Fragment events in 2016 (updated)

The first FBDD-related meeting of 2016 has already come and gone, but there are still plenty of events ahead. Below are several updates as well as a new listing.

April 20-21: CHI’s Eleventh Annual Fragment-Based Drug Discovery, the longest-running fragment event, will be held in San Diego. You can read impressions of last year's meeting here, here, and here; the 2014 meeting here and here; the 2013 meeting here and here; the 2012 meeting here; the 2011 meeting here; and 2010 here. Also as part of this event, Ben Davis and I will be teaching a short course on FBDD over dinner on April 20.

May 24: Development of Novel Therapies through Fragment Based Drug Discovery will be held in Houston, Texas. Despite being only one day, it looks like a great lineup of speakers, so check it out!

June 7-10: Although not strictly fragment-focused, the third NovAliX Conference on Biophysics in Drug Discovery is likely to have lots of relevant talks, and is a good excuse to get to Strasbourg, France. You can read Teddy's impressions of the 2013 event here, here, and here.

July 12-15: FBDD Downunder 2016 will be held at Monash University in Melbourne. This is only the second such event; the first was lots of fun and even resulted in a special issue of the Aust. J. Chem., so definitely check this out if you can.

October 9-12: FBLD 2016 will be held in Boston, MA. This marks the sixth in an illustrious series of conferences organized by scientists for scientists, the last of which was in Basel in 2014. Surprisingly, this also seems to be the first dedicated fragment conference in Boston. You can read impressions of FBLD 2012, FBLD 2010, and FBLD 2009. Early-bird registration is now open!

November 7-9: Finally, the OMICS Group is holding their second Drug Discovery & Designing in Istanbul, Turkey, with a track on FBDD.

Know of anything else? Add it to the comments or let us know!

Fragments vs PRMT6

Epigenetics involves turning genes on or off without changing their sequence. This often relies on modifications to proteins or DNA that are recognized by other proteins. As Teddy pithily observed, this is a big field. However, in the realm of fragments, most of the attention has been on bromodomains; other classes of proteins, such as methyltransferases, have been largely neglected. A new paper in J. Med. Chem. by Masoud Vedadi and Matthieu Schapira and collaborators at the University of Toronto and Bayer suggests fragments are promising for these targets as well.

The researchers were specifically interested in protein arginine methyltransferases (PRMTs), which transfer a methyl group to one of the terminal side chain nitrogen atoms on specific arginine residues. PRMT6 in particular targets histone proteins to modulate transcription and has been implicated in cancer as well as neurodegenerative diseases. A few potent inhibitors have previously been reported for PRMTs, and the team started by deconstructing them to hunt for active fragments.

Ligand deconstruction involves chopping a known ligand into fragments to see whether any of these pieces will still bind. In this particular case, EPZ020411 had previously been characterized crystallographically bound to PRMT6 with the basic amine-containing “tail” in the substrate arginine-binding groove. Testing this fragment 6 by itself revealed a low micromolar inhibitor with a ridiculously high ligand efficiency.

Thus encouraged, the researchers ran a functional screen of 2040 diverse fragments (about half from Maybridge) at 1 mM concentration and retested hits at 0.5 mM. About half the resulting hits were false positives or other uglies, leaving the researchers with 14 fragments with IC₅₀ values from 0.3 – 400 µM. As might be expected given the cationic nature of the substrate, 12 of these have basic nitrogens.

Compound 7 was particularly interesting: at 300 nM this is one potent fragment! ITC revealed a dissociation constant of 970 nM, with a favorable enthalpy and unfavorable entropy of binding. It did hit other PRMTs, but was remarkably selective against a panel of 25 other human methyltransferases.

The researchers also determined the crystal structure of compound 7 bound to PRMT6, which revealed it binding, as expected, in the arginine site, making hydrogen bonds with a conserved catalytic glutamic acid. Weirdly though, it seems to be a noncompetitive inhibitor: increasing concentrations of substrate peptide or cofactor had no effect on inhibition. The team speculates that the noncompetitive behavior could be because the substrate makes strong interactions with the protein outside the arginine-binding site. Nonetheless, the fragment did inhibit PRMT6 activity in a cell assay with IC₅₀ = 21 µM.

Overall then it seems that the PRMTs are amenable to FBLD. They are interesting drug targets, and at the very least having more probes will help to unravel the biology.

Selectivity in STD

Among NMR-based fragment screening methods, saturation transfer difference (STD) came in as most popular in a recent poll. The technique is very sensitive and thus able to identify weak fragments. Unfortunately, it’s a bit too sensitive; hit rates of >30% are not uncommon. Many of these hits interact non-specifically with the protein. These can be weeded out using orthogonal screening methods or competition assays, but it would be nice to make STD itself more discriminating. In a paper published late last year in J. Med. Chem., Olivier Cala and Isabelle Krimm describe how to do just this.

