The major diseases affecting the thoracic aorta are aortic aneurysms and acute aortic dissections. Medical treatments can slow the enlargement of aneurysms, but the mainstay of treatment to prevent premature death resulting from dissection is surgical repair of the thoracic aortic aneurysm, which is typically recommended when the aortic diameter reaches 5.0 to 5.5 cm.
Data from genetic studies during the past decade have established that mutations in specific genes can distinguish patients at risk for the disease and predict the risk of early dissection at diameters smaller than 5.0 cm. This information has the potential to optimize the timing of aortic surgery to prevent acute dissections.
DIANNA MILEWICZ: So today, I'm going to talk to you about the full spectrum of genetic contribution to thoracic aortic disease, and what the genes are telling us about the molecular pathways to this disease. And also, along the way, emphasise how the underlying genes influences the management of those patients, and helps predict the timing of aortic aneurysm repair and what additional vascular diseases those patients may get into. So for many years, I've been working on ascending aortic aneurysms. And the natural history of these aneurysms is to progressively enlarge it over time. That enlargement is totally asymptomatic. It reaches a certain size where it becomes unstable, and can lead to a type A dissection, where the blood enters the wall of the aorta. Type B dissections are also part of the spectrum of this disease. They tend to occur with little to minimal enlargement of the descending aorta. And we still don't understand why type B and type A dissections are part of the spectrum of this disease, and why the dissections always are initiated at two locations in the aorta. So the idea behind sort of understanding the genes is that if you know who's at risk for this disease, you can prevent the deadly dissections. And you can do that by routine imaging of the root and ascending aorta, putting them on medications, like beta blockers, and more recently, there's data to support that you can use losartan. And then avoidance of any isometric exercises. But the mainstay of therapy still is surgical repair. And the standard is to do that repair when the aneurysm reaches 5 to 5 and 1/2 centimeters. Now, we've known for many years that genes play a role in this disease, and we know that for Marfan syndrome. And now we know there's additional syndromes, and I'll talk briefly about that. Our work determined that in people that don't have the syndromic presentation, that 20% of those individuals will have a family history. So beyond even these syndromes, like Marfan syndrome and Loeys-Dietz, there's a contribution to this disease where it's inherited in families. And then finally, we know that bicuspid aortic valves are there in about 1% to 2% of the population, primarily males. And about 20% of those individuals will go on to have aneurysms. But the risk of dissection appears to be fairly low with those aneurysms. Then we know what the environmental risk factors are for this disease. If you look at people that don't have a genetic trigger for their disease, 80% to 90% of them will be hypertensive and have longstanding, poorly controlled hypertension. And then there's also reports of people that do a lot of heavy weight lifting and body building or people that have taken crack cocaine. Any of these factors that will increase stresses on the ascending aorta appear to be a predisposition for aneurysms, and in particular acute aortic dissections. So I want to remind you that the ascending aorta actually has this beautiful design, that to withstand those pressures from the beating heart throughout a life span. So the aorta is made of an inner layer endothelial cells and the outer layer of the adventitia with the vasa vasorum, the arteries that provide nutrients. But that middle medial layer is the layers of elastin, shown here in this cartoon in black, and smooth muscle cells that run in between those elastin fibers. And in humans, there is between 50 and 70 layers in the ascending aorta. And the whole construct of the interaction between the smooth muscle cells is really designed to minimize the forces on these smooth muscle cells. So elastin sends out extensions that link up to the smooth muscle cells through integrin receptors. And then those link up to the contractile units inside the cell. And then every time the blood pressure comes up the aorta, the cells are slightly stretched and they contract in response to that pulse pressure. But as you know, the aorta is very elastic, and it recoils with every pulse pressure. And they've shown that the smooth muscle cells, if you poison them so they can't contract, that the aorta maintains that elasticity. So that elasticity is really dependent on the elastic fibers. And so for many years we didn't really quite understand what these smooth muscle cells were doing, other than they were there during development to lay down the layers of elastin. Because in the aorta, which has a large diameter and lumen, it plays no role in controlling blood pressure or controlling blood flow. The roles of the smooth muscle cells usually play in smaller arteries with smaller lumens. And you can see this construct by electron microscopy. Here are the integrin receptors on the cell surface, shown down here-- the dense plaques. They link up to the elastin fibers, and that link up is done through structures called microfibrils. And then they link up to the contractile filaments within the cell. And the contractile filaments are the same as what you see in the skeletal and cardiac muscle, except for they don't form sarcomeres. But you still have the polymers of actin and then the myosin heavy chain polymers, where the myosin motor moves across the actin filament to contract the cell. And they don't have the sarcomeres, which is where the name smooth muscle cells come. They don't have that striated appearance. And what happens with thoracic aortic disease is this beautiful, elegant structure is destroyed. So here is a Movat's stain of the normal aorta, where the elastin is shown