Video Deconstructing a Neurosurgical Disease Play Pause Previous Next 1 of 200 slides Volume Quality 540P 270P Fullscreen Captions Transcript Chapters Slides Deconstructing a Neurosurgical Disease Overview Issam A. Awad, MD, MSc, FACS, MA(Hon) discusses cerebral cavernous malformations in his presentation on Deconstructing a Neurosurgical Disease. SPEAKER 1: Thank you very much, and I have to echo the previously articulated thoughts by our [INAUDIBLE] visiting professor on the fact that really, it should be our business to undo what we currently do, discover different ways of treating neurosurgical disease, and move on to new neurosurgical frontiers. And that's really what we are about. In the process of deconstructing neurosurgical disease, and I use that term, and this will become evident in a few minutes why I do that, in the process of deconstructing neurosurgical disease, we will learn a lot about how to counsel families about how to predict disease behavior, how to maybe refine our surgical strategies, and maybe design new therapies that would avoid surgery altogether. So I want to take the molecular medicine approach to a problem that is a blood vessel abnormality that looks like a mulberry-- that's why you see the mulberry and picture there. And then move it toward a signaling pathway that's called rock-- that's why the Hard Rock Park sign up there-- and try to put the two together for you, and tell you how this is coming together. First of all, I'll sandwich the talk with a few acknowledgements and disclosures. I also have no financial stake in this game, except for grants. And we have an intense dependence on a great deal of people, and you saw this from every one of the speakers before me, how this is a team sport. This is not anymore your own lab doing something. You have to collaborate, you have to learn how to fight some battles on foreign grounds that you are not familiar with, and get people to help you with that. The main three axes of my research have been the genotype, phenotype, and biomarker work that you will hear a lot about has been mostly my group and several Ph.D. Scientists working with me. And I'll acknowledge them as we go along through the talk. The forward and reverse genetics are done in collaboration with a wonderful team at Duke, and the signaling work is with the biochemistry team at UC San Diego, and now an offshoot of that at UCLA. So this is work that goes back and forth all over the country, and we use it as an opportunity to meet in each other's lab and learn from each other, and so forth. So to start with, the concept of genes and mechanism of disease started of course with that 19th century monk Mendel, who described Mendelian genetics. And that led in the 20th century to the concept of one gene, one disease, and then its molecular underpinning mostly in the 1960s and forward as we understood the molecular basis of genetic information. So as we understood this, it became clear that some diseases can happen where you have the same phenotype, the same end disease, but with different chromosomal locations being disturbed. So what that means is that you have different proteins that participate in the chain that can cause that disease. So if you disturb any one of those proteins, any one of those genes, the endpoint ends up being that final disease. So that's the concept of disease pathway that you hear about, whereby you couldn't disturb that pathway in several different ways. That will come back to help us because that pathway will also give us opportunities to rescue the disease, or to modify the disease by undoing some of those switches. The other concept that's very important is that there are some predisposition genes, and there are some modifier genes. And we don't always understand what is at play. We don't understand in the glioma world, why identical glioma patients can live six months longer or less than the other. There may be both genetic and epigenetic factors that contribute to this, and they are always at play in the background, and we need to be thinking about them. Because again, there if you understand those switches, you could either turn an aggressive disease into a less aggressive disease, or vice versa by understanding these mechanisms that can make a disease more active or less active, or modulate activity. Finally, the concept of germ line versus somatic genes. So unlike many of the tumors, where all the action is somatic, some diseases you inherit, in fact, in every cell in your body, some predisposition to that disease. And then on top of that, something happens in the area where the lesion develops. So you have an interaction between a somatic event that happens where the lesion develops and your germ line, which is the baggage you carry in every cell of your body. So it's very important to understand these concepts, and that's why we're spending some time on them. Now, the way molecular medicine works, and I haven't yet taught anything about CCM, although the slides that are down here are all CCM related, I'm just talking in general about this, when you have a disease, you can take an approach that is called gene discovery approach, which is you take the patients who have