So, my name's Steven Burely. I'm as you can see, see from the screen, I've just joined Rutgers University and the Cancer Institute of, New Jersey. I have had a long, long and varied career despite my young age. I've been to Medical school, I have got a Phd, I have done internship in residency, I had a brief stint practicing medicine which I found very stimulating but not, not for me in the long term. And I, then went from, from the Brigham in Boston to, where I graduated from med school, to Rockefeller University and the Howard Hughes Medical Institute, where I spent 11 very enjoyable years as a professor and an investigator in the institute, studying the structural biology of gene expression. And in 2002, I went to the dark side. I joined a small bio-tech here in San Diego, SGX Pharmeceuticals. We grew that company. Took it to an initial public offering. Took compounds into the clinic. And you'll hear about one of those today. And then, had, had one of those compounds fail and at that point we were acquired in a trade sale by Eli Lilly and company and I spent 4 very interesting and very stimulating years as a distinguished Lilly research scholar shuttling between here and Indianapolis and Shanghai and the UK and Spain various Lilly sites. Learning about the pharmaceutical industry and contributing to a variety of programs at Lily. And I was fortunate enough to be able to retire in the spring of last year, and then subsequently failed retirement, and now I'm back in at work in academia as of Jan one. enjoying this new challenge at at Rutgers. So my topic today is concerns drug discovery for protein kinases, and a specific approach that we took at SGX Pharmaceuticals and subsequently at Eli Lilly and Company, that relies heavily on my PhD training in structural biology. So, the presentation outline is here. I'll describe some of the goals and some background information. And then present some useful tools that are in use at Lilly and other companies by drug discovery teams. I'll introduce fragment based drug discovery. And then show how fragment based drug discovery can be used with a structural biology specifically x ray crystallography. In a case study involving the mass inhibitor and the objective of producing a new cancer treatment. A so as for the goals I want you to understand some of the challenges in small molecule drug discovery. I want you to come away with an understanding of some of the advantages that are offered by the fragment based approach. And probably most important is to appreciate the very real possibility of unforeseen pitfalls in in a process that has a very, very high failure rate. So there are three general approaches to drug discovery today, and those depend on what kind of a hypothesis you are going to take. Are you going to view a biological system as something you want to modulate as you begin to try to treat a particular condition? Why do you want to go after a particular target because you have a strong belief that the target has an important role in the disease in question. The most successful method to date is shown on the right-hand side here, phenotypic drug discovery. You set up a cell-based assay and add small molecules, potentially large molecules for that matter, and look for evidence of an effect on cells that you believe is going to translate to a beneficial effect on the clinic. At the opposite end of the spectrum, one is taking a molecular or targeted approach, where you believe you know enough about the disease and enough about the target biology to try to interfere directly with the action of the target in the human body and build molecules for that specific purpose and this is where a fragment based structure guided approach comes in. Typically using a variety of complementary biophysical tools some of which are a listed here. In, in their acronym form. So, to recap, Phenotypic drug discovery remains the most successful route. but there is tremendous, interest and effort going on in targeted, drug discovery, particularly the fragment based. discovery, which may or may not rely on a bi-, on biochemical screening as an in-, intermediate step. This, but the biochemical approach came earlier this decade and in the 90's with high throughput screening and large combinatorial libraries generally viewed as as having been over hyped shall we say? And typically today both the biochemical and the molecular are combined if one's going to take a targeted approach. So let's talk briefly about the causes of attrition in the pharmaceutical industry. This histogram coming from a paper in Nature Review's Drug Discovery, published by Kola and Landis in 2004 is instructive. Although the snapshots were taken at the beginning of the 90s and again in 2000, they still apply today. So if we look at the reasons why a drug discovery effort is going to fail in the pharmaceutical industry, you can see that there's a significant fraction of and as yet unaddressed fraction of the failures that comes due to lack of efficacy. An by that I mean you engage the, the target. You, you block the effect of the target in the body. Yet, it doesn't cure the disease, or it doesn't help you manage the disease. That's an reflection of the fact that one doesn't understand the target well enough, I would argue. You can see there are, are other causes of attrition where there have been dramatic, advances over the last, 20 or so years. And particularly in pharmacokinetics, and bio availability, and I'll come onto some reasons why I think that's the case. So getting the drug to the target is no longer the problem. The significant 40% attrition rate problem that it was 20, 30 years ago. The two key reasons for failure today are lack of efficacy and toxicity. And by toxicity I mean both animal toxicology, which you can see has actually increased between 1991 and 2000 and clinical safety, meaning human toxicology, which is this, remained essentially unchanged. combined their count for about a third, efficacy accounts for about a third and then there are a variety of other causes, some of them economic, such as cost of goods etcetera which are making up the remaining third. And the question that I'll address with you today is how medicinal chemists can help address this one third of attrition in the pharmaceutics industry that comes from toxicology be it a human or animal. So a statement of the problem here how many of you are familiar with the Lipinski's rule or have you heard of them at least. A, a modest show of hands, that's that's good. So, I, I bring up Lipinski because he and his colleagues demonstrated in the 1990s, using data from a large pharmaceutical company, that it was possible to make certain predictions about whether or not a small molecule was going to be cell-penetrant, was actually going to be orally bio available. And they come up with this rule of 5 that suggests that if your molecular weight's below 5, this measure of lipid felicity c log P is, greater than 5, and then you have an appropriate number of hydrogen bond donors and acceptors. that you're you're going to have a, a compound which is which is absorbed. So this led, this realization and the publication of a series of very influential papers led to the realization that there was this rule of five in the pharmaceutical industry. And that came unfortunately to be a, a rule for a drug. And that's, I'll show you that that's not the case. But bear with me for a moment as we sort of, think about the problem at hand. If you consider all of them, all of the small molecules of molecular weight less than 500, the ones that are going to cross the intestinal barrier readily, according to Lipinsky and co-workers, they contain the typical elements of carbon, nitrogen, oxygen, florine and of course, hydrogen. And, and you obey all of the rules of chemistry. The possible numbers of small molecules is 10 to the 60. Now that's larger than the number of atoms in the universe. So clearly, that's an unattainable number by any, by any technology. If you consider the subset of those 10 to the 60 compounds that are likely to have pharmacological activity, you've got an estimate of approximately the square root around 10, 10 to the 30. The estimate I'm providing here is 10 to the 26. The subset of those compounds that have pharmacologic properties that in turn have good properties that would qualify them as being drugs meaning absorption, distribution, metabolism, excreting and toxicology. The number is considerably smaller. It's probably around 10 to the 12, 10 to the 15. We don't have a very good feel for it. What is clear is that when medicinal chemists are working and they're working with the objective of reducing attrition in the pharmaceutical industry, they want to bias their, all of their design efforts to this green subset. And the question of course, is how do you get there.
