The First Protein-Carbohydrate Interaction Inhibitor

Immunomodulation is all the rage, particularly in terms of cancer.  Hyaluronan (HA) is a component of the extracellular matrix.  CD44 is a major cell receptor of HA and its fragments.  Its differential response to HA or HA fragments leads to the different biologies. However, due to the different biologies there are limitations to what can be done with HA fragments.  A selective inhibitor of CD44-HA would prove an invaluable tool. Also, in the grandest of Grant-Application-ese, a small molecule inhibitor could be useful for inflammatory diseases and cancer.  

In this paper, results are presented towards this goal.  The hyaluronan-binding domain (HABD) of CD44 is competent to bind oligosaccharides, and even better has been crystallized. In terms of the binding site:

From the structural data, one might conclude that CD44 is not an easily druggable target. The murine HABD−HA complex reveals an extended HA binding site with surface area exceeding 800 Å^2 and molecular binding stabilized by a large number of weak interactions involving at least seven consecutive saccharide units of HA. The HABD has no well formed or deep pockets that would serve as attractive binding sites for small molecule inhibitors and is known to undergo small but important conformational changes upon binding HA. In many respects, the protein−polysaccharide complex resembles protein−protein interactions that are difficult to disrupt effectively with small molecules.

Using SPR, they screened 1000 fragments from the Maybridge Ro3 fragment library at 5 mM resulting in a 4% hit rate. 21 fragments were crystallized (Table 3 in SI), resulting in 5 co-crystals.  Cpds 1 and 2 were deemed worthy based upon their binding site. 

They were poor in terms of blocking HA binding to CD44.  So, they then did some Analog by Catalog and some merging, based upon other scaffolds and ended up with 5a, which has improved affinity for the HABD and had a measurable ability to block HA binding to the HABD.  

This is a interesting paper to me simply because of the target choice:protein-carbohydrate interaction.  I believe this is the first example of FBHG against a PCI. 

Metallophilic fragments – or PAINS?

Several years ago we described fragment libraries designed to chelate metal ions. The idea is that these could serve as affinity anchors to target metal-containing proteins. However, in designing any fragment library, it is essential to avoid pan-assay interference compounds (PAINS), molecules that act through pathological mechanisms. A new paper in Bioorg. Med. Chem. Lett. by Amy Barrios and collaborators at the University of Utah and University of California San Diego illustrates one of these mechanisms.

The researchers were interested in a protein tyrosine phosphatase (PTP) called LYP. PTPs contain catalytic cysteine residues and are thus particularly prone to false positives caused by oxidation or adventitious metals such as zinc. The researchers screened a library of 96 metal-chelating fragments against LYP in the presence or absence of zinc to find fragments that could either rescue the enzyme or inhibit it, either cooperatively with zinc or on their own. Not surprisingly, they were able to find several fragments that could rescue LYP from added zinc, presumably by coordinating the metal and removing it from the active site.

The most potent inhibitor of the enzyme was 1,2-dihydroxynaphthalene, with an IC₅₀ of 2.5 µM. However, the researchers quickly discovered that it was time-dependent, showing greater potency the longer it was incubated with the enzyme. This is a classic sign of something funny going on, and the researchers realized that molecules like this can oxidize spontaneously. In fact, the oxidized product (1,2-naphthoquinone) is even more potent, and also time-dependent (IC₅₀ = 1.1 µM after 2 hours). Not only can napthoquinones directly modify cysteine residues, they can generate reactive oxygen species that can in turn modify cysteine residues – this very molecule was reported as doing so more than five years ago.

This is the kind of mechanism you want to avoid, and it is likely to rear its ugly head whenever compounds of this ilk are screened, particularly against a protein with an active-site cysteine. Hopefully this publication will serve as a warning to folks who may be using this screening library.

One could argue that this paper falls into the category of “you probably already knew this.” However, even if you knew it, many others likely did not. At the ACS meeting symposium on PAINS earlier this month, Kip Guy urged people to publish their PAINS stories. They may not be the sexiest papers, but if they inform others what to avoid they may be among the most useful.

Like the Elves to Valinor?

There are some perqs to being a consultant.  I go to work most days in my jammies, I can work from anywhere with WiFi, and get to work with great people.  I also have a lot of freedom with what I can say/do/write.  Back in April at the CHI Drug Discovery Chemistry meeting, I ran into an editor from an ACS publication.  Now, if you have ever been at a conference with me you know I am a social butterfly.  New face, let's chit chat.  This editor was very nice and after some social niceties, I asked who is refereeing your journal?  This was a general query having to do with some poor quality fragment papers recently, and before I even saw this!  She was polite and asked what do you mean?  I said, well I blog on fragments and we come across crap in JMedChem and the like all the time.  Like this.  I think at this point she obviously was thinking of me like this .  Then I mentioned that the blog was Practical Fragments and her attitude changed.  So, this blog has "street cred"!!  She then asked me if I would write a Viewpoint on the future of FBDD.  Sure, I said.  This could be fun.  

