 Good morning, everyone. I am happy to introduce my MD-PhD student, Judd Kahoon. Before I do that, I'd like to share one observation, or a couple of observations, on the track of research productivity in the United States. So when you look at the last 40 years, NIH funding in billions of dollars has gone from about 1 billion to 30 billion dollars from 1970s. 2010. In that same period of time, the number of drug approvals per research dollar has plummeted. Now that's multifactorial, and all of us are familiar with the various issues with the FDA, as well as, of course, increasing complexity of research. But I think a key element in why this has occurred is is not just failures on the part of Big Pharma or the FDA, but also the decreasing proportion of research that's done by physicians. In the last 40 years, the proportion of NIH grants that is pursued by MD researchers has declined from about half, or a slight majority, to less than a quarter. And that itself is multifactorial, but what that leads to, I think, is a big problem in translational research that leads to that value of death that we've all talked about. Right, absolutely. Indeed. So I think we as a profession have failed in encouraging our young people to pursue research-oriented careers and with that, I'd like to introduce someone who I think will be a rising star in the field of ophthalmology and visual science. Jud Kuhun is a fourth year student now, so he's finished his first two years of medical school and is now in his second year in his lab, and he is an investigator who shows a passion for discovery and diligence and dedication. It's a very unique set of traits that I deeply admire him for, and I'm happy that he's able to share some of his really exciting work in diabetic retinopathy, which will make an impact for years to come. Thank you for that introduction. Thanks for the opportunity to speak today. I'd like to pull up the presentation here. Just a sec. Well, we get this started. Hopefully this presentation I'm about to give is not meant to be distracting or motion sickness inducing, but it is supposed to give you a broad overview of the picture, so that we're always going to come back and try to see the big picture of what we're about to present here. And with that, it looks like we're ready to go. So again, thank you for the opportunity to come here and speak, and I'm very excited to present a little bit of our research, which the focus of is stabilizing the vasculature in diabetic retinopathy and measuring the outcomes of retinal function that come from that. So for a broad overview of the talk, we'll give a little bit of background into diabetic retinopathy, of which I'm sure most of you are very familiar. We'll talk about the model that we use in the lab to mimic diabetic retinopathy, kind of the positive and negative side of that. We'll explain briefly about the experimental design and how we're going to intervene here, as well as two key results. The two key results we'll be looking at the vascular structure and function, as well as the retinal structure and function. And then at the end we'll talk a bit about the future directions. So we can get going with that. As an introduction, diabetic retinopathy is the leading cause of blindness, worldwide, in the working age population. This includes about 93 million people worldwide who have some form of diabetic retinopathy. The direct medical costs in treating diabetic retinopathy are half a billion dollars a year in the U.S. alone. Now if you factor in indirect medical costs for this, it's 35 times that amount because due to the young age at diagnosis. This is happening in the working age population. 35% of the people with diabetes in their lifetime will experience some form of diabetic retinopathy. And those numbers, the numbers of people with diabetes are expected to triple by 2030. So this will be of increasing importance as we get more well-fed in our society. Two of the key issues I'd like to address in diabetic retinopathy are the hyperpermeability and the ischemia, the vascular leakage, and then the poor perfusion that leads to ischemic retinal tissue that occurs in diabetic retinopathy. In the proliferative form of diabetic retinopathy, there's an ischemia-induced angiogenesis, blood vessels start to grow, and those blood vessels are not stable, they're not mature, so they leak. A non-proliferative form of diabetic retinopathy leads to edema, specifically macular edema. So just to quickly refresh everybody's mind, we've got some fundus picks here of a nice normal retina with the vessels coming out of the optic disc surrounding the macular there. In the early stages of diabetic retinopathy, the hallmark that occurs is hyperpermeability, and here we can see some exudates that have leaked from the vasculature into the retina, lipid deposits and some macular edema indicating that there is a leaky or a hyperpermeable state in the vasculature. Should the hyperpermeability and improper perfusion persist for a while, this can lead to neovascularization with fragile vessels that can easily hemorrhage. I'm sure you recognize these spots of laser photocoagulation in the background there, which is currently a standard of care for diabetic retinopathy. Should those vessels start to grow out and even attach to the vitreous, they can pull the retina away from the RPE leading to retinal traction, which would definitely be bad for vision. Another way to assess retinal hyperpermeability is with fluorescein angiography. Here you can demonstrate the leakiness of the fluorescein from the vessels, as