In an STD experiment, the protein target is irradiated and transfers some of its magnetization to bound ligands. When these dissociate they retain some of the magnetization, and so the NMR signals of the fragments decrease. The problem is that lots of fragments interact non-specifically with proteins. For example, if a lipophilic fragment dances across greasy patches on a protein surface to escape from water without making any specific interactions, it will still get magnetized. Can such signals be distinguished from fragments that bind in a single, well-defined manner?

Within a ligand that binds to a protein, a proton that binds closer to the protein will show a stronger STD effect than one that is exposed to solvent. This is in fact the basis for STD “epitope mapping”, which allows one to roughly model how a ligand binds to a protein – or at least which parts of a ligand are closest to the protein. In the new paper, the researchers argue that simply observing differences in the STD effect between different protons in a ligand can distinguish whether or not that ligand is binding in a single binding mode.

Several examples support this assertion. For one protein, all the fragments that showed significant epitope mapping could be competed with a known reference molecule, suggesting binding to a specific site; this was less often the case for fragments that did not show epitope mapping. In another example, the privileged fragment 7-azaindole was found to bind to two different proteins with different epitope maps, suggesting different (but specific) binding modes for each protein. The technique also seems fairly robust to the affinity of the fragments (K_D 50 µM to > 1 mM), the details of the NMR experiment (saturation time from 0.5 to 4 seconds), and ligand/protein ratios between 66 to 1 and 400 to 1.

As the researchers note, there are caveats. For example, if a fragment can bind in two different but nonetheless specific binding modes, it may show uniform STD effects and will be a false negative. Nonetheless, comparing STD effects across a ligand does seem a worthwhile exercise. Not only could it help prioritize fragments, it could also reveal which protons are further from the protein, and so suggest growth vectors.

Dihydroisoquinolones as fragments

It’s a common problem: you find a fragment that binds to your target and want to grow it to improve affinity. A search for commercial analogs comes up empty, so you look into modifying the hit, only to discover that you’ve got a six-step synthesis on your hands. Or worse; perhaps there is no precedent at all. The chemical literature is replete with total syntheses of complicated natural products, but seemingly simple fragments are often not well-represented. Last year, researchers from Astex exhorted chemists to develop synthetic routes for attractive fragments, and in a recent paper in Org. Biomol. Chem. David Rees and colleagues take up their own challenge in the case of dihydroisoquinolones.

Dihydroisoquinolone itself is a nifty little fragment. It has just 11 atoms, cLogP = 1.0, and its solubility is > 5 mM in aqueous buffer. Its cis-amide moiety can serve as a hydrogen bond donor and acceptor, and the adjacent phenyl ring provides a bit of grease for interacting with hydrophobic protein residues.

The researchers built on existing methodology using a rhodium catalyst to introduce polar groups (such as hydroxymethyl and dimethylamino) at the R position. Depending on the nature of the R group, regioisomers in which the substituent ends up at the 4-position could sometimes also be isolated.

The methodology is robust and tolerates air, moisture, and various substituents. The alkene starting material is easy to come by, and the aromatic starting material is easy to make. By varying this, the researchers could generate 6- or 7- substituted dihydroisoquinolones, though 5- and 8- substituted versions seem harder to access. The team was also able to use other aromatics as starting materials, including thiophene, thiazole, and pyridine.

Thus, if dihydroisoquinolone comes up as a hit, this paper will allow you to quickly explore most of the vectors. So how often does this fragment show up? It is not clear why some fragments, such as 7-azaindole and 4-bromopyrazole, show up again and again, while others languish so lazily in the library that they might as well not even be there. We’ve highlighted at least one case where a dihydroisoquinolone was a useful hit.

Practical Fragments would love to know your experience. Do you have dihydroisoquinolones in your library? How often do they show up as hits? And what other fragments do you find that are in need of better synthetic routes for further exploration?

Fragment-Based Drug Discovery: Lessons and Outlook

In 2006, Wolfgang Jahnke and I co-edited the very first book on fragment-based drug discovery. Half a dozen books have followed, most of which have been reviewed at Practical Fragments (see right-hand column). These are now joined by a new book edited by Wolfgang and me in Wiley’s Methods and Principles in Medicinal Chemistry series.

At 500 pages and 19 chapters, this is the most extensive treatment since the Methods in Enzymology volume five years ago. In the interest of space I can’t write more than a sentence or two about each chapter, but I would like to thank all the contributors. Although I’m undoubtedly biased, I believe this work will set the standard for years to come.

The book is divided into three sections, starting with The Concept of FBDD. Rod Hubbard (Vernalis and University of York) opens with a chapter on the role of FBDD in lead-finding, which provides an introduction, historical overview, and summary of current thinking and future challenges. One particularly interesting section compares the contents of the 2006 book with the state of the art today, highlighting the fact that many of the basic techniques were already in place a decade ago, but the number of success stories has increased dramatically.

Chapter 2, by Glyn Williams and colleagues at Astex, discusses how to choose targets for FBDD, including concepts such as ligandability. Key principles are nicely illustrated with several important targets including the IAPs and HCV-NS3.

