Bench to Bedside: Lifitegrast, pt.2 – Discovery and development of Lifitegrast – Tom Gadek


Hi, my name is Tom Gadek. I wanted to talk to you today about some of the more technical details — scientific details on how we discovered and developed Lifitegrast as a treatment for ocular inflammation in a disease called dry eye. I have a PhD in chemistry, organic chemistry, actually from the University of California Berkeley. And have worked in the biotech industry for about 30 years now. That’s a shocking number, but about 30 years and about 10 years ago, I cofounded a company called SARcode. And I am the founding CEO and a shareholder in that company, so I do have a vested interest. The company is known as SARcode Corporation, that was the early phase that I was CEO of. And then it transitioned to SARcode Bioscience about the time it exited from clinical trials and sought to then develop the drug further and go for FDA approval. I’d like to talk to you today again about some of the more technical aspects of that discovery journey. At the time we began in about 2006, T-cell mediated inflammation was known to be linked to dry eye disease. And that was in a series of publications showing over time that there was at least an inflammatory component to the disease, if not a driver in the disease process. Numerous publications by 2006 had appeared in the literature. Particularly important was the work of the dry eye workshop in that regard. And what was believed at that time was that there was a cycle of inflammation in the disease and that started with some sort of inflammatory stress on the ocular surface. That then activated T-cells and other things, such as antigen presenting cells to then mature, migrate out of the ocular surface and into lymph nodes downstream, communicate with T-cells in that location, activate those T-cells by the formation of something called the immunologic synapse, and then overall, cycle T-cells out of the lymph node and back into the ocular surface. Where they would then be resident and sort of be on patrol, looking for these stresses and respond in a manner that caused inflammation, in response to inflammatory signals. One thing to keep in mind relative to the disease was that we’re not trying to stop the disease early and prevent the activation of antigen presenting cells, APCs. No, that’s already happened by the time the patient’s shown up in the doctor’s office complaining about their eye. That they have a symptom of dry eye that includes inflammation in the eye, sense of dryness, sense of scratchiness when they blink, and a number of other potential problems that they may be complaining about that leads to ocular discomfort. So all of the T-cell and antigen presenting cell process has happened already. The patients are in some sort of a steady state which is overall inflammatory by the time they show up in the doctor’s office. T-cells also, besides appearing in the ocular surface, show up in periocular tissues. And the idea that we had in starting the company and going after dry eye as a target was to control the function of these T-cells in patients, in their ocular surface and their periocular tissue when they show up in a doctor’s office. And we’re proposing to do that by blocking the adhesion of T-cells to the surface of the eye. Their ability to move out of blood vessels, so extravasation, their migration to a site of inflammation. So they’re going to follow a gradient of inflammatory cytokines to the site of inflammation. And we’re going to try to prevent the formation of the immunologic synapse which allows T-cells to communicate with each other and activate the T-cells for cytokine secretion. So blocking the activation of those T-cells and then we should be able to control the disease from our perspective, prior to treating any patients. So coming back to where did we get the idea of going after T-cells and LFA-1 as targets in dry eye. And this was our view from about 2006, when we began SARcode the company, but also began studies to bring forward Lifitegrast. In about 1999, Allergan presented data to the FDA that they very kindly deposited on the FDA’s website as a Powerpoint presentation that they used in their presentation to the FDA when they were asking for approval in their so-called NDA review process. So NDA is New Drug Application. And there were slides, there were a lot of data that was presented in slides that they put together. I think they had a package of 99 slides that they had on the FDA’s website. That data and that Powerpoint presentation is still on the FDA’s website. One of the nicer things is that in their data packages they had tables of data and you could go in and click on the actual values in any of their graphs and be pointed to the data, including their error bars. Very helpful in the analysis. Overall, they showed that in fact, there was from histology slides, they took patient samples and looked at them and stained for the presence of T-cells, for the