well as image some macular ischemia. So if I've said anything here, it's hopefully that diabetic retinopathy to the elements are hyperpermeability and ischemia. Great. So there's the background on it. Let's look at the model that we're using in the lab to mimic this. We're using the Akita mouse model. Now the Akita mouse has a mutation in the insulin 2 gene. It's a heterozygous for this because the homozygous are embryonic lethal. But that mutation in the insulin gene causes insulin to build up inside the pancreatic beta cells, eventually causing beta cell dysfunction. The mouse can't secrete insulin, resulting in hyperglycemia. And then specific to the retina, some of the things that are seen in this mouse are capillary and pericyte loss, just as they are in humans. Increased retinal permeability, and that permeability is leading to the macular edema in patients, as well as ganglion cell loss. To look a little closer here, what this is, is a trypsin digest. We've taken the retina either from a patient or from a mouse and digested away all the neural tissue, leaving only the vascular tissue behind. Here in our non-diabetic form, you can see these black dots that appear as nuclei just outside of the vessel. Those are the supporting pericytes, which surround the endothelium, offering trophic signals as well as structural support. In diabetes, you lose pericytes, which in turn you start to lose capillaries. With the capillary dropout, you're no longer perfusing the tissue surrounding that capillary, and that can lead to ischemia. The same thing is found in mice here, where you have a capillary dropout and pericyte loss, showing that this is a model of early form or early stage diabetic retinopathy. Here's our model. This is a cross-section through a capillary. We have the lumen of the blood vessel in here. We've got two endothelial cells connected on the outside, and then a supporting pericyte, that perivascular support cell. The perivascular support cell, the pericyte, secretes angiopoietin 1. Angiopoietin is a poetic growth factor, a vascular maturing growth factor, much like thrombo-poietin is for thrombocytes, or erythropoietin is for erythrocytes. Angiopoietin works on the blood vessels. Angiopoietin 1 acts on the T2 receptor. The T2 receptor does a number of jobs on the endothelial cell, but one of the things that it does is to stimulate VECAD here in stability in the membrane. VECAD here in is that structural support in the adherence junction that prevents leakage and permeability in between the endothelial cells. So really only two things I want you to remember from this slide. Number one, angiopoietin signals to support to promote VECAD here in stabilization. Now in diabetes, the first thing that is noticed is you have pericyte dropout. It's hypothesized that the loss of pericytes due to hyperglycemia leads to a decrease in angiopoietin 1 signaling. Without the angiopoietin 1 signaling, the VECAD here in that stabilizing adherence junction protein is internalized. That leads to hyperpermeability and macular edema. After a while, without any pericytes, you start to lose the endothelial cells and you get the capillary dropout. There, you're going to have poor perfusion, which will lead to ischemia, increased vascular endothelial growth factor, and after a long duration, certain patients will develop retinal neovascularization. There's the general microvascular pathology that occurs in diabetes. So what's our hypothesis? Our hypothesis is that endothelial stabilization with compang1, a modified form of that angiopoietin 1, which I'll describe in a minute, will mitigate the effects of diabetes in a mirroring model. To illustrate this point, we see the same graphs or the same figures as we do before. Loss of angiopoietin 1 can lead to VECAD here in internalization. Loss of endothelial barrier integrity, macular edema, or leakage. And what we hope to do is, even in spite of pericyte loss, or try to supply angiopoietin 1, which the pericytes normally supply, to do that, we're providing long-term expression of compang1, which will activate the T2 receptor in spite of the hyperglycemia, promote VECAD here in stabilization, and reduce the ischemia in the retina. That's our hypothesis of what we can do. So let's look at the design of the experiment to show how we are inducing compang1 expression over the long-term. First off, what is compang1? Goeang Ko's group at the Korean Advanced Institute of Science and Technology developed a novel form of angiopoietin 1. Here we see full-length ang1, or schematic here, of native ang1. And a problem with native ang1 is that it has N-terminal superclustering, which brings angiopoietin 1 together, but it quickly falls out of solution, and it's not very soluble. That's a problem for using it as a therapeutic. So they derived cartilage, oligo, matrix protein ang1. All I want you to remember about that is it increases its solubility, makes it more potent, and more stable. So we're dealing with an ang1 variant that still signals the T2 receptor, but it can do so with increasing concentrations. To deliver compang1 into the mouse retina, we utilized an adeno-associated virus, serotype 2. Now these AAVs, or adeno-associated viruses, are used currently in clinical trials for retinitis pigmentoso or AMD. So they are a viable viral vector that can induce long-term expression of your protein of interest. Our basic setup was to take a mouse, either a normal mouse, a non-diabetic mouse, in our case, the C57 black mouse, which is the background for these diabetic akita mice, which I