 OK, we're going to get started today. So I'm pleased to introduce our guest speaker today, who is Dr. Joe Corbaugh. And he's from Washington University in St. Louis. And I haven't known him very long. I met him a little over a year ago when I was a guest speaker at WashU. And we found out we had a lot of common interests related to carotenoids in the eye and working on animal models, such as birds, et cetera. And he's going to be telling us some about that today. He is a Californian, a true Californian, has been, grew up in Los Angeles, went to Stanford, where I learned he was a linguist and a biology major. So he had good diverse. He was very into Latin and was even, I think, a championship Latin person, whatever that means. But he did very well in the state championships in California. Then went on to Stanford, or that's where he was at Stanford. He then did his MD-PhD at UC San Diego and decided to go into anatomic pathology and was at Brigham and Women's Hospital in Boston. And soon after that, he did some post-doctoral work with Connie Sepko, is that right? Which is where his connection to vision came. And so even though he is still a professor of pathology and he is a neuropathologist, not an ocular pathologist, a lot of his research work is focused on vision, especially related to a genetic regulation of vision and that's what he's gonna be talking about today. And also, he is a true biologist also and will be telling us some about vision in animals and birds, I think some too. So with that, I will bring Joe up. It's a pleasure to be here. And thanks Paul for the nice introduction. I'm actually gonna talk about birds, but at my noon seminar, so if you wanna hear about that, please come at that time. So today I'd like to start with a little clinical anecdote about a Turkish gentleman that contacted me a number of years ago and told me about his wife that had retinitis pigmentosa. And they were interested in having a baby and wanted to find out the genetic basis of her retinitis pigmentosa to know the prospects for the baby and whether the baby might be born with the disease. And they were having difficulty finding out the cause of her disease and they had heard about some of our research and he was interested in knowing whether we might be able to help them identify the cause in this particular case. And so I said, well, perhaps we could help. And so I suggested that we start by looking further into the history. And he said, well, can I send you her pedigree or a family tree? I said, yes. And then he immediately emailed me this. And so this gentleman had really gone to town and was very aggressive in pursuing this question and generated this remarkable family tree, which I'll point out a couple of things. And one is that his wife was the only affected member of a very extended pedigree. And the second thing is that this pedigree sort of gave me a sudden cold sweat and a bad memory of medical school trying to figure out inbreeding coefficients for a patient like this. But you can see that there's quite a bit of inbreeding in this family. And I helped him interpret this by saying that maybe this suggests that her retinitis pigmentosa might be an audisomal recessive form of retinitis pigmentosa. And he said, oh yeah, I'm aware of that. And we've already tried to evaluate for what the different causes of retinitis pigmentosa, audisomal recessive retinitis pigmentosa that we're known at this time. And she said, he doesn't have any mutations in any of these genes. So I said, well, okay, that makes it difficult. We also talked about whether she had mutations in ophthalminate forms of the disease. She also didn't have any mutations in the genes that were known at that time. And so I said, well, that makes the problem a little more difficult because retinal disease and photoreceptor generation is a very genetically heterogeneous family of disorders. There are over 200 different retinal disease genes mutated can cause degeneration of photoreceptors. And notably, the majority of these genes are photoreceptor genes in that they're either specifically expressed in photoreceptors or enriched in the cell type. And at the time, a number of genes were known that were involved in ophthalmoresissibritinitis and that this has continued to grow. But still to this day, a sizable fraction of them are in fact unknown. So patients come in, you sequence all these different genes and you fail to find a mutation. And so I said, then the next step would be to try to proceed further with this. And he stopped me again and said, how about if I send you her exome sequence? So he already obtained an exome sequence for this patient, which it was, again, at the time, surprising now it's becoming more and more common and these patients are really taking control of their own genetics. So after he sent us that, we were actually able to very rapidly identify a candidate causal gene that turned out to be the gene upon further investigation that was causing her disease. And so today, in the first part of the talk, I'm going to tell you about this approach we use to read coding variants and identify the coding variant that caused this patient's disease. Unfortunately, I won't have time to talk about one of the major areas of work in my lab and that is how to interpret non-coding variation in the genome. This has been a major interest of my lab for many years, how to map and identify non-coding cis-regulatory regions, the control gene expression. The majority of functional human genetic variation, probably 80 to 90% of it, lies outside of coding regions and for the most part it's thought to affect these non-coding regions and either subtly or in a major way, modify predisposition to disease. I won't have time to talk about that today. Instead, I'll focus the majority of the talk on two different very early stage ideas for possible therapy, one involving the direct reprogramming of raw photoreceptors into cones and the other are efforts to develop an infrared optogenetic approach to therapy. I'll tell you about those at the end. So this audience obviously doesn't need this introduction, I'll go through it very quickly. As we know, vision occurs when light passes into the eye through the thickness of the retina to reach the photosensitive outer segments where the light signal is converted into an electrical signal that's then processed via the inner retina and then passed to the brain as visual information. There are two main photoreceptor cell types who cell bodies reside in the outer nuclear layer, rods and cones. And outside of the fovea, of course, the predominant photoreceptor type is the rod constituting about 95% of all the photoreceptors, extra foveal photoreceptors. Another thing I want to remind you about because it becomes important shortly, and that is photoreceptors are a ciliated cell type. That is, there is a connecting cilium that bridges the inner to the outer segment acts as a critical conduit for the passage of proteins to the outer segment. So some of you may be less familiar with exactly how transcriptional regulation