in black, and the red is the smooth muscle cells in between. And once again, many layers in the human aorta. And this is what you see with anybody that has a thoracic aortic aneurysm, except those with an inflammatory or traumatic cause. You see destruction of the elastin fibers. You see fewer smooth muscle cells and an accumulation of the blue material, which is proteoglycans. So what goes on in the pathology doesn't really tell us that much as to what's triggering this disease. And so as I said, my work really is going after this genetic predisposition, so we can begin to use the genes to figure out who's at risk in the general population. So we can pull out those people, image their aorta, do the surgery, and prevent the dissection. And this is like the many faces of aortic disease. And these are all individuals who've had acute aortic dissections. At the one end of the spectrum, you have people with Marfan syndrome. We can recognize those individuals from the general population because of their syndromic features. And they have a high risk, almost 100% risk of having thoracic aortic disease during their lifetime. Typical age of onset of a dissection, if they are not followed in their mid 40s. And then there's individuals like John Ritter. He had his dissection in his early 50s. And after his death, the family came to Houston. We imaged everybody, turns out his brother had an aneurysm. He went to surgery and is still alive today. So this sort of indicates that this kind of dissection in a family can be a red flag that there's other family members at risk. Turns out that Tex Ritter died. It was John Ritter's father and one of the singing cowboys back in the 1930s. He had chest pain suddenly and dropped dead. And then we recently found out that his brother, Rex Ritter, had actually a documented aneurysm that was being followed in the 1960s and '70s. So these individuals we think all have a single gene that's altered, and that can be passed in the family in an autosomal dominant manner, so that somebody like John Ritter has a 50-50 risk of passing that genetic predisposition onto his children. And then at the other end of the spectrum, we have people like Richard Holbrooke, who is the Under Secretary of State to Hillary Clinton. He had a dissection in his office, he was 68. He died a few days later after they tried to surgically repair it. And he has no family history, he had a history of hypertension. But he probably had some genetic predisposition, but it's not these high risk genes that we see in these familial cases or syndromic faces. And at the other end of the spectrum, we have Michael DeBakey, who had his dissection at the age of 96. It's quite ironic that he had a dissection, because he pioneered the surgical repair of a dissection, and did the first successful surgical repair of an acute aortic dissection, actually in 1956. And he, even at the age of 96, he underwent the surgical repair and survived and went on to live just a few months short of his hundredth birthday. But he may not have had any genetic predisposition, so it may have been almost 100 years of just chronic biomechanical forces on his ascending aorta that led to his dissection. Because we can see these same pathological changes as people age. So Marfan syndrome, everyone in the room knows about this. This due to FBN1 mutations. They encode those microfibrils that-- the major protein in those microfibrils that link up the smooth muscle cell to the elastin fiber. And we know what the complications-- the extra aortic complications in these patients are. We also know that if we do the routine management of the imaging and beta blockers or losartan, that they-- and then surgical repair when they aneurysm is a certain size-- that these individuals can live a near normal life expectancy. We also know after repair of that aortic root aneurysm that they are at a low risk to go on to have intracranial aneurysms or iliac or splenic artery aneurysms. And a small percent will go on to have descending dissections. And we haven't really defined what percent of individuals will go on to have that complication. And we know from elegant work by Hal Dietz that in the Marfan mouse model there's evidence of excessive TGF beta signaling, and if he blocks that signaling, using either an antibody that neutralizes TGF beta or with losartan, that he can block the disease in the mouse model. And that was the basis of a recent clinical trial that showed losartan worked as well as beta blockers in slowing the growth of aortic aneurysms in patients with Marfan syndrome. So once we figured out the subset of patients that truly had FBN1 mutations, we realized that there were people that were coming to the clinic with aortic root aneurysms, with Marfanoid features, but they didn't have a mutation in FBN1. And so over the past 10 years, there's been a number of additional genes that have overlapping features with Marfan syndrome, but they are distinct genes from the FBN1 gene. And these are mutations in the transforming-- once again that TGF beta signalling pathway. So we've got TGFBR1 and 2, which are originally described for Loeys-Dietz syndrome, and then SMAD3, TGF beta 2. And more recently, there's been SMAD2, SMAD4, and TGF beta 3. And all of these genes can cause some degree of skeletal manifestations of Marfan syndrome. They're all inherited in that autosomal dominant manner. They present with aortic root aneurysms. But they go on to have aneurysms and dissections of other arteries. So we recommend that these individuals have head to pelvis imaging routinely to look for additional aneurysm. In addition, they have some degree-- many of them have some degree of arterial tortuosity. But now we're beginning to realize that that is pretty much there with all the genetic syndromes. And then a subset of these patients with TGF beta pathway genes will have additional features beyond the Marfan syndrome, including craniosynostosis, cleft palate, bifid uvula, hypertelorism, and very thin skin. So here's some of the patients that were initially described as having Loeys-Dietz syndrome. And based on these initial patients, that were mostly