17 October 2016

FBLD 2016

Last week the sixth FBLD meeting was held in Cambridge, MA. Like its predecessors in 2014, 2012, 2010, 2009, and 2008, this meeting was an enormous success, mixing more than 230 scientists with excellent (and liberal) food and drink. With 33 talks, more than 30 posters, and several vendor booths and workshops I won’t be able to do more than capture a few highlights.

The most striking feature for me was the number of success stories. This began with Steve Fesik’s keynote lecture, in which he discussed the MCL-1 inhibitors he and his team at Vanderbilt have discovered. When we highlighted his work last year he had reported low nanomolar inhibitors, but these did not have cell-based activity. His group has now optimized the molecules to low picomolar biochemical potency, low nanomolar cellular activity, and good activity in mouse xenograft models. This has not been easy: more than 2210 compounds were made, guided by 60 X-ray structures and dozens of pharmacokinetic experiments. It seems to be paying off though, and the researchers are developing biomarkers with the goal of advancing a compound into clinical testing.

Two other notable success stories about clinical candidates must be mentioned, though I’ll wait until publications come out before going into detail. Kathy Lee described how she and her colleagues at Pfizer chose a fragment that was less potent and ligand-efficient than other hits due to its interesting binding mode and were able to advance it to PF-06650833, an IRAK4 inhibitor with potential for inflammatory diseases. And Wolfgang Jahnke discussed how he and his colleagues at Novartis were able to discover and advance ABL001, an allosteric inhibitor of BCR-ABL, despite having the project halted twice – a reminder that persistence is essential.

Several other success stories have been covered at least in part on Practical Fragments, including inhibitors against PDE10A (presented by Izzat Raheem of Merck), Dengue RNA-dependent RNA polymerase (presented by Fumiaki Yokokawa of Novartis), lipoprotein-associated phospholipase A2 (presented by Phil Day of Astex), and BACE1 (presented by Doug Whittington of Amgen).

Crystallography was another theme, and several of the success stories relied on crystallographic fragment screening. Frank von Delft of the Structural Genomics Consortium described developments that allow screening 1000 crystals per week at Diamond’s Xchem facility in the UK, which include acoustic dispensing of compounds into crystallization drops – while carefully avoiding hitting the crystals head-on.

Several computational talks reported results that run contrary to conventional wisdom. Vickie Tsui of Genentech discussed their CBP bromodomain program (which we recently discussed here). Several water molecules form a highly ordered network in the protein, and a WaterMap analysis suggested that these were high-energy and that displacing them would lead to an enhancement in activity. Unfortunately this turned out not to be the case, though the researchers were able to get to low nanomolar inhibitors by growing towards a different region of the protein.

Li Xing mined the Pfizer database of 4000 kinase-ligand structures to extract 595 unique hinge binders. Not surprisingly, some of these – such as adenine and 7-azaindole – bound to multiple kinases, but 427 were complexed to just a single kinase. Hinge binders typically form 1 to 3 hydrogen bonds to the protein, and while there didn’t seem to be a correlation between the number of hydrogen bonds and potency, more hydrogen bonds did correlate – perhaps counterintuitively – with lower selectivity. To the extent that hydrogen bonds are thought of as enthalpic interactions, this further muddies the argument that enthalpy and entropy can be useful in drug design.

On a more positive note, Sandor Vajda (Boston University) suggested that, according to analyses done in FTMap, perhaps 60-70% of protein-protein interactions may be druggable – as long as we accept that this may require building larger molecules than commonly accepted. And Chris Radoux (Cambridge Crystallographic Data Centre) discussed the computational tool for characterizing hotspots that we previously covered here; a web server for easy search should be available soon.

Library design was also a key topic. Richard Taylor of UCB described his analysis of all FDA-approved drugs, which revealed >350 ring systems. Interestingly though, 72% of drugs discovered since 1983 rely exclusively on ring systems used prior to that date. Clearly there is plenty of untapped chemical real estate.

But getting there won’t necessarily be easy. David Rees stated that 33 fragments recently added to the Astex library required 13 different reaction types. Importantly, many of the fragment to lead successes at Astex have required growing the fragment from the carbon skeleton rather than from more synthetically tractable heteroatoms. Knowing in advance how to do this with every new member of a fragment library should make life much easier in the long run, though it is a serious challenge for chemists.

There is far more to write about, including a great discussion led by Rod Hubbard on how FBLD is integrated effectively into organizations and how it enables difficult targets, but in the interest of space I’ll stop here. If you were at FBLD 2016 (or even if you weren’t) please share your thoughts!

10 October 2016

Tips for high-throughput crystallography

X-ray crystallography is tied for second place among methods used in fragment-based lead discovery, according to our most recent poll. This makes sense, since structures are usually essential for advancing fragments to leads. Faster fragment-finding methods are usually used to triage fragments down to a manageable number of hits to feed into crystallography, but the high incidence of false negatives means that promising fragments might be inadvertently discarded. If structures are key goals at the end of a fragment screening campaign, why not start directly with crystallography?

In fact, this is exactly what more and more groups seem to be doing. The problem, historically, has been throughput. Increasing automation has been solving some of the mechanical issues (such as mounting crystals and collecting data at a synchrotron), but what about the actual processing? A recent paper in Structure by Andreas Heine and collaborators at Philipps-University Marburg and Helmholtz-Zentrum Berlin für Materialien und Energie provides some useful advice.

The protein in question is endothiapepsin, a model aspartic protease that is easy to crystallize and diffracts to high resolution. Earlier this year, we discussed the researchers’ work soaking 360+ fragments against this protein, and a companion paper gives detailed information on how several dozen fragment hits bind. The Structure paper describes an automated refinement pipeline, and highlights some of its most important features.

Determining a crystal structure involves iterative cycles of modeling the protein backbone and side chains into regions of “electron density.” One risk is “model bias,” illustrated memorably in this brief video. This is especially important for small molecules: since they represent such a tiny fraction of the overall structure, it is especially easy to see what you want to see. To avoid this, people often look for regions of electron density – which in addition to a bound small molecule could represent co-solvents, buffer, or an amino acid side chain that has unexpectedly moved – before doing much refinement.