works. This is the bread and butter of what my lab does. And I give a toy example here of a gene, say, that's expressed in a developing mouse embryo, in developing limb buds, in the retina, and say, in the developing brain. And the way typically that a complex developmental pattern of expression will be established is via the action of so-called cis-rigidory elements. These are non-coding stretches of DNA that contain multiple binding sites for both transcriptional activators and repressors, which bind to the DNA and then direct the expression of a nearby gene in a specific pattern. So you might have a non-coding cis-rigidory element driving expression specifically in the photoreceptor. There's another one driving expression in the developing brain, in the limbs, et cetera. It obviously depends on the specific gene. So finding where these regions are in the genome, mapping them, and characterizing the detailed architecture of the sequences and how that mediates gene expression is one of the major efforts of my lab. Transcription factors, as I said, are the main players that bind these elements and control expression. They control spatial and temporal patterns of expression, as I'm depicting here, but crucially, in something that's overlooked sometimes, that they control quantitative levels of expression. A lot of functional genetic variation in regulatory regions is likely to not affect spatial temporal patterns, which can be a real gross effect, but the quantitative levels of fine tuning of gene expression in different tissues. So I'm gonna summarize a lot of data very briefly for the sake of time, but one of the early approaches we took to try to identify the non-coding regulatory regions that control gene expression in rods and cones is we studied a key regulator of photoreceptor gene expression called CRX as a transcription factor that's absolutely required for the development and differentiation of both rods and cones. In human patients that have mutations in CRX, you get labors, congenital hemorrhosis, and you get a failure formation of the outer segment of both rods and cones. This is absolutely critical. It's sort of master regulator of photoreceptor gene expression. We reasoned that that would help us pinpoint many of the critical non-coding regions that regulate photoreceptor gene expression. We'll use a technique called CHIPSI to map the location of where CRX was bound across the genome in mouse photoreceptor cells, and this slide summarizes that data in one slide. So if you look in some, the distribution of CRX bound regions or CDRs across the entire genome, and you place all genes on top of each other, you can see that they occur both upstream and downstream of genes, sometimes even internal to the gene body, but most importantly that there's a real peak of clustering of CRX binding right near the transcription start site of genes. Okay, and again, this is data that's summed over all the genes in the genome. This is in the mouse, but it gave us a clue that this might be a useful tool for prioritizing candidate genes in patients that have a retinal disease caused by mutation in an unknown gene, because we already knew that of the genes that are known to contribute to disease, many of them are photoreceptor specific or photoreceptor rich, we reasoned that genes that have a lot of CRX binding specifically around them would be good candidates to evaluate for novel retinal or photoreceptor disease genes. So that's the approach we took. These data are from the mouse, what we did is use a bioinformatic tool to map the mouse non-coding regulatory regions onto the human genome, and then we did something very simple. We basically went across the human genome and we assigned these CRX bound regions and here's a number of them clustering around the redoxing gene, two individual genes simply based on proximity. The idea is if you have a cluster of CRX bound regions that are known to be in your gene, then it's likely to be regulated by CRX and probably a photoreceptor gene. So we do that and we can then tally essentially the weight of binding of CRX at these different regions for each of these genes and come up with a simple numeric score for every gene in the genome, the extent to which that gene is regulated by CRX. And so in mapping studies where you have a pedigree family with multiple affected members, you can often narrow down to a sometimes sizable stretch of the genome where a particular disease gene may be assigned, but it can be a large region, still hard to figure out exactly which gene is involved. So first to sort of test whether this approach would really work, we looked at genes which were already identified as photoreceptor disease genes and we said, okay, what if we had a pedigree where we mapped a disease gene to this 50 gene interval that included redoxin. But we didn't know that redoxin was the disease gene. Would our approach help us prioritize that and quickly identify that? And obviously this is a straw man because redoxin is such an important gene of photoreceptor, it would be a logical choice to look at, but there are many photoreceptor genes that are really not so obvious. And so when you do this approach and you tally all those CRX binding scores across this region, indeed, redoxin comes out at sort of the top of the list. And again, that's maybe not surprising, but if we took this approach and then try that with all of it, what at the time were known causes of retinitis pigmentosa, in about two thirds of cases, the actual causal gene was in the top three candidates when we ranked them with this approach. So we figured this doesn't work in all cases, but it does seem to work, at least in a significant subset of cases. And so then we tried to apply the same approach to our patient, now the situation is a little different. We didn't have genomic mapping data. We had an exome sequence for a single patient and this is a simplified version of her pedigree, but we looked at it anyway, used this approach and basically found that there were about 40 genes that had homozygous variants predicted from the exome sequencing. We were predicting that she had a lot of some more recessive disease, so that's why we focused on those. And when we ranked them, we got a list like this. We then tried to confirm the mutations in the candidates starting at the top and the first several turned out to not actually have homozygous mutations and that's a common problem with exome sequencing. We need a lot of false positive results, but this fourth gene, MAC, turned out to have homozygous mutation, ultimately proved to be the gene that was causing her disease. So what is MAC? This is a male germ cell associated kinase, this is sort of an odd name for rather odd gene in that it has a very curious expression pattern. As far as I know, it's expressed in only two places in the human body, it's expressed in the photoreceptors and it's expressed in the test as a developed expert. And it contains a napkinase domain at the interminous and we found a mutation here in this patient. We then subsequently worked with