Fragments vs PRC2: ligand deconstruction

Ligand deconstruction is a strategy for early-stage drug discovery in which a known hit is dissected into component fragments and one or more of them is optimized. When successful, it can lead to new and improved chemical series. One such example was just published in J. Med. Chem. by Andreas Lingel and colleagues at Novartis.

The researchers were interested in finding inhibitors of the protein methyltransferase polychrome repressive complex 2 (PRC2). Although some drugs have entered the clinic against this anticancer target, all of these are competitive with the cofactor S-adenosylmethionine (SAM), and resistant mutants are already being detected. Thus, the team sought a molecule that would act through a different mechanism.

PRC2 is actually a complex of four different proteins. The SET domain of the protein EZH2 contains the catalytic machinery, but a protein called EED stabilizes the protein complex and is necessary for activity. EED also recognizes trimethylated lysine residues on histone substrates, allosterically activating methyltransferase activity.

A high-throughput biochemical screen identified compound 1, which has low micromolar activity and is noncompetitive with the SAM cofactor and substrate peptide. Subsequent NMR and crystallography experiments revealed that compound 1 binds to EED, with the tertiary amine binding in the same pocket that normally recognizes trimethyllysine. However, compound 1 is quite complex, with three stereocenters. Thus, the researchers sought to deconstruct it to something simpler. They began by chopping off two of the rings – an unconventional disconnection but one supported by crystallography, which revealed that the terminal rings were not closely associated with the protein.

The resulting fragment 2 was down more than an order of magnitude in potency but had improved ligand efficiency. Crystallography confirmed that it binds in a very similar fashion to compound 1. Initial SAR was conducted around the methoxybenzyl moiety, which is buried within the enzyme. Most changes were not tolerated, but compound 9 did show somewhat improved activity.

Next, the researchers sought to optimize the positively charged portion of the molecule. Replacing the amine with a guanidine improved the affinity but at a cost to cell permeability. This led to a search for less conventional replacements, ultimately yielding the 2-aminoimidazole moiety in compound 16. Not only did this regain the activity of the initial molecule, it also shows good permeability and cell activity. Crystallography revealed that it too binds in a similar fashion to the original hit.

This is a nice example of fragment-assisted drug discovery (FADD), in which concepts from FBDD were used to simplify and optimize a hit from HTS. There is of course much more to do with this series, not the least of which is figuring out exactly how the molecules actually inhibit PRC2. Trimethyllysine-containing peptides that bind to EED normally activate the enzyme, yet the small molecules that bind to the same site somehow allosterically inhibit activity. Despite multiple crystal structures, the researchers frankly acknowledge that they were “not able to decipher the molecular basis for this phenomenon.” A number of conformational changes occur when EED binds to ligands, and perhaps these propagate through the protein complex. A picture may be worth 1000 words, but we may have to wait for the movie to learn the full story. 

New tricks for old methods: STD NMR

According to our last poll, ligand-detected NMR is the most popular method for finding fragments. And among the several ligand-detected NMR techniques, the most popular appears to be saturation transfer difference (STD) NMR. The basic concept behind this approach is to selectively irradiate a protein, which then transfers its magnetization to any bound ligand, thus “saturating” (reducing the signals) for the ligand. Subtracting this spectrum from a reference spectrum reveals which ligand (if a mixture) or individual protons within a ligand are in close proximity to the protein.

Although STD NMR is fast and easy to run, it does have drawbacks. One is the fact that it requires pure protein: if there are other proteins in solution, it will be impossible to tell whether the small molecule binds to the protein of interest or to something else. This shortcoming has been overcome in a paper published recently in J. Biomol. NMR by Tamas Martinek and collaborators at the University of Szeged, the Hungarian Academy of Sciences, and the University of Debrecen.

In a normal STD experiment, the protein protons that are irradiated are far upfield (often around -0.5 ppm) – a region not relevant to most small molecules. These protons then transfer the magnetization throughout the protein and ultimately to any bound small molecules. In order to choose a specific protein, the researchers add an ¹⁵N-labeled antibody selective for the protein. They can then selectively irradiate the ¹⁵N-labeled antibody, which transfers its magnetization to the bound protein and from there to any bound ligand. They call this approach monoclonal antibody-relayed ¹⁵N-group-selective STD, or mAb-relayed ¹⁵N-GS STD.