described. At two months, we'd give them one intravitrious injection. An intravitrious injection would contain a control buffer, PBS, a viral vector encoding a green fluorescent protein as control, or the viral vector encoding our protein of interest, compang1. Over the next four months, we would take in vivo retinal assessments of these mice. Some of those in vivo retinal assessments include visual acuity tests, ERGs, optical coherence tomographies, similar things to what we do in patients. At our endpoint of six months, we would assess the permeability of the vasculature in the retina, take cross sections to look for morphology of it, assay protein RNA content, and try to get some mechanistic data out of that. So the first question you have to ask is, does this work? Can we induce long-term expression of a protein of interest in the mouse retina? If you remember, our control vector was a viral vector encoding GFP. We get GFP expression from a single intravitrious injection of AAV2 encoding GFP, which starts at about one week post-injection. This is a view from the spectralis looking at the mouse retina, and we can see that there is GFP expression in all quadrants of the retina that increases two weeks and persists through six months. An ex vivo flat mount, the retina was taken out and laid flat on a glass slide and also analyzed for GFP content, showing that that one intravitrious injection had viral expression over all the retina, and mainly the virus is in the ganglion cell layer. Here we have a cross section from the coroid, outer segment layer, outer nuclear layer, inner nuclear layer, and the ganglion cell layer. And mainly the virus and its protein expression were limited to that ganglion cell layer. So that's all well and good for GFP. We also confirmed that CompEng1 was expressed in the retina. Here we have RT-PCR just checking for mRNA to make sure that the transcript is being produced, and only in our diabetic mouse treated with our viral vector expressing CompEng1 did we see CompEng1, which was good confirming that we get RNA expression, as well as protein expression. Using an IP and a western blot, we were able to pull out CompEng1 from the mouse retina, and only the mouse retina treated with our protein of interest. Okay, so there's the background that we were able to get expression in the mouse of what we were hoping to get. Now let's see if it had any effects. We'll look at the vascular effects before we look at the retinal effects. So we'll look both for structure and function of the vasculature. Taking out the retina at the end of our six-month study and doing a stain for retinal vasculature, either with isolectin, which is shown in green, which stains in the thylial cells, or alpha-SMA, which stains smooth muscle, which is found in the smooth muscle cells surrounding the endothelium, as well as the pericytes. We'll blow this up a little bit more so you can see what's going on. You can see a nice, beautiful retinal vascular architecture here in our control non-diabetic mouse. We can see two arteries coming out and a big vein leading back in there. What happens in our diabetic mice is that their retinal vasculature starts to disappear. Paricyte loss, endothelial loss, you can see this on a global scale, as we have decreased staining indicating decreased endothelial production as well as decreased pericytes. In both of our control, PBS and GFP treated mice, there was a decrease in the retinal vasculature that was prevented in our Comp Ang1 treated mice. We can appreciate the retinal vasculature of this mouse, even though it is diabetic, looks similar to our non-diabetic mouse. Just zooming in a little bit farther and checking this out, you can see some red staining here for pericytes and smooth muscle cells. A loss of pericyte staining, an example of that here in our diabetic mouse. Loss of capillaries, as we can see there's not a lot of connections going on here, and a preservation of some of the capillary bed in our Comp Ang1 treated diabetic mice. This is just one example, and we quantified this with image J analysis to look for total endothelial staining over the entire vascular area. Here our non-diabetic mouse showing about a 24-23% coverage of the endothelium, just our estimate of that, is decreased in our two cases of control treated diabetic mice, but is preserved in our Comp Ang1 treated diabetic mice. We're preserving the vascular endothelial area. However, this is a pericyte coverage representation here. The treatment with Comp Ang1 did not preserve the pericyte coverage in our diabetic mice, which is interesting. We were still able to maintain endothelial coverage despite persistent pericyte loss. This is interesting because most patients who present with diabetic retinopathy have already undergone the early stages, have already experienced some sort of pericyte loss, as is the case even with our treated mice. Let's look at some functional data to see if that change in morphology actually has any ramifications in the function of the vasculature in the retina. We'll look at an in vitro study using human, microvascular endothelial cells in a cell dish, and an in vivo study. First, how do you assay endothelial function outside the mouse? We're using something called transepithelial resistance, which is just a way to measure the amount of barrier integrity or resistance that a monolayer of cells will offer to a small current. A gold electrode is placed at the bottom of the cell culture dish. A small current is passed through that, not enough to disturb the cells, and cells are plated