Name: Jenny Van Eyk
Position: director, NHLBI Proteomics Group, The Johns Hopkins University School of Medicine, 2003 to present; Hopkins Bayview Proteomic Center, The Johns Hopkins University School of Medicine, 2003 to present; associate professor, The Johns Hopkins University School of Medicine, Department of Medicine, 2003 to present; associate adjunct professor, Queens University, Kingston, Ontario, department of physiology, 2003 to present; scientific advisory board member, Plasma Proteome Institute, 2004 to present; scientific advisory board member, Ortho Diagnostics, 2005 to present.
Background: PhD, University of Alberta, Canada, 1991; postdoctoral fellow, University of Heidelberg, 1992; research associate, University of Alberta, 1995; postdoctoral fellow, University of Illinois, Chicago, 1996.
Jenny Van Eyk and her colleagues at Johns Hopkins University are doing proteomic research on ischemic heart disease. Recently, Alfa Wasserman Proteomic Technologies announced a collaboration with the university to use the AW Promatix 1000 ultracentrifugation system to develop methodologies and protocols for isolating and enriching organelles and sub-cellular particles.
ProteoMonitor recently spoke with Van Eyk about her work.
Tell me about your research.
Heart [work] is probably three-quarters of what we work on. Or [the] cardiovascular system. We should probably include the pulmonary system and vasculature, too. Essentially one-third of my lab works on biomarker development. That’s both working on discovery and validation. We have large projects both in industry and academia. On the other two-thirds, we work on cells and tissue and we’ve taken them apart. We’ll look at the myofilament proteins, we’ll look at the cytosol, we’ll look at the mitochondria, we’ll look at the inner mitochondria, we do a lot of sub-proteomics.
And the strategy that we use as far as the proteomics goes is we do whole protein separation. So we run an awful lot of 2D gels, obviously. And we also run a lot of 1D LC on intact proteins and 2D LC. We have found that by combining the three different protein separation methods, we have very little overlap between them. We have usually less than 10 percent overlap.
So by combining the three technologies, we can pretty much reach the same level as what people get with shotgun [proteomics], but with it, we get all the intrinsic properties, so we do a lot of post-translational modification analysis. If we’re talking about the LC, these fractions contain upward of 100 to 200 proteins. So we can do iTRAQ on that, or we can do shotgun on whatever you want to see on the identification and quantitation. We do the same thing on the gels. We ID every spot, so we know when there’s a modification and then we do things to figure out what those modifications are.
Essentially on the tissue side, we take apart the cell and we use whole intact protein separation using multiple methods to go after post-translational modification. On the development side, we also use intact protein separation, but although the technology is similar, the strategy’s quite different because we do one to three different human cohorts for a discovery. So if we’re going after myocardial ischemia, we’ll have in human populations three different ways of making a heart ischemic and we’ll have temporal profiles of that, and we’ll do the discovery of that.
And then the validation cohorts are independent, they’re a separate cohort altogether. And we usually have two separate cohorts, one we make up, which has a lot of the controls, people with pulmonary disease, people with chronic obstructive pulmonary disease, people with a whole range of heart failure, acute ischemia, acute myocardial infarction and cancer just to see a broad view of general sickness and overlapping disease markers. And then we have a very disease-oriented cohort, too. So we would have people coming into [the emergency room] for example.
What have you found out so far?
Probably one of the things we’re most noted for on the biomarker side is the understanding of the troponin I and troponin T, which are the gold standards now for the detection of acute myocardial infarction. The fact that this is actually a very complicated analyte and these proteins get specifically modified with increasing ischemia, and so you can actually do risk stratification and stuff.
We just had a paper out in Circulation Research on this condition called preconditioning which is a protective mechanism induced in the heart prior to making it have a big ischemia. So if you can protect the heart using these pharmacological agents or a little bit of ischemia, the heart puts up this big defense so when you hit it with a very large ischemia, let’s say [something that would] kill 50 percent of the cell, it would kill only 20 percent. It’s very, very robust. It can be done only experimentally right now. We haven’t figured out enough to be able to [do it in a clinical environment].
We just had a paper looking at two pharmacological agents that use very different [mechanisms], or [what] people thought [were] very different mechanisms, to induce this preconditioning pharmacologically. One activates through PKC kinase, and the other one has a direct effect on this channel in the mitochondria, and what we show and said was, ‘Well, the end effect of this preconditioning should be one of two proteins and they should overlap because they have the same phenotype.’ But, of course, what happens is we found out that they’re not necessarily the same protein but everybody is going directly to the mitochondria.
Along the way we found in both that paper and the paper that we just published on the inner mitochondria, subproteome, that an awful lot of the mitochondrial proteins are actually modified. These are even proteins in the inner mitochondria which is quite remarkable. Within 60 minutes of the drug treatment, we find some of the big proteins involved in ATP production being phosphorylated. These proteins are located away from the cytosol, so they’re located in the inner part of the mitochondria away from any signal transduction that we know of, which means somehow that when we stimulate, let’s say PKC pathway, somehow within 60 minutes, that signal comes to the mitochondria through the outer mitochondria, through a signal pathway that’s inside the mitochondria that’s not known and phosphorylates this protein in the inner mitochondrial membrane.
We also know that many of the subunits within the mitochondria and inner mitochondria are actually modified at the signal transduction level. And this is really just starting to hit the news. There are a few papers coming out now that are indicating that there could be a mitochondria-unique cascade. That’s pretty exciting and opens up a whole new way of thinking about the mitochondria as a signaling response, and also that there could be a whole new bunch of kinases and phosphatases no one ever knew about located in mitochondria. It’s kind of cool. It’s a whole different regulation process people never thought about.
