Avoiding will-o’-the-wisps: aggregation artifacts in activity assays

The phenomenon of aggregation is the drug hunter’s quicksand. A prerequisite for using biochemical assays to study fragments – or any low-affinity molecules – is an ability to sort activity from artifact. Many small molecules, even bona fide drugs, form aggregates in aqueous solution, and these aggregates can non-specifically interfere with biochemical assays. There are several ways to expose these promiscuous inhibitors (see list below), but even with vigilance, researchers can inadvertently stumble onto a route lit by will-o’-the-wisps. The most recent issue of J. Med. Chem. provides a particularly insidious example from Brian Shoichet, Adam Renslo, and colleagues at UCSF.

The researchers were looking for noncovalent inhibitors of cruzain, a popular protease target for Chagas’ disease. After a virtual screen of commercial lead-like compounds, 17 molecules were purchased and tested in enzymatic assays, and compound 1 (below) inhibited cruzain, albeit weakly. However, the compound looked like the real deal: it showed no time-dependence; it was active in the presence of detergent; and Lineweaver-Burk plots revealed that it was mechanistically competitive.

The researchers thus turned to medicinal chemistry, replacing the ester group of compound 1 with an oxadiazole bioisostere and swapping the aryl group for a substituted pyrazole, ultimately arriving at molecules such as compound 21, more than two orders of magnitude more potent than the starting molecule.



So far, so standard: similar stories appear every week in J. Med. Chem., Bioorg. Med. Chem. Lett., ChemMedChem, and other journals, and it would not have been surprising to see this published with a title like “Discovery of a high affinity inhibitor of cruzain.” Only in this case, the researchers became suspicious: most of the molecules were not active against the targeted protozoa, and many of the dose-response curves had unusually steep Hill slopes, a tell-tale sign of aggregation. Looking more closely at their protocol, the researchers also realized that the concentration of non-ionic detergent in their assays was ten-fold lower than they had thought. D’oh!

A series of tests confirmed that, despite interpretable and rationalizable SAR, the series had been optimized for aggregation-based inhibition: compound 21, with an IC50 of 200 nM in buffer containing 0.001% of the detergent Triton X-100, showed no inhibition whatsoever in 0.01% Triton X-100. The compound also inhibited AmpC beta-lactamase, an enzyme particularly sensitive to aggregators, and this inhibition could be reversed with detergent. Finally, dynamic light scattering (DLS) revealed the presence of particles (or aggregates) in aqueous solutions of compound 21.

But the tale gets even more twisted. Some of the aggregators show legitimate, competitive binding to cruzain under high-detergent conditions, albeit at much higher concentrations (with IC50s above 40 micromolar). Conversely, compound 1 actually shows noncompetitive behavior in low-detergent conditions, though again only at fairly high concentrations. In other words, promiscuous inhibitors can behave legitimately under sufficiently stringent conditions, and legitimate inhibitors can behave promiscuously under less stringent conditions.

What’s especially sobering is how easy this promiscuity would have been to overlook: many molecules with good activity in biochemical assays don’t show any effects in cells, and it is easy to ignore steep slopes in inhibition assays. How many of those “Discovery of a high affinity inhibitor of Hot Target X” papers actually report promiscuous inhibitors? The authors, who have been researching this problem for a long time, end on a justifiably paranoid note:

The cautionary contribution of this study is to point out that even within a clear SAR series, one is never entirely free from the concern that non-stoichiometric, artifactual mechanisms are contributing to the inhibition one observes.

This is a serious problem, both for the researchers doing the original work and for anyone trying to follow up on the results. But one can take precautions, summarized below and described more fully here:

• Add non-ionic detergent to the assay (Triton-X 100, Tween-20, CHAPS, others)
• Increase protein concentration – this should have no effect on genuine binders (within limits)
• Characterize the mechanism of inhibition (competitive, noncompetitive, or uncompetitive): competitive inhibitors are normally not promiscuous
• Centrifuge your samples and retest them – this can sometimes remove aggregators
• Examine your samples with DLS or flow cytometry – aggregators can sometimes be directly observed as 50-1000 nm particles
• Look closely at your dose-response curve - unusually steep slopes can signal aggregation

And of course, biophysical methods such as SPR, NMR, and X-ray crystallography can provide more information than biochemical assays and reveal stoichiometric (and – in the case of SPR – superstoichiometric) binding.