At the same time, I was preparing a talk for the Zing FBDD conference.  So, I decided to make that talk (that's a link to SlideShare, I hope it works) a demo for the Viewpoint. Give it a read, its short and sweet (at least I think).  

For those of you without 10 minutes to spare, here are the take home points from the Viewpoint:

• Fragonomics is a key component of most (all?) hit generation processes
• It is a fully mature field.  The current debates amount to quibbling about details.
• It has no future as a stand-alone field.  But there are still challenges for it to tackle.
• Medchemists no longer rule the field.  

"The age of the medchemist is over; now is the time of the biophysicist." 

 

 This got some serious push back at the conference, and I expect (hope?) it will here too.  Of course, I am paraphrasing this. I am not suggesting that medchemists make like the Elves and sail off to Valinor.  They are still incredibly important and can still play a role in early hit generation.  However, the focus, thanks to Fragonomics, is vastly different.  A chemically trained biophysicist can run a fragment-based hit generation project and you don't have to have engage the most precious resource (medchem) until well into the process.  This is a good thing.  

Well, I have planted my flag.  Now its time for you all to weigh in.

248th ACS National Meeting

The Fall ACS National Meeting was held in my beautiful city of San Francisco last week, and a number of topics of interest to Practical Fragments were on the agenda.

First up (literally – Sunday morning) was a session on pan-assay interference compounds (PAINS) organized by HTSPAINS-master Mike Walters of the University of Minnesota. Mike developed his interest in PAINS like many – from painful experience. After screening 225,000 compounds against the anti-fungal target Rtt109, he and his group found several hits that they identified as PAINS, but not before spending considerable time and effort, including filing a patent application and preparing a manuscript that had to be pulled. One compound turned out to be a “triple threat”: it is electrophilic, a redox cycler, and unstable in solution.

Mike had some nice phrases that were echoed throughout the following talks, including “subversively reactive compounds” and SIR for “structure-interference relationships,” the evil twin of SAR. To try to break the “PAINS cycle” Mike recommended more carefully checking the literature around screening hits and close analogs (>90% similarity). Of course, it’s better if you don’t include PAINS in your library in the first place.

Jonathan Baell (Monash), who coined the term PAINS back in 2010, estimated that 7-15% of commercial compounds are PAINS, and warned that even though PAINS may be the most potent hits, they are rarely progressable, advice that is particularly needed in academia. For example, the majority of patent applications around the rhodanine moiety come from academia, whereas the majority of patent applications around a more reasonable pharmacophore come from industry. Jonathan also warned about apparent SAR being driven by solubility. Finally, he noted that while it is true that ~6.5% of drugs could be classified as PAINS, these tend to have unusual mechanisms, such as DNA intercalation.

As we discussed last week, anyone thinking about progressing a PAIN needs to make decisions based on sound data. R. Kip Guy (St. Jude) discussed an effort against T. brucei, the causative agent of sleeping sickness. One hit from a cellular screen contained a parafluoronitrophenyl group that presumably reacts covalently with a target in the trypanosome and was initially deemed unprogressable. However, a student picked it up and managed to advance it to a low nanomolar lead that could protect mice against a lethal challenge. It was also well tolerated and orally bioavailable. Kip noted that in this case chemical intuition was too conservative; in the end, empirical evidence is essential. On that note he also urged people to publish their experiences with PAINS, both positive and negative.

There were a scattering of nice fragment talks and posters. Doctoral student Jonathan Macdonald (Institute of Cancer Research) described how very subtle changes to the imidazo[4,5-b]pyridine core could give fragments with wildly different selectivities. I was particularly tickled by his opening statement that he didn’t need to introduce the concept of fragment-based lead discovery in a general session on medicinal chemistry – another indication that FBLD is now mainstream.

Chris Johnson (Astex) told the story of their dual cIAP/XIAP inhibitor, a compound in preclinical development for cancer. As we’ve mentioned previously, most IAP inhibitors are peptidomimetics and are orders of magnitude more potent against cIAP than XIAP. Astex was looking for a molecule with similar potency against both targets. A fragment screen gave several good alanine-based fragments, as found in the natural ligand and most published inhibitors, but these were considerably more potent against cIAP. They also found a non-alanine fragment that was very weak (less than 20% inhibition at 5 mM!) but gave a well-defined crystal structure. The researchers were able to improve the affinity of this by more than six orders of magnitude, ultimately identifying compounds with low or sub-nanomolar activity in cells and only a 10-fold bias towards cIAP. This is a beautiful story that illustrates how important it is to choose a good starting point and not be lured solely by the siren of potency.