the consortium of European scientists to identify five additional families or small pedigrees that had additional mutations in this gene. And it turned out that all the mutations identified, identified at that time, fell within this interminable napkinase domain. We went on to evaluate the effects of these individual changes on the kinase activity and indeed, several of these completely abrogate kinase activity. So it suggests that whatever this kinase is doing, kinase domain is critical for the function of this gene. And so it's interesting that around the time that this work was going on, we were sort of surprised by a paper that came out of Japan for a Kawa's group where they were studying this gene in the mouse. Now the mutation in Mac in the mouse had been made almost a decade before by a group that was studying its potential role in spermatogenesis and they found only very minimal defects and no decrease in fertility and that might explain that no one's ever observed any changes in fertility in male patients that have mutations in this interesting way. They did see a very curious cell biological phenomenon and that is this is an antibody stain of wild type in macrimutin mice, mouse retinas, looking at the region of the inner and outer segments and green is an acetylated tubular antibody which is the marking portion of the cilium and you can see in the wild type cilium looks like this but in the mutant, there are abnormally long cilia which is a really curious phenotype. But most importantly, over time it was found that there's a progressive photoreceptor loss so photoreceptor degeneration in the Mac mutant mice. So this corroborated the idea that these mutations are actually causal in humans and we're likely to be leading to photoreceptor degeneration. Now one of my favorite aspects of this gene and this story is that Mac is an extremely deeply phylogenetic conserved gene. In fact, there's a remote orthologue that's present in the unicellular algae climbing demonus and these algae have two flagella that in the mutant are abnormally long. So it's even the phenotype is conserved and since this is a plant and we're talking about animals here, this is a gene that's been around for maybe approaching a billion years of evolution. There are other genes interestingly in the pathway that regulate ciliary length in climbing demonus that have been identified. Unfortunately, the evolutionary distance is so great we've never been able to find clear orthologs of those other gene products in higher organisms or in humans. One last point is that mutations in Mac turned out to be the most common cause of heritable retinal disease in Ashkenazi Jewish population. So an interesting and sort of surprising fact that goes along with this story. And finally, back to the patient and her husband. So armed with this information, they felt they sequenced the Mac gene and the husband found that he didn't carry mutations and they went on to have a child that's healthy and unlikely to at least develop this form of the disease. So despite the fact that can't yet offer good treatments to these patients, it's nice that the genetics can sometimes still help inform their decision making process. Okay, now I'm going to turn to the second part of the talk and tell you two stories about, as I said, early stage efforts to develop new treatments for a photoreceptor degeneration. And so I asked you to imagine another scenario where let's say this newborn little girl turned out to have, say, almost igus mutations in Mac or other gene that's involved in that causes retinitis pigmentosa. I think this is going to occur more and more as genome sequencing, exome sequencing and whole genome sequencing becoming more and more common. They'll probably come a day fairly soon where newborns might routinely be sequenced. Their whole genome might be sequenced. You need to be able to interpret that data and if possible, act on it in advance. And some of these diseases obviously have their onset a little later in life. And so there may be a significant therapeutic window in which one could act to forestall or prevent the onset of disease. I mean, this is a major goal. And so if we entertain this scenario, the question is how can we do this? And one approach that's been taken obviously is gene replacement therapy, where you have a gene that is defective, you try to deliver it by a virus, you introduce a good, healthy copy of, intact copy of the gene, and then restore the function. And that's one approach, but it's sort of one disease during that time that has its problems in terms of drug development and so on. And so a bit of a holy grail in the field has been to develop therapies that can be applicable to a broader range of diseases. So subset forms, genetic forms of red lung degeneration. And so we set out with the hope of trying to develop an approach that could be used and could be applied to multiple forms of disease. So just briefly to summarize, the mouse model here of a red knight is Pigtosa. And what you see is a progressive loss of photoreceptors in the outer nuclear layer. But if we look a little more closely, what you find in fact is that there's an early loss of rods depicted in red here and only later a secondary loss of cone photoreceptors. So this is a typical progression of red knight is Pigtosa, right? Early loss of rods and a secondary non-cell autonomous loss of cones. So we know that a lot of forms of red knight is Pigtosa are caused by mutations in rods, specific genes, as genes expressed only in rods, not in cone photoreceptors yet. Nonetheless, you get this invariable secondary death of cones. There are many possible reasons for that. The one is that you have a massive loss of photoreceptors in the outer nuclear layer creates a toxic environment, introgression of inflammatory cells and mediators that probably make it for non-cogino for the survival of cones. And so we had this idea and I still think it's a tad crazy, but we had the idea that you might be able to forestall or prevent this secondary loss of cones if you could take the rods that are going to express a mutated gene and to directly convert them into cone photoreceptors. The idea being that if the disease is caused by mutation in a gene that's specifically expressed in rods, then if you were to convert the cell into a cone, a gene, a cell type that doesn't even express that gene, then in theory you might be able to prevent the effects of that mutant gene on that cell. And to pursue this idea, we leveraged knowledge of the development of photoreceptors where it's well known that a photoreceptor precursor makes a developmental fate choice and depending on whether a particular transcription factor called inarrel is expressed, it either becomes a rod or a cone. So if inarrel is turned on in that precursor it differentiates as a rod, if inarrel is not turned on it differentiates as a cone. And we know in the mouse if you knock out inarrel during early development then cells that would have normally been faded as rods will differentiate instead as cone photoreceptors. And