on top of that electrode. As the cells increase their barrier integrity and increase those adherence junctions, they offer increased resistance to the current passing through, and we can measure that. That'll look something like this. Without any cells plated, you have very low impedance and also very low resistance. As the cells are plated and they grow and form a monolayer, there's an increase in the resistance, and then we can add an agonist or something to increase barrier integrity, and we should see a jump in that. And that is exactly what we saw here. We have four groups of treated cells. These are human retinal microvascular endothelial cells, and here we plate them, watch them grow to confluence, and once they reach their plateau point, we administer our treatments. In this case, we administered Compange 1, hoping to see an increase in resistance. PBS is our control to see the normal course of these endothelial cells, or VEGF, to increase vascular permeability and decrease resistance. Sure enough, in our in vitro assay of barrier function, Compange 1 increased the resistance capacity, increased the transepithelial resistance of our endothelial cells compared to the normal course of PBS, and VEGF decreased the resistance. Interestingly enough, in this experiment, Compange 1 did not rescue VEGF-induced permeability increases. And we'll get back to why that's important as we look at what happens inside the mouse. So there's an in vitro assay confirming that Compange 1 has functional capacity, that it is activating barrier integrity in the endothelial cells. And next we'd like to look at an in vivo model of this, and this is done with the Evans blue dye technique. You take a blue dye injected in the tail vein, it binds to albumin, and wherever the vasculature is more permeable, the blue dye will leak. Here you can see nice blue paws, blue nose, blue ear, but what you shouldn't see are blue retina or blue brain or places where the blood retinal barrier or blood brain barrier should be nice and intact. You shouldn't see a lot of blue in the retina. After an injection of the Evans blue dye into the tail vein and your mouse gets nice and blue, you simply harvest the retina, leach the dye from the retina, and measure the amount of dye that was present. Here's our non-diabetic mouse. We can see an over three-fold increase in our diabetic mice in terms of blue dye leakage in the retina, indicating retinobasal permeability. Comp ang1 prevented that increase in retinobasal permeability back to control levels, which was pretty cool because it's showing that not only do we have structural changes, we have actual functional changes in the vasculature as well. What could be mediating these things? As you remember, I asked you to remember just about angiopoietin 1 and Vecadherin. Vecadherin is that adherence junction protein that stabilizes endothelial cells, and in vitro here, using human retinal, microvascular endothelial cells in a culture dish, we can see that Comp Ang1 treatment increases Vecadherin stability protein-wise. This also happens in vivo. In our mouse model, we have our control non-diabetic mouse here versus our three-diabetic mice and our AAV2, so our viral vector expressing Comp Ang1, increases Vecadherin stability in the mouse. This could be the way it's mediating the increase in barrier function. Now interestingly enough, we looked for VEGF levels, whole VEGF levels in the retina, and found that in our C57 mice, compared to our diabetic mice, we had an increase in VEGF levels, and that increase in VEGF levels was prevented with Comp Ang1 treatment. This represents another mechanism by which Comp Ang1 can be preventing basal permeability and stabilizing the vasculature. Not only is it stabilizing Vecadherin, that integral membrane adherence junction protein, but it's also decreasing retinal secretion of VEGF. With decreased VEGF secretion, there's another factor for vascular stability, decreasing the permeability-inducing effects of VEGF. And that kind of ties in with our ESIS results from that barrier assay that we saw here, wherein Comp Ang1 didn't produce VEGF, didn't prevent VEGF-induced permeability. Okay, that's all well and good that we've shown we can change the vasculature. But if you can change the vasculature structure and function without affecting the retina, it doesn't really matter. So we looked at the same points here for the retina. We'll look at retinal structure and then retinal function. Looking at retinal structure, we can use optical coherence tomography or OCT on our spectralis, and we can pinpoint areas certain distances away from the optic nerve and measure the retinal thickness, or the retinal nerve fiber layer thickness, as demonstrated here, in a nice concentric circle. So what this is, this is a cylinder that's been laid out, and we're just looking at the cross-section of the retina all throughout here. As we look, we can see that Comp Ang1 prevents the retinal thinning that occurs in diabetic mice. Now, here's where mice and men differ in one aspect, is that diabetic macular edema can lead to retinal swelling in patients, but in mice, there is no macula, and they do get retinal thinning. Well, this retinal thinning is not unique to mice that patients can have ganglion cell layer loss due to ischemia and diabetic retinopathy. So the ganglion cell layer loss that is present in patients is also represented here in our diabetic mice. So overall retina, we have a thinning of the retina in our two diabetic control-treated mice, which is prevented with Comp Ang1. The difference