the disease versus the ones who don't. Even family members in the same family that have it, others that don't, and look at how their genomes are different, and find a significant association that can tell you what chromosome is affected, and ultimately what specific gene is affected. And you localize the gene, you know its protein, and that is a genome wide screening to try to get to the disease cause. Another way you do it is some of the work you heard from Dr. [INAUDIBLE] is a lesions exploration, whereby you look at the lesion and non-lesional tissue, and you see differential genes, and differential chromosomal abnormalities that can then enlighten you about what is causing the disease. Differential gene expression is a little bit different. This one applies gene chips that include anywhere from 15,000 up to all of our genomes, 30,000 or more genes on a gene chip, and you do enough differential examination of a lesion versus a non lesional tissue, you will see groups of genes that stand out in the lesion that do not exist in the control tissue. And some of those can then be put together in what we call pathways. So if you find eight differentially expressed genes, and three of them are related to one signaling pathway, then that tells you that that signaling pathway probably plays an associated or causative role with the disease. And finally, once you have a hint about what gene may be involved, that's where the fun starts. Because then you can knock down this gene in either cells, and see what happens to them when you make the cell deficient in that gene, how is it different from the other cells? You could do it in animals, such as earthworms or zebra fish, or flies that are very easy to modify genetically, as they multiply like crazy. And some of them are see through. You could look at the earthworm, and it has only one blood vessel that is also its elementary tube. So it's a wonderful organism to study blood vessel modification by genes, because you could almost look through it with the lights on and see the change. And then you can work towards mammals, and a beautiful to work with is our mouse. You know, the mouse now is a very able animal in terms of us genetically engineering mice with different backgrounds, and seeing what happens to them. And using all of this information in an integrated way to deconstruct the disease. So now let's move to neurosurgery and to the blood vessels of the brain. And about 20 years ago, we didn't really understand very much the difference between these different vessels. We all used to call them, you know, angiomas. And you know, whether it's a venous angioma, cavernous angioma, people use these terms interchangeably. They don't really understand it. But with MRI, and with better analysis of clinical information using pathology, medical records, et cetera, we were able to separate the phenotypes into different phenotypes, ones that have arterial venous shunting. We call them AVMs. Ones that are mostly venous dysmorphous, or malformation or the veins. We call them venous angioma. And then the capillary malformations are the ones where instead of having little thin tubes that are capillaries, you form big bubbles that are also capillaries, but they are abnormal, and we call those cavernous angiomas. So understanding the difference between these phenotypes allows you to separate apples from oranges as you start to explore the disease, because you're never going to find the gene link to a mixed entity. It's like the old doctor study, pulmonary infiltrate as a disease. Could be cancer, could be pneumonia, could be a lot of things. So you can't study vascular malformation as a disease, you have to reduce it more to a particular specific phenotype. So the CCM, or this Mulberry-like lesion is, in fact, made up of dilated capillaries that bleed in the adjacent brain chronically, and it looks like it has some iron signal around each of the lesions. But occasionally it can bleed outside the lesion, causing a hemorrhage, as you could see in those two cases. As it turns out, if you look at the autopsy series, or at consecutive MRI series, this is not a rare entity. About half a percent of the population have one of these in their brain. Most of the time it's silent, and this is doing nothing, but over time, if you follow every one of those lesions, there is about a 30% lifetime risk that it would do something. Headache, seizure, bleed, something. So it is an active lesion, and occasionally, obviously, in an unpredictable way it can do bad stuff. And we don't know the switches that make a benign lesion like this one here turn into something horrible, like this one. And we don't really understand that very much. We can do with statistics and say you have that percent per year bleed, but if you're the one guy that bled, that doesn't do you any good that the other 99% didn't bleed. We need a little bit more certainty, and more mechanistic information as to what makes a lesion misbehave. Now, the one good thing about this disease, and I got very attracted to it early in my career, is that about a third of the cases, where I did my fellowship was in the southwestern US, where about half of the cases in that reservoir