 Good morning. It's eight o'clock so we'll go ahead and get started. We have Dr. Ambadi, one of our faculty members who's going to be giving grand rounds today. And all of us know that Dr. Ambadi was doing things at young ages when we weren't even thinking about important things in life. But he sort of passed that stage in his life. So it's been almost 20 years since he graduated from medical school. 18 years. We won't want to add more. But he does great work here at the Moran Eye Center on the research and the clinical side of things. And he's going to talk to us about some of his research and clinical applications of that. So Dr. Ambadi. Thanks. Good morning. So I appreciate all of you coming in earlier this morning. When my brother and I both decided to do ophthalmology for a few minutes we were thinking about specializing in the right eye versus the left eye. But we decided front and back worked better instead. So that'll be the theme of today's talk. And actually it's going to be two completely different talks. So I'm hoping to finish the retina research part in half an hour and the clinical cornea part in half an hour. But I do invite questions and discussion. So as most of you know about half my time is over on the fifth floor in the research side. What I'd like to share with you today is some of our work over the last 11 years that hopefully will be translational in nature. And all of you or the vast majority of you in this audience have seen this slide more times than you care to admit probably where the corneally vascularity is mediated by a substance called soluble flit which binds Vajrath vascular endothelial growth factor. Those of you who have been here in Derek Holtz years have seen this on more than one occasion. And I promise I'm not going to talk about Raver II today. But the reason I laid this out is it's important for the rest of the talk to just bear in mind the basic dictionary of vascular endothelial growth factor, Vajrath, mediates abnormal blood vessel formation in the eye. Vajrath receptors include membrane bound flit I as well as soluble flit I which is a Vajrath antagonist. It's a soluble receptor so it's hanging out in the interstitial space sucking up Vajrath. So we showed several years ago that it's vital to the cornea to express this molecule. Now how does that apply to the retina? Well in the back of the eye there's also vascular demarcations which are breached in macular degeneration. You have a coroidal neovascularization that grows subretinally and which can bleed and destroy vision. And angiogenesis in general affects so many different parts of the eye and so many different diseases. So this is an important problem. And it's a very prevalent problem. In the US alone there are as many people if not more with macular degeneration than all cancers put together. Okay, so that's the perspective. The first one we established was expressed in the cornea and over the last five years with the help of the EyeBank and as well as collaborators from around the country we found that it's also expressed in the retina. It's stained in on an RNA level it's stained in purple in both the photoreceptors and in the RPE. On a protein level it's stained here in red in multiple layers of the retina and if you focus on the RPE it's pretty highly expressed in the retinal pigment epithelium. In contrast, in macular degeneration specimens it's pretty much absent in the RPE. Both in C and B specimens from human eyes that were donated after death as well as in retinal angiomonas proliferation eyes it's absent in the photoreceptor layer. So this is just circumstantial evidence that S-flit is absent in macular degeneration specimens. Take that over to the lab side. If you inject anti-flit 1 antibody subretinally in a mouse and the first two residents have gotten very good at intravitural injections in people then I would invite any of you who want to test their microsurgical skills to come try to do some subretinal injections in mice and that will hone your skills very quickly. What happens is knock down a flit 1 with an antibody subretinally induces VEGF and C and B as opposed to a control antibody and if you do forcing angiography and OCT in a mice we have the same spectral instrument that's used on the clinical side. We have a mouse rig for it on the research side downstairs. You'll observe that we induce C and B that's much larger than anything induced with a control antibody injection and this is confirmed on OCT evaluation. So that's knocked down a flit 1 at a protein level. We can then proceed deeper in the pathway by knocking it down at an mRNA level. We can express using an adeno associated virus and AAB it's very similar to what's used in clinical trials in labor's congenital amaurosis or other diseases to express a short herpen RNA of type of RNA, a double-stranded RNA that targets selective genes. Now if you recall back to medical school mRNA is single-stranded. Double-stranded RNA occurs in viruses and RNA is a signal for cells to destroy gene transcripts because double-stranded RNA is normally not present in eukaryotic cells. We can