so we developed a mouse that had a phloxed allele of inarrel with the idea that we would wait till adult stage, acutely knock out inarrel and then see what happens. Do as in development to the rods then convert directly to cones. And so with this mouse we would grow it up to around day 42 then we would use for hydroxy tamoxifen to induce knock out of the gene and then sacrifice the animals a certain period of time later. And the first thing to note when we do this is that in the adult inarrel knock out that there's initially a rather good preservation of the outer nuclear layer. So the cells are still there, intact. And furthermore, we did a series of in-situ conversations for both rod and cone specific genes to see what's going on at the level of the transcriptome. And so here I'm showing inarrel itself which is nicely disappears in the adult inarrel knock out as you might expect. But we were first initially very pleased to see that other rod specific genes such as GMB1 and Genet1 and genes that encode components of the rod phototransduction cascade were also completely eliminated. However, as we looked further at other rod specific genes we found that nearly all of them were markedly reduced in their expression when you acutely knock out inarrel. But many of them still had relatively modest residual levels of expression. Now mind you, that's in contrast to the developmental or embryonic knock out of inarrel where all these genes are pretty much completely gone. So that really suggests that the phenotype we're getting when we acutely knock out inarrel in the adult is somewhat different than what you get during development which we were surprised by. If we look at genes in cones we found the opposite. So in some cases we saw a pretty marked derepression of cone gene expression in what would have been rods. But in other cases, particularly in the cone options we saw very little change in expression. Now it looks a little lower here in the adult knock out but if you quantify this by QPCR there really isn't any change in either of the two cone options. Again that's different than the embryonic knock out so in the embryonic inarrel knock out OPN1SW which includes blue cone option is markedly derepressed and expressed at high levels throughout all of the cells that would have been rod photoreceptors. So in other words to summarize the gene expression data knock out inarrel during development you get a true trans fading of what would have been a rod into a blue cone. But when in acute adult knock out you get a marked decrease in many rod genes and an increase in a subset of cone genes but not all of them. And so in other words it's a sort of partial reprogramming of a rod into a cone giving you a sort of intermediate gene expression phenotype. We looked a little further at this phenotype and we used electron microscopy to analyze the structure of the nuclei and the cells in the outer nuclear layer. As you probably know in the mouse and many other nocturnal mammals the rod nucleus has a very special architecture with the majority of the heterochromatin is in the center of the nucleus and forms a very thick ball that is very electron dense like this. We found in the acute adult inarrel knock out there's several changes in the architecture of the nuclei. For one they're not typically a spherical they're larger on average and more irregular which is more of a cone-like feature. Many of the cells have a looser heterochromatin with more U-chromatin around them in the periphery which again is more cone-like but the fundamental chromatin picture is still rod-like for the most part. And the last thing is many of these intermediate or reprogrammed rods have juxta-nuclear mitochondria which is relatively rare in normal rods but fairly common in normal cones. And so we took this as evidence that there is sort of a hybrid phenotype here where the ultra structure reveals some cone-like features in the reprogrammed rods but for the most part they still look pretty rod-like in their heterochromatin pattern. So one other aspect we evaluated in these reprogrammed photoreceptors was their physiology and this requires just a bit of background. So as we all know photoreceptors use an 11-cis retinal chromophore in the opsin that is the key photoreceptive molecule that undergoes an assist to transisomerization upon a receipt of a photon of light. Then that spent chromophore, the all-trans chromophore has to be released from the opsin and the photoreceptor cell. Has to be trafficked up to either the RPE or the mule glia where then it is recycled and then passed back to the photoreceptor as a rejuvenated pigment. So in the case of the visual cycle that passages through the RPE, both rods and cones can access the RPE visual cycle and the RPE hands back 11-cis retinal to both the rod and the cone ready to use for the next round of photoreception. The mule glia in contrast is a more recently characterized pathway for renewal of the chromophore pigment. Seems to be only accessed normally by cones. So cones can pass their spent chromophore to the mule glia which then processes it as 11-cis retinol. So the alcohol form which goes to the cone, the cone then has to convert retinol into retinal to then make it usable by the cell. So rods normally cannot access this cycle. We found that our acute NRL knockout in the adult, the acutely reprogrammed cells make it an impossible for the rod to access this mule glial cycle. So it suggests to you that there is at least some physiologic aspects that are cone-like in these partially reprogrammed rod photoreceptors. Okay, so that was a little disappointing for us actually because we were really hoping for a more complete directory programming of the rod into the cone. It was partial in several different respects. However, we still wanted to go ahead and test our original hypothesis, namely that this type of reprogramming might be able to confer protective effect on the rod photoreceptor, thereby in turn preserving the cells and then in turn decreasing the probability of secondary cone loss in the model of retinitis pingutosa. And so we decided to test that idea and we used the Rodopsin mutant mouse as a model of retinitis pingutosa. In this mouse, there's a progressive death of rods followed by a secondary death of cone. So in that sense, it mimics the human disease. And what we did was we crossed our homozygous floxalil of NRL into the Rodopsin homozygous mutant background which has this progressive degeneration and then at the adult stage, we injected four hydroxy tamoxifen in the mice to acutely knock out NRL, reprogram the rod photoreceptors and then see if a protective effect was conferred on the cells and specifically on the cones. And so this is what we found if you look at the outer nuclear layer of the Rodopsin mutant control versus the Rodopsin mutant that's undergone acute NRL knockout, there's significant preservation of the thickness of the outer nuclear layer and particularly inner segments and that difference in thickness is quantified here. But perhaps even more importantly, we found