had the familial disease. And that is great, because here you have a Mendelian autosomal dominant disease that affects a good fraction of your cases, and allows you to do some very simple genetics, and then go back toward the more complex lesions. In Parkinson's, for example, only 1% to 2% of the Parkinsonians are familial. So it's great to have a disease where you have 30%, 40% of the cases familial. So in this work, we teamed up over the years with great geneticists when I was at Yale, and then more recently with the Duke group. And we sent our families for testing, and we were able to localize three different gene loci that would cause this disease on three different chromosomal arms. And we now know the protein for each of those three. So get used with me to use the word KRIT which is the CCM1 gene. This one I won't use very much. Many people call it malcavernin. Isn't that kind of like a sexy term? Malcavernin causes caverns. And then this CCM3 one is a puzzling one, because it's programmed cell death 10 gene. The other proteins in that class modulate apoptosis, and they really play a very important role in normal biology, and here the tenth over of those proteins happens to cause CCM3. So what we have learned about this, some things that are kind of interesting is that all the Hispanic Americans of Mexican descent share a founder mutation. Isn't that interesting? That means they all came from one family, and we can test for that familial gene very, very quickly. For less than $80, a Hispanic American will tell them whether they have the Hispanic mutation or not. So it's already helped thousands of patients, just this simple discovery of the founder mutation in Hispanic Americans. Even this stuff was kind of missed among the big cases until we established a very special clinic for CCM3 patients, and now we know through this clinic that every CCM3 patient in North America is being flown to Chicago for evaluation. There are not that many. There are less than 2% of CCM patients at large, and what we know about them is that this a pediatric disease, where a huge lesion burden develops very early in life. Bleed rates are immense, very early, before the 10th year of age. This is the mean age of bleed of CCM3 versus the mean age of bleed of CCM 1 and 2. So here you have hyperactive genotype that opens an opportunity for us to ask, what makes it hyperactive? What is it about this protein versus the others on this pathway that makes the disease aggressive? We have no available therapy to alter the disease. We do not know what makes the lesions form or proliferate, and we know that some lesions form without the genetic baggage in the germline, and they often form on a venous anomaly in the brain, or a venous angioma in the brain. And these are invariably not the genetic ones, not the three CCM mutation ones. These that have multiple lesions without the vein, are the ones that will map to one of those three. Now, there are other forms of the disease. This one received radiation therapy for a glioma to both frontal lobes, and you see, this frontal lobe that has some radiation damage has a bunch of cavernous malformations that developed right in the radiation field. So there is something in the brain that can generate these lesions independent of the three CCM familial genes. These patients, by the way, and these patients have been tested in big studies, and they do not map to the familial genes, so that is something else that makes them form the lesions. Hemhorrage is part of the phenotype. I will later debunk the idea that every CCM lesion must have a bleed, but every CCM lesion we recognize clinically has iron with it, and has hemosiderin. And in fact, we can't really see it on MRI if we don't image iron. You know? So some of the tiny lesions, you have to use gradient echo or susceptibility sequences in order to see them, otherwise you only see the largest of the lesions. The lesion is full of thrombus at different stages of organization, but the vascular wall is very interesting. It does not have the organization of arteries and veins that you see in normal vessels and in AVMs. It barely has some actin in the subendothelial layer. Occasionally some caverns will express myosin, also in the subendothelial layer, but all the caverns do not have the contractile smooth muscles, so they do not form vessels the way mature vessels are. They are essentially grossly dilated capillaries. This picture here is very telling, as you could see how you have a single layer of endothelia filled with a bunch of red blood cells, and it's blown up like a balloon. That type of vessel should be a tube that allows only one red cell. That's what a capillary is designed to do. If it turns into a giant balloon, then there is something bad going on. So when you look at the ultra structure of that layer, it does not have the healthy, tight junction that we recognize in normal capillaries, but instead it's got gaps in the tight junctions, and also some vacules in the endothelial cells. So it is a very permeable phenotype that is almost ripping apart these endothelial cells, as opposed to having them form very tight, neat tubes. This now, of course, can be looked at in