the very severe end of the spectrum, it was suggested that the aortic aneurysms be very aggressively managed. That these people, instead of waiting to see the aneurysm grow to 5, 5 and 1/2 centimeters, like you can in a patient with Marfan syndrome, that the initial recommendations was to repair them with just minimal enlargement, once you've confirmed the TGFBR1 or TGFBR2 mutation. And the disease can be very aggressive in these patients. Here is a patient we followed that had his entire aorta repaired due to progressive aneurysms. And then he went on to have an aneurysms of the great vessels of his neck, and went on to have splenic and hepatic aneurysm. In all, he's been through 20 vascular surgeries, done well with those surgeries. And he reflects that if you address this disease surgically, in a very aggressive manner, that you can manage this bad additional vascular disease. I think the surgeons could run their whole clinic on a few of these patients that have very severe disease. So in the initial publications, the survival in these patients was actually very low and actually worse than we see in Marfan patients, with an average survival-- a 50% survival under the age of 30. But now we know, with additional time, we know that these mutations in TGFBR1 and 2 and these additional genes can cause a whole spectrum of disease. So they can cause the Loeys-Dietz, with the craniosynostosis and the hypertelorism. But they can also cause people to present like a typical Marfan patient. But they also cause-- and this is probably the most common presentation for these patients-- they can cause presentation with no syndromic feature, just autosomal dominant inheritance of thoracic aneurysms and dissections. So through the Montalcino Aortic Consortium, which is an international consortium focused on defining the clinical features with each of these genes, we have collected a cohort of almost 176 patients with TGFBR1 mutations and 265 patients with TGFBR2. And we've just completed the initial analysis that was presented at the American Heart meetings, and I'm not going to go through all the data. But here's the survival. Instead of being an average survival at 30 years, which was reported with the initial patients, the survival is much more extended when you look at the full spectrum of this disease. In addition, this is survival free of any vascular feature. And so even though they are presenting with an aneurysm or dissection for surgery, around the age of 30, they are getting that repaired and living longer. And so it may be-- including the milder end of the spectrum-- it may also be that these patients are being recognized, they're being diagnosed, and they're being properly managed. And that may be leading to a much improved survival. And what we've been able to pull out is some initial differences between TGFBR1 and 2 in the medical management. Up to now, they've been grouped together, but they are two distinct genes. And you can see here, if you look at survival free of an aortic event, defined as aortic surgical repair or dissection, that with TGFBR1 there's a big difference in survival based on gender. So men will present at a younger age than women with an aortic event, whereas there was no difference with TGFBR2. And there's additional features that have fallen out in these studies. So right now the field really doesn't quite understand what's going on with TGF beta signaling in these patients. So just to go through briefly TGF beta signaling, the ligen binds to the type 1 and 2 receptors, which are encoded by TGFBR1 and 2. It comes together and activates SMAD3, one of the downstream signaling molecules, that bonds to SMAD4. It moves into the nucleus and changes the phenotype of those smooth muscle cells in the aortic wall. And what we're hitting along this pathway is every single protein in the signalling pathway can be mutated. And when it's mutated, it causes a predisposition to disease. But the real paradox that's come out of this is that all of these mutations universally are loss of function mutations that are predicted to decrease TGF beta signalling. At the same time, if you look at end stage disease, there's evidence of excessive signalling. So we still don't know the role of TGF beta signalling in this disease. But it makes it a little bit scary to direct therapies at TGF beta signalling, because we know that if you lose that TGF beta signalling, you're predisposed to disease. And even despite the fact that there's evidence of excessive signaling in end stage aneurysms when they're removed-- end stage aortas when they're removed from the patient. Now we want to talk about the genes that we found for these people that don't have syndromic features, like John Ritter. And we did a simple clinical study in the late 1990s to show that 20% of people that come in with an aneurysm or dissection for surgical repair, that don't have Marfan syndrome, or Loeys-Dietz syndrome, or any syndromic feature, will have a family history. We defined it as autosomal dominant in the vast majority of these families, and it's associated with decreased penetrance and variable expression. And the decreased penetrance is shown in this woman here. Her father died of a dissection, her brother had a dissection, both of her children had a dissection-- her daughter actually died after giving birth to her twins of an acute aortic dissection. And she's in her 70s now. We imaged her entire aorta routinely, no evidence of disease. So she must carry the mutant gene, based on her position in this family. But she has no-- despite her advanced age-- she has no anatomical evidence of aortic disease. And I'm just going to go through a few families to show you how much variability there is clinically in these families. So here's a large family that we identified in Texas. This individual actually got Lifeflighted to Houston with an aortic dissection. His route was actually 7 1/2 centimeters at the time of presentation. And he gave us the history that his uncle had a big aorta. And so we brought in everybody with an asterisk by their symbol-- there's a symbol in this pedigree. I brought them in for imaging, and