presence of LFA. And they showed that both LFA and ICAM were present in these biopsy samples. And that in diseased versus normal patients, the level was higher in the staining of these histological samples. But strangely enough, they showed that — not strangely enough, they did show that treatment with Restasis for six months lowered the number of T-cells in tissue and also in histological staining sections. But interestingly enough, at the end of the day after they identified these markers of the inflammatory process underneath dry eye, they didn’t go forward in trying to identify any molecules that might affect these markers more directly than Restasis itself. So they did show that Restasis treatment trended to reducing the amount of LFA, reducing the amount of ICAM, reducing the number of T-cells in these biopsy samples. But their data showed trends, but did not reach statistical significance for a dose response in reduction in LFA or ICAM, or the number of T-cells. So we asked ourselves, could we build an ICAM mimetic, have it then bind to LFA-1 and block T-cell function as a treatment for dry eye? We wanted to purpose build specifically a compound that would bind LFA. And we thought that trying to mimic the ICAM epitope was a good place to start. So we began this work earlier and we knew that there was a structure of ICAM-1 and that if you did mutagenesis in ICAM-1, converting individual residues in the first domain of ICAM-1 to alanines, we could show that we had an effect on the binding. We built a model of ICAM-1 based on the generalized immunoglobulin domain homology. This was done by Leonard Presta. We made mutants in individual residues throughout the first domain of ICAM-1, this was done again by Leonard Presta and Sarah Bodary. And we showed that certain residues, when you converted them to alanine, affected the binding of ICAM-1 to LFA by about 10 fold or more. And we highlighted those and went in and looked at them. And surprisingly, rather than being related to a very linear short sequence within the sequence of ICAM-1, they spanned 3 or 4 strands that went up and down the protein structure. But overall, in size and shape, they led to a footprint about the size of a steroid. And within that footprint, we felt pretty confident that we could make a molecule that would be able to suspend in space the chemical functionality of the side chains of the residues that had been identified by alanine point mutagenesis. That’s how we transferred the ICAM epitope onto a small molecule framework. So here shows a small molecule that’s roughly the size of the epitope of ICAM-1 in its binding to LFA-1. So that in itself was a pretty interesting success. We looked at the molecule, Lifitegrast, of the class of molecules that mimicked the ICAM epitope. And we played a game which chemists, medicinal chemists, are accused of. Structure-activity relationships. We divide the molecule into five regions, those five regions are shown in here in red, blue, blue, green, green. And we dealt into each position, slightly different chemical functionality and different chemical space and shape. And we looked at the effects on the binding to LFA-1, the ability to block ICAM binding. But then also other properties that we would want in a treatment for dry eye, water solubility, the clearance of the compound from systemic circulation, which we did not want to achieve, and specificity for LFA-1 over related receptors, MAC-1 and a few others. And we looked at the combinations of — we looked at the effects of individual substitutions within this framework, and at the end of the day, we came back and we made 20 different substitutions at each of these five positions. Five to the 20th is a very large molecule in the hundred billion range. I was too lazy to make a hundred billion compounds, so we went with the presupposition that the best molecule at the end of the day would be represented by the best combination of each individual module within the molecule, combine them and got to predictive properties, where properties of the molecule itself or some of its parts. We then tried to build in desirable properties for the ophthalmic drug potentially given as a drop for the treatment of dry eye. We knew that we wanted high affinity for LFA ICAM, or LFA-1 mimicking ICAM. And we wanted for scientific reasons, we wanted the compound to be a direct competitive antagonist of ICAM binding. Binding directly to the ICAM binding site on LFA-1. That would lead to selectivity for LFA-1. We then played with the molecule and build in high solubility on the order of 200 mg/ml. So roughly the solubility of sugar in water. We thought after an eye drop was applied to the eye surface, that would give us properties that drove the molecule into the ocular tissue, not only the ocular surface but the periocular tissue. We didn’t want to overdo that process because what we wanted at the end of the day was also low exposure to and rapid clearance from the systemic