Alba Macias (Vernalis) talked about their efforts against the anti-cancer targets tankyrases 1 and 2 (we’ve previously written about this target here). In contrast to most fragment programs at Vernalis, this one started with a crystallographic screen, resulting in 62 structures (of 1563 fragments screened). Various SPR techniques, including off-rate screening, were used to prioritize and further optimize fragments, ultimately leading to sub-nanomolar compounds.

The debate over metrics and properties continued with back-to-back talks by Michael Shultz (Novartis) and Rob Young (GlaxoSmithKline). Michael gave an entertaining talk reprising some of his views (previously discussed here). I was happy to see that he does agree with the recent paper by Murray et al. that ligand efficiency is in fact mathematically valid; his previous criticism was based on use of the word “normalize” rather than “average”. While this is a legitimate point, it does smack of exegesis. Rob discussed the importance of minimizing molecular obesity and aromatic ring count and maximizing solubility, focusing on experimental (as opposed to calculated) properties. However, it is important to do the right kinds of measurements: Rob noted that log D values of greater than 4 are essentially impossible to measure accurately.

Of course, this was just a tiny fraction of the thousands of talks; if you heard something interesting please leave a comment.

Intentionally dirty fragments

Practical Fragments has tried to publicize the dangers of pan-assay interference compounds, or PAINS. These compounds show up as nuisance hits in lots of assays. So what are we to make of a new paper in Curr. Opin. Microbiol. by Pooja Gopal and Thomas Dick, both at the National University of Singapore, entitled “Reactive dirty fragments: implications for tuberculosis drug discovery”?

As the researchers point out, several approved anti-tuberculosis drugs are fragment-sized and hit multiple targets; they are “dirty drugs”. For example, isoniazid (MW 137, 10 heavy atoms), is an acylhydrazide that is metabolically activated and forms an adduct with an essential cofactor, causing havoc to the pathogen. Ethionamide (MW 166, 11 heavy atoms), a thioamide, works similarly. The fact that these molecules are so small probably allows them easier passage through the microbe’s rather impermeable cell membrane, and the fact that they hit multiple targets may make it more difficult for the organism to develop resistance. The researchers conclude:

The success of small dirty drugs and prodrugs suggests that fragment-based whole cell screens should be re-introduced in our current antimycobacterial drug discovery efforts.

While it is true that many antimicrobials do have reactive warheads, and it is also true that there is a huge need for new antibiotics, I worry about this approach. Not only is there an increased risk of toxicity (isoniazid in particular has a long list of nasty side effects), it can be very hard to determine the mechanism of action for these molecules, complicating optimization and development. As evidence, look no further than pyrazinamide (MW 123, 9 heavy atoms). Despite being used clinically for more than 60 years, the mechanism remains uncertain.

Fragment-based lead discovery is typically more mechanistic: find an ideal molecule for a given target. Indeed, much of modern drug discovery takes this view. Gopal and Dick propose a return to a more phenomenological, black-box approach. This may have value in certain cases, but at the risk of murky or – worse – misleading mechanisms.

If you do decide to put PAINS into your library, you might want to read a new paper in Bioorg. Med. Chem. by Kim Janda and collaborators at Scripps and Takeda. They were interested in inhibitors of the botunlinum neurotoxin serotype A (BoNT/A), which causes botulism.

Since BoNT/A contains an active-site cysteine, the researchers decided to pursue covalent inhibitors, and the warheads they chose, benzoquinones and napthoquinones, are about as PAINful as they get. However, in contrast to other groups, they went into this project with their eyes wide open to the issue of selectivity and examined the reactivity of their molecules towards glutathione. Reaction with this low molecular weight thiol suggests that a compound is not selective for the protein. Not surprisingly, selectivity was generally low, though a few molecules showed some bias toward the protein.

The researchers also tried building off the benzoquinone moiety to target a nearby zinc atom, and although they were able to get low micromolar inhibitors, these no longer reacted with the cysteine; apparently when the ligand binds to zinc, the protein shifts conformation such that the cysteine residue is no longer accessible.

