 OK. It's a pleasure and honor to be here to tell you a little bit of the story of the relationship between a Mendelian disorder and a common complex disorder. And I start by saying that this story may challenge how we think about the genetics of Mendelian disorders and also how we think about the genetics of complex disease. Because not too long ago, we saw genetic disorders as divided into two major categories, those that were simple single gene disorders, things like Gauchy disease, cystic fibrosis, PKU, and disorders that we knew were genetic, but we considered to be complex multi gene disorders like asthma diabetes, ADHD like you heard last month, and Parkinson's disease. The Mendelian disorders were considered simple because they fit into this single gene disorder category. But we started to note as we realized we started to see this boundary between these two discrete categories becoming blurred when things like environment and epigenetics, there we go, modifier genes and risk factors started to blur the boundaries that distinguished them. In fact, I think today we see these disorders more as a continuum, where on one side you have disorders where there's a single primary gene that has a large effect and then perhaps multiple modifier genes or risk factors that contribute. And on the other end of the spectrum, you have disorders where there's multiple primary genes with small effects, and you might also still have modifying aspects. So today I'm going to I want to introduce the two disorders that I'm going to speak about. The first is Gauchet disease, which is a rare autosomal recessive enzyme deficiency with a variable age of onset and multi organ involvement. On the other hand, I'm also going to speak about Parkinson's disease, which in contrast is a common disorder affecting about 1.5% of the population over 65. It's a complex multi gene disorder that's more late onset. Symptoms we'll discuss. And it is distinguished by the loss of dopaminergic neurons and the formation of Lewy bodies. These are these aggregates that you see on autopsy in the brain. And it mainly affects the substantia nigra and brainstem. I just wanted to review the anatomy where this was. So how are these two seemingly unrelated disorders connected? I think throughout this talk, I'm going to show you how this is being approached by clinical studies, pathologic studies, imaging, genetic studies, cell biology, protein work. And I also hope that you will learn that Mendelian disorders can at times provide a unique window into complex diseases like Parkinson's disease. So to begin a little bit more about Parkinson's disease, start with some definitions. I'm going to use the word Parkinsonism. And that's a term that describes the motor features of Parkinson's disease. Parkinson's disease is clinically and pathologically defined. There's actually rigid criteria put out by the UK Parkinson's Disease Society Brain Bank criteria. It includes bradykinesia, the slowing of initiation of voluntary movements and the reduction of speed or amplitude of repetitive actions, and at least one of the following, muscular rigidity, arrest tremor, and postural instability. Actually, it's more than one entity. We can call them the Parkinson's diseases. First of all is the classic idiopathic Parkinson's disease, which is thought to be largely sporadic and not necessarily running through families. Then there's environmental causes of Parkinson's disease, the classic being the MPTP story or the frozen addicts. There are the Lewy body dementias that overlap some with Parkinson's disease, but pathologically are really distinct entities. And these include dementia with Lewy bodies and the Lewy body variant of Alzheimer's disease. And then there's this category that's lumped as Parkinson's plus, including things like multiple system atrophy and progressive super nuclear palsy, which I'm not going to really go into very much today. The list of genetic forms of Parkinson's disease is growing. They're kind of numbered now as gene park 1, park 2, et cetera. I think they're up to at least 15 park genes, some, but not all, are associated with Lewy body pathology. The pattern of an inheritance of these park genes is variable. There's some that are autosomal dominant. Perhaps the most famous of these is with the first, park 1, first described by Bob Nussbaum across the street at NIH. And this is a very rare cause of hereditary Parkinson's disease. And it's caused by point mutations and multiplications of alpha-synuclein. But its discovery is an example of why it's important to find a gene because it really taught us a lot about the central importance of alpha-synuclein and the pathogenesis of this disorder. Another more common autosomal dominant form is park 8 or lark 2. This is fairly frequent, especially in some populations like Ashkenazi Jews and North Africans. And it's inherited with incomplete penetrates. In a pedigree, you can have the mutation and not have Parkinson's disease. And it has variable pathologic findings. In addition, there's some variants. There's some park genes that are inherited in an autosomal recessive fashion. Park 2, the first one identified, is associated with an early onset, slowly progressive form. And it's more common in early onset patients, where in that group, it accounts for about 10% to 20% to these patients have mutations in this gene. And the other two listed here are also more rare and associated with an early onset. So why is it important to discover these genes? This is a slide that I borrowed from John Hardy. And it's very, very current. Because what he tried to do was he took some of these park genes by their other names and tried to see how they might be interrelated. And it seems that many of them fall into two pathways, either the lysosomal pathway or the mitochondrial pathway. Actually, both are probably critical in PD pathogenesis. And as we're discovering these genes, we're learning more and more about these pathways. In fact, where I'm going to focus is on GBA. So GBA is the gene mutated in Gaussier disease. Going back full circle. Gaussier disease is the most common of the lysosomal storage disorders. It's also the most common inherited disorder in the Ashkenazi Jewish population. It's caused by the deficiency of glucoserebracyase, which is the enzyme that cleaves this glucose moiety off of the lipid glucoserebracy. It's a disorder primarily of the reticular endothelial system, where lysosomes within macrophages become engorged