that if you stained for either of the two different cone options that you see still pretty nice expression of cone options indicating preservation of cone photoreceptors and the acute NRL knockout, whereas almost all cone ops and expression is lost in the controls by the date that we're analyzing them. Furthermore, we did some ERGs to analyze the physiologic response and we found that with increasing flash intensity there's very little response, if photopic response in the control degenerated mice, but that we still got pretty good a rescue here of the electrical response in the rescued, rather than suggesting that yes, indeed, if you acutely knock out NRL and even though you're not getting complete and perfect reprogramming of the rod photoreceptors into cones, you still can confer a protective effect against at least this one form of degeneration, cause of degeneration. So this works in mice and mice that have a phloxed allele, but obviously that's not really an effective approach or viable approach for treating human patients. And so we wanted to see if we could acutely knock down NRL via no associated virus using RNA I or RNA interference against NRL. And so we created an AV virus that carried an NRL RNA I, we injected it sub-retinally and allowed it to spread out and then infect the photoreceptor cells. And so I have here a picture of a retina that was infected with a construct that expresses GFP ubiquitously just to tell us where the virus has infected. And then it also expresses NRL RNA I. So you can see fairly nice and complete infection at least through about two thirds of the retina here. And so then importantly, we wanted to reconfirm that by RNA I we'd get the same kinds of gene expression changes that we saw when we did this with a phloxed allele. And so what we did is we took portions, took pictures here of the retina in the infected region, uninfected region and crucially in the transition zone between infected and unaffected. And then we did again a series of in-situ hybridizations for both rod and cone-specific genes. And you can see very nicely that NRL expression is largely eliminated and that changes right here at this transition zone. Rodopsin, if anything, is even at lower levels when we use the RNA I, then we've lost out the allele. That was a little surprising. And then some other rod genes are down. But overall, we saw a very similar picture with rod genes. So some of them are eliminated, some are reduced, most are reduced, but not all of them are gone. And contrary-wise, the cone genes were, subsets were derepressed, but again, the cone-opsins were pretty much unchanged. So what you have is a scenario where you're really decreasing the levels of rodopsin, but you're not derepressing the cone-opsins. There's not a lot of opsin being expressed in these reprogrammed rods. And so one way to look at them is that they have entered maybe a dormant state. Don't exactly know what that means, other than they're unlikely to be very responsive to light by themselves. And so the last thing we tried, and this is still relatively preliminary, was we tried to treat rodopsin mutant mice with this NREL RNA I delivered by AAV. And we find that we can, in the affected cells, see some cellular preservation relative to controls. But further experiments need to be done. I should mention that a recent study presented at the last ARBO meeting by Xi Jianwu's group from National Eye Institute has taken up this idea and is now using CRISPR-Cas9 knockout of NREL to see if they can get a similar protective effect. They've looked at this, this is unpublished work, but they've looked at this in several different mouse models of retinitis pigmentosa. And they report very good rescue in a variety of different generation models with CRISPR-Cas9. So CRISPR-Cas9 might ultimately be, I think, a more effective approach than RNAI to getting knockdown of NREL in the adult in eventually in humans, perhaps. Now there are caveats to this approach, obviously. A major one being that mutations in NREL itself can cause retinitis pigmentosa. Okay, so you might argue that you're sort of trading one disease, one form of disease with another, and I think that's a very valid point in one that would have to be considered and addressed. Another thing that I should note, however, is that the architecture of the transcription network that regulates raw and blood receptor gene expression in humans is fairly different from that in the mouse. So the best example is that a transcription factor immediately downstream of NREL called NR2A3 when mutated in humans causes enhanced escoma syndrome. So what goes on there, it suggests that the rods are actually being converted into a blue cone or a blue cone-like cell in those patients. If you knock out NR2A3 in the mouse, however, that is not what happens. So there's clearly been some rewiring of the transcriptional network that regulates the final tier of raw gene expression between human and mouse. And so I think we have to proceed cautiously and perhaps if this general concept were ever to be implemented as a therapy in humans, targeting NR2A3 might be another possibility around the NREL. Okay, so I'm going to turn now to the last section, which is quite different from what I've described before and that is our efforts to develop a new approach to infrared optogenetic therapy. But again, this is very early stage still. I'm going to tell you most of the very much pre-clinical types of experiments we've been doing. So when you think about therapeutic strategies for retinal degeneration, you can sort of group them broadly into two categories. Those that could be implemented before degeneration has its onset, say like when you know that that newborn has those mutations that's going to eventually lead to retinitis pigmentosa, or after degeneration has largely progressed and you have very few photoreceptors left. In the first group you have obviously gene therapy rescue or perhaps more broadly applicable approaches like directory programming of rods into cones. After generation, one of the most actively pursued areas is the attempt to generate photoreceptors in a dish from stem cells and then re-implant them in the patients. I think that strategy is brought with a lot of problems, not just the difficulty of creating photoreceptors from IPS cells, but the implantation problem, getting photoreceptors to engraft and be functional in a degenerated retin I think is no small task. And so in the context of those difficulties, people have turned to some other approaches, one of which is optogenetics. And so the idea here is that although the photoreceptors are largely gone, the inner retinal cells are still present and at least to a reasonable extent intact. And so people have thought of the idea of using optogenetic constructs to endow those normally non-photosensitive inner retinal cells like bipolar cells or ganglion cells with photosensitivity. So what is an optogenetic construct? So these are microbial options, but they differ very markedly from vertebrate options in that not only do they have a