Date: December 2016
Authors: Leonidas Stefanis, MD, PhD; Thomas Gasser, MD; Roger Barker, MBBS, MRCP, PhD
Editors: Michael S. Okun, MD, and Stella M. Papa, MD
Over the last 20 years there has been considerable progress in deciphering the genetic underpinnings of Parkinson’s disease (PD). This has led to new insights into PD pathogenesis, the generation of new animal models of the disease, the implementation of experimental therapeutics in such models, and, recently, the design and initiation of clinical trials with potentially neuroprotective agents.
We have asked Drs. Thomas Gasser and Roger Barker to comment on the continuing progress in this field, and how it is shaping our views on Parkinson’s disease pathogenesis and treatment.
Dr. Thomas Gasser: The discovery of the alpha-synuclein gene and what followed out of it in 1997 is clearly the most striking example that at least this particular form of monogenic (single gene mutations) Parkinson’s disease has told us a lot about sporadic PD: it took only two months after the publication of the A53T-SNCA mutation by Polymeropoulos et al. (Polymeropoulos et al., 1997) that Spillantini and co-workers published that the encoded protein is the main constituent of the Lewy-body in sporadic PD (Spillantini et al., 1997). The world of Parkinson’s disease has never been the same since. More than 10 years later, in 2009, it became clear that common variants in this gene were found, in genome-wide association studies (GWAS), to be also the most significant genetic risk factor for sporadic PD, although admittedly with a small effect. Nevertheless this finding clearly supports the causative role of alpha-synuclein itself in sporadic PD. Other genetic loci bearing genes for monogenic diseases, including LRRK2, MAPT and GCH1 (the gene causing dopa-responsive dystonia) have been identified in GWAS in sporadic PD, suggesting that alpha-synuclein is not the odd exception, but rather reflects an underlying general principle, which has been described as the concept of pleiotropic risk loci by Singleton and Hardy. This concept states that multiple variants of different effect strengths can exist within a single genetic locus, affecting both monogenic and sporadic forms of a disease.
Prof. Roger Barker: The monogenic forms of parkinsonism have been helpful at identifying critical pathways linked to this disease state- such as the role of alpha-synuclein (through finding families that overexpress or have mutations in the gene coding for this protein) as well as mitochondrial dysfunction. However to say that this has helped us find a single common mechanism is not the case as, for example, problems with mitochondria in PD were known about well ahead of any genetic forms of PD ever being identified! Furthermore the subsequent list of genetic causes of PD has shown that there is not a single mechanism by which these abnormal genes cause disease rather that they can affect any of the critical intracellular homeostatic systems of a neuron – e.g. PINK1 upsets mitochondria; Fbox7 upsets the ubiquitin proteasome system; GBA mutations affect the lysosome and so on.
Furthermore some monogenic forms of parkinsonism have rather different clinical presentations and progression (e.g. parkin positive patients) and more importantly a different pathology in some instances (or at least no Lewy body pathology of note!). Thus these conditions cannot be equated to the much more common idiopathic PD and as such their ability to inform us about what actually goes wrong in this latter condition is unclear. Indeed knowing that drugs used to treat some psychiatric disorders (e.g. neuroleptics) give parkinsonian side effects helped identify the central role of dopamine in the clinical expression of aspects of PD, but did not in any way help us understand why patients developed PD in the first place.
Are there distinctive clinical features in PD monogenic forms?
Dr. Thomas Gasser: Aside from a strong family history (in some dominant families), there is no single feature that is distinctive for monogenic PD as a group. Early onset (< 40) clearly enriches for recessively inherited cases (most of them caused by mutations in the parkin or PINK1 genes), and the vast majority of the few cases with VERY early onset (< 30) will be monogenic. There also are distinctive features between different groups of monogenic PD: recessive cases usually develop motor fluctuations early, but lack dementia and other non-motor features, while dementia and orthostatic hypotension is quite common in GBA-associated disease. Nevertheless, for each gene, the spectrum of disease manifestations, including age at onset, is wide. Occasional Parkin patients can have an onset age as late as 70, and a consortium accumulating clinical data on more than 1700 LRRK2 mutation carriers showed a range of onset-ages between 20 and 90. So while some clinical features suggest a monogenic form of PD or even involvement of a specific gene, it is difficult to predict.
Prof. Roger Barker: Although some genetic forms of parkinsonism have a distinct clinical course (e.g. the alpha synuclein duplication, triplication families; patients with Fbox7 mutations), the majority have no unique distinguishing clinical features. Obviously very young onset cases have a much higher chance of being genetic in nature, and parkin cases tend to have more problems with walking than perhaps is seen in idiopathic PD. However, in later onset cases, the clinical phenotype in genetic forms can look identical to that seen in sporadic disease, although some genetic variants are associated with a more malignant phenotype (e.g. GBA heterozygotes).
Is there a clear pathogenetic distinction between the mendelian forms of PD?
Dr. Thomas Gasser: If we stick to the image of multiple partially overlapping pathogenic cascades being set off by different mutational events, there would be no need to categorically distinguish between different forms of PD. Rather, one could think of more or less closely or distantly related forms in a continuum. In that world, a classic Autosomal Dominant (AD) form of PD caused for example by a synuclein triplication and a classic Autosomal Recessive (AR) form represented by a PINK1 mutation could be thought of as the extreme ends of this continuum, while others fall in between.
Prof. Roger Barker: The major difference relates more to age of presentation than pathology or clinical course per se, and as we move into a new era of genetic sequencing it is more informative to think about the genes and genetic variants that cause and drive disease rather than whether they are AD or AR – especially given the varied pathology that exists for many of these forms of PD (e.g. LRRK2 patients).