a very sophisticated way. This is ultra structural work that [INAUDIBLE] did in our lab, showing that you can actually measure the gap of the endothelial cells, and find that it is very enlarged in the CCM endothelial tight junctions as opposed to normal capillaries. And when you look at the molecular signature, this is normal, this is CMM, this is normal, this is CCM. What you see is that the normal capillaries have button-like structures that hold the endothelial cells together, and those buttons are undone in the CCM endothelial layer. Now, for a long time, we debated whether this is due to physical pressure of the dilation of those, or whether it was a fundamental mechanistic effect. And that took some biology work to define. These lesions have proliferative activity, so they express all the angiogenic factors, all the receptors, and when you do MIB, which are the proliferating cells, this is like cancer. You know, 7% of these endothelial cells are proliferating. There is no other normal endothelial cell in the brain that has positive MIB. You look at the brain, and only you see it in a little [INAUDIBLE] occasionally. But then nothing else, you know? So you shouldn't have MIB activity. Immunoglobulins were not known to be really involved in this lesion until they were looked at very systematically. And in fact, the predominant infiltrative cell is B cell and plasma cells that are antibody producing cells that are, of course, T cells and macrophages. But the presence of B cells and plasma cells was unexpected. And we pursued that lead, and [INAUDIBLE] has been with me in the lab for about seven years. He's an immunobiologist who's leading this work. He first showed that the CCM lesion has much more immunoglobulins than the serum, so it is enriched in immunoglobulins. And then when he dissected the plasma cells from the lesion and looked at their oligoclonality, at how diverse those plasma cells are, in the blood you see a very wide repertoire of variable regions of the plasma cells in the blood, but the plasma cells in the lesion are very oligoclonal. So they are honed against a specific antigen that we do not understand. And when you look at the clonality of the cells, within a given lesion the cells are clonally expanded. So these inflammatory cells are not a nonspecific injury response, but they are a honed in antigen driven response. So we've done work over a number of years, actually coding the valuable regions of the plasma cells dissected from these lesions. And in fact, now we are at the point where we have enhanced it recombinant artificial antibodies that we have synthesized. We can make as much of it as we want. that is based on those sequences. So that antibody will recognize whatever the antigen is in the lesion, and we can start dissecting the immune response in the lesion, and do the types of things that you guys did with tumors to try to see does immune modification, is that a pathogenetic opportunity, and a therapeutic opportunity for us to target? So let's go back a little bit to the genes, and the CCM proteins, which are the proteins coded by the three genes I told you about, they are expressed in the neurovascular units. So they're expressed primarily in the endothelial cell, but also in the astroglia that are adjacent to the neurovascular junction. So there is something about the neurovascular mileau that favors these proteins. Even though you inherit the gene all over your body, there is some predisposition within the neurovascular environment that explains why don't pop cavernomas all over your body, but you pop them in the brain. So there is something in the brain mileau that needs those genes, needs those proteins in a different way. These proteins are associated with the micro filaments, and with the stress fibers of the cells that help anchor the cells together. And the first breakthrough on this came from the laboratory at UCSD that is now our partner. And there is a theme here. If there is somebody very good, don't make them your competitor. Make them your partners. It's a lot easier, it's more fun. Both of you can have much more fun that way. So usually they welcome that, and you have opportunities to contribute equally to this work. So what this group was doing, they have no idea about CCM as a disease, but they were studying a signaling pathway equal called [INAUDIBLE] in epithelial and endothelial cells, and lo and behold, the KRIT protein was one of those pathways they were studying. That had no idea it caused CCM. But when you knock down this protein in endothelial cells in the dish, those endothelial cells become permeable, and they don't form tubes anymore. So they turn into bubbles, and they leak. You know, you measure permeability in gels, and they leak. So the leak and the bubble formation is the result of the loss of CCM proteins in endothelials. It's phenomenal to have had this come from a biochemistry lab that had nothing to do with CCM. Now, when you knock the gene altogether in mice, this is like two CCM patients marrying their cousins, you know? So the disease is also somal dominant, so we don't really have the total null phenotype in humans. But if you create the total null, meaning CCM