26 March 2018

Acceptable tradeoffs: From fragment hit to fragment lead against mGluR2, without structures

Membrane-bound proteins such as GPCRs are often ignored by practitioners of FBLD in part because – Heptares notwithstanding – they are usually difficult to characterize structurally. This seems like a missed opportunity. A large fraction of drugs target GPCRs, and the vast majority of these were developed without crystallographic information, so why is the fragment community so fixated on structure? A paper just published in J. Med. Chem. by György Szabó, György Keserű and colleagues at Gedeon Richter, the Hungarian Academy of Sciences, and Mitsubishisi Tanabe shows how much can be done without strcutures.

The researchers were interested in metabotropic glutamate receptor 2 (mGluR2), a popular target for schizophrenia. In particular, they sought positive allosteric modulators (PAMs), which act outside the main ligand binding site to enhance signaling. A functional screen yielded compound 4 as a fairly potent fragment-sized hit. Comparison with other larger reported inhibitors suggested growing could be productive, leading to molecules such as compound 5, with sub-micromolar activity. Further optimization for potency and ADME properties led to compound 29, with low nanomolar potency.

Unfortunately, this molecule is very lipophilic (cLogP > 5), resulting in poor solubility, high plasma protein binding, and thus limited efficacy in a mouse pharmacodynamic model. All attempts to reduce lipophilicity came at the cost of potency.

To determine which elements of compound 29 were most important for binding, the researchers turned to group efficiency analyses; that is, they systematically removed different chemical groups and weighed the loss in binding energy versus the reduction in size. Even though they could not visualize precisely how each group interacted with mGluR2, the researchers could measure it. This effort revealed that the biaryl moiety was not particularly efficient, and although trimming it came at a cost in potency, this was compensated for by improved ligand efficiency. Substitution at another position off the initial fragment led to a satisfying boost in activity (compound 30). Further optimization for pharmacokinetic properties led to the fragment-sized compound 60, which is considerably less potent in vitro than compound 29 but which has better brain penetration and also better efficacy in two mouse models.

Several lessons can be drawn from this story. First, as Mike Hann warned seven years ago, molecular properties should not be ignored in the push for potency. Indeed, despite the 25-fold decrease in potency for compound 60 compared with compound 29, the smaller molecule is more effective in vivo. This is reminiscent of the Merck verubecestat story, which also involved optimization of a fragment hit to a potent but lipophilic lead that was ultimately abandoned in favor of an initially less active but more ligand-efficient series. The second lesson is that in vitro models can only take you so far. And finally, creative chemists are able to advance fragments even in the absence of structural information. Hopefully more of them will give it a try.


Peter Kenny said...

Hi Dan,

While having structural information available certainly makes life a lot easier, I have never been convinced that this is absolutely essential for exploitation of a fragment hit. In general, I would always want to test analogs of hits to generate initial SAR and, possibly, better starting points for structural studies. That said, 29 is one or two orders magnitude more potent than a typical fragment and this makes it a less convincing as evidence that one can take fragments forward without structural information.

It is vital to be able to quantify affinity (or potency) in order to do FBLG without structures and one will also need to synthesize more compounds than would be typical in F2L campaigns. For aromatic carbon atoms, scanning with methyl, chloro and aza would be expected to be particularly informative. Phenolic hydroxyl groups may identify hydrogen bond acceptors in the protein that can be exploited. Pendant aromatic rings can be exchanged for (cyclo)alkyl groups and linking CH2, O, N, S can be exchanged for each other.

I’m guessing that the bromine in 30 may be halogen bonding to the target and the IC50 for the chloro analog is likely to shed some light on this.

Dan Erlanson said...

Hi Pete,

These are good points, and I agree that the authors were starting from an enviable position in terms of initial activity. They also benefited from other published lead series against this target. The more difficult challenge is when, as you said, you start from a much weaker fragment against a completely novel target: how long do you work on a series when modifications don't improve affinity?

Peter Kenny said...

Not knowing when to put a project out of its misery is a problem throughout drug discovery. If neither methyl scan nor screening analogs (including bioisosteric replacements) led to an increase in potency then I would certainly be worried. A lot would depend on the value of the target and what other options were available. In the early days of HTS at Zeneca, a project manager asked me if I thought we had a lead and my response was, “depends how desperate you are”.