Difficulty sorting true low-affinity binders from false positives stymied fragment-based approaches for decades, and in fact the nature of promiscuous inhibition caused by aggregation wasn’t even characterized until earlier this century. We now have techniques to sort deceptive aggregation from true but faint affinity. Let’s make sure these tools are consistently used.

DCC and FBDD

Dynamic combinatorial chemistry (DCC) has grown alongside of and often intersected with FBDD. In a recent issue of Angewandte Chemie, Jörg Rademann and colleagues at the Leibniz Institute of Molecular Pharmacology describe the latest example.

Put simply, DCC generates new molecules with some desired property by allowing smaller molecules to assemble reversibly under selection pressure. If the selection pressure is binding to a protein target and the molecules undergoing reactions are fragments, DCC can be used for FBDD. As we previously noted, Huc and Lehn published one of the earliest demonstrations of this. DCC has also been used at a few companies, including Astex and Sunesis, and even formed the basis of the (sadly) short-lived Therascope.

Rademann’s approach, "dynamic ligation screening", is based on labeling one fragment with a fluorescent probe and then screening it in a fluorescence polarization assay with other test fragments. If the labeled fragment binds competitively with a test fragment, this implies that the two fragments bind to the same site. However, if the fluorescence polarization signal increases in the presence of the test fragment (indicating increased binding of the labeled fragment), this suggests that the test fragment and the labeled fragment are binding cooperatively.

The researchers applied dynamic ligation screening to the protease caspase-3, a key mediator of apoptosis relevant for many diseases. As their labeled “fragment,” they chose a high-affinity tetrapeptide containing an alpha-ketoaldehyde: the ketone interacts covalently with the catalytic cysteine of the enzyme, while the aldehyde can form imines with amine-containing fragments. Interestingly, this strategy selects for fragments that bind in the S1’ subsite of the enzyme, which has not received as much attention as the tetrapeptide binding sites S1-S4.

A fluorescently labeled version of the tetrapeptide was screened against a library of 7,397 fragments, of which 4,019 contained primary amines. Of these, 78 fragments caused a decrease in the fluorescence polarization signal, suggesting that they compete with the tetrapeptide for binding. These were tested in an enzymatic assay: 21 of them were active at 10 micromolar concentrations, and four had Ki values from 3.1 to 5.5 micromolar; these four molecules have electrophilic carbons, making it likely that they bind to the catalytic cysteine residue.

Of greater interest, 176 fragments were cooperative, increasing the fluorescence polarization (FP) of the labeled tetrapeptide fragment by at least 20%. 50 of these were tested in an enzymatic assay, with the amine shown below emerging as the most potent FP enhancer and a Ki of 120 micromolar alone. A series of experiments guided by mathematical modeling suggested that the protein was templating the formation of an imine bond between the aldehyde of the tetrapeptide and the amine. Moreover, the reduced (amine) version of this conjugate exhibits a very high affinity for caspase-3, with a Ki of 80 picomolar.



Of course, affinity is not everything: with a molecular weight of 767 Da and a clearly peptidic nature, the pharmaceutical properties of this molecule, and even its cell activity, are questionable.

This study is reminiscent of some work we did at Sunesis, using caspase-3 to template the assembly of a non-peptidic inhibitor using Tethering. In that case we built molecules in the S1-S4 pockets, but did not do much work to extend into the S1’ pocket. It would be interesting to see if the fragment Rademann and colleagues discovered also boosts the potency of the molecules we identified.

For dynamic ligation screening to be general it needs to surmount at least two major potential limitations. First, it remains to be seen whether the technique will work with actual fragments, which are likely to have far lower affinities than the 25 nM tetrapeptide used in this study. Second, cooperative binding of the fragments does not translate to synergy in the final molecule: the conjugate has a lower ligand efficiency than either of the fragments, despite the apparent cooperativity of the two fragments binding to the target. This could be because the conjugate contains an amine, whereas the two fragments in solution presumably were linked by an imine; the differences in geometry and chemical nature between these two moieties are profound, and one could imagine that many amine-linked compounds would not be selected as imines, and vice versa.

Still, this is an interesting approach to tackle the long-standing challenge of linking fragments, and it will be fun to watch for new developments.