with that we found a large number of asymptomatic aortic root aneurysms. And in this generation here, these individuals all had aortic aneurysms around 5 to 6 centimeters over time. We sent everybody in this generation for aortic repair. But I don't know if we're actually managing this family correctly. I mean, here we have a huge family, only one individual presented with dissection, and he had a very large aneurysm. So whatever is going on in this family, the underlying genes predicting that they're going to form large aneurysms that are associated with a low risk of dissection. And we actually had 10 years of imaging from this individual's medical record. And his aorta actually sat about 5 centimeters over that 10 years with little to no growth on beta blockers. But the exact opposite is true with this family. This is a family where everybody who died of dissection went to autopsy and had no evidence of enlargement of their aorta. In fact, we were following this woman in Houston. We imaged her aorta and three months later, she was actually died of acute type A dissection. So for this family, until we find the gene-- which we've not found the gene yet-- until we find it, we have nothing to offer them as a marker for who's at risk for sudden death due to dissection. And so that gives you an idea how much differences there are in the aortic disease presentations in our family. But we also find that there's differences in the associated features, like some families will have intracranial aneurysms. Like this family, where these two women, who are in the pedigree and at risk for the gene, presented with intracranial aneurysms instead of thoracic. We can see patent ductus arteriosus associated with thoracic and aortic disease. About 15% of our families will have that association with BAV, and so on. So there's a lot of clinical heterogeneity, which predicted early on there'd be a lot of genetic heterogeneity-- that there'd be a lot of genes, just like in the genes in the TGF beta pathway-- a lot of other genes that would predispose to this disease. And I also had to say that where the aneurysms occur in the ascending aorta is also different. So we have patients with aortic root aneurysms, like you see with TGF beta pathway genes and FBN1. But we have other genes where the root is spared and the ascending aorta's involved. And then we still have other genes that present with more of this fusiform enlargement, involving both the root and the ascending. So over the years we've collected a cohort of patients with multiple members with this disease. And we currently have over 800 families in our study. And almost all of them show that autosomal dominant inheritance. And we've gone through and our work and other people's work have now identified a whole list of genes for this condition. And I've divided them into those that are affecting the elastin fibers in the connection up to the smooth muscle cells, those that are involved in TGF beta signalling. So as I said previously, you can have mutations in these genes and present with no syndromic features. And then finally these down here, all genes that are involved in smooth muscle cell contraction. And these genes down here are genes that we've identified, but we don't quite understand how they're causing disease at this time. So what we're finding with these genes that are disrupting smooth muscle contraction, we're hitting the major structural elements of that contractile unit, but also both the kinases that control contraction and control relaxation of the smooth muscle cells. So as I said before, the actin filaments-- actin's made as a monomer. It polymerizes into actin filaments. And the gene that makes those actin filaments is ACTA2. And we have mutations in that gene and I'll talk more about that in a minute. We also have mutations in MYH11, which enclosed that thick filament, the myosin heavy chain. And then we have mutations in the kinase that phosphorylates the regulatory chain, activating the myosin head movement. And we have mutations also in the kinase that controls the phosphatase that leads to relaxation of the unit. These are loss of function mutations and these are gain of function. So in both cases, mutations in these kinases are predicted to decrease phosphorylation of the regulatory light chain and lead to less contraction of the smooth muscle cells. And it's important to point out that the cells can't contract properly unless they're holding on to something. They have to be holding on to the matrix. And that linked up to the matrix is through microfibrils and FBN1 is the major protein in those microfibrils. And we have mutations in some of the other proteins in the microfibril. And they've shown the Marfan mouse also has decreased contractions. So this decreased contraction may also contribute to Marfan syndrome. So once again, it really highlights that this structural integrity from the elastin filaments to the contractile unit is important for maintaining the structural integrity of the aorta. And if you mess with any of these components, you end up with aneurysms and dissection. So instead of focusing on TGF beta signaling, we're now proposing that the smooth muscle cells serve as a mechanosensor in the ascending aorta. And if there's anything defective in that mechanosensing, that they activate smooth muscle repair pathways, which can include upregulating TGF beta, angiotensin signaling, MMPs, proteases that will degrade the elastin fibers and increase proteoglycan synthesis. And that's what's leading to thoracic aortic disease. And the nice thing about this model is you can-- also explains hypertension. In the case of hypertension, the cells aren't sensing something's wrong but instead the pressures- the forces are increased. And that they may be upregulating these same exact pathways. So that's the genes for thoracic aortic disease in the high risk genes. And I just want to tell you quick story, a little bit about ACTA2, since it is, right now, the major gene that causes familial aortic disease. It can explain up to 20% of familial disease. And then there's an interesting story that came out of ACTA2. So