To demonstrate the approach, the researchers observed the binding of 2 mM lactose to galectin-1 (Gal-1) using an ¹⁵N-labeled antibody against Gal-1. Lactose binds to Gal-1 with a dissociation constant of 0.155 mM, which is a relevant affinity for fragment screening. Gal-1 was present at 20 µM and the antibody was present at 10 µM, both of which are reasonably low. Control experiments established that both Gal-1 and the antibody were necessary, and the experiment was successful even in a cell extract.

So, as Teddy would ask, is this approach practical? You need an ¹⁵N-labeled antibody against your target, and it is important that the antibody is specific and does not compete with your ligand (ie, that it is non-neutralizing). Also, the amount of time required to acquire the spectra appears to be more than an hour. If this could be reduced, would ¹⁵N-GS STD assume a useful niche in the NMR toolbox?

Enthalpy revisited – and retired

The relative importance of enthalpy versus entropy for protein-ligand interactions has been a subject of considerable attention. In a 2009 post we suggested that it might be worthwhile to focus on fragments that bind predominantly enthalpically, and in 2011 we highlighted a paper suggesting that enthalpic binders may be more selective than entropic binders. But the universe has a way of confounding pet models – as we acknowledged in 2012 (twice). The best way forward is often with lots of data, which is exactly what we have in a new paper in Drug Disc. Today by György Keserű and collaborators at the Hungarian Academy of Sciences, Astex, and AstraZeneca.

The data in this case are sets of 284 protein-ligand interactions with thermodynamic binding data from the literature, 782 from Astex, and about 230 from AstraZeneca. Commendably, these data are provided in 103 pages of supporting information.

In order to analyze the data, the researchers developed a new metric, the Enthalpy-Entropy Index:

I_E-E = (ΔH+TΔS)/ΔG

If I_E-E = 0, it means that enthalpy and entropy both contribute equally to the free energy of binding; if I_E-E > 0 it means that enthalpy dominates, and if I_E-E > 1 it means that enthalpy needs to overcome an unfavorable entropy. Similarly, negative values mean that entropy dominates – completely so when I_E-E < -1. (Note that, unlike enthalpy efficiency, this is a dimensionless ratio, which should please our friends over at Molecular Design.)

As it turns out, the vast majority of fragments bind to their targets with favorable enthalpy, and almost all of those that don’t are charged compounds in which desolvation of the charged bit could entail an enthalpic cost. The researchers also examined a set of 94 neutral fragment-sized and 44 larger molecules binding to 17 targets and found that, statistically speaking, enthalpy plays a more important role in the free energy of binding for fragments than for larger molecules. But things can change quickly: in one case, adding just two non-hydrogen atoms to a molecule improves the affinity by more than 4000-fold and changes the I_E-E from -1.5 to +0.5.

The paper does an excellent job describing the challenges of collecting high-quality isothermal titration calorimetry (ITC) data. In a typical experiment, the heat measured with each injection is the same as “would fall on an A4 sheet of paper in 1 second when illuminated by a 40 Watt bulb placed nearly 5 kilometers away.” Errors can be caused by inaccurate concentrations, heat of dilution, and changes in buffer concentration or protonation state. An analysis of replicate measurements at Astex found that, while most of the replications were within 1 kcal/mol of each other, some were off by nearly 3 kcal/mol. However, these larger values were all associated with different forms of the protein, and so may not be considered true replicates, though they do indicate how changes in the protein far from the active site can have an effect on what is often considered (erroneously) a local interaction.

This also emphasizes the fact that, as the researchers note, “the measured binding enthalpy is a net value and the dissection of the individual contributions might be ambiguous.” Or, as Pete has previously stated, “the contribution of an individual protein-ligand contact is not strictly an experimental observable”.