When mapping SAR, one can use what has been already been observed to eliminate (or at lease deprioritize) options. If, for example, methylation at C4 leads to five-fold increase in affinity then the 4-aza analog drops a long way down the priority list. Introducing hydrogen bond donors into molecular structures is typically more difficult than introducing hydrogen bond acceptors and one probably needs to rely more on these being present in some of the screening library fragments. Analogously, I would assume that ionized fragments in screening library will locate anion/cation binding sites and I would not usually recommend the introduction of an ionizable functional group to a fragment hit.

While I would always want structural information (and direct affinity measurement) for any project, the belief in the fragment community that structural information is absolutely essential is may be something that is encouraged by biophysicists.

Unknown said...

Hi Dan,
you have raised a very interesting point, that parallels a theme that Peter and I had just been discussing, which is the option of changing the fragment core ("scaffold hopping"). There are certainly examples of the latter in the literature, although the balance (which may reflect successful campaigns?)is certainly in favour of retaining the "core" of the fragment, as with the benzotriazole in the mGluR2 work. Would a benzodiazole have worked? Peter (above) does suggest point changes at an aromatic carbon to CH to CMe, C-Cl or N. Of course the non-binders in the original fragment screen may also rule out some permutations.
In a recent example McDaniel at Abbvie J. Med. Chem. 2017, 60, 8369-8384, did swap from a moderate affinity (160uM) pyridazinone to a pyridone core, but they were guided by an X-ray structure.
Perhaps the broad question as you suggest is enthusiasm for ever continuing (or even starting) fragment screening without the promise of structural data? No doubt some people have the luxury of de-prioratising a project until that is available. So the question of 'scaffold hopping" sits as a subset of that?

regards, john

Peter Kenny said...
This comment has been removed by the author.
Peter Kenny said...

I think the drug discovery community would benefit from a broader discussion of the feasibility of doing FBLD without structural information (John suggested this in an email about a month ago and we were still discussing how best to generate useful discussion when Dan posted). A related question concerns how useful biochemical assays are for screening fragments and this article is particularly useful in that it shows how assay interference can be assessed and, if it not too severe, accounted for.

The bioisosteric modifications discussed by John are relatively conservative. Aromatic nitrogen is polar but the N->CH transformation tends to strengthen HB acceptors that are part of the same pi-system while weakening the HB donors. It is possible to calculate both HB acidity and basicity and this can be useful when assessing bioisosteric relationships in the absence of protein structural information.

A key question for me would be whether both N2 and N3 of the triazole accept HB donors from the protein. As discussed here one might expect formation of two proximal hydrogen bonds to be especially beneficial for affinity (frustrated hydration argument). If this were the case (indazole and benzimidazole suggested by John would be informative) then moving N1 to the adjacent bridgehead position would be an obvious thing to do.

Dan Erlanson said...

I completely agree that doing FBLD in the absence of structure can be successful but is underappreciated. Indeed, the reliance on crystallography in drug discovery is a relatively new phenomenon; even as late as the 1990s folks would sometimes joke that by the time you had a crystal structure of your protein, the compound was already in the clinic.

As to the question of scaffold hopping, what do you think of Astex's Fragment Network approach?

Peter Kenny said...

The fragment network looks like it could be quite useful and Rich has done a good job. Nevertheless, I would see this as nice-to-have rather than can’t-do-without.

I can certainly remember the early days of FBLD when crystal structures were hard to come by although everything became so much easier and clearer after Ro3 was published (I used to genuflect twice daily to a printout of Ro3 that I’d stuck to the wall over my desk). I can even recall the difficulty of obtaining crystal structures for fragments shown to bind being invoked as evidence for the need to do crystallographic screening. I remember thinking maybe not…

To FBLD without crystal structures, it would be necessary to have synthetic capability that is flexible and can react (no pun intended) rapidly to project requirements. Assay capability, compound managemen, cheminformatics and protein science all have to be first rate because you’ll need synergies between these to compensate for lack of protein structures. Also need to run each project flexibly on its individual merits rather than shoehorning each project into a template dreamed up by grinning LeanSixSigmoids. Target selection is likely to be just as important (possibly more so) than when protein structures are available.

I wonder what FBLD would look like for RNA targets?