To return to the premise of Gopal and Dick, there can be a therapeutic role for dirty molecules. The fact that dimethyl fumarate is a highly effective blockbuster drug for multiple sclerosis calls for a certain degree of humility. However, if you do decide to pursue PAINS, you should do so in full awareness that your road to a drug – not to mention a mechanism – will likely be much longer and more difficult.

Measuring 3D Fragments

3D fragments is still a topic up for discussion/debate.  Of course, determining what is a 3D fragment is also open to debate.  As presented last year, nPMI seems to be the current "best" method.  FOB Chris Swain has done some neat analyses around nPMI and included it in his overview of commercial libraries.  Last year at the NovAliX Biophysics conference, Glyn Williams from Astex presented the "plane of best fit" (PBF) as a superior way to analyze 3D-arity.  At the recent Zing FBDD conference, Chris Murray presented the "plane of best fit" and argued again that it was superior.  So, of course, I asked Chris Swain if he had compared them, and he said, but wait a minute.  

Well, Chris was able to compile it with the help of his son, Matt (obviously the apple does not fall from the tree).  So, go check out his comparison of nPMI with PBF.  He ran 1000 fragments through both and concluded this: 

Whilst both descriptors are intended to provide information on the 3D structure of the molecule it looks like the PBF provides more granularity which may be particularly useful when looking at small fragments.

 So, I present this as a "go get 'em, folks!".  I am particularly interested to know if people are currently using PBF and if their results jibe with Matt's. 

Hydrogen/Deuterium Exchange Mass Spectrometry (HDX MS)

Mass spectrometry does not come up frequently in the context of fragment-based lead discovery (though see here, here, and here). A new paper in Bioorg. Med. Chem. Lett. from Matthew Carson and colleagues at Lilly, with collaborators from Scripps, seeks to change that by describing a technique that can elucidate binding sites for fragment hits.

The researchers were interested in the vitamin D receptor (VDR), an osteoporosis target. Upon binding to ligands such as the D vitamins, this nuclear hormone receptor changes conformation and binds to another receptor, retinoid X receptor (RXR), to control gene expression. The biology quickly gets complicated, but suffice it to say that there is a need for ligands that behave differently than the D vitamins. Enter fragments.

The researchers assembled a collection of about 10,000 compounds, most of which had fewer than 23 non-hydrogen atoms. These were screened at 0.1 or 1 mM concentration in a fluorescence polarization binding assay, resulting in 417 hits. These were then tested in an AlphaScreen assay for compounds that would enhance or decrease binding to RXR (ie, agonists or antagonists). The screen came up empty for agonists. VDR is a member of the same family as the PPARs, for which fragment screens have delivered agonists, so this result was a bit disappointing. The researchers speculate that the fragments may not be large enough to induce the required conformational changes in VDR.

The researchers were more successful finding antagonists: 247 fragments with “lean values” > 0.25 (corresponding to ligand efficiencies > 0.35 kcal/mol/atom). About 2000 analogs of these were then tested, leading to more hits, some of which were quite potent, and 13 of which are shown in the paper. Although some of these look structurally reasonable, one is a toxoflavin-type molecule with a catechol attached that looks disconcertingly similar to a molecule I proposed as an April Fools’ joke. Presumably they are keeping the more interesting structures confidential.

The ultimate goal is to find agonists. Antagonists could potentially be grown into agonists, and to do so it would be helpful to know how they bind. Unfortunately, co-crystallography proved unsuccessful, so the researchers turned to hydrogen deuterium exchange mass spectrometry (HDX MS).

In HDX MS, a protein-ligand complex is exchanged into D₂O for seconds to minutes, allowing exchangeable protons (such as those in the amide backbone of the protein) to exchange with deuterium. The reaction is stopped by lowering the pH, the protein is digested into individual peptides, and these are analyzed by mass spectrometry to assess the extent of exchange. If an amide makes a hydrogen bond in a highly structured region of the protein it will be less prone to exchange than if it is in an unstructured region of the protein. Therefore, if a ligand induces structural changes in the protein these should manifest themselves by altering the exchange rates, and if the crystal structure is known this provides a rough map of which regions of the protein are affected by ligand binding.

The 13 fragment hits were tested in HDX MS with a 200-fold excess of fragment to protein. Of these, 6 stabilized the protein, as assessed by decreased H-D exchange. The stabilized regions overlap with the regions stabilized by the natural ligand, vitamin D₃, though the extent of the regions and degree of stabilization is considerably less, consistent with the fragments binding within a smaller footprint.

Of course, since we don’t have co-crystal structures, it is difficult to interpret the HDX data precisely. Still, it is nice that fragments can produce a signal in this type of assay. It will be interesting to apply this technique to better characterized systems to see how general it is.