From a molecular recognition standpoint I find all this quite interesting and even intuitive in a hand-wavy sort of way. As the researchers suggest, fragments, being small, have minimal surface area with which to make (often but not always entropically-driven) hydrophobic interactions. Instead, much of the binding energy comes from hydrogen-bond interactions, which are (again often but not always) predominantly enthalpic. Moreover, since the entropic cost of locking down any ligand onto a protein is on the order of 3-5 kcal/mol, fragments are already fighting against entropy, and this is exacerbated by low affinity.

But from a practical perspective, my earlier suggestion to focus on enthalpic fragments may have been simplistic: if you’ve found a fragment, its enthalpy of binding is almost certainly favorable, and even if it’s not, this could change with the slightest tweak. So unless we see something truly new, don’t expect many new posts on this topic.

Cussed curcumin

Teddy’s retirement from the blog has cut down on the number of PAINS-shaming posts, and truth be told there are so many candidate papers that they could easily swamp fragments, which I suspect would drive away most of the readership. That said, I did want to highlight an exhaustive Perspective about a particularly diabolical natural product just published today in J. Med. Chem. by Mike Walters and collaborators at the University of Minnesota, Brigham and Women’s Hospital, and the University of Illinois (and also covered in a news story in Nature.)

We’ve previously discussed some of the types of artifacts that can plague small molecule screens: aggregation, covalent adducts, redox cycling, fluorescence, photoreactivity, and more. Curcumin is a jack of all trades in that it is capable of all of the above. It’s also unstable even at neutral pH, and can decompose into other reactive species. It is the quintessential chemical con artist: if you have an assay, curcumin will probably be active in it.

The new paper is a thorough investigation (18 pages, with 164 references) of the chemistry and biology of curcumin, covering in gruesome detail all the many ways it can deceive. After discussing the history and physicochemical properties (and liabilities), several literature case studies where curcumin is proposed as having biological activity are explored and thoroughly demolished; one of these has been retracted but continues to be cited uncritically years later.

One might expect that something which hits so many assays would be toxic. This turns out not be the case: curcumin is present at 1-6% in tasty turmeric and only seems to show any adverse events at very high doses – several grams per day. The reason, the researchers show, is that curcumin’s pharmacokinetics are lousy, with oral bioavailability of less than 1%. This is a very literal example of the cliché “garbage in, garbage out.”

Sadly, these properties have not dampened interest in testing curcumin in people. The researchers identify 135 registered clinical trials, only eight of which have reported study results, with 49 either recruiting or not yet recruiting. The few examples where results have been reported are not particularly encouraging.

Typing curcumin into PubMed pulls up close to 10,000 papers, with more than 150 published in J. Med. Chem. alone. Will this devastating exposé help? For honest and diligent researchers, it should serve as a flashing warning to be extremely careful with any data gathered using curcumin. Unfortunately, some in the scientific community may not care as long as they are able to pump out papers. Indeed, according to Wikipedia, at least one prominent curcumin researcher had to retract several papers because of questionable “data integrity”. And there may be still darker motives: type curcumin into Google and the top results are ads touting the stuff. There’s money to be made, and even more if you slap on some scientific lipstick.

And despite specific J. Med. Chem. author guidelines to be cautious about “interference compounds” and “provide firm experimental evidence in at least two different assays that reported compounds with potential PAINS liability are specifically active and their apparent activity is not an artifact”, the journal recently published a paper fully devoted to the synthesis and SIR of rhodamine derivatives, with no consideration of mechanism nor mention that they can be problematic. (Indeed, the researchers do not even bother to include detergent in their enzymatic assay!)

All of which is to say that it’s easy to publish crap. But hopefully now, more people will recognize it as such.

Fragments in the clinic: verubecestat

Of all the fragment-derived drugs in the clinic, perhaps none is so closely watched as verubecestat (MK-8931), Merck’s BACE1 inhibitor in phase 3 clinical trials for Alzheimer’s disease (AD). With tens of millions of cases worldwide, few other diseases in the developed world are as simultaneously widespread, expensive, and terrifying. And despite billions of dollars thrown at the problem, failure rates are nearly 100%. A recent open-access paper by Jack Scott, Andrew Stamford, and collaborators at Merck and AMRI in J. Med. Chem. provides an excellent overview of this latest contender.

