Introduction
Parkinson’s disease (PD) is currently incurable. By the time of diagnosis, people with PD already have substantial and irreversible neurodegenerative pathology. For this reason, much PD research has started to focus on preventing or delaying rather than curing symptoms [
1]. The interventions under study are targeted on the primary pathophysiological processes of PD, such as α-synuclein aggregation, or the glucocerebrosidase (GBA) or leucine-rich repeat kinase 2 (LRRK2) pathways, which start years before clinical symptoms appear [
2]. Accordingly, prevention trials are being considered that aim to recruit relatively healthy research participants, with no or only mild symptoms, in the hope of halting the pathological process and thereby delaying or preventing the onset or progression of clinical symptoms.
With the exception of rare cases, determining in advance with certainty who will develop PD is impossible. There are also no reliable ways to measure the pathophysiological processes, e.g., α-synuclein aggregation, that are believed to cause PD. For this reason, potential research participants for trials aimed at early disease modification or prevention may be identified based on algorithms that integrate an individual’s risk profile [
3,
4]. This risk is calculated by summing up the risk factors a person has for PD, ranging from sex, smoking behavior, and clinical features such as hyposmia or REM sleep behavior disorder (RBD) [
4], to more advanced testing, including gene mutations, biomarkers of neurodegeneration, and subthreshold parkinsonism symptoms [
3]. Those with an overall high risk of developing PD are referred to as having ‘prodromal’ PD [
3,
5]; terminology which implies that they are in an early stage of PD. Yet, it may well take over 20 years before people with ‘prodromal’ PD will develop symptoms of PD, and some will never develop these at all [
6].
Following the same line of reasoning, the Alzheimer’s disease (AD) label has also been expanded to ‘preclinical’ and ‘prodromal’ stages in the preparation for early disease modification and prevention trials [
7]. Similarly, ‘preclinical’ and, for those with mild cognitive impairment, ‘prodromal’ AD indicate an increased risk to develop AD dementia. This risk status is based on biomarkers of amyloid-β and tau [
7,
8]. Allegedly, these biomarkers reflect the pathological root of AD, analogous to the presumed role of α-synuclein in PD [
9]. Therefore, in AD trials, people with elevated levels of amyloid-β and tau in the brain are identified as having an early stage of AD, even if they do not have dementia. The aim of these trials is to prevent or slow down later cognitive symptoms in these individuals by lowering levels of amyloid-β and/or tau [
10,
11]. So far, however, none of the trials aiming to prevent or delay AD dementia in pre- or early symptomatic persons has led to detectable clinical effects [
12,
13].
Several reasons for the lack of success of AD trials have been suggested. Suboptimal trial design (e.g., inappropriate outcome measure selection; short study duration) and lack of sufficiently robust phase 2 trial results to support a subsequent phase 3 trial are commonly listed reasons. Even though the clearance of amyloid-β did not result in cognitive benefit—undermining the hypothesis that amyloid-β is directly causal for AD dementia—the expansion of the AD label to people without dementia based on biomarkers of amyloid-β is slowly gaining momentum for implementation in clinical practice [
14]. This expansion of the AD label may, however, do more harm than good in the absence of an effective treatment, especially for those who will never develop dementia [
15]. This demonstrates how developments in research may tacitly impact clinical practice.
Looking back at the previous two decades of developments in AD research, we want to critically appraise the recent development to expand the PD label to ‘prodromal’ stages and, more specifically, the attempt to prevent or delay PD in these disease stages by pharmacological interventions. With several immunotherapy trials planned or ongoing in early PD [
16,
17], it is vital to assure that the PD field acknowledges the lessons learned in AD immunotherapy clinical trials (see Table
1). In this paper, we aim to draw parallels between PD and AD prevention, hoping to reduce the chance of encountering repeated neutral trial results as has occurred in the AD field.
Table 1
Lessons learned from AD research for PD early disease modification and prevention trials
Target population | A combination of early clinical features associated with PD and (biological) risk factors could be used as recruitment criteria, because disease progression is too advanced in people who already have PD and (biological) risk factors on itself have not enough predictive value |
| If the intervention targets abnormal aggregation of a putative causal protein for which an accurately measurable biomarker exists, this biomarker status should be adopted in the eligibility criteria |
Disclosing at-risk or early disease status | Apply risk terminology in communication towards research participants rather than a ‘diagnosis’ of prodromal PD |
Legal safeguards are required to protect participants against privacy violation and discrimination based on risk status |
| Changes in research disease criteria may tacitly impact clinical practice |
Pre-trial evidence for prevention trials | Consider targeting several putative pathological processes simultaneously—thereby not relying too heavily on one potential trigger of the pathophysiological process, provided that the underlying evidence is sufficiently supported by pre-trial evidence |
Recruitment strategies | Transnational recruitment registries with clinical and biomarker information may tackle current recruitment issues, but are subject to several ethical challenges |
Outcome measure selection | More sensitive outcome measures may increase the chance of finding a clinical intervention effect, but can only serve as proof of concept if not directly translatable into a clinically relevant effect |
| Immunotherapy may have adverse effects and requires continued scrutiny, also in phase 3 trials |
Advancing to phase 3 trials | Use sub-group analyses guidelines to prevent over-interpretation of post hoc analyses in phase 2 trials that lead to misleading expectations for phase 3 trials |
How promising are immunotherapy PD trials targeting α-synuclein?