with the storage product. And most of you from medical school probably might only remember this one image of Gaussier disease, because it frequently appeared on board exams. This is the classic appearing Gaussier macrophage. And it's defined as having sort of a wrinkled tissue paper appearing cytoplasm. And it's so chuck-full that the nuclei often get displaced laterally. When you look at these cells by EM, it's quite remarkable also. A normal lysosome, if you remember, is usually a small circular organelle. But in these patients, these become so engorged with lipid that they totally get distorted. And this is a Gaussier lysosome. So the history, I think, is kind of unique. And I always show the slide when I speak to audiences with students, because it was described by Philippe Gaussier while he was a medical school student. In fact, it was his medical school thesis. And this was in 1882. It took almost half a century before they realized the nature of the stored material, that it was glucose-reverside. And much of the progress in research, this may be my personal perspective having spent a career in the clinical center, but I do think that this is sort of a poster child for research that happened at NIH. Because in the last 40-some years, 50 years, many of the major advances came out of the clinical center. Beginning with identification of the enzymatic defect by Roscoe Brady in the 60s, the gene was first cloned by Ed Ginz in 1981. Enzyme replacement, the first effective therapy was developed and went through clinical trials at NIH. First animal models were created at NIH. And this association with Parkinsonism also. Well, what kept me interested in Gaussier disease now over two decades has been that there's really vast clinical heterogeneity in this presumably single gene disorder. It's been grouped into three types, type one, type two, and type three. They're very generalized groupings and they're basically based on, type one has absence and type two and three, the presence and rate of progression of neurologic manifestations. So by far, type one Gaussier disease is the most common form. It is very heterogeneous in itself. It can present at any age with variable progression. Common signs and symptoms are hepatosplenomegaly, bone involvement, anemia and thrombocytopenia, though there can be other organs involved. And while some of our patients are fairly incapacitated even in childhood, on the other end of the spectrum, we know there are many asymptomatic individuals. In fact, just calculating back by the gene frequency, we suspect that more patients have Gaussier disease that are not recognized than those that reach medical attention. Like all forms, it's autosomal recessive and panethnic, but it is more common among Ashkenazi Jews where the carrier frequency is about one in 14 to one in 16. The most common finding is splenomegaly, most often painless splenomegaly, and these spleens can be enormous, sometimes brimming over the pelvic rim. The most, one of the more common causes of morbidity is skeletal disease, and there can be fractures, osteopenia, a whole gamut, but the classic finding is this Erlenmeyer flask deformity of the distal femur, which radiologists consider almost bathenomonic. In contrast, type II Gaussier disease fortunately is a little more stereotypic and far more rare. It usually presents at a few months of life, though we also determined that it could also be a disorder of the neonate, and almost uniformly it results in death by two or three years of age. The common signs and symptoms include a pattern of splenomegaly, but babies also develop seizures, neurologic deterioration that's progressive and devastating, and this is pan-ethnic and fairly rare. Usually I imagine there's a couple dozen cases each year in the United States. Type III, again, is heterogeneous because it's sort of a waste basket of anybody who has any form of neurologic manifestations that didn't die in early infancy. So the neurologic involvement is pretty diverse, but pretty uniformly, most patients with type III Gaussier disease have slowing and looping of the horizontal saccades, which is very characteristic. They also can develop a taxia dementia and myoclonic epilepsy. The degree of visceral and bone involvement is variable, but it can be vast, as you can see from this young man here. And again, this is pan-ethnic, though there's thought to be a geographic isolate in the northern most regions of Sweden where it's seen a little bit more frequently. So having worked on this disorder for some time now, I've actually begun to think of it much more as a continuum of phenotypes, ranging from, again, asymptomatic octogenarians that are just diagnosed by accident when they come in for something else late in life or at autopsy even, to fetuses that succumb in utero and a whole gamut of presenting phenotypes that come out in between. In fact, perhaps the major distinction is that some patients have brain involvement and some don't. And what we're going to focus on today is this part of the spectrum, those that develop Parkinsonian manifestations. So the gene I mentioned was cloned in early 1980s. It was pretty quickly realized that there was a very homologous sequence known as a pseudogene, which is a copy of the gene that was located only 16 kilobases downstream on chromosome 1q21. The pseudogene sequence shares 96% of the sequence of glucose ribosidase. And this is important for two reasons. One, when you're doing the molecular diagnosis, you need to make sure that you're looking at the gene and not the pseudogene. And also, because having this sequence nearby has been the source of different mutant alleles that we encounter in patients. To date, more than 300 mutations have been identified in patients. They include point mutations, frameships, deletions, and insertions. I tried a few years ago to graph the distribution of these mutations according to, it's a fairly small gene, so it spans 11 exons and the mutations encounter almost throughout the gene. Some are more common and some are private. Some of the more common ones I've shown here. The ones that are denoted in pink are ones that are actually sequence that you would see in the pseudogene that somehow has been incorporated as a mutation into the gene itself. And two that you'll hear of more frequently are M370S, which is the most common Ashkenazi Jewish mutation and probably the most common Type I mutation worldwide. And L44P, that's often associated with some neurologic