minus minus in any mammal, they don't form a heart or an aorta. So that diseases, where one copy makes you form cavernomas, if you lose both copies, you don't form a heart. That's very important. Now, what happens in the heterozygous state, as most of us do not marry our cousins? The heterozygous state is where it counts. So heterozygous state, you inherit one copy in all cells of your body. The second copy gets you away. So you are living an apparently normal life with that, no overt pathology, until or unless that second copy in one cell somewhere also loses the equivalent, the gene, and then you have a homozygous state right in that spot. And that that's where the lesion forms. That is called [INAUDIBLE] hypothesis. It was first proven in autosomal dominant retinoblastoma, and it's also operating in polycystic kidney disease, where every cell in your body has one copy of polycystin, but then in this cyst, both copies are gone. So is that how CCM forms? So it took us about five years to prove this by doing a lot of mining of lesions for DNA, and looking at the DNA in the germline. And indeed, this is now proven both in CCM 1, 2, and 3, and there is a paper currently out that shows that many of the spontaneous, non-familial lesions have also the somatic mutation of the gene, so that this idea that the sporadic lesion, or a lesion is the result of a somatic head that magnifies the baggage that you bring to the table. We already have one hit in your germline, and then the second hit causes lesions. So to prove that, can you recreate this in the mice? So the problem with the mice is that brain is very small and they don't live too long. So if you depend on stochastic somatic mutations, you have to wait a long time for the mouse to form a spontaneous lesion. So you have to kind of encourage the mice to have stochastic mutations. So in order to do that, you could breed the mice in a background that causes them to have lots of mutations. So one of them is, of course, P53, which is a very bad gene. It's a tumor suppressor gene. If you inhibit that, about 20% of the animals get sarcomas and lymphomas, and so forth. But every animal will get the CCM if they are heterozygous to CCM genes. So basically what happens is if you're heterozygous to CCM and you have lots of somatic mutations, you start forming these CCM lesions in the mouse brain. The mouse brain is as big as a centimeter here, and this is how we encourage it to form. Now, the problem is that these animals still form turmos if they live a long time, so we moved into another predisposition gene called MSH, which is a point mutation repair gene. So that's like the word processor thing, where you type a T instead of an S, and it corrects it. So that's the gene that corrects point mutations normally. If you lose that gene in a human, you get some colon polyps. There are many redundancies in that gene. But if you knock that gene out in CCM heterozygous mice, they don't form tumors, and they form lots of CCM lesions. And these lesions now you can look at and study very, very effectively. So you have more mature lesions that look exactly like the human ones. B cells galore, plasma cells, iron leak. Every feature we have seen in the human lesion is recreated in that little tiny mouse lesion. Every molecular signature that we have seen. Now, on top of that, the mouse allows you to look at the earliest lesions. Something the brain of the human, you don't have access to that. So the mouse tells us that the first lesions are these gross single capillaries. They have no iron leak yet. They have no inflammation yet, but they have 100 red cells in a ballooned capillary. So this is really that pre-lesion, or the earliest lesions. So if you want to be a philosophical arguer, is this really a CCM lesion or not? So the fundamental test would be does it lack the CCM protein? Because if it lacks the CCM protein, like the mature lesion, then that's the lesion. If not, then that's a pre-lesion. So we did this, and it's really cool, because the stage one lesion, half the cells, just like the mature lesion, do not exhibit CCM proteins. So they have somatic homozygous loss, as expected. So the CCM starts as a somatic loss in a capillary that becomes a balloon, and then it matures into the multi-cavernous form. It's a wonderful opportunity, because now you have stages of the disease that you could modify. If you could take a family and keep all their lesions at that stage one, they'll never have clinical sequeling. So this is important. Now, nothing in medicine is discovered new. You know? So we thought this was a great new discovery, but Wyburn-Mason-- anybody heard of Wyburn-Mason? So he is a very prominent British neurologist, lived around World War II, worked on many classical pathologic entities. Wrote in a book on vascular malformations and tumors of the spinal cord that cavernous angioma developed from telangiectasias. Both may occur in the same case, hence it seems that there is no justification dividing this disease. You know, this is unbelievable that somebody 50 years ahead, with no molecular information, would essentially describe what I just told you we discovered in mice. That the first CCM lesion