Fragments of Life shut down LTA4H

A couple months ago we highlighted research suggesting that natural products are a fruitful field for finding fragments. One company, deCODE, has taken this idea very seriously, and has constructed their fragment library based largely on molecules (or close analogs) that actually appear in nature. Their strategy is described in detail in the most recent issue of J. Med. Chem.

The “fragments of life” (FOL) screening library consists of three sets of molecules:

• 218 “molecules of life,” which are known metabolites from some living organism
• 666 synthetic derivatives and isosteres of known metabolites
• 445 synthetic biaryl molecules, which mimic peptide turns (biaryls have also previously been reported to be privileged pharmacophores)

This gives, in total, a 1329-fragment screening set. Naturally, given their origin, some of the fragments are slightly unusual, including the dipeptide bestatin and the trendy resveratrol. However, with the exception of a somewhat higher polar surface area, the molecules conform to rule of 3 guidelines, with an average molecular weight of 182.5 and ClogP of 0.96. All fragments are soluble up to 50 mM in methanol, and in fact stocks are made in this solvent rather than the more conventional DMSO.

The fragment library was tested against Leukotriene A4 Hydrolase (LTA4H), an enzyme with two functions: it has an aminopeptidase activity whose biological relevance is unknown, and an epoxide hydrolase activity that converts leukotriene A4 to the inflammatory leukotriene B4, which is implicated in heart disease and inflammation. Both activities map to a single active-site, a long cleft containing a catalytic zinc.

About 200 of these fragments were screened by soaking crystals of LTA4H in pools containing 8 compounds. Although all compounds in a given pool were structurally diverse, in some cases electron density was ambiguous, necessitating subsequent soaks of individual fragments to confirm hits. Ultimately 13 fragments were found to bind LTA4H, a hit rate of 6%. These fragments were tested in functional assays and found to have IC50s as good as 178 nM for bestatin, though the next best was mid-micromolar. Most of the fragments bound in the active site, although one fragment bound on the surface of the enzyme. Considerable structural data are presented in the paper, and all the structures have been deposited in the protein data bank.

Interestingly, the researchers also found that some of the fragments only appeared to bind when they were soaked in the presence of another fragment, bestatin. Bestatin also caused the binding mode of another fragment to shift compared to its binding mode without bestatin.

Based on the crystal structures available, some of the fragments were elaborated to provide more potent inhibitors, increasing affinity by some four orders of magnitude, as well as improving ligand efficiency (see figure). Crystallography revealed that these more potent compounds bind in a similar fashion to the fragments.

Compounds 14 and 18 also bind in a similar manner to DG-051, which has recently completed phase IIa clinical trials. There is apparently another manuscript in the works focused exclusively on this molecule. We look forward to reading the full story.

Hsp90 and fragments – part 2: NVP-BEP800/VER-82576

Proving again that Hsp90 is tailor-made for fragment-based approaches, the latest issue of J. Med. Chem. has a thorough article describing the development of an anti-cancer candidate targeting this protein. The researchers, mostly from Vernalis but also from Novartis and the Institute of Cancer Research, used a combination of fragment-based methods, computational screening, and medicinal chemistry. This is likely to be representative of a coming wave of reports in which a fragment approach supplements other techniques (or vice versa).

The fragment effort started with a library of fragments grouped into pools of 10-12 each and screened using three different NMR methods (saturation transfer difference, water-LOGSY, and T2 relaxation filtered 1D). Compounds were tested in the presence and absence of PU3, a compound known to bind to the ATP site of Hsp90. Of the 1351 fragments tested, 59 of them (4.4%) were confirmed in all three NMR experiments and competed with PU3. Interestingly, a further 158 compounds were found to bind to Hsp90 but could not be displaced by PU3, suggesting that they bind outside of the ATP binding site; these were not pursued.

Two of the fragments identified, along with their IC50s in a fluorescence polarization (FP) assay, are 10 and 11 (see figure – click to see larger image). These fragments were also characterized crystallographically. An interesting aside is that the binding mode of fragment 10 differed depending on whether it was soaked into Hsp90 crystals or co-crystallized with the protein. (This is reminiscent of the different binding modes, both productive, observed for two fragments using crystallography versus NMR on Hsp90 by researchers at Abbott.)

In addition to the fragment work on Hsp90, a virtual screen of 700,000 (non-fragment) compounds was conducted, leading to the purchase of 719 commercially available molecules that were tested in the FP assay. Two hits, 12 and 13, are shown in the figure and were also characterized crystallographically.