manifestations. So how can you explain such heterogeneity in a single gene disorder? It's long been known that neither the amount of lipid that's stored in these patients nor the amount of residual enzyme activity and all patients that live have at least some residual enzyme activity, but the amount doesn't correlate well with the patient's phenotype. So when I came to NIH in the late 80s, I saw that some gene had just been cloned and sequenced and what I would do is learn how to find mutations and then I'd be able to explain that the severe mutations cause the severe disease and the mild mutations where there is some dramatic patients and it would all be wrapped up and I'd finish my fellowship and move on. But it turned out that it became a lot more complicated. There were some correlations, but we found that clinically different patients can have the same genotype. Conversely, clinically similar patients can have many different genotypes. In fact, siblings with the same genotype can have different clinical manifestations and responses to therapy, even identical twins. So it kept me in business for many more years to come. One of the examples is a paper that we published in 2005 where we evaluated at the Clinical Center 32 patients that were all homozygous for this mutation, L44P. And among the 32, they all had that slowed horizontal, saccotic eye movement so they would all be considered type three. But the degree of systemic and neurologic manifestations was highly variable. We had patients with severe psychomotor delay, even autism, and yet on the other end of the spectrum we had successful college students. In fact, one of the patients actually played a jeopardy for several seasons on television. So the residual enzyme activity also was quite variable among these patients that all shared the same genotype. These are just four of the 32, just showing the spectrum of what we encountered. So I think that by focusing on cohorts where they have the same genotype and also siblings where there's a difference between genotypes, it may give us some insight into some of these factors that modify a single gene disorder. So just to rehash, the distinguished, the two different groups that were once considered simple Mendelian single gene disorders and complex disorders are blurred by many different things that could include molecular chaperones, post-translational processing of protein, modifier genes, et cetera. So now I'm going to go and begin my story of the relationship of Gauchier disease and Parkinson disease. And I was asked to begin a little bit with a case report. So this was actually the first patient that I saw back in 1996 that began the whole story for me. She was diagnosed with Gauchier disease in her late teens, was mild, she did have a splenectomy but had few problems with it and pretty much was not treated and had not really thought about her diagnosis. And when she was 42, she developed a tremor with progressive rigidity, mast facies, difficulty initiating movements and rapid deterioration of her gait. She was followed in New Hampshire. They actually tried a paladotomy with no improvement and they also tried the enzyme replacement therapy for Gauchier disease, again with no improvement of her Parkinsonism. We saw her first at NIH, I believe when she was about 49, I'm gonna see if I can get the, you have to go back to, there you go. This is an old video when we first started doing this but you can appreciate her mast facies, her rigidity, her difficulty initiating movements, rapid tremor. So we followed her for another five or six years. She developed progressive dementia and she died at age 54. So we first thought that this could very well be a coincidence. After all, having a rare disorder does not make you immune for having another more common disorder. But we started, when I went through the literature and talked to other colleagues around the world, it seemed that all the Gauchier centers seemed to have more cases of Parkinson than you would have imagined. So we started to try to collect those cases and in 2003, we published a series of 17 such patients. They included Ashkenazi Jewish patients and non-Jewish patients from countries around the world, males and females. And they had many different genotypes. That in 370S was common but it was not necessarily shared by everyone. Clinically, when we looked at these patients that had Gauchier and Parkinsonism, when we looked at their Gauchier manifestations, most had relatively mild Gauchier manifestations but their Parkinson manifestations were more classic, always including tremor, bradykinesis, rigidity, mask, facies, and some actually had cognitive decline or dementia. When we looked at the age of diagnosis of Gauchier disease, it was relatively late at 35 as the mean. When we looked at the age of onset of Parkinsonian symptoms that was relatively early, 48 years of age, the dopest response was mostly favorable with somewhat variable. None who received the enzyme replacement therapy had any improvement. And interestingly, some of these probands had siblings or parents that had Parkinson disease. Well, that observation made us think that we should explore what's going on in these families. We asked whether carriers could also be more susceptible to Parkinsonism. So for a period of about 18 months, we did a prospective study closely interviewing all Gauchier probands that came to the clinical center. There were about 45 of them. And interestingly, about 12 had relatives with Parkinsonism. Often this was a parent or a grandparent who was either an obligator known to be a Gauchier carrier. And then this was explored in other centers, especially the Gauchier clinic in Jerusalem, where they found about the same finding, about 25% of Gauchier probands had a first degree relative with Parkinsonism. So this led us to propose that heterozygos might also be at an increased risk for Parkinsonism. One of my favorite stories was when we interviewed the parents of this child who was a child with about seven with type three Gauchier disease. As I was taking the history from the mother, I started to ask about the father and the mom said, oh, he's here at NIH, but while we were here, he decided to go enroll in a familial Parkinson disease protocol that Bob Nussbaum was conducting. And so Bob enrolled him and actually went out to visit his family. And there were several generations of Parkinsonism. And when he went to visit, he collected DNA on the people that have circles