take advantage of this by expressing a double-stranded RNA, a short herpen RNA, a herpen forms a double strand. We can trick a cell into destroying the S-flit mRNA and when you do that you knock down S-flit 1, we would predict VEGF would be released and that would induce neovascularization. Now does that happen? Indeed it does. So we can knock down S-flit 1 protein and this correlates very nicely with expression on staining. Most importantly with C and V formation on forcing angiography and on histology. And you can quantify this. VEGF will be increased as this coil neovascular volume compared to any of the controls. So we've knocked down F1 on a protein level, we've knocked down S-flit 1 on an mRNA level and can we knock it down on a gene level? I'll get to that in just a moment. I just wanted to show briefly what these spectral images look like on the mouse. Let's see if we can play this. This is a high quality resolution even in the mouse eye and you can see the C and V lesion distorting the entire mouse right now. So the third arm of this process was genomic knockdown of F1 using what's called the crelox system. So how does the crelox work? You can obtain a transgenic mouse that has the F1 gene flanked by what are called the flanking sequences. And if you insert an enzyme called creerocombinase, that will excise F1 from the gene. So it removes a target gene from the chromosome. If you inject a plasma that expresses crea-enzyme into this transgenic F1-lox-speed animal you do of course knock down S-flit 1 expression. Hey, Brett. And induce C and V and induce C and V that's an observed enforcement in the mouse. Now all of the models I've shown you so far have involved the sub-retinal injection process. Now a criticism that you could throw at me would be, well sub-retinal injection is not a good test. You're damaging the retina with your needle no matter how good your injector is. And that's a fair criticism. So how do we get around that? I've knocked down protein, I've knocked down MRNA, I've knocked down gene, all using a sub-retinal injection delivery system. Well one way we can get around that is to take advantage of this transgenic animal. The way we can do that is develop new transgenic animals. Now there are other transgenics that express Cree in selective areas of the retina. So on the left-hand side is a mouse that expresses red fluorescent protein known as Rosa throughout every cell in its body. So if you have a six or eight year old son a mouse that glows red in the dark, that's that mouse. It glows red in the dark, it does. Don't want to know. If you make that with another mouse that expresses VMD Cree, now if you recall VMD, the teleform, macular dystrophy, that's an RPE specific gene and this is a Cree enzyme driven by this RPE specific gene, we can observe this Cree expression shown in green just in the RPE. If you made this red glow-in-the-dark animal with a different transgenic called ICree75, we can induce Cree just in the photoreceptor layer. So that Cree enzyme that I showed can excise particular genes, we no longer have to inject subretinally. We can get the animal to express it either in the photoreceptors or in the RPE. So that drives our next experiment. We can make that VMD Cree animal with that flit-lox animal so you get a new transgenic resulting in selective knock-down of flit just in the RPE or you can make ICree75 with the flit-lox animal and knock-down flit just in the photoreceptor layer. Does that make sense to everyone? So we're knocking down S-flit1 just in the photoreceptors or just in the RPE compared to control transgenic animals. And when you do that, C and V occurs spontaneously in the VMD Cree flit-lox animal. Recall this is where S-flit1 is reduced just in the RPE and this is confirmed on OCT, force in angiography and on electron microscopy. Conversely, if you made it with the ICree75 flit-lox animal, very interesting lesions. You don't get C and V, but you get these neovascular lesions in the inner retina at the border with the outer retina, which remind us of RAP lesions, retinal angiomonas proliferation. So C and V is the most common form of AMD where the sub-retinal angiogenesis is growing up from the coin and then in RAP we have inner retinal vessels growing down into the photoreceptor layer. So essentially what we believe we're observing is induction spontaneously without any injection of C and V or RAP. So that's what we've shown so far. Knocking down this protein at the protein level, mRNA level or DNA level induces different phenotypes of AMD. Who cares? Why does that matter? So the second branch of our research effort is on developing therapies where hopefully we can knock down VEGF intracellularly. Avastin, Lucentus, Ilea all of these wonderful drugs that have made a difference, all of those are anti-VEGFs that bind VEGF outside of cells. However, many vascular endothelial cells express their own VEGF and their own VEGF receptors. So imagine trying to do a drug introduction project when the junkie is the pusher. If you have your own receptor and growth factors it is often very difficult to disrupt