circulation. So we didn’t want drug building up in the total body, we wanted to have the effect in the eye where we desired it. We did not want to find out later that the drug was overall immunosuppressive to your entire body. So in addition to treating your dry eye, you could go get a heart transplant. That turned out to be a problem, that excessive systemic circulation turned out to be a problem that was experienced with some of the antibodies that were developed against LFA-1 for the treatment of some T-cell mediated diseases. And as I mentioned earlier, we wanted high water solubility, to give us an osmotic gradient. And we wanted chemical stability in the compound so that we could achieve a shelf-life as a commercial product. Manufacture it, put it in a bottle, put it in a box, put it on a shelf in a pharmacy, and it would be able to stay there for two years. That’s your expiration date on your bottle when you go get it at the pharmacy. So we were able to do that. At the end of the day, we had to ask the question — we picked Lifitegrast on those properties, is it the right molecule to treat dry eye? We thought so. But we had to ask the question. So we took it into some preclinical pharmacology analyses and looked at its ability to bind to LFA, block ICAM binding, but also bind to LFA on the surface of T-cells in a tissue culture media and block the binding of ICAM-1. So in this experiment, ICAM-1, recombinant ICAM-1 is coated on the bottom of a tissue culture plate in the well, and we allow cells to settle down. A stable cell line, Jurkats, which is a human stable T-cell line. We allowed them to adhere to the ICAM in the presence of our compound and we showed that we’re able to block that adhesion with an IC50 of about 3 nanomolar. So this is then a multivalent, almost velcro interaction of the T-cell LFA-1 with a velcro level of ICAM coated on the plate. Very, very high avidity of individual high affinity interactions. Compound is potent. We then looked in vivo, and in this case we had a model where we induced inflammation in the retina of a rat’s eye by giving it chemically induced diabetes. And in this process, what would happen in the diabetic process is that T-cells would adhere in the microvasculature of the retina and then they would start to induce inflammation and cause bad events downstream. Ultimately, resulting in loss of eyesight in the animals. This is an animal model of macular degeneration, diabetes induced macular degeneration. So what we’re able to do is give our compound as an ophthalmic drop, every day for 3 months. We gave it three times a day. And we showed that we’re able to induce a statistically significant reduction in inflammation, in black, and the white bars represent the normal levels that you would expect to see. Numbers of cells in the vasculature in response to generating the disease. So normally the cells would become adherent in the diabetic vasculature, here we’re blocking that and the cells are able to keep moving through the vasculature. So we then said, after the T-cells are then activated, they then secrete cytokines, so let’s go in vitro and stimulate some T-cells in tissue culture. We then took human T-cells, these are primary cells, that were induced to be inflamed and secrete cytokines by introduction of staph enterotoxin B. And in this table of data, there are some gray columns. Those are columns of cytokines. So cytokines in those columns, in fact, are cytokines that are known to be involved and expressed at high levels in the tears of patients that have dry eye disease. So they’re linked to dry eye. And we’re able to reduce most of them at a level of drug 1mM or so. That we could achieve in people when we administer the drug. So what we also saw was if you crunched this data and looked at it as a combination of the level of cytokine in tears versus the level of drug we can get to in tear when we administer it, you can get to an effective concentration of the drug or effective potency of the drug. And we can then show that Lifitegrast or SAR 1118, as its research name was known, is in fact more potent than cyclosporine in many of these cytokines. We then looked at the effect of ICAM-1 and the T-cell receptor, TCR, in the formation of the immunologic synapse. This is data from Jay Groves and his group at the University of California Berkeley. And what they have done is that they have made a situation where T-cells, mouse T-cells, are allowed to settle onto a microscope slide that’s been coated with a layer of ICAM in a fluid membrane. Pretty sophisticated, and not only is the ICAM there, but it’s in fact tagged with yellow fluorescent protein. So they can look at the end of the day, either a green circle or a yellow circle where T-cells have arrayed LFA-1 and bring in ICAM as color at that location as the central focus of formation of the immunological synapse. They’re also able to stain the T-cell receptor, or TCR, with an antibody against T-cell receptor that is