around their symbol. And those that had the Parkinsonian manifestations were all L4-4P carriers, whereas those without clinical symptoms did not have any mutations in gluca cerebrosidase. So this encouraged us to continue in this direction. So then we decided that the next chapter sort of went on in patients with Parkinson disease that had no known Gauchier disease. And again, this was sort of a serendipitous finding. Several years ago, a student who had been a Howard Hughes medical student in our laboratory back in the, I believe early 90s, called me and said that she was now chief neuropathology fellow at Mass General and she was doing an autopsy of a patient with Parkinson disease. She read on the chart that the patient also had Gauchier disease. She googled them, discovered that that was something I was interested in and was calling me to ask if there was any tissue or anything that she'd like me to say. And I was thrilled and I said, but meanwhile, could you get me a couple age-matched Parkinson subjects so that if I do any studies I'll have a comparison? So the three tissue samples arrived in the laboratory and we just decided that we should confirm just to make sure that nothing got mixed up. And when we measured the enzyme activity, one of them was very, very low, but the other two were lower than we expected. So I said, what the heck, go ahead and let's go get the brain samples and extract a little DNA and we'll sequence it just to confirm which is which. Well, it turned out that the patient with very low enzyme activity had Gauchier disease. It had two copies of N370S, which is a common mutation. But the two that were supposed to just be plain Parkinson disease both also had one copy of a mutation in glucosyriversidase. So we got very excited and we went back to this brain collection and then to several other brain banks that are collected around the country and in a few months had assembled 57 brain samples for patients with the pathologic diagnosis of Parkinson disease, extracted DNA and sequenced them. And eight of the 57 carried mutations in glucosyriversidase and we went back and said, you know, is this a coincidence? Could this disorder be more frequent? So we actually asked for controls from the same brain banks, didn't find any mutations among the controls. Now I have to say that this was not an easy finding to publish. Everybody had objections to it. One of the problems was we didn't know ethnicity of the brain donors and they thought there might be a bias towards Ashkenazi Jews or I can't even remember all of the complaints that I had but I think I sent the manuscript out about to spy for six different journals. But eventually the finding has come accepted and it really has propelled the community studying the genetics of Parkinson disease in a new research direction. And there have been many different replication studies published around the world. Part of the skepticism of the initial studies I think was because when they said there wasn't enough power, people that study common disorders want to have large, large, large cohorts and also because large genome-wide association studies that were being conducted around the country on Parkinson disease had never picked up this gene so they thought it must be some kind of mistake. But as I said, replication studies kept coming on. These are some that just were in a few years, some of them quite large studies and in all of them you can appreciate that the percentage of Parkinson patients that had mutations in Glucose rebrisidase was much higher than in their control populations. In fact, I was really gratified one day to be Googling Parkinson genetics and I got to the Michael J. Fox website which apparently kept a list of the top 10 Parkinson genes of the time and Glucose rebrisidase was now on the number one place and this is still before many people believe the whole story. In fact, one of our initial skeptics, John Hardy, who's very well known in the Parkinson genetics field, wrote an editorial where he conceded that the Glucose rebrisidase example was an illustration of how an important genetic risk factor for a complex disease can actually evade detection by systematic analysis. It only came into the radar because of clinical observation and for us clinicians, those kind of statements are incredibly rewarding. If you just go through Medline right now, PubMed and look at the number put in Glucose rebrisidase in Parkinson, when we started this endeavor, there were one or two case reports, but right now there's over 200 papers that's almost rising exponentially. Part of what propelled this and got it defining to be so widely accepted was a study that we put together a few years ago where because everybody was complaining that we didn't have power in any of these individual studies, we assembled a collaboration of 16 centers around the world that had been looking at Glucose rebrisidase mutations in Parkinson disease and putting the data together and this included centers from four continents, marked some of them on the map, some of them had a few cases, some of them had large numbers, some of them sequence DNA, some of them screened for a few mutations, but when we put them all together, we had over 5,600 cases and we had about 5,000 controls. And the bottom line was that around the globe, subjects with Parkinson disease are at least five times more likely to have a mutation in this gene, giving an odds ratio of 5.43. Now, I mentioned that the way they identified mutations varied from place to place. There was a subset of almost 2,000 subjects where the entire Glucose rebrisidase gene was sequenced and in that group, actually the mutation frequency was higher, it was almost 7%. And interestingly, when we looked at those mutations, we got different distribution of mutations depending on ethnicity, but importantly, a lot of groups that only screened for N370SNL44P because it was easier to just screen for two than to have to do whole gene sequencing on large numbers, but if you only focused on those two, you would fail to detect at least 42% of mutations in non-Ashkenazi Jewish cohort and this is important because I think that the results of this study were probably an underestimate. We also looked at the clinical characteristics where we were provided the clinical information and that was a subset of about 3,000 patients. When we looked at different aspects of Parkinson disease, including whether it was a symmetric onset, whether there was Bradykinesia