The problem is that the electron density might be very spotty and easy to overlook. This is especially true for fragments that bind weakly and which are small by definition. Some initial refinement can thus improve the quality of the electron density maps. The researchers find that adding water molecules and including these in the refinement is the single most important step. Adding bound hydrogen atoms to the protein model is also helpful: even though each hydrogen only contributes one electron to the overall density, there are more than enough to make a meaningful difference. Finally, for very high resolution structures (better than 1.5 Å), it can help to treat each atom of the protein individually (anisotropic refinement of B factors, or atomic displacement parameters). However, at lower resolution, doing this can lead to overfitting. Incorporating these steps into the automated process revealed that 25% of fragments would have been missed had conventional methods been used.

The paper includes lots more detail that will be of interest primarily to crystallographers. Moreover, the data for all 364 fragment soaks has been uploaded to the protein data bank. This is a very high-quality data set: all the crystals diffracted to better than 2.0 Å resolution, with the mean being 1.35 Å, and should be a useful resource for those of you establishing your own automated processing system.

03 October 2016

Poll results: affiliation, metrics, and fragment-finding methods

The latest poll has just closed, and the results are quite interesting – I’ll get to these in the next paragraph. First, a quick note on methodology. The poll ran from August 27 through September 30. Due to issues with polling in Blogger, we began running polls in Polldaddy in 2013; its interface gives the total number of votes for a question but not the number of individual respondents. Thus, for the questions on metrics and methods, I assumed that the number of respondents was equal to the number of people who identified themselves as practicing FBLD in the first question, or 123 out of a total of 154. The true percentages for the metrics and methods that people use could be higher or lower if not everyone answered all the questions.

Readership demographics have been remarkably stable since 2010 and 2013, with just over half of respondents from industry, and around 80% of all respondents actively practicing FBLD.

The next question asked about screening methods, and here things get more interesting.

The first thing to notice is that, as we also saw in 2013, nearly all fragment-finding techniques are being used more, with the average user employing 4.1 distinct methods today compared with 3.6 in 2013 and 2.4 in 2011. Ligand-detected NMR has jumped to first place in terms of popularity, with SPR and X-ray crystallography tied for second, followed closely by thermal shift. MST, while still in the minority, has had the largest percentage increase. The use of crystallography has certainly jumped since 2011, which fits with recent publications.

Finally, with regards to metrics, ligand efficiency (LE) continues to dominate, followed by LLE (or LipE), though overall usage of both is down compared with 2014. Only one of the other metrics broke the 10% mark. 
Again, if some practitioners answered the first question of the poll, but not the next two, the use of all methods and metrics could be underestimated. Still, these results seem to fit with what I’ve heard talking with folks – any surprises?

26 September 2016

Fragments vs DOT1L, two ways

This past July Practical Fragments was devoted almost entirely to bromodomains, an important type of epigenetic protein. Protein lysine methyltransferases (PKMTs) are another significant class: 51 human enzymes that transfer a methyl group from the cofactor S-adenosylmethionine (SAM) to the side chain amine of lysines, typically in histones. In two recent papers in ACS Med. Chem. Lett., researchers from Novartis describe how they discovered inhibitors of DOT1L, a target for certain leukemias.

The first paper, by Frédéric Stauffer and colleagues, started with a fragment screen of the DOT1L catalytic domain using surface plasmon resonance (SPR). This led to the discovery of compound 1, which is teetering on the edge of molecular obesity (at least for a fragment) but did show activity in a functional assay as well as binding by NMR. Moreover, a co-crystal structure revealed that it binds in a new pocket near the SAM binding site, primarily through hydrophobic and stacking interactions.

Replacing the potentially unstable pyrrole with a quinoline led to compound 3, and subsequent structure-based design led to compound 5. Interestingly, while the methoxy substituent on compound 5 was installed to form a hydrogen bond with the protein, this instead caused a shift in binding mode – a reminder that fragments don’t always retain their original orientations during optimization. This new binding mode provided a vector to grow through a narrow channel into another pocket, ultimately resulting in compound 8, with low nanomolar activity.

The second paper, by Christoph Gaul and colleagues, started with a high-throughput screen (HTS). One low micromolar hit turned out to be a (felicitous) regioisomeric impurity from a commercial supplier. Crystallography revealed that this binds in the same pocket as the fragment in the previous paper, and subsequent medicinal chemistry led to low nanomolar inhibitors such as compound 3’. Unfortunately these turned out to have low permeability, probably due to the high number of hydrogen bond donors and acceptors. Fragmenting compound 3’ led to compound 4’, with a dramatic loss in potency, but structure-based design ultimately led to potent molecules such as compound 12’. This compound is also selective against other PKMTs, cell active, and orally bioavailable in rats.

These two papers provide a nice window into the complexity of lead discovery. In contrast to other examples, the fragment made largely hydrophobic interactions, while the HTS hit made numerous hydrogen bonds. Both hits bound in a new pocket, a reminder that secondary ligand binding sites are common. And in both cases, extensive medicinal chemistry was necessary and led to molecules that scarcely resemble their starting points. Interestingly, a previously described clinical candidate against this target, EPZ-5676, was identified by yet another approach: structure-based design starting from the cofactor SAM. All of which is to say that there are lots of ways to find inhibitors, and they don’t always fall into neat categories.

19 September 2016

Fragments vs GSK3β via DOS

Diversity-oriented synthesis, or DOS, enables the rapid and systematic synthesis of multiple related compounds from small sets of molecules and reactants. By creatively choosing the chemistry, DOS practitioners can selectively generate all diastereomers and produce more complicated molecules than are usually found in commercial screening collections. While much of the attention has been focused on larger molecules, DOS offers clear applications for addressing the chemistry challenges of FBLD. This is illustrated nicely by a recent paper in ACS Med. Chem. Lett. by Alvin Hung, Damian Young, and collaborators at the Broad Institute, Harvard, the Albert Einstein College of Medicine, A-STAR, and Baylor College of Medicine.

