An Interview with Mark R. Cookson, Ph.D.
Investigator, Cell Biology and Gene Expression Unit
Laboratory of Neurogenetics
National Institute on Aging
Q: Most of what is known about the genetics of Parkinson’s disease (PD) has been discovered in the past few years. What’s driving the current interest in this area?
A: The main reason for the lack of knowledge about genetic forms of Parkinson’s is that for a long time everyone “knew” that PD was never inherited. This has turned out to be flat-earth thinking—because everyone “knows” the earth is flat, why would it be round? In fact, there are a small but significant number of cases in which PD is inherited, but this was underappreciated for two important reasons. First, many of the inherited cases have unusual features; for example the disease is very mild but starts very early in some recessive families, enough of a distinction for these cases to be considered atypical. But in other inherited cases, the phenotype [clinical syndrome] is very close to that of sporadic PD.
Second, there are several genes for PD or similar conditions but they show different patterns of inheritance and are still very rare. Added together, all genetic cases account for less than 10 percent of all PD, so it would be easy to miss these cases in the population of patients. So genetics provides important clues to the disease process, but it doesn’t yet explain all cases, or even the majority of cases.
What is driving current interest is a complex mixture of motivations that often drive science. Once people realize the world isn’t flat, many want to know how round it really is. After the first genes for PD were found, many scientists started to look more seriously at families in which PD may be inherited. This has led to several genes having been identified for certain and a couple of candidates that are less clear—often these are the ones that are very rare.
Because there are now several recognized PD genes, we can begin to build a better picture of what can trigger the disease process. Having one gene is rather like trying to plan a road trip with only one point along the way. Having several genes might, if one can put them in the right order, tell you how to get there.
Q: Your laboratory focuses in part on LRRK2, a large family of genetic mutations associated with hereditary Parkinson’s that has been called the Rosetta Stone of the disease. Why is LRRK2 important?
A: LRRK2 is important for its frequency, as much as anything else. LRRK2 cases account for between one and five percent of all Parkinson’s disease, depending on which population is studied and how the patients are selected. Although one percent doesn’t sound like a significant proportion, we have to remember that this is the archetypal “non-genetic” disease as discussed above. So, to find a significant proportion to have a single genetic cause changes our view of etiology quite substantially, even if it doesn’t explain everything. It is also worth pointing out that LRRK2 is an excellent candidate for developing new therapeutic ideas.
Q: How can the study of rare mutations like LRRK2, which affect a relatively small proportion of people with Parkinson’s, contribute to the development of therapies?
A: Another way to ask that question is, if some people have a mutation that causes their disease and other people have a similar disease but with a different cause (i.e., not genetic), is it possible that what would work for one group would also work for the other? You might think the responsible response would be that we don’t know for sure whether a therapy based around LRRK2 will work for the 95 to 99 percent of people with Parkinson’s who don’t have a mutation in this specific gene. However, several characteristics of the people with these mutations make us more comfortable with the idea that such a therapy may be more generally helpful.
First, in most cases with LRRK2 mutations the clinical syndrome is very similar to common forms of Parkinson’s. The disease starts in midlife (from the age of 50 onward) and progresses over time, just like typical Parkinson’s. The patients respond to Levodopa—perhaps not a great drug but as good a drug as we have right now—which suggests that the symptoms have a common root cause in terms of which areas of the brain malfunction. Second, the pathology of cases with mutations overlaps with typical Parkinson’s. In post-mortem studies, most cases with LRRK2 mutations have small protein deposits called Lewy bodies in the surviving cells of the brain. This may seem a bit obscure, but it isn’t true of all cases with inherited Parkinson’s, and therefore suggests that there is something specific about LRRK2 and its relationship to the typical pathology of PD.
If patients with LRRK2 mutations and patients with typical PD share the same clinical and pathological outcomes, then it seems reasonable to infer that the things that happen along the course of the disease are also similar. And if so, then the logical extension is that what works in LRRK2 patients might also work in more typical PD. We don’t know this, and it’s hard to test in detail, but it’s a useful prediction. If the logic holds, then we can identify people with these mutations and have a high-risk group that has the same underlying cause of disease. Anything that is helpful to those people in terms of new therapies would then be worth trying more generally.
Q: Your basic research work on LRRK2 has revealed a drug target that your lab is pursuing. Where are you in your kinase inhibitor program and what are the goals of this research?
A: We’re at a very early stage. LRRK2 is one of a large of proteins collectively known as kinases. Kinases are switches that control signaling within each cell. As such, they are involved in most if not all aspects of biology. A well-known example is that kinases control the rate at which cells in the body divide to form new cells; when this goes wrong, one outcome can be cancer. In fact, mutations in kinases are found in several forms of cancer, which has contributed to a large body of work on ways to manipulate kinases, including switching them off with small molecules.
At this point we don’t have an inhibitor for LRRK2, so instead we have tried to understand how this kinase activity is regulated and how it affects neurons. We have found that active forms of LRRK2 will cause the death of neurons, but inactive forms do not. This implies that a small molecule inhibitor would do the same and prevent damage, assuming it was sufficiently specific and potent. The limitation here is that our work has mainly been done in a petri dish using cultured neurons, not in an intact brain. The next step is to develop an animal model where things are hopefully a better representation of what would happen slowly over time in the human brain.
The long-term goals are to move this into something that is transferable from laboratory settings into human beings. In one sense this is rather simple conceptually: (A) make an inhibitor for LRRK2; (B) make an appropriate model; then add A to B. If the model is “better” with the inhibitor, then we have proof of the concept that inhibiting LRRK2 is beneficial. Of course, neither is quite as simple as that. To test the hypothesis, we need a predictive model and an inhibitor that is potent, stable and selective; right now we have neither, but we think it can be done. The main point—that kinase activity is important—is a hypothesis and can be proven or disproven, so long as the test is careful enough. Even if it turns out that kinase activity is unimportant clinically, the molecule may offer other clues about Parkinson’s.
Q: You have said that recent genetic discoveries in Parkinson’s suggest “the problem is tractable, with the right kind of logic.” Please expand/explain this viewpoint.
A: The essence is that kinases, in particular, are molecules in which there is a lot of prior knowledge, and so figuring out what they do—and what goes wrong in Parkinson’s—ought to be possible. In principle we need “just” two pieces of information. We need to know what the immediate substrate(s) of LRRK2 is/are and we need to know how mutations in LRRK2 affect that activity.The word “just” is dangerous, because there have historically been many kinase-substrate pairs that work in one context but aren’t predictive in all situations. At the same time, there are equally many great examples of signaling pathways that have been elucidated based on the rather simple logic of using disease-causing mutations to tease apart the proximal events. So the hope is that we can start to do this for Parkinson’s.