Taken end to end, I think that this development program might be able to lay claim to being the first and oldest of the modern rejuvenation research initiatives, starting sometime back in the early 2000s. It began at the Methuselah Foundation as LysoSENS, the first of the SENS programs to get underway with modest philanthropic funding. Some of you may remember gathering dirt from graveyards to send in for analysis, in the hunt for microbial species that consume the molecular waste that our bodies cannot remove. Researches knew that those microbes existed because graveyards do not accumulate this waste - it is being broken down by something in the environment. The program carried forward into the SENS Research Foundation when it spun out from the Methuselah Foundation, and a portion of it was later licensed to Ichor Therapeutics, and became LysoClear.
Thank you very much for the kind introduction and for the invitation to speak to all of you at this event. It was a great event last year and I'm very excited that we extended it to a two day event this year. I am thrilled to be back. We brought a whole army of our staff, up there in the back, so I'm very excited. Before I dive in, a couple of housekeeping things. Everything I do is for-profit, so financial disclosures: I have a financial interest in everything I am about to talk about. I also need to acknowledge the wonderful team that assisted with building out the research program that I'm going to describe, in particular one of our grad students is in the audience, and a lot of the figures in the work I'm going to be presenting were the result of his extraordinary efforts, so thank you. Also this program is a spin out of Aubrey de Grey's SENS Research Foundation, and received founding investment from Kizoo Technology Ventures - I think Michael Greve is around here somewhere - but thank you guys for getting this off the ground.
Just to give a very brief overview of Ichor Therapeutics in general: Ichor is a vertically integrated biopharmaceutical company, and we, since 2013, have focused exclusively on diseases and mechanisms of aging. Within the Ichor umbrella we have a variety of portfolio companies. Some are platform technologies, some are single asset plays, all of which are focused on different indications and contributing to the field of anti-aging research. Today we're going to focus on our longest-standing program, LysoClear, which is an enzyme therapy that we're developing for age-related macular degeneration, and the subject of my thesis work as well. Age-related macular degeneration is the leading cause of vision loss in individuals over the age of 50. About 200 million people worldwide have this disease. One of the major problems with this disease is that it robs patients of their high acuity central vision that gives them the ability to interact with the world in a meaningful way. If we think about our own lives, vision is really how we interact. We can hear, we can smell, we move around, and are involved in our everyday lives, but vision is a primary means for that. When you look at a geriatric population, that isn't as mobile, can't hear as well, can't smell as well, the devastating effects of macular degeneration are very real. We hope to put an end to that.
Now when we look at the origins of where macular degeneration originates and manifests, anatomically we're talking about the posterior segment of the eye, the retina, which is responsible for all of our vision. At the very back of the eye there is this little indented part of the retina that is called the macula. The retina lets us see, and the little indented part, the macula, is what gives us that high acuity central vision. If we dive in just a little bit closer, we can appreciate that there is a great microanatomy, a lot of crosstalk between a variety of different cellular layers that are responsible for taking light and allowing us to see.
Starting in the vitreous and moving posteriorly, we have a variety of different neural layers that are responsible for the electrochemical signal transduction that gets sent to the brain and we interpret as vision. Further on, we have a variety of different photoreceptors. Those include the rods and the cones. Those are the cells that get hit by light and can kick off that entire cascade. Now a general rule in biology is that if you have a cell type or a system that is repeatedly stressed, in this case the photoreceptors with light, you need to turnover those systems in order to eliminate the accumulation of damage. So the photoreceptors are supported by very critical phagocytic cell called the retinal pigmented epithelium or RPE. These cells are essential for photoreceptor function and survival. The photoreceptor outer segments are constantly growing towards the upper RPE and the RPE are constantly gobbling off little bits of the photoreceptors in an effort to turn them over.
Like any cells in the body, the RPE cells' phagocytic potential is really based on the lysosome, the organelle that is responsible for degrading things. LIke any other cell in the body, like any other lysosome in the body, a normal healthy lysosome has a plethora of different enzymes that are able to break down all of the little bits of stuff that make up us. However, there are cases where this doesn't function properly. Canonically we're thinking about lysosomal storage diseases. These are congenital diseases where patients are born missing one of those essential lysosomal enzymes. A lot of the time these diseases are lethal in utero, but for the patients who are born with these diseases, the clinical manifestation can be very, very severe. The patients accumulate molecular junk uncontrollably and can't do anything about it. Our central hypothesis for how macular degeneration works is that age-related macular degeneration is an evolutionarily silent lysosomal storage disease, driven by a very specific sort of junk accumulation called lipofuscin. In the context of retinal lipofuscin, it is a combination of retinoid derivatives and lipids. When I say "evolutionarily silent" I mean that in our evolutionary history we haven't lived long enough to accumulate enough of this junk to evolve enzymes able to break it down. But in our view, the onset and progression of macular degeneration is very akin, mechanistically, to what we see in conventional lysosomal storage diseases, and this is going to be a central theme for how we approach treating the disease, and our translational pathway.