those autocrine loops by an extracellular treatment. So how can we attack VEGF intracellularly? For that we borrowed a page from our friends in the HIV literature. David Baltimore and colleagues about 15 or 18 years ago did a very nice experiment where a few multiple stromal derived factor with a 4 amino acid peptide called KDL which serves as a retention signal for the endoplasmic reticulum. What you develop is a SDF-KDL recombinant that stays in the endoplasmic reticulum. All proteins are expressed in the endoplasmic reticulum. Now the receptor that binds SDF is CXCR4. Now what they were able to show is that this recombinant protein, SDF-KDL sequestered CXCR4, kept it from ever getting to the cell surface in macrophages. And by doing that they could keep these macrophages and these mice with the stromal treatment from getting HIV. CXCR4 is a very important receptor for HIV entry and by knocking down its expression you can by this intracellular sequestration you can prevent HIV entry and they call that an intracellular chemokine or an intracine. Why is that important? Well in Ghana and certain other places there are certain populations of humans that are CXCR4 knockouts people who don't have this gene and they're resistant to AIDS no matter what they do. Now getting back to the track with our work we sought to knock down VEGF by developing an intracellular receptor as opposed to an intracellular chemokine. So in a cartoon model you have VEGF receptors that are hanging on the cells, VEGF binds, and that leads to signal transaction inside cells. We proposed to make a VEGF receptor that would have that KDL sequence amino acid endoplasmic reticulum retention signal which would keep the receptor within the ER and also sequester VEGF within the ER. Thereby disrupting intracellular production and release of VEGF. Now about eight years ago we did experiments in a mouse conial injury model and this is where you can put sodium hydroxide on the surface of a mouse eye and induce aggressive neovascularization which progresses over time, but in treated eyes we can regress neovascularization. So we can prevent aggressive K and B while the corny is not clear we can make while those vessels go away. Now can we achieve relevance to the retina and why would we why is there relevance to the retina? Now no offense to all or Emmy or Al, I would submit to you that our retina clinics look like this, where we have patients coming in again and again and again for their individual injections. I don't think our patients like this, I don't think our doctors like this. It is the best that we have to offer but we should do better. So in the lab what we've been developing are these nanoparticles that are biodegradable. PLGA is a polymer that degrades to lactic acid and glycolic acid those are degraded by the Krebs cycle and this nanoparticle contains the plasmids that express our interceptors against VEGF. PLGA is that biodegradable polymer and this is studded with a peptide called RGD. And RGD is a homing peptide. It allows selective guidance just to abnormal new blood vessels. So rather than carpet bombing an area, it's more of a guided missile. It homes just to abnormal blood vessels and stays away from normal blood vessels and the way it does that is it homes to alpha V beta 3 integrin which is part of the molecular signature of neo vessel tissue beds. It's not expressed in normal blood vessels. Alpha V beta 3 is selective to growing new blood vessels. And we first showed in 2009 pilot study that these nanoparticles were injected by tail vein into a rat that sustained laser induced coital neo vascularization. All of you know if you shoot an argon laser too hot or too long into a person's retina you can break Brooks membrane and induce CNV. And you can do the same thing in a mouse or a rat and induce CNV. You have the CNV lesion breaking Brooks and entering the retina. When you treat these rats with RGD coated for 2, 3K nanoparticles you can suppress the CNV lesions relative to any of the controls. And naked plasmids nanoparticles without plasmids or untargeted nanoparticles. We then followed up over the last few years in both mouse and monkey models. In that mouse model where I showed you we could knock down S-flit with that double-stranded RNA, we can regress CNV over the course of a month as opposed to control animals which actually develop secondary lesions due to this knock down of S-flit. In a quantified fashion relative to any of the controls the treatment is the only one that's able to shrink CNV. And this is just a higher mag view, CNV lesion in a mouse, 4 weeks after injection it's almost completely gone. This is just one particular mouse but this effect was statistically significant. I told you that the RGD nanoparticles are selective and to give proof of that if you take a mouse eye that has a CNV lesion you can label these nanoparticles with Nile Red and on a high mag you can see those Nile Red particles in the CNV lesion but not in the unaffected retina and not in a healthy eye after a systemic injection. Separate model laser model of CNV we observe a very similar