here's the family that we initially used to map and identify the gene. Everybody shown in orange has the mutation. And everybody in black has thoracic aortic disease. So you can see it's segregating with disease in the family, this particular arginine 149C alteration in the gene. But there's decreased penetrance. And it's actually the least penetrant gene we've identified to date. If you look at this generation, half the people that carry the mutation-- half the people that have the mutation-- don't have disease and half do. So it's only about 50% penetrant in this particular family. And so this family, when they contacted us about participating in our study-- and since it was a large family, I actually flew out to where they live to meet the family. We met in an auditorium like this in their local hospital. And I went through what we were doing and how if we found-- even without finding the gene, we could image everybody and repair aneurysms before they caused dissection. And then explained, the more people that participated, the more likely we would be able to find the gene. And actually this guy-- at the end, I said, are there any questions? And this guy got up and he said, you know there's a lot of early onset coronary artery disease, heart attacks in our family. And we don't have-- nobody smokes, we don't have high cholesterol. And there are so many people in our family with early onset coronary artery disease that the Rockefeller Institute has already been here to collect DNA from family members with coronary artery disease. And he said, could it be-- you know we have this dissection and early onset coronary artery disease in this family, could it be the same gene? And I said, no, there are two different processes. One makes the artery get bigger, and the other-- the coronary artery disease is due to a blockage or occlusion of the arteries. So I don't think it's the same gene, and it's good that you're participating in both studies. Because I think you've got two bad genes running in your family that are causing both diseases. And after we found the mutation in that individual, we went back to that family, we went back to all our families and found missense mutations throughout the gene in different locations. And I want to say that acted ACTA2 mutations cause both PDS and BAV as associated features. And they lead to both type A and type B dissections. Here's a woman with an ACTA2 mutation. You'd never pick her out at risk for disease. And just the presentation is a little bit distinct from what we see in FBN1 or TGFBR2. And this is a study that we did of the first 277 patients with ACTA2 mutation. And if we looked at presentation with an aortic event, defined as an aneurysm or dissection, the vast majority of people with ACTA2 actually present with an event. And the most common event is an acute aortic dissection. So they're not being picked up because of syndromic features or family history to the same extent that FNB1 and TGFBR2 patients are because they present with much lower percent-- actually show up with an aortic event. And particular, less than half will show up with an aortic dissection. Another interesting feature of ACTA2 is that type B dissections are much more common than they are with FNB1 or TGBFBR2. And individuals with type B dissections are actually presenting younger than the type A dissections. And we don't typically see that with other genes. So if you see somebody that's under the age of 30 or with a type B dissection, no syndromic features to date, at least in our population. That's been 100% ACTA2 mutations. So keep that in mind. And the individual's will-- as I said earlier-- will present with fusiform aneurysms of the ascending aorta. And if we look at the size of the aorta, at the time of type A dissections, we can see that about a third of them will present with dissections under a diameter of 5 centimeters. So we're actually recommending repair at 4.5 centimeters. And people presenting with aneurysms are still being, on average, repaired at 5 centimeters. So they still tend to follow the guidelines for Marfan syndrome. So going back to this family, once again. When I initially went to see them, I asked if afterwards if I could examine them. And it turned out that everybody who had survived dissection had a particular rash. And it was a rash called livedo reticularis. After we found the gene, I went back to the family-- back to where the family was-- and I could show that livedo reticularis was there in everybody that carried the mutation, down to two-year-old children. And so, you know, initially I thought, well that's interesting, when I was examining the patients with that initial meeting. But later when I was flying home on the plane, I started thinking, well, livedo reticularis, that's a rash due to occlusion of the small arteries in the skin. And that got me thinking that maybe it was the same gene causing both the early onset coronary artery disease in this family and the aortic aneurysms. And the possibility became greater when we actually looked at the pathology of the aortic walls. So in that medial layer, it looked like your typical degeneration that you see with any aneurysm, any mutation. But when we looked at the small arteries in the outer wall, the vasa vasorum, the arteries were distinctly different from what we typically see. And here is a control aorta, where these arteries are very small. They have one layer of smooth muscle cells around them. But when we looked at different ACTA2 mutations and the vasa vasorum, those arteries were greatly enlarged, and they were either completely occluded or stenotic. And these cells that were occluding these arteries stained positive for smooth muscle cell markers. It actually looked like they had little tumors of smooth muscle cells in these small arteries. So that, once again, was another piece of the puzzle to say, maybe it is the same gene. And we actually had a set of smooth muscle cells explanted from patients with ACTA2 mutations. And we could tell in culture-- when we were culturing these smooth muscle cells-- they actually proliferated more rapidly, and we were able to