The researchers started with a very small (86 fragment) library, which Damian is in the process of expanding to 3000 compounds. Differential scanning fluorimetry was used to screen the molecules against the kinase GSK3β, which is implicated in cancer and Alzheimer’s disease. Three related fragments slightly increased the melting temperature of the enzyme, of which the simplest was compound 1S.

One nice feature of DOS is that – by design – analog synthesis is straightforward. Thus the researchers made a dozen or so derivatives to flesh out the SAR. This revealed that the enantiomer, compound 1R, stabilized the protein even more than the initial hit. STD and WaterLOGSY NMR confirmed binding, and isothermal titration calorimetry (ITC) revealed modest but measurable affinity. Synthesis of a few additional analogs led to compound 15R, with low micromolar affinity as assessed both by ITC and an enzymatic assay. Ligand efficiency was also good, though the ligand efficiency by atom number (LEAN) values of the molecules do not quite meet Teddy’s Safran Zunft Challenge – a wager due to be settled at FBLD 2016 in a few weeks.

A key selling point of DOS is that, by accelerating chemistry, it enables optimization even without structural information. In this case the researchers suspected that the fragment binds in the hinge region of the kinase, and subsequent crystallography revealed that this was indeed so. Interestingly though, the quality of the crystal structure was insufficient to unambiguously place compound 1R; perhaps it binds in multiple conformations. The crystal structure of compound 15R, on the other hand, was clear.

Of course, there is still a long way to go for this series, and it remains to be seen how broadly applicable DOS will be for FBLD. I look forward to seeing additional examples.

12 September 2016

Improving FBLD at AstraZeneca

FBLD started early at AstraZeneca (AZ). The first conference Practical Fragments covered was held at their erstwhile Alderley Park site; Pete Kenny is an AZ alum, and the company has put at least four fragment-derived drugs into the clinic. Clearly their scientists have learned plenty about what works and what doesn’t, and much of this wisdom is distilled into an excellent recent review in Drug Discovery Today. The authors include Nathan Fuller, Joe Patel, and Lorena Spadola, the first two of whom are organizers for the upcoming FBLD 2016 – for which there is still just barely time to register.

Things weren’t easy in the beginning: of the 63 FBLD targets screened between 2002 and 2008, a mere 10% led to tractable lead series with interpretable SAR. This improved to 37% of the 19 campaigns conducted between 2009 and 2011, and to 64% of the 11 projects between 2012 and 2014.

What accounts for these improvements? Target selection certainly played a role. In earlier years many targets were dropped due to portfolio reasons or lack of validation – nearly half for the period 2009-2011. Often FBLD was tried in desperation when all else failed, and chemists were not always available for fragment-to-lead efforts. Today, fragment screening is considered for all water-soluble targets at AZ, and fully integrated teams are brought into the process earlier. In 2012 the company established a team of medicinal chemists dedicated to FBLD – a strategy that has also been used at other companies.

But many of the improvements are technological rather than organizational. Biophysical screens are displacing high-concentration biochemical screens, which are particularly prone to false positives and false negatives. 1D and 2D NMR remain mainstays, but SPR and X-ray crystallography are increasingly being used in primary screens.

Another major effort was revamping the fragment library, which currently stands at 15,000 members. Each fragment was experimentally confirmed to be soluble to at least 0.5 mM in water and 100 mM in DMSO, and the rule of three was used more as a guideline than a rule. The collection was designed to include a good proportion of “three-dimensional” fragments, as assessed by plane of best fit (PBF) and principal moment of inertia (PMI). About a quarter of the fragments are proprietary, and the company also has another 750,000 molecules within their corporate collection that could be classified as fragments, greatly facilitating follow-up studies.

A 15,000 member library is atypically large, but in practice smaller subsets of the library are deployed: 384 for crystallographic screening, 1152 for NMR screening, and 3072 for SPR screening. Each subset is optimized for the technique. For example, because the crystallographic subset is so small, it is designed to sample chemical space as efficiently as possible. This is done by maximizing the diversity of the fragments and choosing the smallest fragments possible – less than 17 non-hydrogen atoms, as at Astex. In contrast, the NMR and SPR subsets contain fragments having up to 21 non-hydrogen atoms, and the SPR set also contains close analogs of some fragments to improve confidence and provide preliminary SAR. There is some overlap between the sets to facilitate confirmation; for example, a 768-member “ligandability set” is shared between the NMR and SPR screening libraries. Finally, AZ has built a customized set of 800 covalent fragments.

For the most part, fragment hits from each subset tend to have similar properties as the subset in general, suggesting that each sub-library is well-suited for its technique. Importantly, this is true even for three-dimensional fragments, which comprise nearly half of the hits across 19 targets. The researchers also examined how effectively fragments were able to fill the volume of a given binding pocket for five targets with multiple crystal structures. They found that shapely fragments were at least as good as – and sometimes better – at filling the pockets, even with fewer three-dimensional fragments.

Finally, the article summarizes eight projects in which fragment hits were progressed. Dissociation constants for the hits ranged from 50 to 3230 µM; these were advanced to leads with affinities ranging from 1.5 to 180 nM. In half these cases the ligand efficiency improved, and in all cases the three dimensionality increased as defined by PBF. Two of the targets, phosphoglycerate dehydrogenase and mInhA, are discussed in some detail, complete with chemical and crystal structures. Hopefully all will be covered more fully in upcoming publications.

There’s lots more in this paper than I can summarize in a blog post, including multiple figures and tables, so definitely check it out.

27 August 2016

2016 polls!

We're heading into election season here in the United States, which reminds us that we haven't run any polls recently at Practical Fragments. How has the community changed in the past few years? To find out, please answer the three questions in the poll on the right-hand side of the page, under "Editors." Also, please note that you need to hit "vote" for each question separately.

The first question asks whether you are in academia or industry and whether you practice FBLD.

The second question asks what methods you use to find fragments. For purposes of this poll please choose all that apply, whether primary or secondary screens. You can read about these methods in the following links.