tremor, all kinds of findings, those in blue are those that had the GBA mutations, those in red are those without, and the profile was really not that different. Where we did see some differences was, we heard that there was a family history of Parkinson disease more frequently among subjects that had mutations. I guess that could be expected. Response to Aldopa was slightly less good, but statistically significant and there was a tendency to see more cognitive changes among those that had mutations and in fact, I think that many subjects with cognitive decline might have been excluded from this initial study because the diagnostic criteria for Parkinson disease might have, they might have been excluded and had been considered to have other forms of dementia. The other interesting finding was that the patients that had mutations had an earlier onset of Parkinsonism mean of about 4.3 years. So the conclusions of the study was that subjects with Parkinson disease have at least five times more likely to have mutation in this gene, that the screening techniques were crucial, that you can't just get away with screening for two common mutations. The age of onset was earlier, but in general, the Parkinsonian symptoms varied a lot and didn't differ greatly from subjects that didn't have mutations. And I think that by the end, everybody was willing to conclude that mutations in this gene are the most significant genetic risk factor for Parkinson disease identified to date. Well, it did come out of the New England Journal, but I didn't impress my at the time 19 year old son who was a computer guy. And my findings only reached his attention when it was discussed in Wired Magazine, which had an article about 23andMe. This is a direct to consumer website where they were collecting subjects with Parkinson disease and doing screens and they looked at Lucas Reversides and found that in a matter of minutes they could replicate our findings and actually had very close findings to what we had initially seen. So I always like to pause at this point and say yes, but the majority of patients that we see with Gauchier disease and the majority of Gauchier carriers do not develop Parkinson disease. So somehow this mutation in this gene must be a risk factor, but not predictive of developing Parkinsonism. And this for patients, especially our patients that are taught about recessive and dominant inheritance, it becomes very complicated to explain. And what I try to say is that when we look at the patient and that's examining the phenotype and we usually think of a model and say in Gauchier disease, mutations in a single gene result in the phenotype, but it's got to be a lot more complicated than that. There's got to be modifier genes, sometimes multiple genes, health factors, lifestyle and environmental exposure that all compose the kaleidoscope of what makes up what we see clinically in our patients. So I mentioned how this has become a finding that's difficult to talk to the Gauchier community about because it's still a little bit hard to know what it means for them. So what the question I'm frequently asked is what is the frequency of Parkinson disease in patients with Gauchier disease and Gauchier carriers? And it was kind of hard to assess. There is an international Gauchier registry run by the company that treats patients and they tried to look at their patient group. I think they have about 5,000 patients registered. They found Parkinson disease and reported in 68 of 1130 patients over age 60, deciding that the probability was about 5.7% by age 70, nine to 12% by age 80, but that still says that you can get to age 80 and 80 some percent or plus may not never develop Parkinson disease. And there was no predictive Gauchier profile. There's been other studies from individual clinics and I think that we still need better and bigger studies, but right now we can sort of assume that the lifetime risk is about five to eight times higher than the general population. So in our group at the NIH for over a decade now we've been following clinically patients with Gauchier disease and Parkinsonism. We find that these patients can have classic Parkinson disease, a more aggressive early onset form or familial form and we kind of have seen it all. These are three patients that I picked out. One professor who was only diagnosed in his late 60s. Oops, we've followed him now for over a decade. He hasn't exhibited any neurocognitive changes. He's a patient with an early onset and then one who onsetted in his 50s who pretty rapidly started to develop cognitive changes. We see multiple genotypes, so N370S is somewhat common. Most of them are responsive to Aldopa. None of them have improved when they were treated for Gauchier disease. Smell, olfactory testing and cognitive testing indicate that impairment is relatively common in this group. Among the first 12 that we published, the mean age of diagnosis of Parkinson disease was about 49. The mean UDPRS score was about 26. So now we are prospectively following these patients both studying, looking at their clinical features and doing some PET imaging. The goals of the study and the physician and coordinator who are working on the study are here in the audience today if people have more specific questions. It's being conducted in collaboration with Karen Berman and NIMH. And our goals are to better characterize the Parkinson phenotype associated with glucosuribrosidase to look at fluorodopa uptake and evaluate PET as a surrogate marker in these subjects and to see if we can establish the earliest signs of Parkinson disease and at-risk individuals. So with that in mind, we recruit patients that have both Gauchier disease and Parkinson disease. We have some controls that have Parkinson disease alone. And then we're looking at patients with Gauchier disease who don't have Parkinson disease but have a family history either in a parent or sibling and also some carriers that also have the strong family history. When patients come to us, they undergo a pretty uniform physical and neurologic examination. They undergo neurocognitive assessment by a single neuropsychologist. They do the scratch test for all factory testing and they're given multiple surveys looking for different non-motor symptoms of Parkinsonism. They also have imaging studies including MRI for structural abnormalities. They're all given fluorodopa PET scanning and water studies looking for cerebral blood flow patterns and also transcranial ultrasonography