We first wrote about Merck’s BACE1 program almost exactly seven years ago, describing how an NMR screen had provided a weak hit that was optimized to nanomolar inhibitors of the enzyme. However, the molecules could fairly be called molecularly obese. This led the researchers to trim back portions of the molecule, losing affinity but gaining cell-based activity and permeability, ultimately resulting in compound 5 (below) – which is itself a fragment. The current paper describes the optimization of this molecule.

Growing compound 5 and expanding the heterocyclic ring led to compound 7, with low nanomolar biochemical and cell-based activity. The iminopyrimidinone core was becoming increasingly crowded from an intellectual-property standpoint, so the researchers replaced this with the iminothiadiazinane dioxide in compound 9, which modeling suggested should have a similar conformation – a result confirmed by crystallography. However, the alkyne moiety appeared to be metabolically unstable. More importantly, compound 9 was only 47-fold selective against the enzyme cathepsin D (CatD). An earlier Lilly BACE1 inhibitor with a similarly modest selectivity had failed due to toxicity possibly associated with CatD, and the researchers were keen to avoid a similar fate. This led them through additional rounds of optimization, ultimately resulting in verubecestat.

In addition to having low nanomolar biochemical and cell-based activity against BACE1, verubecestat is >45,000-fold selective against CatD, has good pharmacokinetics, is orally bioavailable, and is highly soluble (1.6 mM!) It does not inhibit CYP enzymes and has good brain penetration. Rule-checkers might be surprised at this later point given the high calculated polar surface area (115 Ų), a fact the researchers attribute to an intramolecular hydrogen bond between the amide and the pyridine nitrogen, effectively masking these moieties from the point of view of membranes.

A couple potential liabilities stood out. First, one metabolite is an aniline, and anilines can be mutagenic. Reassuringly, an Ames test on this particular aniline showed no mutagenicity. Also, verubecestat is a 2.2 µM hERG inhibitor, and inhibitors of this channel can cause cardiac arrhythmias. However, this concentration is significantly higher than the highest expected in humans, and studies in primates revealed no safety issues. All of which is a useful reminder that, in our business, rules are at most guidelines, and data is king.

The paper also includes some human data demonstrating that the compound is safe at doses up to 550 mg (!) and causes a dose-dependent reduction in β-amyloid levels. With the results of the first Phase 3 trial expected later this year, we could know soon whether this is a billion dollar molecule or yet another massively expensive failure. If the former, verubecestat could be one of those transformational advances in drug discovery that comes along once in a generation. But even if fails, this is the clearest test yet of the amyloid hypothesis. And fragments made it possible.

Fragment events in 2017

What better way to start the year than with a list of upcoming conferences? Here's what we know about so far.

March 5-7: The UK may have voted to leave the EU, but that's not stopping the Royal Society of Chemistry from holding Fragments 2017 in Vienna, Austria. This is the sixth in an illustrious conference series that alternates years with the FBLD meetings. You can read impressions of Fragments 2013 and Fragments 2009. Registration is already open!

April 25-26: CHI’s Twelth Annual Fragment-Based Drug Discovery, the longest-running fragment event, will be held in San Diego at a brand new venue. You can read impressions of last year's meeting here; the 2015 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 25. Early registration is open until January 27.

June 6-9: Although not exclusively fragment-focused, the fourth NovAliX Conference on Biophysics in Drug Discovery will 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. Registration is open now.

July 23-28: Finally, Australia is coming into its own as a destination for fragment experts, many of whom will be participating in a symposium (on July 27) that is part of the Royal Australian Chemical Institute's Centenary Congress in Melbourne. The entire event should be huge - think ACS with wombats - so if you've been looking for yet another reason to travel Down Under, this is it. Early registration is open through April 23.

It looks like the year is largely front-loaded thus far - know of anything else? Add it to the comments or let us know!