Affinity chromatography, capillary electrophoresis, or ultrafiltration
BLI (biolayer interfermotry)
Computational screening
Functional screening (high concentration biochemical, FRET, etc.)
ITC (isothermal titration calorimetry)
MS (mass spectrometry)
MST (microscale thermophoresis)
NMR – ligand detected
NMR – protein detected
SPR (surface plasmon resonance)
Thermal shift assay (or DSF)
X-ray crystallography
Other – please specify in comments

The third question asks what metrics (listed below) you use. Again, you can choose multiple answers.

Antibacterial efficiency
BEI (binding efficiency index)
Enthalpic efficiency 
FQ (fit quality)
GE (group efficiency)
LE (ligand efficiency)
LELP (ligand-efficiency-dependent lipophilicity)
LLE or LipE (ligand lipophilic efficiency)
PEI (percentage efficiency index)
SEI (surface-binding efficiency index)
SILE (size-independent ligand efficiency)

Finally, are there other topics you'd like to see polled? Please let us know in the comments.

22 August 2016

Crystallographic screening of a nuclear receptor

Crystallography as a primary screen seems to be gaining traction. As the old cliché goes, a picture is worth a thousand words. And as Andrey Grishin recently commented on an earlier post, the increasing speed and capacity at synchrotrons lowers the barrier for data collection. A new paper in ChemMedChem by Yafeng Xue and colleagues at AstraZeneca provides yet more support for starting with crystallography.

The researchers were interested in the retinoic-acid related orphan receptor γt (RORγt), a potential target for autoimmune diseases. The protein is a nuclear hormone receptor, and like many members of this family, ligands tend to be lipophilic with poor physical properties. Also, work by other companies around this target had created a thicket of intellectual property claims. To find new and attractive chemical matter, the researchers turned to fragments.

The ligand binding domain of RORγt was crystallized and soaked against a library of 384 fragments chosen on the basis of maximum diversity and previous success in crystallography. Fragments were screened at 75 mM concentration in pools of four, with members chosen to have different shapes. This process did require “extensive optimization”, and even then about 15% of the datasets were not usable. But the effort paid off, resulting in 21 hits from 18 pools. Hits were then tested by SPR, revealing that the best had an affinity of just 0.2 mM (though with an impressive LE of 0.42 kcal mol-1 per heavy atom), while some were > 5 mM.

As expected, many of the fragments bound in the large and lipophilic ligand binding pocket, accessing various binding modes previously seen with other ligands. This is a nice confirmation that fragments are able to sample chemical space very efficiently, as shown five years ago for HSP90. Indeed, for one particularly productive pool, three of the fragments bound simultaneously at different subsites within the ligand binding pocket!

Of course, proteins are often highly dynamic in solution, and one concern with crystallographic screening is that the protein crystals may not allow much movement. In this case the researchers did observe several cases of induced fit, with one side chain residue shifting more than 3 Å to accommodate a fragment. This revealed a type of interaction that was not predicted using a computational approach: a victory – for now – for the power of empiricism.

As discussed earlier this year, secondary ligand binding sites appear to be common, and indeed five fragments bound outside the ligand binding pocket. Three of these bind at what seems to be a protein-protein interface for other receptors, which could lead to highly selective molecules.

It’s a long way from a 0.2 mM fragment to a useful lead series, but having a structure (or 21) dramatically improves the odds – as demonstrated here and here. The paper ends by suggesting that such a series has indeed been identified, and it will be fun to watch as the story unfolds.

15 August 2016

Dynamic combinatorial chemistry and fragment linking

Dynamic combinatorial chemistry (DCC) sounds incredibly cool. The idea is that libraries spontaneously form and reform. Add a protein and Le Châtelier's principle favors the formation of the best binders. In other words, not only does cream rise to the top, more cream is actually created.

The applications of DCC for fragment linking are obvious, and indeed early reports date back nearly twenty years to the dawn of practical FBDD. The latest results are described in a new paper in Angew. Chem. Int. Ed. by Anna Hirsch and collaborators mostly at the University of Groningen.

The researchers were interested in the aspartic protease endothiapepsin, which is a model protein for more disease-relevant targets. This is a dream protein: it is easy to make in large amounts, crystallizes readily, and is stable for weeks at room temperature. Readers will recall that this protein has also been the subject of multiple screening methods. Previous efforts using DCC had generated low micromolar inhibitors such as 1 and 2. These acylhydrazones form reversibly from hydrazides and aldehydes. Crystallography had also previously revealed that compound 1 binds in the so-called S1 and S2 subsites of endothiapesin while compound 2 binds in the S1 and S2’ subsites. In the current paper, the researchers enlisted DCC to try to combine the best of the binding elements.

To do this, the researchers chose isophthalaldehyde, which contains two aldehyde moieties, and nine hydrazides, which could give a total of 78 different bis-acylhydrazones. They incubated 50 µM of isophthalaldehyde with either four or five of the hydrazides (each at 100 µM), with or without 50 µM protein, and in the presence of 10 mM aniline to accelerate the exchange. Reactions were allowed to incubate at room temperature at pH 4.6 for 20 hours, after which the protein was denatured and the samples were analyzed by HPLC to see whether some products were enriched in the presence of protein.

Biologists may want to consider whether their favorite proteins would remain folded and functional under these conditions, and chemists may also balk at molecules containing an acylhydrazone moiety – let alone two. Leaving aside these concerns, though, what were the results?

As one would hope, some molecules were enriched over others when protein was present, though only by a modest two or three-fold. Two of the enriched molecules – both homodimers – were resynthesized and tested. Compound 13 was quite potent, and crystallography revealed that it binds in a similar fashion to compound 1, though electron density is missing for part of the molecule. Compound 16, on the other hand, is only marginally more potent than the starting molecules. Unfortunately the researchers do not discuss the activities of molecules that had not been enriched at all.

The paper ends by stating rather hopefully that DCC “holds great promise for accelerating drug development for this challenging class of proteases, and it could afford useful new lead compounds. This approach could be also extended to a large number of other protein targets.”

I’m not so sure.