document that increased proliferation in our patients with ACTA2 mutations. So it looked like that ACTA2 was associated with this hyperproliferated phenotype in the smooth muscle cells that could lead to complete occlusion of an artery with no evidence of a lipid deposition. So we went back and collected all the additional vascular disease in these families. And here's that initial family, where now everybody with an early onset MI, defined as under the age of 55, or stroke under the age of 55, is now in either red or green. And now you can see that all these individuals that didn't have aortic disease instead had early onset coronary artery disease, and few of them actually had strokes in addition to that. And these were early events. This individual had his first event at the age of three with no cardiovascular risk factors. And for this particular mutation, we saw a lot of early onset coronary artery disease. And other mutations, we actually saw primarily premature strokes. And many of these strokes were diagnosed as Moyamoya disease. And that's a disease where the carotid arteries are wide open until the very distal portion. And then they become completely occluded or stenotic bilaterally. So this young girl had her first stroke at the age of 16. And this is her imaging. And you can see the carotid artery is completely open to the distal part. And then it's occluded on this side and stenotic on this side. And these are really early onset strokes. This young boy had his first stroke at the five. Actually underwent bypass procedures for these occlusions, and went on to die of a second stroke at the age of eight. And when you look at the arteries in Moyamoya disease patients, they look just like the arteries in the vasa vasorum. They're filled with smooth muscle cells with little to no lipid deposition. So we went back and got all-- as i said, all this vascular disease. And we were able to show both through linkage and familial association that ACTA2 was causing aortic disease, early onset coronary artery disease, and early onset stroke. And found some correlations that they arginine 149 alterations lead primarily to coronary artery disease, whereas this particular change leads to cerebrovascular disease. So right now we're working on a hypothesis that ACTA2 mutations cause loss of contractile function and in large arteries, like the aorta, exposed to high pressures, that leads to an aneurysm. So it follows along the same pathway that we see for all the genes that are disrupting smooth muscle contraction. But at the same time, for some reason when you disrupt the actin filaments inside the cell, that there's a gain of function, and the cells are more hyperplastic than they are typically. And then the small arteries exposed to low pressures, that that can lead to vascular occlusive disease, presenting as coronary artery disease or stroke. But I think this process goes on diffusely in all arteries in the body, as evidenced by the livedo reticularis that we saw on the skin. And actually we recently, unfortunately, had the opportunity to do a complete autopsy on a young woman with one of these mutations who died of a stroke at the age of 30. Every single artery in her body had evidence of intimal hyperplasia with smooth muscle cells and no lipid deposition. So skip that. So in the last few minutes I'm just going to go through what we know about the Richard Holbrooke end of the spectrum, the people that don't have a family history. But what exactly was going on genetically in those individuals. And once again, this is the majority of patients with thoracic aneurysms and dissections. So about five years ago, we assembled a cohort of about 800 patients without a family history, who did not have Marfan's or one of these syndromes, and did some genetic studies on this cohort. So the average age was 65, which is what you expect for people with this sort of sporadic disease, non familial disease. In our cohort, about half the patients presented with acute dissections. The other half presented with aneurysms, and about 20% of our cohort had bicuspid aortic valve. So this is your typical thoracic, aortic disease cohort. And we genotyped them on a SNP array and did a case control association study. So we take patients with the disease and controls that don't have the disease. We do those SNP arrays where we're looking at common variance throughout the genome. And these are just marked along every single chromosome and spaced throughout the genome. And then we can do a case control association and see is there any one of these genetic variants that are there in the population that are more common in our patients than in the controls. And so what you do is you map all these thousands of SNPs and the p value for association with disease in this Manhattan plot. So down here are the SNPs on each individual chromosome, 1 to 22, in the x chromosome. There's no important genes on the y chromosome, by the way. And then this is your LOD score for association about SNP with disease. And this is called the Manhattan plot, and you're looking for skyscrapers-- big peaks, where the association is tight with the disease with a very low p value. And with this, we only had this one peak on chromosome 15. And we went in to fine map it. So here's all our SNPs associated with disease, and here's the LOD scores, between 10 to the minus 8 and 10 to the minus 12, so this is very tight association. And it all fell right on top of FBN1. So now we know that there is a common variant, or one or the more common variants in FBN1, that increases the risk in the general population for this disease. But this is only 1.6 to 2 fold increase risk for thoracic aortic disease. So it's not clinically relevant, but it begins to tell us that it is the same pathways that are causing disease in the syndromic and familiar patients as in the general population. We also did copy number variant analysis of this cohort, and that's where you're looking for big chunks of regions on those chromosomes that are either deleted or duplicated. So people either have-- everybody has two chromosome 1s, for instance. So if you have a big chunk