looking at the midbrain structures. We've recently published our first set of PET studies. This was done on about 44 patients and we looked at both fluorodopa PET, which this is a summary of what we do but they're not individual studies here, they're composite studies where you put the entire cohort together. So this is a group that just has Parkinson disease and this is a group that has Gauchier disease and Parkinsonism. And if you don't see differences that's the correct observation. It turns out that the pattern of dopamine loss in these two groups seems to be similar. Where we did see some differences was in the flow studies. These are rendered templates of the brain where we show areas where there's a difference in the two different groups. And it turns out that there's less activity in areas that are affected in neurodegenerative disorders like Alzheimer's, which may in part explain some of the cognitive impairment that we're seeing in the patients. Interestingly in 14 patients with Gauchier disease and seven carriers who had no Parkinson disease themselves but of positive family history only two showed any evidence of dopamine loss. So that might be somewhat reassuring to some of our patients and it definitely needs longitudinal follow up, excuse me. We've also been able to perform some neuropathologic assessments. In fact, that first patient where I showed you the video was the first autopsy that we were able to get done at the NIH. These patients do have Lewy bodies. They stay in positive for Alpha-Synuclein as you'd expect. And they also have Gliosis of hippocampal layers CA2 to CA4 which are regions that are also affected in some of the Lewy body disorders. When we used some of these brain samples to explore mechanism, we did this immunochemistry studies where we stained the tissue both with antibody to Glucose Reversidase which would appear as red and Alpha-Synuclein which is green. So when we looked at brain samples from patients with Parkinsonism that had mutations in Glucose Reversidase, we saw staining in the Lewy bodies both for Alpha-Synuclein and Glucose Reversidase. When we looked at Lewy bodies and subjects that had Parkinsonism that didn't have mutations, we didn't see the staining or saw very little of it staining for Glucose Reversidase. So we think that may be an important clue as we explore pathogenesis. Now another, clinically we were observing, as I mentioned a few times, that there seemed to be some more cognitive changes in the group that had mutations. This prompted us to go back and do a second multi-center analysis this time of patients that had the diagnosis of dementia with Lewy bodies because this disorder is more associated with neurocognitive changes. This is much rarer. We went back and collected data from 11 centers. There were 720 cases with dementia with Lewy bodies and a large number of age-matched controls. Of the 721, 450 were autopsied and had autopsy confirmation of the finding. And 80% had full Glucose Reversidase sequencing. And the association was really quite significant. The odds ratio this time was over eight. When we looked at this cohort again, the age of diagnosis of mutation carriers was about five years earlier than those without mutations. And the mutations were associated with higher PD scores. So it seems that mutations in this gene may even play a greater role in dementia with Lewy bodies than straightforward Parkinson disease. There have been pilots that he's published in multiple system atrophy and essential tremor but they have not shown the similar association. So why would being a carrier for one genetic disorder put you at risk for an unrelated illness, the million dollar question? Well, at first I thought that this was an extreme situation but thinking about it and talking to people we sort of came up with a list of other situations where having a mutation in one gene that was known to cause a disorder could put you at risk for another complex disease. And I think that, thinking about this, it may not be such a rare phenomena. Glucose Reversidase now fits in that list and those of you who follow the New England Journal just last month there was a report of TREM2 which is a gene that is associated with a neurodegenerative disorder and bone cyst disorder being a risk factor for Alzheimer's disease. Mutations in this gene are seen four times more commonly than in the general population. So how might this be happening? Well, most of our mutations lead to a misfolded protein and this conformational change in protein can have important consequences. First of all, it can lead to an unstable protein which is degraded leading to the functional defect which is what we see in Gauchat disease. You have degraded protein, you don't have the enzyme, you have the phenotype. But it's also possible that you could have a change in the protein and come up with another stable protein with a new function. This is known as the dominant negative effect or having this mutation could lead to alterations in transport or protein-protein interactions leading to organelle dysfunction, traffic jamming as the newly formed protein is processed through the cell or it could cause aggregation of a protein, all three of these leading to what we consider a gain of a toxic function. So in this example of glucosireberosidase, we know that the formation of insoluble alpha-synucreate aggregates leads to the neuronal cell death that's seen in patients. So perhaps having this mutant protein around could increase aggregate formation or lead to organelle dysfunction, particularly in the lysosome, leading to decreased clearance of aggregates. And this is what I'd consider a gain of function hypothesis supported by the fact that mutations are seen in heterozygotes as well as homozygotes. And also by what I showed you where we saw that there was glucosireberosidase present in Lewy bodies in these patients. But the alternate hypothesis is that having mutation can lead to the unstable or deficient protein, which is degraded. You don't have enough enzyme. You have lipid accumulation. And then it's this accumulation that provokes neuronal cell death. And this is more a loss of function hypothesis. And it's supported by the fact that some of our patients actually have no alleles. They actually have mutations that don't make protein. So it's hard to describe a misfolded protein when one's not made it at all. And also when you use inhibitor to glucosireberosidase called CBE, you see differences in alpha-synuclein. So