is a dilated capillary. This modeling allows us to knock the gene and other pathways as well, so the CCM3 knock out genes have many, many more lesions than the CCM 1 and 2. So this is a fundamental, the hyper-aggressiveness of CCM3 is fundamental at the genetic level, and you recreate it in mice. Now, what the hell is RAK? So RAK is Row Associated Kinase. And this is a kinase that has downstream effect on cyto-skeletal proteins that anchor cells together. So if row A is decreased, it's not activated by the kinase, the cells are tied together. If row kinase is activated, then the cells get stress fibers, and they separate. So our biochemistry colleagues took this to the next step, and we helped them with human lesions and with mouse lesions, and we showed that effectively when you knock down KRITS, CCM 1 with SIRNA, you have activation of RAK. Not just permeability, but activation of RAK. But what's really exciting is here, is that that activation of RAK, if you give a RAK inhibitor, still the gene is lost. There's no gene there. If you give a RAK inhibitor, you rescue the phenotype. The blood vessel goes back to forming tubes and to being impermeable. So in the absence of CMM genes, you rescue the phenotype by inhibiting RAK. This is a principle in molecular medicine whereby you can't change your genes, but you can change what they do. So you don't have to do gene transfer to cure this disease if you understand what it does, and alter those steps. So now, does that really matter? So people have, including us, and resulted in editorials in nature medicine and whatever that postulate that the loss of CCM3 causes row activation. That causes permeability and abnormal vessel development. Now, what happens in the heterozygous state if you only have one copy of the gene? Are you really normal? Well, we found out by testing brain and lung permeability, and the group in Utah tested skin permeability, that these heterozygous animals who have no lesions have, in fact, a hyper-permeable vascular bed that is rescued by RAK inhibitors. So every heterozygous mouse has a hyper-permeable vascular bed, even in the absence of lesions. In fact, we have a colleague on campus that studies pulmonary edema, and he uses my CCM1 heterozygous animals because they form all kinds of pulmonary edema. So you're not as healthy without lesions if you're heterozygous. So that's an interesting thing, and it allows you to perhaps measure permeability and understand the background vulnerability to lesions. And wouldn't it be nice to know if radiation causes lesions by also triggering permeability, or if the venous anomaly causes lesions because of focal permeability problems? So here the biology of the gene is teaching you about how to understand the disease at large. Now, when you look at RAK activation in the lesions, it is amazing. The lesions are all filled with RAK activity. And human lesions of all genotypes have that. So that asks, then, begs the question is if I inhibit RAK, do I prevent lesions from forming and maturing? So there is a RAK inhibitor that's available for chronic use in mice. It's not approved for humans, but we can give to mice for effectively the first two decades of a human life. So the first four months of a mouse's life. So upon weening, we start a group of mice on this drug, and another group of mice just have the regular chow. And the ones that are on this drug mixed with their chow, with their drinking water, in fact, do not form as many mature lesions. And it is remarkable that the lesions that do form have very little iron leakage, very little inflammation, and they're kind of chilled down lesions. When you look at the mice that form lesions versus the ones that don't form lesions, the background brain has higher RAK activity in the mice that form lesions. So what this may tell us is that there is some variation among us in our RAK activity. Some things we do increase our RAK activity, degrees our RAK activity. And if we have higher RAK activity, you are more likely to form lesions. At least that's what the mice tell us. So if that's true in humans, we will never be able to prove that unless we can measure RAK activity and permeability in man. We're getting there. This is another thing is when you give this drug, even when you control for lesion size, the iron leak is gone. So it restores permeability in these vessels, and that tells you that if you could measure iron in the human lesions, you might be able to measure therapeutic effect. So you might have a lesion that now leaks a lot of iron, but then you chill it down so it doesn't leak iron. In order to measure that effect, you'd want to be able to measure iron in man. So we have a grant over the next few years that's tackling this issue of do all the genotypes respond, or only CCM1, and could it possibly be accomplished by the effects of statin, which has a row along with other kinase inhibitors? So statin will inhibit RAK, but it does other things as well. But it would be very nice if statin were to work, because that is a drug widely available, known to be safe. You don't have to go to pharma and ask them to develop a new RAK inhibitor for a disease that's an orphaned disease. So in