missing, you may only have one copy of that chunk. And if you have a duplication on one of the chromosomes, then you have three copies. And what we found was that in the general population, there is duplications of a region on 16 p, and it's there in about 0.1% of the population. And we found this duplication increased the risk for aortic dissections over 12-fold. And when we mapped the genes in this region, MYH11, a gene when mutated causes familial disease and encodes that myosin heavy chain, fell right into the middle of this duplication. And so, once again, that's just telling us that the same genes that are increasing risk in the general population, we're hitting the same genes that we're finding for familial disease. So in the last few minutes, I just want to talk about how we're trying to look at low risk variants and then adding on environmental insults to see if we can start mimicking what's going on in a mouse model, what's going on in sporadic patients in the population. So MYH11, I just said, duplications where you've got three copies of that gene, increase your risk for disease. And if there's big deletions in parts of the gene, you end up with familial disease, with that autosomal dominant inheritance. But there's lots of rare variance in this gene that appear to be enriched in patients with sporadic disease. And so we wanted to take one of those variants where we think it would disrupt that myosin motor, but wouldn't cause a very strong risk in the mouse model. So we engineered into a mouse model, a missense, just a change in one base pair, that was actually in that myosin motor head, and we predict to decrease the movement of that myosin head. And this same mutation in MYH7 causes familial hypertrophic cardiomyopathy in the myosin motor that's the primary motor in cardiac cells. So we had a fairly good idea this would decrease the movement of that motor head. So we did studies first, just in vitro. We made the myosin motor and showed that this mutation disrupts the ability of that myosin motor to move. Then we engineered it into a mouse and took out the aortic segments, and could show that they didn't contract as well. So we had a model where we really were disrupting contraction. But, once again, this variant in humans does not cause familial disease. And it did not cause disease in the mouse model. Here's the aorta at a year in our wild type, in our mutant. And you can see that there's no evidence of aneurysms, the mice lived a full life, and there was no pathology that we could see. So then we took the mice and made them hypertensive by feeding them a nitric oxide inhibitor and a high salt diet. So with that, we can take the systolic blood pressure, which is typically about 120 in a mouse, way up to about 160 over time. And what we found was very acutely. Within the first two weeks of starting this, that in our mutant mice, there were about 20% to 30% died acutely with this treatment, within the first two weeks. Actually, when the blood pressure was just starting to build up. And when we looked at the cause of death, what we found in all these mice was pericardial tamponade, which is the major cause of death in patients with the type A dissection that don't even make it to the hospital. If we go into the morgue and look at people that are presenting to the morgue with the type A dissection 90% have pericardial tamponade. And we could show the blood in the pericardial sac-- they actually call this black heart because the blood surrounds the heart-- and when sectioned out the aorta, we could find the rupture in the aorta leading to this pericardial tamponade. In addition, in some of the mice that survived, we could see dissections of the descending aorta, shown by the blood in the aorta, here. And the tearing of the medial layer that we found on histopathology in this mouse. But these were not causing deaths, which is exactly what we see typically in humans, that type B dissections aren't as lethal as type A. So here we had a model where the mice were fine, until we made them hypertensive. So we're beginning to mimic what you see with sporadic disease. And we've actually used this mouse model to see if we could therapeutically manipulate dissections. So in black is the basal dissections of 20% of the mice that we see. And with this, we've been using drugs that can completely prevent these dissections. And then we can use other drugs and increase the rate of dissections, by manipulating molecular pathways in the aortic wall. So this nicely mimics that so-called sporadic disease with low risk variance and an environmental insult. And it's really given us an opportunity to have a model to manipulate therapeutically, in hopes of finding a drug that specifically addresses the dissection. Nobody dies of an aneurysm, they die of the dissection. So we really want to focus on what is triggering the dissection. So just in summary, the familial thoracic aortic disease, or heritable thoracic aortic disease, is due to mutations. And that tells us a lot about the molecular pathways. And we know that the gene not only helps inform when to do the surgical repair, but the risk for additional vascular diseases. And just finally I'd like to acknowledge everybody that I collaborate with, in particular the people here at Washington University. Alan's been great, a great collaborator over the past 15 years. And I'm working with Nate Stitziel and Bob Mecham on the lysyl oxidase mutations that we found recently. So thank you very much, and I'm happy to answer any questions. [APPLAUSE] SPEAKER 1: That was an excellent talk, Dianna. Any questions? I'll start with a question. DIANNA MILEWICZ: Yes. SPEAKER 1: I'm struck by how heterogeneous these disorders are and how the penetrants are so variable, for a family member in their 20s might be affected, and another family member with the same gene in the 60s or 80s don't have that, or have a cerebral aneurysm instead of a thoracic aneurysm. So what do we know about modifiers or other factors or other genes that are interplaying with the expression of these disorders? And then with that correlate, how do you follow a patient who has a family