the jury is still out on how this plays to be. There was an important paper that came out about a year ago from Boston showing that in cell that we actually collaborated with where they showed using all kinds of things, silenced RNA, neuronal cells, IPS cells, the elegance, human and mouse brain studies, that they came up with a hypothesis that there's a bi-directional feedback group where having this lipid glucosal ceramide promotes the formation of soluble alpha-synuclein oligomers. Having increased number of these oligomers blocks the trafficking from the ER to the Golgi, which would in turn lead to less enzyme and more lipid causing a self-propelling cycle. I actually think that this is interesting but it still may be more complicated. One of the findings that we published in collaboration with Jennifer Lee's group in NHLBI was that at least in the test tube there seems to be a molecular link between alpha-synuclein and glucosireberosidase. And they showed this by a lot of fancy techniques including fluorescence measurements where you see that when you put the two proteins together there's a shift, NMR, where you see the shift actually indicating that the interaction is at the C terminus of alpha-synuclein. And then we actually did some pull-down experiments and brain extracts showing that glucosireberosidase can be immunoprecipitated together with alpha-synuclein. Now all three of these findings were only seen at pH 5.5, which is closer to the lysosomal pH, and not at pH 7, not with mutant glucosireberosidase and not with other lysosomal enzymes. So we sort of put this together as a model with this cartoon where we postulate that perhaps the C terminal of alpha-synuclein interacts with glucosireberosidase, that's this glowing region, whereas the N terminal helix could bind to the glycolipid rich vesicles in the lysosome. Jennifer's gone on to establish that membrane-bound alpha-synuclein interacts with glucosireberosidase and actually inhibits its activity. So we speculate that maybe this binding at lysosomal pH could facilitate alpha-synuclein degradation, have a beneficial effect, or prevent aggregation. What we found is that's really needed to propel these studies or some better models. One of the problems is that the mouse models that we have there don't really mimic the human phenotype. Skin fibroblasts, which we have over the years collected on our patients, are useful for some things, like extracting DNA and RNA and protein studies, but they don't show the lipid storage that you would need in a model. So that prompted us to explore using patient-derived induced pluripotent stem cells or IPS cells. IPS cells are reprogrammed skin fibroblasts that can grow for extensive periods and be used to form any type of cell. And this was our first IPS cell colonies from a patient with type two Gauchier disease. We've worked to, working to develop this, and it's right now in progress. We've managed to make our first IPS cell lines and differentiate them into first Gauchier macrophages. And to our delight, we were able to show a storage phenotype. This is a control IPS cell. This is a Gauchier IPS cell. When you feed them ghost cells that are labeled, you can see the lipid accumulate only in the Gauchier macrophages and not in the controls. We've now begun to generate neurons and we're in the process of establishing cell lines from patients that have both Parkinson disease and Gauchier disease. And these show that markers for neurons are positive in these IPS derived patient neurons. So how do I fit this all together? I really still don't know for sure. We know that alpha-synuclein is degraded by the proteasome or mitochondrial pathways and also by lysosomal pathways. And I think since this is a lysosomal enzyme, our part of the story has to do with the lysosome. So let's speculate that when the lysosome is functioning normally, you somehow need to have wild-type glucosaribrosidase and alpha-synuclein and somehow they come together and the net result is that the alpha-synuclein is degraded like it needs to be. However, when you have mutant glucosaribrosidase around, speculate that you have, of course, diminished glucosaribrosidase activity and less protein around. And somehow the net result here is that less alpha-synuclein is degraded and it accumulates in aggregates contributing to neuronal cell death. Now if this is the model, why doesn't every patient that have Gauchier disease go on to develop Parkinson disease? Well, it's gotta be a lot more complicated than that. But we know that Parkinson disease is a disorder that occurs as we begin to age. And with aging, the number of lysosomes that cells have just go down, they function less well. And also the amount of alpha-synuclein we produce goes up. So perhaps there's this homeostasis necessary and having mutations around the stirubs that I think that will still have a lot of research ahead. And I just want to, for a few minutes, turn to implications for therapy of these findings. So I mentioned a few times therapy for Gauchier disease. There is an enzyme replacement. It has been very successful in improving quality of life, especially in patients with type one Gauchier disease because it improves hemoglobin and platelets and it shrinks spleen and liver size. But it's incredibly costly, ranging from about 100 to 300 or $400,000 per year per patient. It's a treatment, it's not a cure, so you have to stay on it. The response is highly variable. It's an IV preparation, so it's inconvenient. And importantly, it doesn't cross the blood-brain barrier, so it doesn't help any of the neurologic manifestations associated with the disease. So we've been working with the new, not new now, but the then new NIH Center for, the NCGC, the NIH blocking on the name, Center for, somebody help me, what is it called? NCGC, chemical genomics, gotcha. Okay, so when we've been working with Chris Austin and Wazing several years ago, we developed our first screen and had a nice paper in PNAS. What we were looking for are what's known as chemical chaperone, thinking that chemical chaperone therapy may be very advantageous, first for Gauchier disease. Because in Gauchier disease, we know that glucosuribrosidase is synthesized in the ER. It's glycosylated and folded, but it doesn't attain its functional tertiary status structure until it gets into the lysosome. However, mutin glucosuribrosidase, when you have a mutation, can be, as I mentioned, just misfolded and degraded. So that if you