effect. CNV we can regress with our treatment much more than any of the controls. Confirmed on pathology where CNV lesion is much bigger in a control animal than in a treated animal. How does this compare to Avastin? Well you can't really use Avastin in a mouse because Avastin is humanized but there are anti-mouse anti-vegeta antibodies which are essentially the correlate of Avastin in the murine model and we didn't observe a statistically significant difference but there's a trend that this RGD treatment is better but it's not statistically significant but certainly it's not inferior to intervitual anti-vegeta injections. Can you test a vision? Well you can't stick a mouse in front of a Snellen chart but I can steal from Judith and use the optokinetic nystagmus model. If you stick a person in front of a series of graded stripes that move their eyes beat and you can do the same thing to a mouse. You can stick a mouse in a drum, keep the mouse from escaping and this mouse has to see this series of black and white stripes moving. And the thinner the stripe, if the mouse can track, the better the mouse can see. Normal mouse acuity is about in an untreated, unlasered animal by this test is about 0.38 cycles per degree that's a normal mouse. When you treat it an induced CNV it falls down to the 0.3 range, 0.28, 0.3 range. We can restore much of that with treatment with our targeted nanoparticles that regress CNV and none of the controls induce a statistically significant improvement. So unlike any of the control populations our treatment can indeed restore not all but some visual function in a mouse. Mice are nice, do mice really matter? Let's look at monkeys. You can induce laser CNV in a monkey and the monkey eye resembles the human eye much more and in a control animal treated with buffer the CNV spots get worse over time in all the control groups whether treated with PBS buffer, like nanoparticles or untargeted nanoparticles. The CNV spots over time get worse. However when we treat the monkey with RGD targeted for 2, 3K nanoparticles we can actually induce regression of CNV and this is confirmed on immunofluorescence where control animals have much more vascularization shown in red and fibrosis shown in green. Hologram is a fibrosis marker, collagen one is a rascular marker and on pathology where the monkey CNV lesions are considerably larger in the control than in the treated animals. And so we're able to regress both CNV and fibrosis in a statistically significant fashion relative to all controls. Why is any of this important? I think it's important to you that in this one's ophthalmology where the 7-up trial is published that over the next several decades we're going to have a serious problem. Anti-vegeta therapy administered to AMD patients for seven years almost two-thirds, three-fifths of them develop subretinal macular fibrosis and 90% of them develop foveal atrophy. Paul and Emmy and others will know better than I but what I recall from the MPS studies in the 80s is that this is much higher rates of fibrosis in atrophy with anti-vegeta therapy than natural history of AMD. So while anti-vegeta therapy is a boon to new master AMD, there are significant costs over the long term. If we can develop therapies, Paul, sorry. I was asking if they see that because I had a lot of believe we don't see it. Fair enough, that's a fair point. This just came out in the November issue of ophthalmology so I'd really welcome your thoughts on it once you've got a sense to review it. But assuming this is true, and this was a fairly large cohort then I think it is important to develop an anti-vegeta therapy that is targeted that is not globally suppressive of VEGF in the retina because VEGF is not just vascular growth factor, it is also a neurotrophic agent. Photoreceptors deprived of VEGF and ganglion cells do suffer. I think that's the relevance of targeted therapy. So you might ask if you administer these nanoparticles systemically, are there systemic side effects? In mice, in monkeys, in rats, we did not observe any. Indeed, we did not observe any nanoparticle deposition in kidney, lung, skin, or liver, all of which are high vascular flowbeds. So in summary, these targeted nanoparticles enhance selective localization, regressed fibrosis, improved vision, and may, in the future, serve as an alternative or adjunct to monthly intervitural injections. So to wrap up, what I've shown you in this first talk is that understanding the basis for vascular zoning can serve as the basis for a novel, non-viral gene therapy delivered by targeted nanoparticles to reverse CMV. So I've just taken the credit for ten years' worth of work by a lot of different people, funded by several agencies that we're very grateful for, and any questions? Novartis has asked us to test it in their models, so we'll see what comes of that. We'll see where it goes. I wish Meg were here. Meg has some interesting data. I don't think she's published on it yet, but Meg does have some SNP data on some FLIT variants. All right, with that, we'll completely switch gears now that the back of the eye people have...