many ways this could be a very good approach, and NIH agrees we should test this extensively in mice. I can't show you the pictures, but it looks like statin doesn't work as well as the RAK inhibitor. So we'll figure out why, and maybe it's those, or whatever, but it's not going to be an easy, quick jump. What about B cell depletion? So we depleted B cells that are effectively gone from the bloodstream and the spleen, and these mice live without B cells. They're protected from infection, and they also have less lesions, and the effect is very similar in magnitude to that of RAK inhibition. So does that mean is RAK working by inhibiting inflammation, because permeability does not allow inflammation? Or is it the other way around, and dissecting the relationship of inflammation to what RAK does is very important, and will be pursued? Now, that means the last two to three minutes, a jump to the humans again, because it's ain't complete till you go back from the lab to man. So we can look at CCM lesion burden using gradient echo and see many more lesions than you can see on conventional MRI. But if you apply these new sequences, they're called [INAUDIBLE] or susceptibility sequences, you in fact can see many more lesions. So it's almost that you can measure the burden of disease more accurately with these high sensitivity methods. And in fact, these correlate with age retrospectively. We have an ongoing prospective study that is telling us how many new lesions per year do you form in each of the genotypes. So it looks like in the CCM3, it's very measurable. You know, four to six new lesions per year, or per hundred patients. So it's a very measurable number that you could alter. In the more benign genotypes, it's maybe less than one lesion per hundred patients per year. So the new lesion genesis is very slow in some genotypes, but not in others. Can we measure permeability in man? So we have a fellow in our lab that has spent two years doing this, and you use dynamic permeability of gadolinium to calculate permeability rates in lesions, or in background brain using this dynamic imaging of the gad leak. So you give gadolinium, and you take-- is it six to eight images very, very quickly, and then you look at the amount of gadolinium enhancement, and you calculate the convolution curve. Within any region of interest, you will have this type of permeability numbers, and we looked typically at the mean or median for our analysis in given our OIs, or regions of interest. So this allows you to look at permeability of lesions, but also permeability of background brain, which we know is impaired in mice. So this is our work in humans, and it's just been submitted for publication. And what it does show is that the familial cases have higher permeability than the sporadic cases in white matter of normal brain, which is exactly what you would predict in this disease based on every cell in your body being heterozygous. And in fact, the patients on statin that happened be in the familial group don't have the high permeability. And some patients in the familial group have high, others don't have as high permeability. And when we look at those, the ones with high permeability have more lesions. This is exactly like the mice. So the ones that have higher permeability in normal brain have more lesion burden, and the ones with lower permeability have less lesions. So it's a very interesting way now to measure in humans what we are studying in the animals, and give us biomarkers of the disease activity. The other thrust in our lab that we started based on the laboratory work is the work of [INAUDIBLE], who is the first engineer I ever recruited as a faculty. And he is phenomenal because he specializes in the imaging of iron. And this is Juan's work, and you are in fact able to quantify iron using a technique called quantitative susceptibility mapping, where the amount of iron signal can be measured in the lesions. So if we are going to treat lesions or follow them based on iron leak, this will be a measure that allows us to do that. Look at this work that Juan has done. So basically, this is validation work in the mice. The area that have iron is the area that is positive on QSM, and these are human lesion specimens that I gave Juan. He measures QSM, and then he does mass spectroscopy. And the amount of iron is exactly what he would predict on mass spectroscopy and on QSM. So this is a very good and accurate method of measuring the amount of iron, if that is a relevant end point. Iron and permeability correlate with each other. We just published this. So the more leaky lesions have more iron. These are two independent techniques altogether that show a biologic effect that you would expect together. So I think we're on the right track, measuring what is happening with these lesions. So to conclude, I hope I showed you a disease that you didn't know very much about, that is in some ways a paradigm disease. Because it's the common phenotype that is often quiescent, but if activated, it can cause bad stuff. And we have now some triggers and mechanisms to understand how it forms, and what makes it active. And for the tumor people, is this really a tumor, or is it a