can find a small chemical that binds to the active site and enhances folding, you may be able to deliver it to the lysosome where it can do its work. And we think that this might correct the enzyme deficiency in patients with Gauchier disease. So we approached it with high throughput screening where our goals were to identify activators or inhibitors of the enzyme and that would work as pharmacological chaperones. When you work with the NCGC, what you have to do is first have an assay that's miniaturizable. They all have to be done in a 1536 wealth format, which takes some time. Fortunately, we could miniaturize the assay because we had the commercially available enzyme and we had a fluorogenic readout. So once we had the assay, we provided the assay and the enzyme to the NCGC. And they, with robotics, can screen seven concentrations of these libraries. When you've started with 60,000, I think we're up to half a million compounds. And robots screen plates of compounds that have different concentrations. And it's really amazing. This is something that a postdoc would probably not be able to do in a decade. And we can get results in a matter of hours to days. We then came up with a new screening approach that was published in PLOS this year where instead of using the commercially available enzyme, we thought we should screen mutant enzyme, but we didn't have that available. So what we did is we took a frozen patient spleen sample and used that as the source of mutant enzyme and performed the screen of about a quarter of a million compounds and this time we actually found some really interesting compounds, both activators and inhibitors. Our lead compound is considered a non-inhibitory chaperone. And I'll show you, we believe that it can reverse lipid storage and enhance enzyme activity. So we hope to be able to develop this further. And I suspect that the small molecule therapy may stabilize or activate glucosyriversidase to treat Gauchatia disease, but it could also have implications for glucosyriversidase-associated Parkinson's disease. And actually some of the companies are now looking to this. Now the screens are just one step of the process. First then they have to, the compounds have to be confirmed and then you have to optimize the chemical structure and the medicinal chemists become involved and then you provide functional studies to determine cell toxicity from the compounds and then you need to have a phenotype that's corrected by the compounds and we've been working our way along this process. So with our lead non-inhibitory chaperone we show with four different assays that we increase the enzymatic activity. When we go to patient fibroblasts, these are patient fibroblasts treated with DMSO and these are patient fibroblasts treated with this compound. Green is glucosyriversidase and red is LAMP, which is a lysosomal marker. So in our patients we don't see very much glucosyriversidase going to the lysosome but when we add the compound we see this yellow which shows that we are seeing transfer to the lysosome. The compound appears to be selected for glucosyriversidase and it didn't activate other hydrolysis that we tested and to our delight we were able to make use of our new models, both macrophages collected from patients and IPS derived macrophages. So these are non-treated. This is, oh, sorry. This is a control line. This is patient macrophages and this is IPS derived macrophages. So we're looking at storage of the lipid and you don't see it in controls. We add these fluorescently tagged ghost cells to increase the amount of storage that we see. We still see very little in controls but we see a lot of it in the patient macrophages and then when we add our compound we're seeing that the storage is cleared which is nice and reassuring and hopefully that will help us get to the next stage of development. So I wanna go back and say that understanding the links between these two disorders is very rewarding and frustrating but I think that it will ultimately teach us about the pathogenesis of both disorders leading to improved genetic counseling and better therapeutic strategies. So I hope that you can appreciate now that the complexity that we're seeing in Mendelian disorders like Gauchier disease may ultimately give us a nice window into other complex disorders. So I wanna acknowledge a lot of different people. First of all, my group who's done all the work. Many of whom are shown here are listed. Many close collaborators both at NIH and around the world and I always like to give special thanks to the patient's family members and referring physicians who have contributed so much to these studies. Thank you. Thank you. Thank you. We're gonna have a number of questions about one and that is environmental exposures as a possible destination in Parkinson's disease. So that's a difficult question in Parkinson's disease and it's being challenged in major studies. If I could come up with DNA from a cohort that was known to have a strong environmental exposure, I would really like to be able to sequence glucose-reversidase to see if there's some association but at this point it hasn't been done. Other comments or questions? You can compare and raise the question. Oh, okay. And individuals have a strong family history and yet have no manifestations of either the Gauchies or the Parkinson's. And is it not only because they don't have the abnormalities or have you studied them, perhaps they have a resistant gene that could protect them from this and use that observation? So the question is patients that have the mutation but don't have either disorder, is there a protective phenomena? I won't say that patients that have two mutations in the glucose-reversidase gene don't have Gauchier disease, they just might have a very mild form because by definition if they have two mutations, they have the disease. But the other question is really important, is there some kind of protective gene? And one way we're hoping to look at that is we've been trying to collect Sib pairs where both patients have Gauchier disease but only one developed Parkinson's disease to see if we can tease out what's protecting the one that didn't get it. But at this point, we don't know. Can you tell us what chemical class or compounds that you've identified? Well, I'm not the chemist, I don't even remember the name of it, but it has been published. I can give you the reference. Yeah. Other comments or questions? If not, please join me in thanking Dr. Walsh.