This is an interesting study; the work was carefully done and thoroughly documented—but I’m less sanguine about whether DCC will actually ever be practical for lead generation. Indeed, the very fact that the experiments were done well yet are incapable of distinguishing a strong binder from a weaker one argues that the technique is inherently limited. I would love to see DCC work, but it seems to me that, even after two decades of effort, DCC has not been able to move beyond proof of concept studies. Does anyone have a good counterexample?

08 August 2016

Metallophilic fragments revisited

Way back in 2010 we highlighted work out of Seth Cohen’s lab at UC San Diego on “metallophilic fragments”, which are specifically designed to bind to metal ions. As long as one avoids PAINS, the approach could be useful for targeting metal-dependent enzymes. Indeed, multiple drugs derive much of their affinity by binding to metals; these include HDAC inhibitors (for cancer) and integrase inhibitors (for HIV). In a recent paper in J. Med. Chem., Cohen and colleagues describe work against an influenza target.

The researchers were interested in the so-called “PA subunit” of RNA-dependent RNA polymerase, which is both essential and highly conserved among influenza strains. The endonuclease in the PA subunit requires two metal ions, either Mn2+ or Mg2+, and in fact previous publications had demonstrated that metal chelators could inhibit the enzyme. In the current paper, the team screened about 300 fragments at 200 µM in an activity assay; those that inhibited >80% were retested to produce dose-response curves. Compound 1 came in as reasonably potent and impressively ligand-efficient, as is often the case with metal-binding fragments. Docking studies suggested that it could bind to both of the metal ions in the active site.
Initial SAR around compound 1 led to compound 10, with a significant improvement in potency that the researchers attribute to increased basicity and thus stronger interactions with the metals. Taking pieces from previously published molecules led to another increase in potency (compound 63). Separate fragment growing efforts off compound 1 led to sub-micromolar inhibitors such as compound 35. Combining both series led to compound 71, which is the best of the bunch with low nM activity, though it fell short of the hoped-for additivity of binding energies.

Compound 71 was also tested in cellular assays. Happily, it was able to protect cells from a lethal dose of influenza virus with an EC50 in the low micromolar range, about 100-fold below the cytotoxic dose observed in the same cell line. Of course, there is still a long way to go: no pharmacokinetic data are provided, and selectivity against other metalloproteins may be a challenge. Still, it will be interesting to watch future developments, both with this series and with the approach in general.

01 August 2016

Lead Generation: Methods, Strategies, and Case Studies

Lead generation refers to that point in drug discovery when initial screening hits against a target are wrought into compelling chemical matter. This chemical matter is often plagued with deficiencies in terms of potency, pharmacokinetics, or novelty, yet it provides a starting point for further optimization. This is the subject of a massive (800+ pages!) new two-volume work edited by Jörg Holenz (GlaxoSmithKline, formerly AstraZeneca) as part of Wiley’s Methods and Principles in Medicinal Chemistry series. Readers of this blog will not be surprised to find that fragments play a major role; indeed, the molecule on the cover of the book came out of FBLD. I won’t attempt to summarize all 25 chapters here, but will simply highlight those most relevant to FBLD.

Mike Hann (GlaxoSmithKline) sets the stage in chapter 1 by briefly describing the characteristics of successful leads. He emphasizes the importance of physicochemical properties and avoiding molecular obesity, and how judicious use of metrics can help navigate away from perilous chemical space. He also summarizes internal programs that again demonstrate that fragment-derived leads tend to be smaller and less lipophilic than those from other lead discovery techniques.

In chapter 3, Udo Bauer (AstraZeneca) and Alex Breeze (University of Leeds) discuss the concept of ligandability – the ability of a target to bind to a small molecule with high affinity. Fragments are ideally suited for assessing ligandability, and the researchers briefly describe fragment-based experimental and computational approaches to do so. They also include a nice 11-point summary of factors to consider when starting lead generation on a new target, ranging from the presence of small-molecule binding sites to the number of patent applications.

Chapter 6, by Ivan Efremov (Pfizer) and me, is entirely about fragment-based lead generation. I'm undoubtedly biased, but I think it provides a self-contained and fairly detailed guide to FBLD, including topics such as screening methods, hit validation, metrics, hit optimization, fragment growing vs fragment linking, and case studies on vemurafenib, BACE, MMP-2, LDHA, venetoclax, MCL-1, and GPCRs.

Helmut Buschmann and colleagues at RD&C Research, Development, and Consulting, focus in chapter 9 on optimizing side effects of known molecules to develop new drugs, but they also discuss some interesting older work reporting that 418 of 1386 drugs contain other drugs as internal fragments.

Chapter 12, by Dean Brown (AstraZeneca), is devoted to the hit-to-lead stage, and much of his advice is applicable to FBLD. Dean also includes a fantastic metaphor to illustrate the size of chemical space: "if a typical corporate screening collection were to fit on a postcard, the rest of the earth is the amount of available drug-like space." This assumes a million-compound library and a conservative estimate of 1023 drug-sized molecules, so if anything it is an understatement.

Molecular recognition is critical for both FBLD and lead generation in general, and this is the topic Thorsten Nowak (C4X Discovery Holdings) tackles in chapter 13. He covers key areas such as thermodynamics, emphasizing the importance of enthalpy while acknowledging the difficulty of prospectively using thermodynamic data. The role of water and halogen bonds are covered, along with some freakishly high ligand efficiency values. There are a couple errors: one paper is categorized as using dynamic combinatorial chemistry when in fact it actually used static libraries, and Tethering is confused with Chemotype Evolution, but overall there's lots of good stuff here.

Biophysical methods are covered in chapter 14, by Stefan Geschwindner (AstraZeneca). These include NMR, SPR, ITC, thermal shift assays, native mass spectrometry, microscale thermophoresis, and more.

Chapter 16, by Ken Page and colleagues at AstraZeneca, discusses "lead quality." This often entails various metrics, from simple ones such as ligand efficiency and LLE to more complicated attempts to predict clinical dosages. Although it is easy to poke fun at metrics, most thoughtful scientists find them useful for making sense of the reams of data generated in lead optimization campaigns.