Lipofuscin accumulates in the lysosomes of these essential support cells, the RPE. In the very earliest stages of macular degeneration progress, we don't see anything clinically at all, because it is all happening inside of the cell. But at some point in time, lipofuscin accumulates to a level sufficient to promote pathology. It crosses a threshold and problems start to emerge. I've got two types of figures here. The ones on the left are fundus images, which is when you take an ophthalmoscope and look right into the back of the eye and snap a picture. We can see the optic nerve coming through and this little darkened area here is the macula. If we were to take a cut cross-section through the back of the eye and then flip it and look at that cross-section, then that's what these images on the right are. It is called optical coherence tomography, OCT.
As lipofuscin accumulates in the RPE cells, eventually the RPE cells start to become dysfunctional and choke on the accumulating lipofuscin. To cope with the stress, they do what cells need to do - they dump the junk. So the brightest white line here is the RPE layer, and we can see these bumps underneath it. Those are extracellular junk deposits that the RPE are laying down. We call those drusen. So lipofuscin: junk in the cells; drusen: the junk outside the cells that the cells are depositing. We can also see this very easily clinical via fundus imaging, we just look and we see all these little white specks. All of those are drusen in the back of the eye.
Now what is important to note about this, in this early-stage mild form of age-related macular degeneration, we maintain a nice concave morphology of the macula. We also have integrity in the RPE layers, and in the photoreceptors above them. Even though we have this junk being deposited, even though we see these morphological changes happening, usually there is no clinical presentation of this because the photoreceptors that are required to see are not yet disrupted. As the disease progresses, however, eventually the RPE cannot handle the stress burden and the cells start to die, and with them the photoreceptors that rely upon them. So we see by OCT the thinning of the RPE and photoreceptor layers, we see a collapse of the nice morphology of the macula. Because of this thinning, when you look by fundus imaging you can see the choroid, the blood vessel layers at the back of the eye. That presents as geographic atrophy.
This is the intermediate form of age-related macular degeneration, and, collectively, mild and intermediate are termed dry or atropic macular degeneration. Which suggests that there is a wet form. As these cells are dying and having all kinds of problems, it creates a pro-inflammatory environment in which you have complement activation, an immune response, and, importantly, hypoxia. Part of the way that the body attempts to cope with this mess is to create new blood vessels. New blood vessel formation occurs at the choroid, and protruding into the eye. When you have new blood vessels in the body, frequently they are very leaky vessels, so you can have exudate and outright hemorrhage, either into the subretinal space, which can lead to retinal detachment, or into the vitreous itself, which can lead to total blindness. This process of neovascularization is called neovascular or wet age-related macular degeneration, and this is the most severe and advanced form.
So our disease model hypothesis is that lipofuscin, which accumulates inside of the RPE cells, so intracellular stress, drives the accumulation of extracellular drusen. This in turn causes ROS, inflammation, complement activation, hypoxia, and then these stressors eventually lead to the disease state. Our goal at LysoClear is to develop an enzyme therapy that targets lipofuscin at the earliest stages of disease onset and progression. How are we going to go about doing that?
It is first of all important to note that there is no good standard of care for this disease, and, by the way, the dry form is about 90% of patients. Only 10% have the advanced wet form. So for 90% of patients there are no FDA approved drugs. They do have a vitamin formulation that will mildly reduce your risk of progessing, and if you are a smoker you should not be smoking, for this among many other reasons, but it is really not until you get to the really severe form that there is any sort of really efficacious treatment regimen. These are mostly monoclonal antibodies that target the VEGF pathway to inhibit further neovascularization. So there is a huge overwhelming need for more effective treatments for this condition.