With these SAR and X-ray crystal structures in hand, the researchers added elements from the larger compounds (a phenyl from 12 and 13 or the amide from 13) to their fragments to generate the more potent fragments 14 and 15. Further structural examination led to the 2-amino-thienopyrimidine scaffold 16, which formed the basis for subsequent optimization.

Appending a phenyl group onto compound 16 led to the expected boost in potency, with the dichloro compound 21e having good biochemical as well as cell-based activity. Addition of a solubilizing group led ultimately to NVP-BEP800/VER-82576, which, in addition to potent biochemical and cell activity, also showed tumor regression in mice following once-daily oral dosing, along with pharmacodynamic biomarkers consistent with Hsp90 inhibition. This compound also showed good antiproliferative effects in a number of human cancer cell lines. Crystallographic characterization revealed that the molecule binds in a manner consistent with the previous structures (and with the co-crystallized structure of fragment 10).

This could be seen as a nice example of what has been dubbed fragment-assisted drug discovery: fragments were not the sole drivers of the project, but they did play an important role in guiding the overall strategy to develop optimal molecules.

In related news, Vernalis just announced a multimillion dollar collaboration with GlaxoSmithKline around an undisclosed cancer target – another indication that fragment-based approaches have not just scientific value, but monetary value as well.

Fragment-based conferences in 2009 and 2010

Hard to believe, but 2009 is more than halfway over. As far as I know there is only one more event this year focused primarily on fragments, but it’s a biggie, and conferences are already being planned for 2010. Here’s what might be of interest over the next few months.

August 16-20: The fall ACS meeting this year will be held in sweltering Washington, DC. Although there are no sessions devoted exclusively to fragments, a number of FBLD talks and posters are sprinkled throughout the conference.

September 21-23: Much-anticipated FBLD 2009 will be held in historic York, UK. There will also be a one-day workshop on September 20 to provide an overview of fragment-based drug discovery for newcomers to the field. The draft schedule for the conference has just been released (pdf) – looks like a great lineup, so if you missed the earlier conferences this year, don’t miss this one!

2010
February 3-5: CHI’s Molecular Medicine Tri-Conference is being held in the beautiful city of San Francisco, with a track on medicinal chemistry that will probably have some fragment talks, and a fragment workshop on February 2.

March 21-25: The spring ACS meeting will be in foggy (but beautiful) San Francisco. No schedule or link yet.

April 20-25: The Keystone Symposium on computer-aided drug design will take place in brisk Whistler, British Columbia. Although not exclusively devoted to fragments, the schedule shows plenty of talks on FBLD.

April 26: The CHI fragment-conference will be held in summery San Diego. No web link yet, but I’ll have this when it becomes available.

As always, let us know if we’ve missed anything and we’ll get the word out!

Hsp90 and fragments

Some targets seem particularly amenable to fragment-based approaches. Protein kinases are one example. Another is the N-terminal ATP binding domain of Hsp90, a widely pursued anti-cancer target: at least two fragment-derived compounds against this target are currently in the clinic. At FBLD 2008 in San Diego last year, so many talks discussed this protein that it became a running gag (one speaker promised at the outset not to talk about it, then slipped in a few slides). A recent paper in ChemMedChem provides a particularly clear example of a multidisciplinary fragment-growing approach against this target.

The researchers, mostly from Evotec, started with a high-concentration biochemical displacement assay to screen 20,000 fragments against Hsp90. A relatively potent aminopyrimidine (compound 1, below) was characterized crystallographically, and this structure was then used to run a virtual screen of 3.8 million commercially available molecules using the program GOLD 3.0.1. Some of the resulting hits were purchased and tested, including compound 3, which showed sub-micromolar biochemical activity but no cell-based activity. Subsequent modeling and medicinal chemistry led to compound 19, which, in addition to mid-nanomolar biochemical activity, also displayed submicromolar cell activity in A549 and HCT116 cancer cell lines.


In addition to compound 1, a number of other fragments containing the aminopyrimidine substructure were also identified as hits. This moiety seems to be a privileged pharmacophore for Hsp90: for a fun read, check out this 2007 paper from researchers at Abbott, in which fragments are linked together in a couple different ways as well as grown. As in the more recent paper, the protein displays a remarkable degree of flexibility to accommodate small molecule binders.