Are there common pathways of pathogenesis, so that drugs, for example, that work for genetic synucleinopathy cohorts may be applicable to LRRK2 cohorts?
Dr. Thomas Gasser: Clinical and pathologic heterogeneity suggests that even in the most “simple” monogenic cases, there is probably not a single one-to-one unidirectional path from mutation to disease. Maybe a better representation would be a network of interconnected cascades and feed-back loops, where the underlying mutation would be an initiation event in monogenic cases. This suggests that common or overlapping pathways are likely. Which of them could be effectively used for treatment may depend on the initiating mutation(s), but also on any number of modifying factors in a given case.
Prof. Roger Barker: It is certain that in sporadic PD a number of pathways are affected which ultimately combine to give the distinct pathology and pathway failures seen in this condition However what these pathways are, and which are critical is unknown and while the genetic forms of parkinsonism have revealed likely players, none have been shown to be dominant in sporadic disease. Thus any therapy that comes out of a pathway as a treatment for a known genetic form of PD is unlikely to work as well in patients with sporadic PD – although may have some disease modifying effects. Similarly drugs developed for one genetic form of PD will not necessarily help patients with other monogenic forms of the condition. So for example, knocking down/out alpha synuclein production in families with alpha synuclein overexpression is not likely to help patients with LRRK2 mutations as the latter condition may have a primary pathology in mitochondria that may lead in some patients to a secondary problem with protein aggregation, which in this case may not be pathogenic but simply a biomarker of mitochondrial dysfunction in certain nerve cells.
Is it useful to know the genetic background of a patient as we move to clinical trials of disease modifying therapy?
Dr. Thomas Gasser: Clearly, gene mutations can almost never be seen as absolutely deterministic by themselves. Even the strongest alpha-synuclein mutations (A53T), which lead to a rapidly progressive and severe form of PD with onset in the 40s in most cases, are not completely penetrant so that some mutation carriers live healthy to old age. So there must be modifiers, probably many of them, some weak, some strong; some genetic, some environmental, and any genetic effect must be considered on this complex background. Having said that, there is no Parkinson patient (or, for that matter, no person with any other disease) in whom the genetic background is completely irrelevant. So it is clearly useful to know the genetic background of any patient in a clinical trial; if not for stratifying inclusion (this is still an exception), but definitively for post-hoc analyses, which will help to generate new hypotheses for stratified approaches to PD treatment.
Prof. Roger Barker: The advantage of using monogenic forms of disease to identify the group to be treated in a clinical trial is that they should all be roughly the same- i.e. rather homogenous. While this is true compared to an unselected collection of patients with idiopathic PD, it does not actually mean that the patients will all behave the same. For example, patients with another genetic disease of the brain affecting the basal ganglia, Huntington’s disease (HD), all have the same genetic defect but present and progress very differently in the clinic. Indeed in PD, patients with LRRK2 mutations do not always develop disease and when they do, this can be at different ages and in different ways, with different pathologies. Thus while knowing the genetic status of patients may help in the design of clinical trials, it should not be the major factor in any trial design unless the agent being tested is targeting the specific genetic problem. Thus in trials using strategies to knock down the production of alpha-synuclein, it may be better to recruit families where the patients have a duplication/triplication of the alpha-synuclein gene. Finally using specific genetic forms of PD may also be misleading for most patients with PD. For example, one could use a therapy that very specifically targets the kinase that is affected in LRRK2 patients – it could work really well for such patients but it may not work in patients with idiopathic PD who have no problems with LRRK2(or even conceivably could make them worse).
Will genetic approaches change how we do drug trials in the future?
Dr. Thomas Gasser: Monogenic forms of PD are rare. While they have given us unprecedented insight into the molecular networks involved in disease pathogenesis, the development mutation targeted treatment is hampered by the simple fact that it is close to impossible to collect sufficiently large patient cohorts powered to unequivocally prove a therapeutic effect. So it is unlikely that these rare disease-causing mutations will be targeted in clinical studies. This is probably different for variants that have a frequency and effect strength that lies between the rare bona fide disease causing mutations and the common, low effect risk factors as detected by GWAS. At present, two of this intermediate type of variants are known: the LRRK2 G2019S “mutation” (which is in fact more a strong risk factor), and, collectively, a considerable number of pathogenic variants in the GBA (Glucocerebrosidase) gene. Those variants are common enough to allow to collect meaningful cohorts, and the variant effect is strong enough (odds ratios roughly between 5 and 15, which translates to an age-dependent penetrance of 30 to 40%) to suggest that a drug targeting directly the consequences will have a clinically relevant effect. In fact, my prediction is that we will see the first mutation-based trials in these populations in the very near future.
Prof. Roger Barker: I think this will happen as we get to know more about how common variations in genes affect how drugs are broken down as well as drive aspects of the disease process. In this respect there are now common genetic variants (which are not mutations) that seem to change the course of the disease and stratifying around such factors may be helpful in such trials- e.g. GBA heterozygote patients (a gene which when mutant on both genes causes another condition – Gaucher's disease) with PD tend to progress more quickly and develop an earlier dementia and thus may be suitable for trials looking at disease modification. I think also better recording the genetic make-up of patients will become a more popular way to set up clinical trials as we also get to better understand what drives the changes in the expression of some of these genes through epigenetic profiling (i.e. looking at changes that are important for the expression of genes). However, the identification of more monogenic forms of PD will have limited impact on the development of new disease modifying therapies. Instead, it will give us some confidence that the pathways we are currently studying (which are most, if not all of those, that work normally inside any cell!) are relevant at some level to the core events that define PD pathogenically.