetic variations that in their own right do not cause disease. By focusing on the obvious latest mechanism of interest, we run the risk of missing potentially rewarding new avenues. As multiple mechanisms may contribute to disease, it is likely that some ‘cocktail’ that addresses the multiple factors may be required to slow progression. It is additionally entirely possible that the mix of mechanisms may vary from one individual to another, thus the way the cocktail is blended may accordingly need to be personalized.
2. In order to gain insight into disease mechanisms and to test new treatments, we depend upon animal models, but these have been poorly predictive of neuroprotection
Until recently, most animal models used for Parkinson’s were based upon the acute administration of a toxin (6-hydroxydopamine or MPTP) that relatively selectively damaged dopamine neurons. While these models remain enormously helpful for predicting the response to symptomatic therapies, and have shed some important insights into understanding why dopamine neurons die in Parkinson’s, they do not mimic the progressive nature of Parkinson’s, they do not provide a complete picture of the disease, and they have not to date been successful in predicting neuroprotective benefits in humans. Newer models based on overexpression of α-synuclein using either genetic manipulations or by exogenous administration of abnormal α-synuclein fibrils may prove more robust, but their use and applicability are not fully established and this remains a major unmet need for the Parkinson community.
3. In order to assess the effectiveness of new therapies we need better biomarkers
The gold standard outcome measure of any clinical trial is the clinical outcome. However, variability in and slow progression of these measures generally means that very large groups of patients need to be studied over relatively long periods of time. Furthermore, the traditional clinical outcome measure is motor function, which we increasingly recognize is not the major source of disability for most people living with Parkinson’s, who may be much more interested in a treatment that prevents falls, cognitive decline, mood changes and problems with bladder and bowel function. Additionally, most clinician-rated motor scores are subjective and prone to unconscious bias in the examiner and placebo effects in the person with Parkinson’s. Some degree of objectivity may be provided by the use of wearable sensors, that can also capture a more representative slice of the person with Parkinson’s life compared to the snapshot obtained during a standard clinic visit. However, given these constraints, it would be extremely helpful to have some test of blood, saliva, urine or spinal fluid that might provide an objective readout of disease progression. A robust test of this type does not yet exist. To date, the most promising markers have been those derived from brain imaging, but even here, considerable caution should be used in interpreting the findings. Traditionally, some form of dopaminergic imaging has been used. In addition to technical issues with interpretation, as already noted, dopamine loss may not be the only outcome of interest and changes in other systems of the brain may be of at least equal relevance to overall benefit. Other imaging markers of interest might include measures of brain inflammation and ideally of α-synuclein deposition. While progress is being made on all these fronts, we are not there yet. It should be remembered that biomarkers can provide extremely useful adjunctive information, but they do not replace the need for comprehensive clinical evaluation. There is unfortunately a history of people overinterpreting the results of studies demonstrating that a particular treatment ‘cured the scan’, even though the person with Parkinson’s did not benefit. This is nonetheless another important area of development. Better biomarkers would allow for more efficient testing of the efficacy of emerging neuroprotective agents. By demonstrating whether or not the new therapy engages its proposed target, biomarkers can also help weed out treatments that are unlikely to work. Given the highly variable progression of Parkinson’s from one individual to another, biomarkers may also help stratify those participating in a clinical trial into different categories of expected decline.
4. Well designed and rational treatments may still fail
Despite the best planning, treatments that make great sense on theoretical grounds may fail when put into people. There are many reasons for this. The treatment may be consumed outside the brain, converted into an inactive metabolite or fail to cross from the bloodstream into the brain. Sometimes a completely unpredicted mechanism may interfere with the ability of the treatment to work. This may have been the case with Glial Cell-line Derived Neurotrophic Factor (GDNF). While GDNF promotes the function of dopamine neurons and was protective in classical toxin-induced animal models of Parkinson’s, the deposition of abnormal α-synuclein interferes with the signaling pathway for GDNF, so that it may fail to work in people with Parkinson’s (the final word on this is not yet resolved, as not all clinical trial findings are yet reported). Finally, it is estimated that by the time people experience the symptoms of Parkinson’s, they have already lost close to ½ of their dopamine neurons and close to 80% of dopamine levels in the brain, thus most of the damage has already been done and it is too late for neuroprotective treatments to be useful. This has spurred increasing interest in early detection or in studying people who do not yet have Parkinson’s, but who are thought to be in a high-risk category.
5. Designing and conducting a clinical trial is not easy!
Many people, the scientific community included, tend to think that the major intellectual requirement for developing a new therapy ends once one has identified the underlying mechanisms and designed an appropriate drug. This underestimates the enormous skill required to design a trial with clear outcome measures that are not subject to multiple conflicting interpretations. A number of treatments thought to have neuroprotective effects in the past had unanticipated (and longlasting) symptomatic effects. Some trials have placed undue emphasis on the effect of the treatment on a biomarker whose expression may be affected by the treatment, in the absence of a true impact on disease progression. We need better clinical and biomarker estimates of global disability rather than focusing on motor scores and dopamine. Even if the symptomatic effects (including side effects) of a treatment are not directly related to an impact on parkinsonism, they may lead to unblinding in clinical trials and thereby augment placebo effects. Even without this additional ‘help’, the placebo effect plays a major role in Parkinson’s disease and failure to adequately account for it will predictably lead to ultimate failure. No two people with Parkinson’s are exactly alike; this inherent variability must be taken into consideration in planning and interpreting studies. Finally, it must not be forgotten that those who participate in and complete a study are highly motivated and, apart from the demands that are placed on them, may not be representative of everyone else with PD.
Are all these reasons to give up? Absolutely not! The challenges outlined above are worth remembering as some pitfalls can be avoided. It is hard for most people to remember the days before levodopa became routinely available. The impact has been unimaginable. We are now on the threshold of the next great promise, to have an impact on the disease itself, rather than just symptom management. Now is the time to focus on resolving these challenges, and also to remember that the development of new treatments depends upon support for fundamental research, as well as a close partnership between basic researchers, clinicians and those who live with Parkinson’s. By walking together, we will make greater strides.
A. Jon Stoessl, CM, MD FRCPC is the President of the Board of Directors of the World Parkinson Coalition. He served as Co-Chair for the Third and Fourth World Parkinson Congress and is currently a Co-Chair of the Fifth World Parkinson Congress. He is currently the Professor and Head of Neurology and Co-Director of the David Mowafaghian Centre for Brain Health, University of British Columbia.
Ideas and opinions expressed in this post reflect that of the author(s) solely. They do not necessarily reflect the opinions of the World Parkinson Coalition®