I mentioned previously that we're viewing age-related macular degeneration in a manner akin to conventional lysosomal storage disease. So we are wondering if we can borrow some aspects of how we will treat a normal lysosomal storage disease, and with that motif put together a strategy for going after AMD. What is that strategy? Well if your lysosomal storage disease is based on the idea of missing an enzyme in the lysosome, then at the high level it is a simple process: you make the missing enzyme, and you introduce it in such a way that it goes to the lysosome of the target cells. This has been developed and has been approved clinically, and used very successfully by Genzyme. It has been a conventional way to treat lysosomal storage diseases since the 1990s.
The pathways that are used: you make your enzyme, you decorate it in such a way that it gets recognized by a receptor on your target cell, mannose 6-phosphate glycosylation is the canonical delivery path for lysosomal enzymes to reach the lysosome. Mannose receptor is another pathway that is used, and then through an endocytosis pathway your enzyme is able to be selectively delivered to the lysosome. So this is how lysosomal storage diseases are treated. The problem is that we don't have any good human enzymes that are able to break down RPE lipofuscin. We aren't supposed to. We've never evolved them. So we can't rely on human enzymes to facilitate this junk removal. Instead we have to look elsewhere in the world.
That is where SENS Research Foundation (SRF) came in. Some years ago SRF said that this is a problem, and this is the canonical LysoSENS paradigm: here is a junk accumulation, let's look for enzymes that can break it down and remove the junk. By way of wonderful research by John Schloendorn and his team, SRF identified identified a variety of fungal peroxidases, among other enzymes, that are capable of breaking down certain lipofuscin components. The specific component that they looked at was a molecule called A2E - again, lipofuscin is comprised of a variety of retinoids and lipid derivatives. A2E is one of the best studied and most resistant to degradation, and so that is what SRF used for their assays. They identified a bunch of enzymes that can break down A2E from wild-type sources.
This is where we came in. Our goal was to take this very encouraging proof of concept research and build a program to establish a proof of concept that we could actually remove lipofuscin both within cells and within animals to pave the way for moving into patients for the first time. The first thing we needed to do is to get out of a wild-type system - there is just a lot of variability when dealing with wild-type enzymes. So we moved into a recombinant system to make the best performing enzyme, manganese peroxidase. We expressed that enzyme in the yeast strain Pichia pastoris. The reason for that is that Pichia preferentially mannosylates unlinked glyosylation sites, so Pichia automatically adds those sugars to the enzyme that we need for it to be delivered into the cell.
When we make our enzyme and run a gel, we see a beautiful big smear, which is characteristic of hyperglycosylation, and when we treat our enzyme with PNGase F, an enzyme that selectively cleaves off all of your sugars, we see a nice tightening of the band at the expected molecular weight of about 54 kDa, suggesting that we have an enzyme that is indeed glycosylated. Of course we wanted to know how many sugars we have, so we sent the sugar trees out for glycoanalysis and identified 26 mannose residues on the sugar chain. For context, for those of you who aren't glycobiologists, and neither am I, you probably want two to five mannose sugars minimum to achieve uptake optimally - though depending on the enzyme you can go more or less than that.
So, ok. We have an enzyme, it has the sugars that are required for delivery. The next step is that we need to figure out its selectivity against lipofuscin targets. I mentioned to you that lipofuscin is a hodgepodge of different things, and the hardest part of it to break down are these retinoids, vitamin A derivatives. Unfortunately, it is not just one thing but many things. So we scoured the literature and synthesized all of the major lipofuscin components that we could find. We tested our recombinant enzyme using our local HPLC, and we showed that our enzyme is actually able to break down every single lipofuscin component that we tested. Now this isn't entirely surprising, because these are all chemically related species, so we would have expected the enzyme to work similarly on these different species. We have EC50 variants of 1 μmol all the way up to 20 μmol; interestingly, A2E, the molecule that SRF did their screen with, was the most resistant to degradation by the enzyme. As I move forward and present both the cell based work we did and some of the in vivo work, we are going to use A2E as the readout, quantitatively, and the reason for that it is the easiest to analyze, it is the most resistant to degradation, so we believe we are under-reporting the effectiveness of the enzyme by using the hardest target to degrade.
One of the problems that we ran into very very early with this program is that, go figure, cells outside the body behave a little bit differently than they do when they are in the body. So although RPE cells have mannose receptors in vivo, and this is very well established, the second you take those cells out and grow them in culture, the receptor disappears, which makes it really difficult to use that as a model for update and delivery of your drug. So we had to split the efficacy and the uptake into separate experiments. For efficacy, we used a canonical line of RPE cells, ARPE-19. The graph on the left shows them to be negative by flow cytometry for mannose receptor, that is what CD206 is. So what we did is that we had a layer of these cells, untreated they are high viability. When we added excessive levels of A2E as a stressor, we see that the viability drops, and then we saw that our manganese peroxidase is able to rescue some of that viability when we had A2E loaded cells and treated them with the enzyme.