chromophore and can sense light, but they have in the same polypeptide an ion channel. So what you can do is with the expression of this single gene product, you can control the flux of ions across a membrane with light alone. And this has been a huge boon in neuroscience because it allows direct interrogation and control of neural circuits. And they have optogenic actuators that can both activate or turn off gene circuits with light. So you get very rapid control. And so this has now been attempted in some mouse models of blindness and have been able to get some rescue of visual function by introducing optogenetic actuators either into the ganglion cells or into bipolar cells. I think bipolar cells are probably the preferred approach because then you still allow the retina to retain some of its downstream processing of the visual information. So one of the difficulties of this approach is most of the optogenic actuators that have been tested so far are maximally sensitive to fairly short wavelength blue light. And so the idea is, and there are some clinical trials that are already about to begin in Europe, is to use gene therapy to introduce the optogenic channel into bipolar cells but then have the patient where some sort of amplification goggles that take in the visual scene and then project very bright blue light onto the retina. And that's a problem because blue light is pretty energetic and it has to be very bright because you're cutting out the middlemen, you go straight from the ox into the channel, you lose the amplification cascade that you get in normal bipolar receptors. And so therefore you have to apply very bright light to activate these optogenic actuators. And so that's what widely recognized as a problem in the field. And so one of the desideratas is to create optogenic actuators that are red shifted. That is ones that can be activated with much longer wavelengths of light. The idea being that this will minimize phototoxicity. It will also minimize induced pupillary reflexes because those are dependent on activation of melanopsins which are maximally sensitive to blue light so you can get really redshift your optogenic actuator. Then you can induce gating with longer wavelength light that won't activate the melanopsin and therefore won't induce a pupillary reflex which will obviously limit the amount of light that's being delivered to the retina. You can also minimize photophobia with longer wavelengths of light and even facilitate ossable macular therapy because as we know of course the normal macula has lots of pigments that like to absorb in the blue region of the spectrum. So again if you can use much longer wavelengths of light then you could potentially use such optogenic actuators in that region without having the deactivating light absorbed. And so we came at this from a very sort of different and unusual channel and that is we got interested for pretty much completely other reasons at a very old, almost 100 year old problem in vision science. So this is a problem that was first appreciated in the 1800s where it was noted and at that time I should point out that many vision scientists thought that there was only a single photosensitive pigment and that it was the same in most animals and had a peak around 500, peak absorbance around 500 nanometers. And that was shown to be the case in mammals and in reptiles and amphibians for the most part but then a group came along in 1896 and found if you look at freshwater fish the absorption is redshifted up to about 530, 540 nanometers. And at the time this was unexplained until about 40 years later in the 1930s George Wald who later went on to win a Nobel Prize for this work discovered the basis of this mechanism and what he found was that these freshwater animals use a different chromophore that is redshifted relative to 11 Cisretanelle. And our favorite example of this is migratory salmon. So when salmon are in the open ocean they use the typical 11 Cisretanelle chromophore but when they migrate to inland streams they encounter a redshifted photo environment. Some murkier water absorbs short wavelengths of light and red shifts the overall spectrum of life that the animals encounter. And so we're well aware that salmon undergo a suite of physiologic adaptations to adapt to the new environment to prepare for spawning and all. But it's less widely appreciated that they also dynamically redshift the sensitivity of their entire visual system to adapt to this redshifted photo environment. And so just a little more background on how that works just to remind you the visual pigment as I've mentioned several times consists of both of a apopso protein as well as a covalently bound chromophore 11 Cisretanelle and then when a photon of light is received the 11 Cisretanelle is summarized into all trends right now that leads to a change in the overall configuration of the protein eliciting the phototransduction cascade. And we're all familiar in the course of evolution opsoons can change and lead to different tunings and absorption properties that's how we get blue, green and red chromo opsoons in humans. They all have the same chromophore but the exact amino acid side chains that surround that chromophore change the electrostatic and steric environment around the chromophore tweaking its exact configuration and thereby changing its light absorption properties and shifting it either toward the blue or the red depending upon the specific array of amino acids that are in the binding cleft. But another way animals can shift their absorption is via changing the chromophore. And this is how salmon do it and it's how many freshwater fish do it. And what they do is they take the A1 or vitamin A1 form of retinol and they simply add an additional double bond in the ring here. And it's a very simple modification but because it affects all the chromophore in the retina it can lead to a red shift of this in studio all of the different photoreceptors and opsoons. And the reason for that is that the number of conjugated double bonds here that is alternating single and double bonds is what determines the light absorption properties of the chromophore or for that matter any molecule of this type. And so by adding that extra double bond you significantly increase the length of that conjugated chain. Thereby red shifting its absorption. And so this was known then in the 1930s when Wal and a series of very careful experiments established it but the identity of the enzyme that actually added this double bond remained unknown until recently. And that's what I'd like to tell you about today. Just one more comment. I said that this was originally found in freshwater fish it turns out that probably 15,000 or more species of freshwater fish use this switch to a greater or limited extent. Salmon of course have a dramatic migration where they use it but other animals change their A1 to A2 content throughout the different seasons as the quality of light changes from summer to winter. Other species like lamprey when they migrate also go from A1 to A2. In contrast many