malformation? And I love that argument, because there are features of this that is like a hamartoma, or a tumor. So it has proliferative index, it grows. It's got angiogenic and growth factors in it, like [INAUDIBLE] would have been proud to call these angiomatus tumors. But then other people would say they're malformations. They're deformed vessels. And maybe the argument is an important one, because going from a malformation to a tumor is something very important for us to understand. What makes a lump a growing lump, or a stable lump? Something that the Germans talked about a century ago, and now we have tools to re-dissect that argument. This is my final slide, that if I spent career working on CCM, I hope that the people in my lab will take this work to other diseases where similar events are taking place. So aging of the brain and hemorrhagic micro-angiopathy in fact has very similar features to CCM disease. You have RAK activity, you have permeability, and you have iron. Except they happen an old patients with aging. So what we understand in this very special orphan disease may open some windows in medicine, hopefully to other diseases, just like we use tools from other diseases like [INAUDIBLE] and Mendel to dissect what we learned about this entity. At the end of the day, this cannot happen if you didn't have a big group really working like family together and understanding that each one around the table sharing a meal actually has family, has a life, has a career, and their time with us on CCM is a brief stop along that long journey. We want to make that a pleasant stop, a positive stop, and understanding how this works together that allows us all to be much more productive, and is an important part of doing collaborative research. Thank you very much, everybody. So, questions? [INAUDIBLE]? AUDIENCE: CCM1 heterozygous animals? SPEAKER 1: Yes. AUDIENCE: And radiate half of the brain. [INAUDIBLE]? SPEAKER 1: So those animals are done, and I don't have the answer, but we've done the experiments. So we're looking at various sensitizers, right, including radiation, to see if that will happen. What surprised us is that the majority of radiation induced cavernomas do not map for any of the familial genes. I wasted $60,000 on that project, and not a single case mapped to the CCM genes. So there must be other predisposition in addition to that. The other sensitizer is, of course, the animals that over-secrete RAK. So there are animals that are engineered to over-secrete RAK, and we have those being processed. So this is really a key. These kinds of things that you can work with to move this forward. AUDIENCE: Is there a way to collect retrospective data from the familial cases among those who took statins for other reasons? SPEAKER 1: Yeah. That is being done, and in fact, the clinical part is a little bit confusing, because the patients on statin are different from the patients not on statins. They're older, they have vascular risk factors. So even permeability is something we have to interpret very carefully. But it surprised me that the older patients with [INAUDIBLE] on statin had, in fact, lower permeability than the other familial cases. So yeah, that is being done. But I think there are enough confounders that the effect is going to have to be prospective. AUDIENCE: You said the [INAUDIBLE] homozygous knockout has no heart or aorta. Is that right? SPEAKER 1: Yeah. All three genes, actually. CCM1, 2, and 3, homozygous don't form a heart of aorta. AUDIENCE: Do they form vessels in the brain? SPEAKER 1: They don't get that far. They don't get that far. So they really malform at a much earlier stage, fourth or fifth day of embryogenesis, and they basically don't get there. Now, I want to tell you about a model that I didn't talk about today, is if you take the animals that are just born, heterozygous, and knock conditionally the second copy, what happens? So these are post-natal heterozygous. The heart has formed, so they're OK there. What happens to them? So if you do this experiment in the first three days of life, post-natal, they form CCMs all over the brain. And retinas. Because the vessels are still maturing in that part of the brain. If you do that homozygous knockout after day four or five post-natal, they don't form. So if you knock out CCM in the period of active angiogenesis, which they hit the brain and the retina are still active in the first two, three days of life, then you'll need at least one copy, or you're going to be bad. AUDIENCE: What if you knock it out at a point in time when vessels are forming elsewhere in the body? Do they get CCM? Do they get-- SPEAKER 1: Yeah. So the group at [INAUDIBLE] is looking at reaction to injury in induced homozygous loss. Because there what you would say is you're recruiting a bunch of capillaries, but those capillaries have no CCM. So what does that look like? Does it look like cavernoma every time you're injured? And that work is not out yet. But it opens up those types of windows to things that are very critical when you've got a protein that looks like it's so benign for some strange disease, but yet is so fundamental to life in another way. Published April 2, 2014 Created by