history like this? DIANNA MILEWICZ: So great questions. First of all, the modifiers can be other genes that the person has, be genetic modifiers. Or it could be environmental. So it could be that they have more forces on their ascending aorta. And so we're trying to look at that by collecting all the monozygotic and dizygotic twin pairs in our cohort. Because if they're monozygotic, they share 100% of the gene. So we can say, how often do they present about the same time with an aortic event, or are the aneurysms the same size. Now, keep in mind that those people share a lot of environmental risks, also, because they're being raised, most typically, in the same environment. And what we find in those monozygotic twins is like the presentation is amazingly similar. Like one twin will dissect one day, and the next one the next day. It's absolutely amazing. So we're still collecting all the data, going back and getting all the medical records. But this makes me think it's primarily genetic modifiers. And that the environment can play a role, but it's overwhelmingly the other genes that we have in our genome. So stay tuned. And that I think those other genes will also predict, like if you have that ACTA2, arginine 149C, whether you're going to present with early onset coronary artery disease or thoracic aortic disease. And it's kind of bizarre, because very few people have both disease in that family, they have one or the other. And so we're in the process of actually trying to get a large cohort, specifically with that mutation. Because we can actually map the modifiers. The modifiers should be common variants and maybe some contribution from rare variance. So, and then the second question was how do you follow somebody? Well, if you don't know the actual mutation, what we recommend is that everybody be imaged that's at risk, based on where they are in the family. And then, also, be imaged for any other vascular disease that's running in the family, such as intracranial aneurysms, and so on. And then the timing of surgery, if you can get history on people that have died of dissection, base the timing as to what information you can get on the size of the aorta at the time of dissection. But we still have families, like the one I talked about, that have no enlargement. And in those families, we actually sit down and say, you know your options are to go ahead and prophylactically repair your aorta, at least until we find the gene. And then, if there is no information on the family members who died of dissection, then just default to the regular management, which is the management for Marfan patients. [INAUDIBLE] SPEAKER 2: Have you looked at the ethnic distribution of the SNPs that are associated with [INAUDIBLE] DIANNA MILEWICZ: So that's a great question. These initial studies were actually limited to Caucasians. But we, more than likely, had individuals of Jewish Ashkenazi descent in our cohort. But-- and I think that's a great question. We know that thoracic aneurysms and dissections do occur in Israel, and it would be interesting to go back and look at that sub-population to see if this FBN1 peak is also there in that population. [INAUDIBLE] DIANNA MILEWICZ: Sudden death of unknown causes, I don't-- I mean obviously, you're going to screen them for cardiovascular risk factors and, in particular, hypercholesterolemia. But I think that what I would recommend is actually send the family back to investigate further for death certificates and so on, to see if it's actually something other than coronary artery disease. The vast majority will still-- if they are dying suddenly with chest pain-- it's still going to be coronary artery disease. Now if it turns out to be thoracic aortic dissections, then you can do genetic testing for these genes. And all the genes that we've identified are on diagnostic panels worldwide and being used to determine who's at risk in the management of the disease. But you really need to be testing somebody you know is affected. So you want to image everybody and find somebody with an enlarged aorta to do the testing. [INAUDIBLE] DIANNA MILEWICZ: No, I've heard this speculation before after presenting these data, that perhaps there is a small vessel arteriopathy in those patients. And if you look at-- after we found this, we actually assembled a cohort of Moyamoya disease patients. Almost uniformly all of those patients have livedo reticularis. So even with your typical Moyamoya disease patient that's presenting with a stroke, they will have suggestions that there's occlusive arteriopathy going on. So I think that's a great question. [INAUDIBLE] DIANNA MILEWICZ: So our data on this mutant mouse suggests that some acute dissections are triggered with rapid increase of the blood pressure and pathways that get altered acutely. And we know that hypertension's a major risk factor, so we know that controlling that hypertension will hopefully decrease aneurysm formation and the risk of dissection. And that's why we tell our patients not to do isometric exercises. Because when you go to lift a heavy weight, your blood pressure-- the very heavy weights-- your blood pressure can shoot up to 300 over 150 when all your muscles are contracting, and your heart still needs to perfuse the muscle, despite that contraction. But at the same time, many of the genes that predispose to disease, the majority don't cause-- there's no associated hypertension. And in fact the ACTA2 patients with the most severe mutations and aortic disease starting in their teens actually are hypotensive. The ran about an 80 over 50 as an average blood pressure. So there-- I mean, I think that there are genes that are causing it-- that they will lead to disease without the hypertension, without the environmental risk. Just like if you have cystic fibrosis, you're going to get the disease no matter what environment you put yourself in. That's true with most of these mutant genes. But then the sporadic disease, I think there's much more of a role for protecting the aorta by modifying the blood pressure. SPEAKER 1: Thank you very much for a wonderful presentation. DIANNA MILEWICZ: Thank you. [APPLAUSE]