amphibians start with A2 in the aquatic tadpole stage and then go to A1 when they become a terrestrial or semi-terrestrial adult. And so it wasn't easy to get our hands on the right stages of migratory salmon. So we turned to an interesting result that had been published about 10 years ago showing that if you simply treat zebrafish with thyroid hormone you can sort of induce many of the physiologic changes that salmon undergo. So we use zebrafish as a little miniature model of migrating salmon. And we treat them with thyroid hormone for three weeks and we found that we can completely convert their chromophore from vitamin A1 to A2 and get a red shift. So then what we did was we reasoned that that's thyroid hormone treatment is probably inducing the expression of the enzyme that adds the double bond to A1 to create the A2 chromophore. And so we took RPE from wild-type fish and fish treated with thyroid hormone and we compared their transcriptomes by RNA-seq. And so what you see here, each of these dots represents a different gene transcript. Those that are on the diagonal of course are genes that are equally expressed between the control and the thyroid hormone treated. But we found that there was one outlier here that was markedly induced upon thyroid hormone treatment in the RPE. Now in parallel with these zebrafish experiments, we were also doing similar studies in another organism, the American bullfrog. And I said that most amphibians when they go from a tadpole stage where they're expressing A2, they switch to vitamin A1 in the adult. But the American bullfrog is a little special in that in the adult stage he likes to sit in the water in the pond like this. And the idea is that he can scan below the surface into the murky red-shifted aquatic environment for prey down below, but simultaneously be scanning the aerial environment for bits and pieces if he might want to eat above. And as far as I know, American bullfrogs will eat just about anything. You can find pictures of this on the internet. So they're very much an omnivorous creature. And so it's very important that they feed themselves and keep an eye on both of these environments. And again, George Wald, in his fourth decade of studying this system in the late 19th, in the early 1970s, and it's one of his last publications showed beautifully that there's a difference in the A1, A2 content in the dorsal and ventral RPE of the bullfrog. We reconfirmed this chemically and we found that in the ventral RPE, that is the portion that's feeding the photoreceptors looking up into the aerial environment, you only have the traditional vitamin A1-based chromophore leavens as right now. But in the dorsal RPE, where you're feeding the photoreceptors looking down into the murky aquatic environment, you have sort of a mixture of both A1 and A2. Suggesting, again, that the enzyme that converts A1 to A2 is expressed dorsally in the RPE, but not ventrally. And so we did the same experiment by Ronnie Seek in bullfrog, and we found indeed that the dorsal RPE expresses a transcript that doesn't appear to be expressed at high levels in the ventral RPE. And these both in the zebrafish and the bullfrog turned out to be orthologs of the same gene. This gene is CYP27C1, which is a member of the cytochrome P450 family of oxygenases. These, this is a very large family of enzymes present in many diverse organisms. I think there's 57 different homologs or members of this family in the human genome. And they're familiar to many people, including physicians, because these are the family of enzymes that are primarily responsible for metabolizing drugs in the liver. And pharmaceutical industry always has great interest in this because polymorphisms in these genes change the rate at which people metabolize various drugs warfarin in the drugs. And also other xenobiotics that come in when we eat plant material and whatnot. So just an aside real quick, but I won't get into this. This is a CYP27C1 gene. It is actually present in the human genome. It is absent from the rodent genome and from a variety of other mammals. So, even though no mammal has ever been shown to use vitamin A2 in the eye, naturally, this gene is present. So it doesn't seem that it's expressed in the eye, maybe expressed elsewhere in the body. I can go into that in the question answered if you're interested. But obviously we were focusing here on fission amphibians where it very much is used in the eye to great effect. But we wanted to pursue that question further and verify this. And so the first thing we did is raise an antibody against CYP27C1. We then put that antibody on extracts of dorsal and ventral RP from the bullfrog and found very nicely that high levels are expressed in the dorsal, but absolutely none is expressed in the ventral RP correlating with the expression pattern of the vitamin A2 chromophore. We then did the same thing and by standing in zebrafish and found that the RP specifically induces expression of CYP27C1 upon thyroid hormone treatment. It's not present in the untreated controls. We then went to look at the actual activity of the enzyme. And so when you transfect tissue culture cells with a construct expressing CYP27C1, feed the cells vitamin A1, you get production of vitamin A2 in contrast to a mock transfective control. So that shows that indeed this gene, this enzyme is sufficient to convert vitamin A1 into A2, but actually is it necessary in vivo and to address that question, we use talent technology to engineer a zebrafish knockout of CYP27C1. We created a number of alleles that led to short deletions that resulted in really frame shift mutations and premature stop codons. So I believe these are null alleles. If you take one of these trans-heteroconite psychic mutants, treat it with thyroid hormone. In the wild type case, as I said, we can induce expression of CYP27C1, you get none of the protein in the mutant indicating these are true knockout. So I mentioned this before, just to recap, you treat wild type fish with thyroid hormone, you get a direct and quantitative conversion of A1 into A2. If you treat the CYP27C1 mutant fish with thyroid hormone, you only have A1, so there's no shift to A2 indicating that CYP27C1 is absolutely required for the production of vitamin A2 in people. And so next we collaborate with Vladimir Kafalov's lab at WashU to do single-cell suction electron recording of the red single cones of the fish to see if they fail to shift in the mutant. And so when you do this in wild type or mutant red single cones that have not been treated with thyroid hormone and you flash different wavelengths of light and then look at the electrical response, you see that they're very similar and they both have a peak sensitivity around 561 nanometers. However, when you treat with thyroid hormone for three weeks and the wild type in the mutant, the wild type undergoes a dramatic red shift up to 618 nanometers. So it's almost a 60 nanometer red shift that's achieved when you switch from vitamin A1 to A2 but the mutant fails to shift completely as we expect because there's no production