Chapter 17, by Steven Wesolowski and Dean Brown (both AstraZeneca), is arguably the most entertaining. Entitled "The strategies and politics of successful design, make, test, and analyze (DMTA) cycles in lead generation," it is replete with pithy quotes and even an original (and highly geeky) cartoon. Along with multiple examples, the chapter formulates plenty of questions to consider during lead optimization, and ends with a particularly relevant quote by Billings Learned Hand: “Life is made up of a series of judgments on insufficient data, and if we waited to run down all our doubts, it would flow past us.”

In chapter 23, Sven Ruf and colleagues at Sanofi-Aventis Deutschland describe a success story generating leads against cathepsin A, a target for cardiovascular disease. HTS yielded three different chemical series with sub-micromolar activities, each with different liabilities. Crystallography revealed their binding modes, and this allowed the team to mix and match fragments across the different series to generate a molecule that ultimately went into the clinic. Although this may not be classic FBLD, it does seem to be a good case of using concepts from the field, or fragment-assisted drug discovery.

A similar, if less directed, approach is the subject of chapter 25, the last in the book. Pravin Iyer and Manoranjan Panda (both AstraZeneca) describe "fragmentation enumeration," in which known drugs or clinical candidates are fragmented into component fragments and recombined. On some level the fragments themselves are likely to be privileged; the researchers cite the famous quote by Sir James Black that "the most fruitful basis of the discovery of a new drug is to start with an old drug." Most of the work is computational, although one molecule derived from the approach has encouraging cellular activity against Mycobacterium tuberculosis.

There's far more to this book than could be listed even in this relatively long post, including multiple case studies, so for those of you who are interested in lead generation definitely check it out!

25 July 2016

Multiple bromodomains, multiple methods, and even more fragment hits

All this month Practical Fragments has been focused on bromodomains, highlighting chemical probes against BRD9, CBP and EP300, and family VIII bromodomains. Today’s post covers three earlier-stage programs on three different bromodomains.

In Acta Pharm. Sinica, Bing Xiong, Nai-xia Zhang, and colleagues at the Chinese Academy of Sciences discuss their work on BRD4, an anti-cancer target about which we’ve written previously. The researchers describe the construction of a fragment library designed for NMR screening; this is a good resource for people undertaking similar efforts. Interestingly, of 800 compounds purchased, only 539 were soluble to at least 100 µM in aqueous buffer. These were pooled into 56 groups of 8-10 compounds and screened at 200 µM (total fragments) using STD and T1ρ. This yielded 10 hits, of which three had measurable IC50 values from 110 to 440 µM. Five of the hits were characterized in more detail using two dimensional NMR (1H-15N HSQC), and three by X-ray crystallography. Some of these fragments are less-precedented as bromodomain ligands, and could be useful starting points for further work.

In contrast to BRD4, for which multiple ligands have been reported, the bromodomain on BRPF1 is less explored. In a recent paper in J. Med. Chem., Jian Zhu and Amedeo Caflisch (University of Zürich) provide 20 new co-crystal structures, all of which have been deposited in the protein data bank. The researchers performed a computational screen of 24,133 molecules using a program called SEED, which was able to crank through the entire set in just a day. Crystal soaking was attempted with thirteen of the top 30 hits, resulting in five structures, of which three bound in the manner predicted. Crystal structures of another 15 analogs and other bromodomain inhibitors were also determined. Some of the molecules are reasonably potent, with double-digit micromolar affinities and good ligand efficiencies.

Finally, while most bromodomains have a conserved asparagine residue that makes hydrogen bonds to the substrate (or inhibitor), 13 of the 61 known human bromodomains do not, and these tend to be more difficult targets. The second bromodomain of the pleckstrin homology domain-interacting protein (PHIP(2)), which has been implicated in melanoma, is one of these “atypical” bromodomains. Researchers at the Structural Genomics Consortium (SGC) led by Frank von Delft (Diamond Light Source) and Paul Brennan (University of Oxford) took a crystallography-first approach toward this target, as they report in an open-access paper in Chemical Science.

The researchers started by assembling what they call a “poised fragment library”. This is essentially a library designed for rapid follow-up chemistry, in which each library member can be deconstructed into individual components, which can be systematically varied. For example, a fragment might consist of two moieties connected by an amide bond, so that analogs could be easily made using parallel synthesis. The initial 2347 fragments were a subset of the 11,677 fragments available in-house or through collaborators, but the researchers also identify a set of 10,448 commercially available poised fragments. Commendably, they also provide full identities of both sets of fragments, which could be useful for folks building or adding to their own collections.

The Diamond Light Source is able to crystallographically screen 1000 fragments per week, but in this case only 406 diverse fragments were tested. Rather than using the nearly universal DMSO as a solvent, the researchers dissolved their fragments in ethylene glycol, since DMSO actually binds to bromodomains. Previous solution-phase screens of PHIP(2) at the SGC had come up empty, so the crystallographic screen was done at the very high concentration of 200 mM. Not surprisingly, this yielded just four hits.

Each of the hits bound in the acetyl-lysine recognition pocket, and three of them even showed high-micromolar activity in an AlphaScreen assay, with impressive ligand efficiency values. A few dozen analogs were made, which led to slight increases in activity in all cases, and measurable activity for analogs of the fragment which had shown no activity by itself. Although there is still a long way to go to find chemical probes for PHIP(2), at least there are now good starting points.

And that concludes bromodomain month. The number of papers and chemical probes that have come out just this year are a testament to the power of fragments to tackle this class of targets, perhaps equaled only by kinases. And while I'm not aware of any clinical candidates targeting bromodomains that started as fragments, I'm sure these will be coming soon.

20 July 2016

Fragments deliver a chemical probe for Family VIII bromodomains

Today’s post continues the theme of July as bromodomain month at Practical Fragments. The 61 human bromodomains (found in 46 proteins – some proteins have more than one) have been divided into eight families based on their sequences. Family VIII contains ten members, some of which are involved in keeping stem cells from differentiating. Two papers describe chemical probes that target some or most members of this family.