wasn't great, it was statistically significant. To focus in a little bit more and see which areas of the retina were changing specifically to look at the ganglion cell layer to see if that was changing as well, we took cross-sections and stained them. Dappy, a nuclei stain is in blue, VE cat here and is in red. Here's our C57 mouse retina, there's the photoreceptor layer, the bipolar cell layer and the ganglion cell layer. And you can see a nice thick ganglion cell layer here that disappears in this diabetic mouse, treated with PBS, and this diabetic mouse treated with control viral vector, but is preserved better in this Comp Ang1-treated mouse despite his diabetes. If you look a little closer at the actual ganglion cell layers, we can see ganglion cell layers from a non-diabetic mouse, lots of nuclei in the ganglion cell layer, decreased nuclei in the ganglion cell layer in our diabetic mice, and preservation of those nuclei in the ganglion cell layer in our Comp Ang1-treated mice. We quantified those results and did show statistical significance between our control non-diabetic mouse, our two control-treated diabetic mice, and preservation of the ganglion cell nuclei in our Comp Ang1-treated mice. So we are preserving some of those ganglion cells, as well as the thickness associated with those ganglion cells. We're currently doing estimates now on the retinal nerve fiber layer to determine if that thickness is responsible for the preservation of whole retinal thickness. So the take-home points from that are we were able to alter the retinal structure. Now, how does that relate to function? We did two assays of retinal function, one being electro-retinography, or the ERG, and mice, just like humans, were given a full-flash ERG at increasing intensities, and here is a representative ERG trace at a lower-intensity light flash. You can see our Comp Ang1 or our control C57 mice. Those are the black layer up here. They have a nice B-wave amplitude, and the measure we used to quantify each ERG was the B-wave amplitude. But our two control-treated diabetic mice have a lower amplitude here. You can see those waves are a little bit different than our control wave, and the Comp Ang1 prevented that decrease and looks closer to the control mouse. Quantification of this, looking at B-wave amplitude, shows you have kind of two groups of mice. You have our diabetic mice down here with lower amplitudes, either treated with PBS or GFP, our control diabetic mice. In black you have our C57, our normal non-diabetic mouse, and our diabetic mouse treated with Comp Ang1, showing amplitudes and ERG waves similar to that of our control mouse. So retinal electrical function is intact. Let's look at visual acuity in a mouse. You can't just hold up a Snellen chart and ask them which one is better, so you test their kinetic tracking response. And the way that you do this, as was recently described by Pearson in a nature paper that just came out, is to place a mouse inside of a box and have these screens that go all the way around. These screens are going to show bars of black and white of increasing frequency and thinness there. What the mouse will do is it passes in front of it, it will track it automatically, the optomotor response. By increasing the frequency of these bars, we can determine a visual spatial frequency threshold or the threshold at which the mouse stops responding that correlates to the mouse visual acuity. Just to give you an idea of what we're actually looking at, I have a little video here of optometry measuring visual acuity in a mouse. Here's a C57 mouse, and these red bars are just superimposed, and you can see its head start to track where those red bars went. Here's an Akita mouse, not offering much in any sort of tracking response. You can see the bars spinning around it look like this, but this mouse doesn't seem to notice or maybe care. Here's another mouse. This is a diabetic mouse treated with PBS offering poor tracking responses once again, maybe a little bit there as it follows, and here's our Compange 1 treated mouse. You can see that nose following and tracking along. So that's when the person doing the experiment would say, yes, he's tracking, versus no, he's not. Now after many, many hours of doing this, we were able to show that the frequency threshold of the optomotor response is about .35 cycles per degree in our non-diabetic control mouse. This spatial frequency threshold or visual acuity analog decreases in our two diabetic mice. Compange 1 prevents that decrease. So Compange 1 is not only preserving the retinal structure, but it's preserving retinal function as assayed by ERG and our measure of visual acuity in the optometry. All right. We've reached the end of some of the data and I'd just like to talk a tiny bit about the future directions where we'd like to go. We would like to determine whether Compange 1 promotes better blood flow in diabetic versus control mice and how do you assay blood flow? Well, the retina is awesome because you get to see vasculature and neuronal cells in vivo. So one way we're going to take advantage of that is through something called Lucography. And if I can, I'll just show you a couple quick videos. Okay. So what we've done here is we've injected the tail vein with a dye that selectively stains the nuclei, but is permeable to white blood cells. And you can see white blood cell transit through the capillaries and it jumps back in here. You can see it traveling down the vein here. We're able to follow those white blood cells as they pass through their