Before exposing relatively healthy people to the risks and burdens of PD prevention or early disease modification research, particularly in the case of immunotherapy, and before investing substantial resources, it should be reasonably plausible that the intervention under study will lead to a health benefit. How strong is the pre-trial evidence of immunotherapeutic agents currently selected for PD prevention trials in people with no or only mild complaints? And how does that compare to the presumed high plausibility of previously tested anti-amyloid-β interventions in AD research, which so far have failed to lead to a tangible health benefit?
The most promising target for early intervention is α-synuclein aggregation in the brain, which is strongly associated with most motor symptoms of PD [
42,
43]. A causal role for α-synuclein aggregation in the disease process of PD seems highly plausible, since genetic variants strongly associated with PD determine α-synuclein levels and folding [
44]. Moreover, in a mouse model of PD, anti-α-synuclein immunotherapy which reduced α-synuclein aggregation also reduced neurodegeneration [
45]. For prevention trials specifically, α-synuclein seems a suitable intervention target, because it plays a crucial role in the stages of PD prior to neurodegeneration [
9]. On the other hand, clinical disease severity in PD is not directly linked to reduced α-synuclein levels in CSF [
46] and, similar to amyloid-β in AD, aggregation of α-synuclein may also be an epiphenomenon rather than the pathophysiological cause of neurodegeneration [
44].
The safety and tolerability of anti-α-synuclein immunotherapy have been established in humans [
47]. Recently, the PASADENA study results showed that the anti-α-synuclein antibody Prasinezumab (RO7046015/PRX002) did not lesson symptom worsening after 1 year in participants with early PD (NCT03100149) [
48]. Other immunotherapy phase 2 trial results in recently diagnosed PD patients are underway (SPARK Study, ClinicalTrials.gov identifier NCT03318523). Even if these phase 2 trials suggest beneficial clinical effects, expectation management will be vital. Successful removal of a presumed causally related protein does not necessarily lead to improved functioning, as we have seen in AD trials. Similarly, highly promising phase 2 trial results may not result in clinically detectable effects in phase 3 trials [
49].
Previous AD immunotherapy trials have been criticized for their strong reliance on amyloid-β to define and diagnose AD [
50], especially after their results showed that the removal of amyloid-β had no positive clinical effect [
51]. In light of the disappointing results of these trials, it seems wise for upcoming PD immunotherapy trials not to rely too heavily on the role of α-synuclein alone in causing the symptoms of PD, and continue to focus on co-investigating multiple other intervention targets. A key argument for the latter is the strong inter-individual variation between biomarker levels and PD symptoms [
30], which increases the likelihood that other biological processes are being involved.
Another likely contributor to PD is dysregulation of the immune system, in which α-synuclein may also play a central role. Possibly, the immune response is triggered in PD by pathogenic forms of α-synuclein [
52]. In AD, anti-amyloid-β treatment led to sometimes severe inflammatory responses in the brain, ranging from MRI changes without symptoms to fulminant fatal encephalitis. Whether such an inflammatory response may occur following anti-α-synuclein treatment is uncertain, but it has been suggested that these side-effects might be avoided in future PD trials with the right participant selection [
53].
Infectious triggers in the gut microbiome are another, more recent focus in the search for potential intervention options linked to the immune system, given its link with brain inflammation in PD [
54,
55]. How these links with PD may translate into a potential intervention for PD is, however, still uncertain. Rather than aiming to modify the biological root of PD, future interventions may also aim to foster functioning neurons and protect them from the damaging effects of α-synuclein [
54]. Interventions following this strategy can be sought in repurposed drugs [
56,
57], such as Exenatide [
58]. Given a strong genetic link between LRRK2 and PD, LRRK2 kinase inhibitors are also investigated as a potential treatment. Phase I study results showed that a LRRK2 inhibitor can substantially lower kinase activity, which might be neuroprotective [
23]. Challenges for this treatment strategy include the lack of measures for LRRK2 activity so far, and the potential for adverse peripheral side-effects [
23]. Potential biomarkers for target engagement are currently being explored, which may also lead to the selection of patient subgroups that may have a greater benefit from treatment [
2]. Ambroxol, a repurposed drug known for respiratory disease treatment, may become an important therapy for those with a mutation on the glucocerebrosidase gene (GBA) that increases one’s risk to develop PD [
59]. However, only the small minority of PD patients who carry this genetic risk factor may benefit from this therapy and phase III trial results still have to measure Ambroxol’s impact on PD motor features.
Conclusion
Similar to the movement in AD research, PD research is now focussing on an earlier (even prodromal) diagnosis in the hope that intervening in an early stage may slow down or possibly even arrest the disease process in those with no or only mild symptoms. This strategy shift is accompanied by new challenges that have hampered progress in the field of AD in recent years, where a similar research strategy led to a series of disappointing trial results. In this paper, we have provided guidance on how we can capitalize on lessons and experiences from AD research in the field of PD, such as how to inform people of their risk status and how to deal with the ethical challenges of trial-readiness cohorts. We also draw attention to the possible impact that PD-risk algorithms—developed with good intentions for research purposes—may have on persons in clinical practice. Taken together, we anticipate that consideration and implementation of these lessons and experiences will accelerate progress for people at risk of or living with PD.