of vitamin A2. Did you have a question? Okay. And so the last asset we did was to test whether this actually affects their behavior because the ultimate evolutionary pressure that probably led to this interesting enzymology that allows these animals to redshift their visual system is that it actually permits them to see further and more deeply into these murky, redshifted aquatic environments because otherwise why would it be selected in evolution? And so what we did was create a behavioral chamber where we could directly observe the response of the fish to light. So adult zebrafish are positively phototactic to light so they swim toward a light source that they can see. And so what we did was put a very, very long wave length infrared 940 nanometer backlight underneath a tank and then we had a night vision camera above so we could directly observe the fish swimming around in the dark as well as in the light. And then we had another LED light source, a diffuse light source on one end of the tank which we would intermittently turn on or off to see if the fish would swim toward it and spend more time at that end of the tank. And so just to give you a direct idea of what the data looked like, this is how the fish behaves when the lights are off and this is speeded up a bit, these quick fish aren't quite this frenetic. But the fish goes around and around kind of feeling its way along the sides of the tank using a lateral line system but this is completely in the dark and the fish can't see anything. In contrast, if you turn the lights on down at this end of the tank, if it's a wave length the fish can see, it will spend a lot more time down at that end of the tank. And make some forays away from it but then we'll come back and stay down there. Okay, so this is a very simple assay and then what we wanted to do was quantify essentially the amount of time that the fish spent within one inch of this end, a lit end of the tank. And so we did this with two different wavelengths of monochromatic light, 590 nanometer light which is about a yellow orange light that we predict the red single cones from both the wild type and the mutant ought to be similarly sensitive to this wavelength of light. So both wild type and mutant fish ought to respond and swim toward this light and a much longer near infrared wavelength of light, 770 nanometer light which we predict the wild type fish will be much more sensitive to but the mutant fish will be very poorly sensitive to this wavelength. So when you do this first with the yellow light in the dark, all of the different conditions spend about 25% of their time down at that end of the tank but when you flip on the light they spend between 50 and 60% of their time down in the tank. And it doesn't matter whether a wild type or mutant whether they've been treated with thyroid hormone or not they all respond to the light. In contrast, when you use that longer wavelength near infrared light, only the wild type fish that have been treated with thyroid hormone appear to be able to see the light, respond to it and spend more time down at that end of the tank. So these acids like any behavior acid are a little variable. So there's some noise in the data but we're pretty confident that these fish can indeed see these wavelengths and you can play with the intensity and the wavelength of the light and be pretty confident that indeed the thyroid hormone wild type fish are the only ones that can see this light. So that was the first and very comforting demonstration of the adaptive value of this vitamin A1, A2 switch that had been studied for a hundred years. And I think fairly convincing evidence that CIP27C1 is the enzyme that mediates this switch. So that's just a summary of what I just said. And back to the therapeutic point. So our idea here is that co-expression of CIP27C1 might be a great way to be able to create localized pools of vitamin A2 in target cell populations that are being used for obstrogenic therapy. So it's a paper came out a couple of years ago showing that you can substitute vitamin A2 for A1 interchangeably in these obstrogenic actuators and indeed just like vertebrate photoreceptors, they will, vertebrate options, they will redshift their sensitivity. And so our idea is then to develop therapies where we co-express CIP27C1 with an object actuator, usually actuators that have also been engineered to be redshifted and with the ultimate hope that we can implement optogenics using near infrared light. So very, very far redshifted from the current therapies that are under development and that are actually being trialed in humans. And then there's all these attended benefits that I already mentioned. So to just briefly summarize what I've talked about, my lab is developing diagnostic approaches for reading both coding and non-coding variants and how they contribute to human retinal disease. And we're developing a number of early stage therapeutic approaches, both directory programming of photoreceptors to prevent retinal generation in multiple forms, genetic forms of the disease and infrared optogenetic therapies. To credit some of the people that participated in this work, the human genetics, a big role was played by Annika Den Holender, Front's Cramer's Group in the Netherlands, who are part of a very large European consortium that allowed us access to their precious patient data and allowed us to identify additional patients with mac mutations. In my lab, Cindy Montana is a talented MD-PhD student now an ophthalmology resident at Washington University who did the rod re-programming work. Jennifer Inright, also an MD-PhD student that just graduated from her lab, is also going to be going into ophthalmology. And she did the vitamin A1, A2 work. And I thank you for your attention. We have just a minute or two for questions, any? Long talk, sorry. Yeah. Any questions, yes, Cindy? So, I thought that was a great talk. I was wondering if you wrote in your cells regarding the CYP27C1 form. Yeah, so in fish and in bullfrogs, it's only expressed in the RPA, it seems. But there are indications from other species, I should say in zebrafish. In other species of fish, there are some suggestions that it might, the enzyme might also be expressed in miliglia. And the data that exists for that is that in a single retina, sometimes the rods and cones actually have a different ratio of A1 and A2. And one way to get that, of course, because only cones can normally access the retinal-visual cycle, is that if you express this enzyme in miliglia. But we haven't confirmed that in those species yet. Any other questions? Any other questions? I do want to emphasize, this work is very translational and, Joe, you didn't show the cartoon that was done about CYP27C1, do you know? Oh, yeah, yeah, I saw Sherman's Lagoon. I'm not a cartoon reader, but apparently this is a syndicated cartoon and they picked up on this and did a series of cartoons about CYP27C1. It's a whole two-week series. I'm the enzyme, yeah. I looked it up, it's very impressive. Thanks for your attention. Yeah, go ahead, please.