tagged with a red dye. So they then can watch the formation of the immunologic synapse in real time on a living cell in a microscope. And if you just look at a cell on the microscope in the bright field, at the left here, you just see sort of these wavy surface images of some waves on the surface of the membrane of the cells. But once you start to then look for the ICAM yellow fluorescent protein on the surface of the microscope slide, it’s drawn in by binding to LFA-1 and forms a circle. And then smack in the middle of that circle is your T-cell receptor. so at the end of the day, you end up with almost a bullseye and target ring. Groves and his group have looked at this over time and found that synapse formation at 37C on a mouse lymphocyte takes about 15 minutes, and it comes together initially as a preformed cluster at the molecular level that then arrays into a much larger circle pattern. Very, very interesting and very cool to watch. We then took our compound and treated cells in the immunologic synapse formation assay on the microscope with Lifitegrast at increasing concentrations from 100 picomolar out to 1 millimolar. And we saw over time that the morphology of the T-cells in the first row here changed. They stopped looking like T-cells by about 1 micromolar in the brightfield. We then looked at it by interference microscopy, the RICM image. When it’s a very, very tight adhesion as you have over here in the control, you can see very defined edges to the cell. It depends on the closeness to the cell, if it gets more than about 10 nanometers above the surface of the slide you can’t really see it very well. But again, increasing concentrations of drug across here by about 100 nanomolar have really messed things up. Cells are harder to find and they’re definition of their image is not clear at all. They’re not rounded up, they’re not down close to the glass surface where the ICAM is. Similarly, if you look at the yellow fluorescent protein image of ICAM-1, again shown here in black and white. So to not fool your eye, you’re not more sensitive to green so it’s just converted to a black and white image. By 100 nanomolar or so, you’re seriously screwing things up. Fewer cells are able to find. Similarly, even surprisingly, T-cell receptor’s binding and formation of the immunologic synapse is severely impacted by 100 nanomolar or 1 micromolar drug levels. We know from phase 1 clinical trials that 5% Lifitegrast solution, a single drop of that put on the eye gets to levels above 1 micromolar in tear, and maintains them for 24 hours or so. So it seems that we’re having an effect that we would expect based on mimicking ICAM and binding to LFA-1, we’re blocking the binding of ICAM-1 clearly, in the fluorescence studies with yellow fluorescent construct of ICAM-1. Interestingly, if you look at this over time, here we’ve picked out six cells that have formed the immunological synapse, they’re adhered to the slide surface, and then we administer — I believe it’s 100 nM Lifitegrast as SAR 1118, which is what it was called at the time. And we watched it at one minute, two minute, three minute. So if you look here and compare this to time zero, you see that the background on the slide is just generally brighter. That’s because the ICAM has been released and has diffused away in two dimensions on the surface of the microscope slide. You see great disintegration, largely at least 50% of the cells are gone here. So very, very quickly, we’re able to reverse a preformed large protein-protein interaction and disintegrate the preformed immunologic surface. So the earlier slide showed that Lifitegrast can prevent formation of the immunologic synapse. This data shows that Lifitegrast can disintegrate this pre-existing immunologic synapse at concentrations that are sustained in human tear after administration of a single drop. And interestingly, LFA-ICAM binding is reversible in real time. So within the literature, it may be the first example of that. And was a controversial topic coming into the execution of this program, as to whether a small molecule could inhibit a large protein-protein interaction. We then said okay, let’s go see if it works in an animal. So we took dogs and we first did a quick safety study and we showed that the Lifitegrast as administered as a single drop and then multi-drop administrations, and then multi-day administrations in dogs was safe and well tolerated in normal dogs. And that went from a concentration up to as high as 10% solution of Lifitegrast in PBS. We showed that in dogs that had dry eye, a naturally occurring disease in dogs, that tear production in those dogs as measured by Schirmer’s tear test, was increased in as little as two weeks. In contrast to cyclosporine, which in people takes up to six months to reach statistical significance. We were able to reach statistical significance in the production of tear, tear volume, as measured in the Schirmer tear test, in as little as two weeks and