[BLANK_AUDIO]. So, let's go on and talk very briefly about fragment based drug discovery. The goal quite simply is to go from a hit that looks like this, a small heterocycle perhaps with multiple, chemical handles, sites of diversity on it. To something, like the compound at the bottom, which is a very tight, specific binder that has the the legs to become a, a drug candidate. And be put into animal toxicology studies and ultimately into human tox studies if warranted. The these hits, these fragment hits can come from many, many different sources, and I just mentioned a few here. It's important that you use structure in the process to guide this transition, to do it efficiently. And it's also im, important that one drives potency and selectivity towards getting a candidate. And there's some perceived and real challenges to to this process. Many chemists don't like low affinity starting points because they're very weak binders. there's also a concern that the number of fragments in a typical fragment library is very small, not the millions of compounds you'd find in high throughput screening libraries, but only thousands. There's the whole issue of the hydrophobic effect, which I talked about Lipophilicity and Size do Matter. Having referenced the structure is important, and there's this whole question of whether or not fragments can be selective, and I'll touch on that briefly. The reason that the fragment based approach works at all is, because when a fragment does bind to a target, it actually binds quite tightly. And the reason I say that is that every, every molecule when it binds to a target has to undergo the loss of rotational and translational entropy and the penalty that has to be paid for that by the system. And that penalty is independent of molecular weight. So if a small thing binds, you know it binds very tightly because you've overcome that very substantial penalty in the. For those of you that are interested the arithmetic is laid out here in the, in the slide that you can see perhaps most emblematically is that when you go from say an 11 atom fragment to a 30 atom drug candidate. There's been[COUGH] a, actually, here, quite a modest increase in, in the potency, in the affinity. And particularly when one considers the intrinsic strength of binding. I won't, I won't belabor this point but I, I do want you to come away with with an understanding that although the fragments appear to bind quite weakly. As you know, in terms of, of the measured affinity, they're actually forming a very good anchor to the target. And that helps you as you do the chemistry. It also helps you actually detect them and, the way that we pioneered to, this detection at SGX Pharmaceuticals was with x-ray crystallography. And it, the way it was done was really quite simple. We would mix fragment pools together, usually ten, there are nine in this particular box. You incubate these compounds at high concentration with a preformed protein crystal. Protein crystals have got large solvent channels that these small molecules can freely diffuse through. And if they are capable of binding to the target they'll absorb to the surface of the protein in the crystal. Then you can do an x-ray crystallography experiment. And you can visualize both the protein. You can see here the atomic model in this chicken wire mesh which is the electron rich areas of the crystal, meaning the protein. And then this area unaccounted for by our knowledge of the protein sequence which is the small molecule fragment bound. And in this particular case the feature from the small fragment molecule corresponds to the binding of that compound. So the beauty of this is that, when you get a hit, you can see it, and you can decide whether or not you want to do something with it, in terms of the chemistry. And with modern technology, high throughput x-ray crystallography at a third generation Synchotron source which is the one available outside Chicago operated by the department of energy. You can do this experiment in a matter of just a couple of days, screen an entire library against the target and get starting point. So, the hit rate is substantial. You, you start with 2000 fragments, you got a hit rate of anywhere between 1 and 5% meaning you got anywhere between 20 and 50 starting points for chemistry. Most chemists won't touch more than five starting points. I mentioned that there are only typically a few thousand compounds in these fragment libraries. They look like this. They've got multiple chemical handles on them, places of which you can do chemistry. And there are commercial available R groups that you can attach to these chemical handles to explore the diversity of of each of these fragments or scaffolds. And when you start to consider the potential diversity here and the fact that there are literally thousands and in some cases tens of thousands of R groups available. The potential diversity is on the order of 10 to the 15 compound for a 2000 compound library. That's a lot better than a million compounds from a high through put screening library. Unfortunately it doesn't come very close to the ten to the 30 compounds that contain Carbon, Hydrogen, Fluorine, Nitrogen, Oxygen that are a molecular weight of less than 400 but that's the way it is. And the great thing about ca-, about carefully chosen fragments is that you can do chemistry on them very rapidly. sometimes in parallel 48 or 96 at a time using automated approaches. In our experience at SGX Pharmaceuticals and subsequently at Lily and in other companies, we know that you can use this approach for many different target classes. When you do find scaffolds or fragments that bind to a target, they bind to multiple sites, and I'm showing here MEK1, a protein kinase. Where we showed that fragments will bind both to the active site, shown here and an allosteric site near to the active site. When you do find small molecules that bind, fragments that bind, you see multiple chemotypes, meaning different starting points for chemistry. Diversity right in, in the core, right in the primary anchor to the target. All this traits, translates into a wealth of opportunities for doing medicinal chemistry. The example I want to tell you about, is the MET Receptor Tyrosine Kinase, this is the receptor for hepatocyte growth factor or scatter factor. It is of significance in the drug industry today because we know that there's a series of activating mutations. That can be found throughout the length of the polypeptide chain depicted here schematically. That are causative of, of various cancers. This first came up in the context of familial cancers, the Hereditary Papillary Renal Cell Carcinoma. That was where MET first came to attention. And we now know that sporatic mutations that occur in the tumor and ultimately give rise to the tumor can cause a disease in such as non small cell lung cancer. We also know that MET is very important in metastatic disease. When a tumor becomes metastatic it frequently does so because of the addition of the, the accretion of additional genetic abnormalities. And some of those genetic, genetic abnormalities are activational so, of MET through through point mutations. So, the final way in which we, we know that MET is important for, for cancers is that there's evidence of gene amplification in certain cancers. Particularly gastric in the Orient, and in EGFR inhibitor-resistant lung cancer in populations that have been treated with drugs like Tarceva. So, very interesting looking target, pretty well-defined target biology, unlikely there's going to be an efficacy failure here. So a good, a good gamble for a biotech company. You're always going to have the risk of a tox failure and you'll see one in space. But what you want to do as a small company in particular is to reduce the likelihood of the efficacy failure. and so let's, let's, let me show you the path that we took to getting a very potent very selective inhibitor of the MET Receptor Tyracine Kinase, and then what happened when we took it into the clinic. So what I'm showing here is the active site of the enzyme, the tyracine kinase with an SGX molecule bound. This is, this came from, originally from a, a fragment hit. This, this particular molecule has had a little bit of chemistry done on it. And what you can see is that it's got a very interesting mode of binding to the target protein. It's, it's not a very tight binder. It's a hundredfold less type, than one would want for a drug. On the right track but not there. It doesn't have particularly good penetration into a cell as you can see with this GTL16IC50 value of worse then a micro mollar. But the good news is the molecular weight's very low and the lean is a good value you remember I said you want to be above 0.27. Where there are problems is that the cLogP of this molecule, is is about is right on the borderline in terms of the values that I suggested. And the LLE, the Lipophilic Efficiency, is poor, we want it to be greater than five. So, what we what we did subsequently to address that issue was to begin exploring a series of modif-, modifications at this position in the active site of the enzyme. And I'm showing you the surface of the protein surrounding that drug in the, would be drug in the active site. And the fact that there's a cleft here that one can explore by putting different R groups into the cleft. An by ex, by exploring the chemistry here at this position, what we're able to do is to reduce the cLogP. an at the same time raise the lipophilic efficiency without compromising the LEAN value. So you can see how we're using these indices as a way of guiding the chemistry. Not focused exclusively on potency but trying to get potency at the right price.