capillaries, measure their transit time and then also observe any leukocyte adhesion or rolling leukocytes that go around here. In diabetes, there's an increase in rolling leukocytes and leukocyte adhesion. So this will offer us an avenue to see if our treatments are preventing that at all. And then just one one more. If we come here, you can actually see a leukocyte start to roll down. There he is right there. And he's adhering to the vein as he starts slowly traversing down. We're working on ways to quantify this to show us this is from a diabetic mouse so that we can be able to quantify retinal blood flow that occurs in diabetes. Another mouse we're currently working with is we all know type 1 diabetes only accounts for a small fraction 5% of all diabetes in the patient population which is what the akita mouse is an example of the type 1 diabetic mouse. We're using another mouse. This mouse has a gene mutation in its satiety receptor. It doesn't know when it's full. The leptin receptor knockout mouse and it keeps eating and keeps eating not unlike many people. And it can be up to 3 times the size of a normal mouse. It exhibits the full suite of the metabolic syndrome. Hyperglycemia, hyperinsulinemia as well as we've started to observe some of the same retinal, microvascular and functional defects. Finally we'd also like to look at actual hypoxia levels in the retina measuring the oxygen tension and saturation of the retina. We looked at VEGF levels which are a good marker of oxygen tension in the retina as well as early studies looking at HIF1alpha, one of the transcription factors that is increased in the diabetic retina. We're also looking for other measures of that. I'd like to acknowledge the people who have made this work possible namely from the Crizzai lab Peter's been a huge help showing me the optometry as well as the ERG and some of those functional studies of mouse vision. From Dean Lee's lab Chris Gibson showed us how to do some of the ESIS or those in vitro barrier function assays as well as all the people in the neuroscience program the T32 training grant that supported me during the year as well as my committee members who have been very patient and understanding and helping me work through this phase of my training. Most importantly I'd like to thank the body lab specifically HERO for helping get this project started for being there for any scientific question that anyone might have and knowing the exact right answer and the exact thing to do to get the project moving. He's been a very big help in that. Tad Mia, excellent lab tech and future medical student who's done imaging helping out the Lucography specifically Paul Olson and Vi who've done amazing work on helping with the optometry and Courtney Walker an exceptional undergrad who's helped a lot with the in vitro stuff. So thank you very much we invite any sort of input you might have on modulation of VEGF or other mechanisms pathways to explore and I'll take any questions. And the great thing about this AAV is it is a long term expression so far in our model in an introvitrious injection we've gone out to ten months and still seen the same GFP fluorescence protein expression without any drop off. It looks pretty good. In some clinical trials they've used the same vector for Parkinson's trials and to administer dopamine and I believe those have gone out in terms of years. And so if you know anything better than six months doing a monthly injection is probably an improvement. And so we're saying at least six months for this study we've seen expression for ten months and I'm guessing it'll continue for a year. So also thinking about one of the critical reasons I've sat through just a few conferences and started to barely learn about that sort of stuff. So we have thought about it I think about it now there have been some attempts using modified ANGE1 but in terms of small molecules to stimulate the type II receptor I don't know how long you can get that expression that activation of the receptor if you're using a small molecule so I don't know what types of small molecule approaches have been used so far we went with the viral just to gain long term expression. That's a good question and so we are not saving the pericytes and we're only providing one of the factors that the pericytes provide in addition to providing ANGE1 there's probably unidentified factors in addition to the structural support so we haven't looked at our studies out long enough or we haven't looked in the right areas to see if there is something missing because the pericytes are still missing and so pericyte preservation itself has been related to hyperglycemia so probably the best way to prevent all of this obviously is to prevent the hyperglycemia you keep your pericytes around it's not as easy as it sounds which is why we're looking at this pathway but that brings up an interesting point to see what kind of deficits still remain how does it work? I think a lot of the same microvascular pathologies that occur in the retina are occurring in the glomerulus as well our treatment at least when I've looked for GFP expression to see if this viral vector is getting out systemically and infecting other areas I haven't been able to see it and it's not contained to stay inside the eye specifically the retina but it would be interesting and I know there are other groups that are looking at angiopoietin signaling in the kidney and showing better working kidneys after that we haven't looked at that but there are studies out there that do show better kidney function thank you very much