The first paper, which actually came out last year in Science Advances, is from a multinational group including Thomas Günther (Universität Freiburg), Stefan Knapp and Susanne Müller (both University of Oxford) and collaborators at Pfizer. The researchers started by screening libraries of acetyl lysine mimetics that had yielded inhibitors against other bromodomains. These came up empty; even promiscuous bromodomain inhibitors failed to hit Family VIII members. As is so often the case, when all else fails, the researchers turned to fragments. A thermal shift assay revealed that salicylic acid – the polypharmacological metabolite of aspirin – binds to the bromodomain PB1(5). Isothermal titration calorimetry (ITC) confirmed this result, providing a dissociation constant of 250 µM.

The researchers were also able to obtain a crystal structure of PB1(5) bound to salicylic acid in the acetyl lysine binding site common to all bromodomains, with the carbonyl making the usual hydrogen bond with a conserved asparagine. But whereas most other bromodomain binders make a water-mediated bridge to a conserved tyrosine, the phenol makes a direct hydrogen bond. The benzene ring also binds deeper in the pocket, displacing four highly conserved water molecules.

The subsequent medicinal chemistry optimization of this fragment is described in a paper published earlier this year in J. Med. Chem. by Dafydd Owen and colleagues at Pfizer, along with collaborators at the University of Oxford, DiscoveRx, Eurofins, the University of Massachusetts Worcester, and Johann Wolfgang Goethe University. Testing commercial and proprietary analogs of salicylic acid quickly revealed that uncharged enamides such as compound 2 were more effective at stabilizing PB1(5) against thermal denaturation than salicylic acid, and crystallography confirmed a similar binding mode.

Two rounds of library synthesis were conducted, first with 130 amines and then with 320 amines, with physicochemical properties of target compounds chosen in advance such that cLogP would range between 1 and 4. Seven family VIII bromodomains were screened in parallel, and compounds were identified with differing specificities. Some of the compounds were unstable in water, but introducing steric hindrance around the amine improved stability and led to compounds such as PFI-3. This is potent against the family VIII bromodomains PB1(5), SMARCA2A, and SMARCA4 and did not hit at least 40 other bromodomains tested. A related compound is active against more of the family VIII bromodomains while still maintaining good selectivity against other bromodomains.

Both of these probes are able to bind to family VIII bromodomains in cells and were used to explore the proteins’ biological roles. A variety of cellular phenotypic assays showed minimal changes, and the compounds do not appear to be toxic. They did attenuate myocyte or adipocyte differentiation, while PFI-3 caused embryonic stem cells to differentiate. One gets the impression that the researchers were hoping for more profound effects, but that’s why you make chemical probes in the first place. Whether or not these compounds will ultimately prove useful as drug leads, they should help to unravel some fiendishly complex biology.

15 July 2016

Fragments in the clinic: 2016 edition

There’s a new FBDD review out today in Nat. Rev. Drug Discovery. I know - there are lots of reviews each year - but this one is written by a who's who list of luminaries, including Steve Fesik (Vanderbilt), Rod Hubbard (Vernalis and University of York),  Wolfgang Jahnke (Novartis), and Harren Jhoti (Astex). I'm also an author so I'm undoubtedly biased, but I think it provides a nice overview of the field, especially for those who don't have time to read the recent book.

The review distills hard-won wisdom from two decades of work and covers practical decisions needed when using fragments: library design, screening methods, protein-ligand interactions, hit to lead strategies, and applications. Another useful feature is what I believe to be the most complete and up-to-date list of fragment-derived drugs that have entered clinical development. Where possible these include chemical structures, so definitely check it out.

The drugs themselves are listed below. Although it has not even been two years since the last compilation, it is exciting to see several promotions and new entrants. This table includes compounds whether or not they are still in development (indeed, some of the companies no longer even exist). A few compounds from earlier lists have been removed because their fragment origins could not be confirmed. Drugs reported as still active in clinicaltrials.gov, company websites, or other sources are in bold, and those that have been discussed on Practical Fragments are hyperlinked to the most relevant post.

Drug Company Target

Vemurafenib Plexxikon B-Raf(V600E)
Venetoclax AbbVie/Genentech Selective Bcl-2
Phase 3

PLX3397 Plexxikon FMS, KIT, and FLT-3-ITD
Verubecestat Merck BACE1
AZD3293 AstraZeneca/Astex/Lilly BACE1
Phase 2

AT7519 Astex CDK1,2,4,5,9
AT9283  Astex Aurora, JAK2
AZD5363 AstraZeneca/Astex/CR-UK AKT
Erdafitinib J&J/Astex FGFR1-4
Indeglitazar Plexxikon pan-PPAR agonist
LY2886721 Lilly BACE1
LY517717 Lilly/Protherics FXa
Navitoclax (ABT-263) Abbott Bcl-2/Bcl-xL
NVP-AUY922 Vernalis/Novartis HSP90
Onalespib Astex HSP90
Phase 1

ABL001 Novartis BCR-ABL
ABT-518AbbottMMP-2 & 9
ASTX660 Astex XIAP/cIAP1
AT13148AstexAKT, p70S6K, ROCK
AZD5099AstraZenecaBacterial topoisomerase II
BCL201 Vernalis/Servier/Roche BCL-2
PF06650833 Pfizer IRAK4

The current list contains more than 30 clinical-stage drugs but is certainly incomplete, particularly in Phase I. If you know of any others (and can mention them) please leave a comment.

11 July 2016

Fragments deliver a chemical probe for CBP and EP300

As we mentioned last week, July is bromodomain month at Practical Fragments. Today we’ll start by looking at two closely related bromodomains, one found in cyclic-AMP response element binding protein (CBP) and another from adenoviral E1A binding protein of 300 kDa (EP300). Both proteins have been implicated in a variety of diseases, particularly cancer, so a chemical probe would be very valuable.

Alexander Taylor and collaborators at Constellation Pharmaceuticals, Genentech, and WuXi, describe such a probe in a recent paper in ACS Med. Chem. Lett. The researchers screened about 2000 fragments in a thermal shift assay using 0.8 mM of each fragment. Compounds that increased the melting temperature of the CBP bromodomain by at least 1° C were validated first by time-resolved fluorescence resonance energy transfer and then by 15N HSQC NMR, ITC, and X-ray crystallography. Compound 1 was one of the more attractive hits, in particular because it was considerably less active against BRD4, whose inhibition causes all sorts of changes to cells.