It sounds like you’re looking for panels of biomarkers rather than one biomarker.
It depends on what you mean by panel. If you mean by panels, having different sets of markers that are able to correlate to different aspects of the disease, yes. If you’re talking about having to have multiple markers to have a single diagnosis, then no. We may end up with multiple proteins, but it’s always better and more simple when you actually want to make a diagnostic marker to have a single protein — just like the troponins that are perfect for a single diagnosis.
However, there are going to be cases where we may not be that lucky to find those. That we may need to maybe combine two or three really good markers to get that really special diagnosis. But what we want to do is to make a panel that’s actually able to reflect the whole cardiovascular system in a single thing.
So when you have heart failure, you just don’t want to care about your heart. You need to also know about the other systems that influence it. So the panels that we’re going after are actually of a panel that allows you to dissect different clinical questions about that one disease that gives you, then, more knowledge. So it’s a panel because it’s multiple markers, but each kind of subset of question that you’re answering within that disease context could be one or more markers depending on what you do.
A perfect example is stuff we already have in cardiac muscle. You can use a troponin I as a detection of whether or not your heart’s dying. You can use BNP [brain natriuretic peptide], a hormone that’s very good for looking at hemo-dynamic stress, and you can use something like CRP [c-reactive protein] which is to look at your immune response. That’s a panel because it’s three different markers, but it’s looking at three different parts of the disease. One’s the heart itself, one’s the hemo-dynamics, and one is your inflammation response. That’s the type of panel we want, ones that actually give you information about different parts of the disease as opposed to just a panel to try to get one that would look at the heart.
Are you looking for prognostic biomarkers or diagnostic markers?
Both. For example, we have a large project in ischemia. Ischemia is when you have very little blood, or no blood flow, to your heart. If it’s long enough, that’s when you’re going to have a heart attack. But there’s a period of time before your cells actually start to die where it’s ischemic. So when you come into the emergency [room] with chest pains, for example, they’ll look for this cardiac specific protein. If you don’t have it, it doesn’t mean you’re not going to have a heart attack in the next six hours. It just means your heart may be ischemic.
One of the holy grails in the acute setting is to find the ischemic markers that come up prior to the troponin, so that you can diagnose people coming in [and have] earlier intervention.
What would you be looking for under those circumstances? What kind of proteins would you expect to find?
They’re either going to be one of the following classes: Either it’s going to be a response that’s a whole body response. Your heart’s not pumping as well because it’s ischemic. So your kidney’s going to react, your cardiovascular system’s going to react. So it could be coming from anywhere.
Two, it could be something that’s shredded from the outside of the cardiac myocyte or secreted in response to ischemia. So it’s either going to be secreted protein, a protein that’s shaved off the cardiac muscle, or something in response via the whole cardiovascular system.
It’s not going to be trivial to get this. That’s why we have three different human cohorts, and animal cohorts, and a very large validation set, because we’re going to have to wean it away from things like: Why would it be different when you go out running and you have your skeletal muscle undergoing ischemia because it starts to hurt? You have lactic acid build up. That’s an ischemic event, too.
How do we differentiate that? How do we differentiate it from ischemia that happens to your kidneys? That specificity means that you’re going to need to know an awful lot, and have a lot of ways of teasing out the ones that are real from the ones that are just your body responding to ischemia regardless of what organ [may be causing it].
In that way, that’s a diagnostic, but in heart failure, we have cohorts where we’re trying to do prognosis. Right now, the other thing that we have is 10 years to 20 years before you have a heart attack, you start to build up atherosclerosis. And we have the Framingham risk scores which are just lifestyles. For example, do you smoke, do you have a family history [of heart disease]? Then you have odds, ratios of being more likely to have a heart attack within the next 20 years.
We’re trying to move that up to a one-year serum marker so that we could predict someone having a one-year event rate, and we’re trying to do it at the one-month [event rate], so that if you’re going in yearly for your check-up, we can monitor you and find out whether or not you’re getting worse or getting better. Those are the kind of prognostic indicators that we’re going after.
What we want to be able to say is that we believe that as you have more carotid artery disease, as you start to go toward having a heart attack, that your body changes, your heart changes as it tries to adapt.
Your heart has to do one thing in this life, this is the only thing that it will do — to maintain the same amount of blood all the time. It has to continuously have the same stroke volume, so it always has to pump out the same amount of blood. It’s its one job. So it will do anything to make sure it does that. And so it changes along the course of disease. The heart changes, the vasculature changes over time to make sure you can pump blood to your body, and all your body gets it. So as you start to build up carotid artery disease, and your coronary arteries start to be not as good, your body has to adjust, and it will.
And that’s why heart disease, by the time it’s actually diagnosed, because of shortness of breath, you’ve had the disease for many, many, many years because it’s a long process. It starts in your 20s and as your plaque grows in your arteries, it goes from being kind of stable to this thing which is ready to explode and cause a clot somewhere.
If heart disease begins in your 20s, what would you be looking for at that point?
They would look for carotid artery disease, so they would be looking for plaque build-up and stuff like that. So it’s all imaging. It’s not a blood test.
And are you trying to develop a blood test?
Do you know what you’ll be looking for?
We assume that it’s going to be, to get specificity, that it will be a disease that will come uniquely from the plaque for that particular one.
So you’re looking for proteins?