The way that we delivered the enzyme for these studies is with a lipid reagent called BioPORTER, a way of forcing the enzyme artificially into the cell. So this isn't our proposed delivery system, this is just asking the question of if we could get the enzyme into the cells, would it be able to protect against some of this retinoid toxicity, and indeed it seems that it can. Then of course we did manganese peroxidase by itself just to make sure it wasn't toxic, and we have a bunch of other studies along these lines as well, and we see that it doesn't different in any meaningful way from the untreated cells.
We then had to ask the separate question, which is whether the uptake is able to occur through a mannose receptor dependant manner. The cell line that we used for this is a mouse monocyte line called RAW 264.7, which is known to have an intact mannose receptor and mannose receptor endocytosis pathway. For this study we started off with mannosylated bovine serum albumin that was flourescently conjugated and we added that to the cells and looked by flow cytometry for an increase in fluorescence. That is the solid pink line here. So we see over time an increase in fluorescence, which we interpret as internalization of the enzyme by our target cell. When we introduce the competitive inhibitor mannan, which also binds to mannose receptor, we see an attenuation of that effect size, suggesting that we're impairing the uptake, and that this is in fact a mechanism by which the enzyme is entering cells. We then did the exact same thing for our recombinant manganese peroxidase, again we think it has mannose because we checked that, and we also fluorescently labelled it. In much the same way we see an increase in the fluorescence of the enzyme, and we can reduce that signal by competing with mannan, suggesting that the increase in fluorescence is mediated by mannose receptor endocytosis.
Next we really wanted to get to this proof of concept point: are we able to remove or degrade existing lipofuscin components like A2E in vivo. There has been a lot of work showing that you can kind of shovel it around or maybe reduce the rate of its accumulation, but no-one to my knowledge has ever shown that you can actually get rid of it and break it down. So that is really what we wanted to get to as a primary inflection point. We ran an intravitreal pharmacokinetics study, and this is just where we injected the vitreous with our enzyme, did sampling measured by ELISA, and identified a half-life of about 9.6 hours. That is great because this is supposed to be a highly targeted enzyme that just goes into these highly phagocytic RPE cells, so we would expect a pretty short half-life. Indeed, we don't want an enzyme that is hanging out there for no particular reason. We want it to be internalized rapidly - versus monoclonal antibody therapies, which are acting on extracellular targets, and where you'd be looking for a half-life measured in days. So we're very excited by that.
We then asked the big question: can we actually have efficacy? To do this, we used a mouse model of age-related macular degeneration, and the juvenile onset form Stargardt's disease. This is an ABCA4-null mouse, and these mice accumulate accelerated levels of A2E. So what we did, we had our mice and we injected them with six doses of either phosphate buffered saline (PBS), low dose enzyme, or high dose enzyme. We see a little bit of a trend, maybe, on the low dose, but nothing significant, and then at the high dose we saw a significant reduction in A2E burden as measured by analytical HPLC.
We wanted to do an intermediate dose, so we redid a separate experiment, where we treated with PBS and an intermediate dose of the enzyme. We did see a statistically significant drop there. When we pull this data together, we see a nice dose-response in our efficacy model in terms of reduction of our molecular target when treated with increasing doses of our enzyme. So collectively we're really excited about this. We've shown that manganese peroxidase is able to break down all lipofuscin fluorophores that we've tested, we've shown that it can be taken up into cells by way of mannose receptor endocytosis, and, most importantly, for the first time to my knowledge, we're actually able to eliminate existing lipofuscin in an in vivo system.
These results were published in December 2018, and later on that same month we successfully closed a financing round for our LysoClear program, to take these very promising lead series, and engineer them into clinical candidates. All of that work is ongoing at Ichor Therapeutics, I'm very excited about it. Our rate-limiting step right now is actually that we need to move into large animal models for safety, toxicity, and so on, to inform a pre-IND meeting with the FDA. I'm super-excited that my large animal vivarium is scheduled to complete construction in the next three weeks, so hopefully we'll have some very exciting data coming out for our next presentations on this topic.