Crystallography of the racemic compound clearly showed that only one of the enantiomers bound, and this was confirmed in functional assays when both enantiomers were tested separately. The active enantiomer makes some of the same interactions typical of all bromodomains with the natural ligand (N-acetylated lysine). Fragment growing was attempted off the aromatic ring, and although several vectors were tolerated, most decreased selectivity against BRD4. However, close examination of the structures revealed a promising vector that led to compound 14, with good selectivity against BRD4. Further optimization ultimately led to CPI-637, with low nanomolar activity against both CBP and EP300 as well as good cell-based activity. Crystallography revealed that this compound binds in a similar manner as the initial fragment.

The selectivity of CPI-637 against other bromodomains is also good (> 700-fold less active against BRD4), though it does hit BRD9 with sub-micromolar activity. Just as with the initial fragment, the opposite enantiomer of CPI-637 is considerably less active. Although no pharmacokinetic data are provided, at the very least this should be a useful probe for cell-based studies.

Switching gears to another aspect of CBP, the multidomain protein p300/CBP-associated factor (PCAF) has a bromodomain that may bind to CBP, though the biology is not entirely clear. PCAF is known to bind an acetylated HIV protein, and has been proposed as a target for AIDS. Obviously this is another opportunity for a chemical probe! The first steps are reported in a paper by Stefan Knapp and collaborators at Goethe University Frankfurt, University of Oxford, Leiden University, ZoBio, and University of Cambridge, published in J. Med. Chem (and open-access).

The researchers screened two separate fragment libraries using either thermal shift assays (at 1 mM fragment) or TINS. Hits were confirmed using SPR and crystallography, resulting in seven structures. As expected, all the fragments bound at the site where N-acetylated lysine normally binds. The PCAF bromodomain appears to be quite rigid, with little movement in structures with the different bound fragments. A few elaborated molecules were tested, with the best showing low micromolar affinity as assessed by ITC; crystal structures with these molecules are also reported and deposited in the protein data bank. It will be fun to see whether their potency can be improved.

We’ll have another post on bromodomains next week, but first stay tuned later this week for an updated list of fragment-derived drugs that have entered the clinic.

05 July 2016

Fragments deliver a chemical probe for BRD9

Bromodomains have nothing to do with bromine. Rather, they are small (~110 amino acid) domains that recognize acetylated lysine residues, a common modification on histones, and are thus key epigenetic “readers”. Humans have more than 60 of them, and as you can imagine selectivity is not assured. However, fragments have proven very useful in targeting these proteins. Since the first mention of bromodomains on Practical Fragments back in 2011 the number of posts has been growing rapidly, so for the first time ever we’ve decided to devote an entire month to the topic.

In other words, July is bromodomain month! We’ll start with two papers against the bromodomain BRD9, part of the SWI/SNF chromatin remodeling complex that seems to be important for acute myeloid leukemia.

The first paper, in J. Med. Chem. (and open access), is published by Laetitia Martin and collaborators at Boehringer Ingelheim, University of Oxford, and Cold Spring Harbor. The researchers used three orthogonal biophysical screening methods: differential scanning fluorimetry (DSF), surface plasmon resonance (SPR), and microscale thermophoresis (MST). A library of 1697 fragments was screened at 0.4 mM (DSF), 0.1 mM (SPR) or 0.5 mM (MST), and hits were then validated using 15N HSQC NMR. The 77 hits that confirmed were taken into crystallography, producing 55 structures.

Validation rates in the NMR secondary screen were excellent for DSF (94%) and SPR (84%) but less so for MST (31%). That said, of the 38 validated hits from MST, 29 were not found in either of the other techniques, and 14 of these produced crystal structures. This is a useful reminder that while screening cascades can whittle down many hits, they do run the risk of throwing out the proverbial babies along with the bathwater.

In parallel with the biophysical screens, a virtual screen of ~73,500 fragments was conducted using Glide to identify 208 fragments that were then tested using SPR and DSF. This led to 23 hits, 11 of which produced crystal structures.

Two of the more potent fragments were the structurally related compound 3 (from the biophysical screen) and compound 4 (from the virtual screen). Optimization started with compound 4 by adding electron donating groups to the phenyl ring to try to improve a stacking interaction observed in the crystal structure. This led to compound 10, and building out the other ring to make it more similar to fragment 3 led to BI-9564.

BI-9564 has low nanomolar activity in both a biochemical assay as well as isothermal titration calorimetry (ITC). It is also quite selective: among 48 other bromodomains, it only hits the closely related BRD7 and CECR, and it is >10-fold more potent on BRD9. None of a panel of 321 kinases were inhibited with IC50 < 5 µM, and only 2 of 55 GPCRs were inhibited. The compound is also cell active, reasonably soluble, has good pharmacokinetics in mice, and orally bioavailable. In short, BI-9564 is an excellent chemical probe – and is in fact being offered as such.

While we’re on the subject of BRD7 and BRD9, it’s worth noting another recent paper, this one in ChemBioChem from Ke Ruan and colleagues at the University of Science and Technology of China. The researchers screened their library of 890 fragments against BRD7 using three different ligand-detected NMR techniques: STD, WaterLOGSY, and CPMG. Fragments were screened in pools of 10 with each fragment present at 400 µM. This yielded just 10 hits, of which 5 confirmed when tested individually. Protein-observed NMR was then performed on these, suggesting that they all bind in the acetyl-lysine recognition sites; they have similar affinities for both BRD7 and BRD9, with dissociation constants between 22 and 600 µM. Crystallography confirmed the binding mode for one of the fragments bound to BRD9. Interestingly, this showed quite a bit of plasticity in the protein compared to the un-liganded structure. Indeed, the BI researchers suggest that different degrees of protein flexibility between BRD7 and BRD9 could account for the selectivity differences observed for BI-9564.

Stay tuned next week for more fragment-screening against a different class of bromodomains!