The last two chapters in this section focus more on numbers. Chapter 3, by Jean-Louis Reymond and colleagues at the University of Berne, covers the computational enumeration of chemical space, with a special emphasis on the contents and uses of their GDB-17 set of the 166 billion possible molecules with up to 17 non-hydrogen atoms. And chapter 4, by György Ferenczy and György Keseű at the Hungarian National Academy of Sciences, provides an overview of various metrics (such as ligand efficiency and LELP) and how these can be useful for fragment optimization.

The next nine chapters comprise the longest sub-section of the book, Methods and approaches for FBDD. To start screening fragments, you need a library, and designing one is the subject of chapter 5, by Martin Drysdale and colleagues at the Beatson Institute. This chapter also touches on concepts such as molecular complexity and “three-dimensional” fragments.

Screening techniques are best used in combination, and in chapter 6 Ben Davis (Vernalis) and Tony Giannetti (Google[x]) describe the synthesis of results from SPR, NMR, X-ray, ITC, functional screens, and other techniques to overcome challenges in several discovery programs. They emphasize that universal agreement among different methods is not always necessary, but carefully analyzing discrepancies can reveal unexpected problems with the screening conditions, target, or hits.

Differential scanning fluorimetry (DSF) – or thermal shift (TS) – is perhaps the most controversial screening method, and in chapter 7 Chris Abell and colleagues at the University of Cambridge cover this approach in depth. The chapter starts with a thermodynamically detailed yet nonetheless lucid discussion of the theory behind DSF, including the interpretation of negative thermal shifts. The chapter also includes plenty of practical advice and case studies, some of which we’ve covered briefly (for example here and here).

Chapter 8, by Sten Ohlson and Minh-Dao Duong-Thi at Nanyang Technological University, covers three emerging fragment screening technologies: WAC, native MS, and MST. And Chapter 9, by Sandor Vajda (Boston University) and collaborators, does an excellent job of summarizing computational approaches.

As others have noted, some of the biggest challenges are not technical but organizational, and in chapter 10 Michelle Arkin and colleagues at UCSF describe how to make FBDD work in academia. The chapter also includes some interesting polling data, concise but cogent summaries of fragment-finding techniques, and case studies on p97 and caspase-6. And in chapter 11, Jim Wells and colleagues – also at UCSF – describe using Tethering to find allosteric sites in proteins.

One area that has grown dramatically since 2006 is the use of FBDD in complex systems (such as membrane proteins), the subject of a chapter by Miles Congreve and John Christopher at Heptares. Chapter 12 also includes successful case studies, some of which we’ve covered. But finding fragments against these targets is still not easy, as illustrated in the final figure: of 18 fragment hits on 15 targets, almost all have ligand efficiency values > 0.3 kcal/mol per atom, and most of them are relatively potent, with affinities in the mid-micromolar range or better. While everyone wants to find strong binders from the start, such numbers suggest many weak-binding hits are overlooked.

Chapter 13, by Jörg Rademann and colleagues at Freie Universität Berlin, covers protein-templated fragment ligation methods, both reversible and irreversible. The chapter is wide-ranging and includes methods such as dynamic libraries and various types of “Click” chemistries.

The last section of the book, which was mostly absent a decade ago, is entitled Successes from FBDD. This starts with a chapter by Daniel Wyss, Andrew Stamford, and colleagues from Merck on BACE inhibitors. As we’ve noted, fragments have had a major role in most of the BACE inhibitors to enter the clinic, with phase III results from Merck’s verubecestat expected next year.

Epigenetics has also been strongly influenced by fragments, and in chapter 15 Aman Iqbal (Proteorex) and Peter Brown (Structural Genomics Consortium) survey the field, with case studies on several proteins that modulate epigenetic marks. These include BRD4, ATAD2, BAZ2B, SIRT2, and others.

One of the original selling points of fragment-based methods is the ability to go after difficult targets such as protein-protein interactions, and this is the subject of chapter 16, by Feng Wang and Stephen Fesik (Vanderbilt University). In addition to general guidelines, the researchers describe a number of case studies, including RPA, MCL-1, and K-Ras.

Some enzymes can be just as difficult as protein-protein interactions, and in chapter 17 Alexander Breeze (University of Leeds) and former AstraZeneca colleagues describe programs to find inhibitors of LDHA (see here and here). They also discuss how some previously reported inhibitors turned out to be artifacts.

More than two dozen kinase inhibitors have been approved by the US FDA, including the first drug derived from FBDD. In chapter 18, Gordon Saxty (Fidelta) surveys a number of kinase programs, including most of the fragment-derived inhibitors in clinical trials.

And finally, in chapter 19 Simon Rüdisser and colleagues from Novartis present an extensive discussion of renin, with special attention to their campaign, which involved a combination of HTS and fragment-based approaches.

While it may not be possible to judge a book by its cover, the cover of this book does illustrate some of the fruits of the field, with structures of three fragment-derived drugs that have entered the clinic. These are just a small fraction of the 30+ drugs working their way through the pipeline, and of the many more that will spring from the research described and informed by the work presented.