maintain that out ot twelve weeks. We also showed, we took biopsies from the animals, conjunctival biopsies before and after treatment, and showed that the number of T-cells infiltrating the conjunctival was reduced at week 12. And that’s presented here as this table. And that’s published data that I gave the reference for. So in summary, the preclinical pharmacology studies have shown that the compound Lifitegrast was designed as an ICAM mimetic, it mimics the epitope of ICAM-1 that is used in the binding to LFA-1. As such, it binds to the same binding site as a direct competitive antagonist of LFA binding to ICAM-1. We’ve then optimized compounds that are capable of binding in that manner as a larger family. We’ve optimized this particular one, Lifitegrast, as a pharmaceutical for delivery to the eye as an ophthalmic drop for the treatment of dry eye. In particular, what I’m saying is that because you drip this right onto the surface of the eye, that’s where there are T-cells in residence within a hundred nanometers or so of the ocular surface. A very, very short distance. It’s very similar to the in vitro studies where we dropped it on cells in culture, and we would expect to see the same effects in there. So overall, the compound performs like an ICAM mimetic, both in vitro and in vivo. Its high affinity for LFA-1 and it blocks T-cell adhesion at very, very high affinity or low concentrations. It’s able to block the binding of ICAM-1 to LFA-1. It’s able to block the formation of the immunologic synapse, which is important for T-cell activation. It is able to reverse the formation of a preexisting immunologic synapse in a very short period of time. So that’s actually pretty interesting, scientifically. It indicates that the drug might act pretty quickly in the clinic, in terms of blocking immunologic synapse formation, which we are not going to look at, but it should effective in reducing cytokine secretions in inflammation at the surface of the eye. It does block T-cell activation in vitro and we’ve shown that it can block T-cell adhesion in vivo and in vitro. It blocks cytokine release. So this whole combination of immunological synapse formation, blocking ICAM, blocking synapse, blocking T-cell activation, blocking cytokine release, this flow of events makes a lot of sense. And those events are related to the presence of dry eye disease. And overall, we’re able to block T-cell adhesion and infiltration of T-cells into inflamed tissue. We’ve shown in vivo that we can block canine dry eye in what was essentially a small clinical trial in 12 dogs that took about a year to execute. But we were able to show that in fact, we were able to treat canine dry eye. At least the clinical sign of canine dry eye. And overall, in animals, it’s safe and well-tolerated, so that actually is very important as well. We tested the safety and tolerability in both dogs and rabbits in vivo. A little bit of studies in mice. A couple different species have been shown to be safe and well-tolerated. So if you sum it up at the end of the day, the pharmacology preclinically indicates that Lifitegrast appears to be a promising treatment for dry eye. Certainly it works in dogs. And we’ve exhausted the things we can do preclinically, so we think the data looks very positive and it’s time to test it in human trial. So we’ve bundled up all of the preclinical data and put it into an IND submission or investigational new drug, and we’ve sent it off as a roughly 10,000 page document to the FDA and have indications that they are happy to let us go forward into a human clinical trial. At this point, I’d like to acknowledge the people who have contributed to everything I’ve told you about I did not do this by myself. In particular, I’d like to acknowledge John Burnier, who was my partner in this conspiracy from the beginning. And he demonstrated his commitment very early on with a similar donation of his own personal funds to keep the company alive until the venture funding came in, in the late 2006. A number of people who also helped along the way by letting us have their time and expertise and opinion as we thought about how to position Lifitegrast as an anti-inflammatory T-cell mediated treatment for dry eye. And these in particular included Rod Ferguson, Geoff Duyk, who are venture capitalists. And we used them as a sounding board to try and craft the investment opportunity. They were very, very helpful in giving their time and thoughts along the way. On multiple occasions. And then from the clinical expert side of things, Gary Novack and Janine Smith, were very very helpful in discussing how to position this and how to think about this as an actual clinical treatment for dry eye. How you wanted to find patients, how you wanted to treat them. I can’t thank them enough. And legal advice, as well, is always greatly appreciated. So I want to thank you for listening.

Leave a Reply

Your email address will not be published. Required fields are marked *