Why NOT AMW

I have been thinking a lot lately about library design, especially after the roundtable at breakfast in SD. I found a rant from an old friend/colleague, from the very early days of FBDD for me (2002 or 2003).   With his permission, but no citation for obvious reasons, I am reprinting it here.  I find this particularly interesting as AMW is still being used as recently as last year in papers describing fragment library design. 

The Average Molecular Weight [Ed:AMW] is currently an accepted orthodoxy within the Medicinal Chemistry community, a role reinforced by the recent popularity of things like the Rule of Five. To be sure, molecules with large molecular weights are not typically observed to be successful drugs.

In all the rest of this note, I’d like to focus on the common scenario of selecting molecules for purchase or testing. I’ve often seen people apply AMW cutoffs or scalings to these processes. I’d like to show why this may be sub-optimal.

First, I contend that the number of heavy atoms may be a much better proxy for “size” than AMW. Certainly, if you want to discriminate against heavier elements like Phosphorus, Sulphur, Chlorine, Bromine and Iodine, then by all means, use AMW. But with an AMW contribution of 126, a molecule with a single Iodine atom would be considered heaver/less desirable than the same molecule with a C6H9N2O substituent! Again, if you have good reasons for wanting to suppress these elements, AMW is a (very) good way of doing that.

The case of the third row and beyond elements is fairly straightforward, but what happens within the organic group Carbon, Nitrogen and Oxygen. Does AMW vs natoms mean anything in there?

Imagine we are selecting molecules for an assay and imposed an AMW cutoff of 180 – admittedly very low. We would then ignore Aspirin, with an AMW of 180.66, and instead we would test the “lighter” .

But it gets worse. Imagine if the Lilly chemists looking for antidepressants had imposed an AMW cutoff of 308. That would have excluded Prozac, with an AMW of 309,

and instead they would have tested the “more desirable lighter molecule”with an AMW of just 306.

Again, an AMW cutoff or bias could see us miss Zyprexa, AMW 312.4 and instead use the “lighter” molecule AMW 308.

Or the even more lighter still Oxygen variant! Now, of course, no medicinal chemist faced with the choice between these molecules would choose the second one, that’s obvious. Logp estimates would probably eliminate molecules that are all or mostly carbon atoms.

But what the examples above show is that in any automated system that is basing decisions on AMW, there will be a systematic, albeit small, bias towards Carbon atoms and against Nitrogen and Oxygen atoms. Why? A single CH2 group contributes 14 to AMW, whereas an NH contributes 15, and a two connected Oxygen atom 16. As terminal groups, CH3 is 15, NH2 is 16 and OH is 17. A Pyridine Nitrogen contributes 14, but an aromatic Carbon contributes 13, t-butyl groups preferred over CF3. Carbon wins the low AMW contest every time.

Now, how significant is this. Probably very small in most practical applications. I’d say that if you are setting up some kind of automated process, and you have equivalent access to AMW and the number of heavy atoms, use the number of heavy atoms in order to eliminate any small, pro-Carbon bias.

Conclusion

We see that automated procedures using AMW instead of natoms, will not only systematically suppress elements like P, S, Cl, Br and Iodine, but may also work to drive out N and O atoms as well!

Fragments versus Ras

The protein Ras is one of those cancer targets that’s been around forever and has rebuffed countless attacks by many researchers using multiple strategies. The most obvious ligand pocket is the one where GTP and GDP bind. Unfortunately, these molecules bind with picomolar affinity, and they are present at very high concentrations in cells. To try to find an alternative small molecule binding site, Guowei Fang and colleagues at Genentech took a fragment-based approach, and have reported their results in a recent issue of Proc. Nat. Acad. Sci. USA (as well as at a recent meeting).

The researchers used 1D NMR screening (STD) to screen 3300 fragments in pools against GDP-bound KRas; 240 hits were retested as single compounds and further validated by 2D NMR (HSQC). This resulted in 25 confirmed hits. Surprisingly, all of them appeared to bind to one region of the protein some distance from the GDP-binding site. Subsequent crystallography confirmed that these fragments bind to a small pocket about 250 ų in size. However, there could be more here than meets the eye, as this is a fairly flexible region of the protein, and the pocket changes shape in response to different fragments.

 At least one of the fragments, DCAI, not only binds to Ras, it also inhibits the association of Ras with the protein SOS, thereby blocking nucleotide exchange and Ras activation. Interestingly, this blocking activity was distinct from binding activity; the fragment BZIM has comparable affinity as judged by NMR, yet does not inhibit the interaction with SOS.

Somewhat surprisingly given its low affinity, DCAI is also active in cell-based assays. And although the molecule is still a long, long way from a drug, the results are encouraging. Perhaps a fragment-based approach will finally succeed against this target. Or perhaps Ras will yet again reveal its intransigence.

Fragment linking leads to nanomolar leads for LDHA

The metabolic changes seen in cancer cells were first observed decades ago, but only recently have companies gotten serious about exploiting these changes for new therapies. One potential target is the enzyme lactate dehydrogenase, in particular the LDHA isoform. In a paper recently published in J. Med. Chem., Richard Ward and colleagues at AstraZeneca describe a fragment linking strategy to generate nanomolar leads for this enzyme.

An internal HTS campaign gave a hit rate of about 1%. Worse, none of these hits confirmed in ligand-observed NMR assays, and follow-up studies suggested that the activity was often caused by heavy metal impurities, particularly silver. (This is yet another form of false positive of which to beware.)

When HTS doesn’t succeed, fragments start looking more attractive, so the researchers used their NMR assay to screen about 1000 fragments in pools of 6, with each fragment present at 0.2-0.4 mM. This resulted in 44 hits; of the 27 of these chosen for follow-up 13 gave quantifiable binding, with affinities ranging between 0.3-4.2 mM. Happily, some of these (such as compound 12, below) could be soaked into crystals of LDHA, and structure determination showed these fragments bound in a pocket that normally binds the adenine portion of NADH. An SPR assay was developed and used to screen 350 analogs of some of these hits, and although some improvement in potency could be seen, ligand efficiencies remained more or less the same. (All Kd values in the figure below are taken from SPR data.)

The active site of LDHA is quite long, and so the researchers sought to span its length to obtain decent affinity. The adenine pocket where the identified fragments bind is located some distance away from the site where LDHA’s product lactate binds, and so linking fragments from both pockets would generate molecules spanning the desired region. Finding fragments that bind in the lactate pocket posed a challenge, however, as none of the first set of fragment appeared to bind there. Because lactate is negatively charged, the researchers assembled a specially-tailored 450 fragment library with a high proportion of acids and screened compounds at 2.5 mM using SPR. This screen resulted in 40 hits, and although many of them were nonspecific (they also bound to denatured LDHA!) some hits were specific, including compound 20.

A crystal structure revealed that fragment 20 bound, as expected, some distance from fragment 12, so the researchers generated libraries around both fragments to try to help bridge the gap. About 150 analogs were made around compound 12 and about 70 analogs were made around compound 20, resulting in compounds 24 and 25. Although not necessarily more potent than the initial fragments, crystallography revealed portions of these that were positioned more closely to one another, and linking them to form compound 26 gave a very satisfying boost in potency. This was actually the first linked compound made, and it was also the first compound to show activity in the enzymatic assay. Subsequent optimization was able to drive the potency down to low nanomolar (compound 34). Not surprisingly, the acidic nature of the compounds precluded cellular activity, but some diester derivatives showed sub-micromolar activity.

This is a thorough and engaging account of how fragment-based methods can tackle a difficult target. Although the compounds still need work, they represent good starting points for further optimization and for better exploring the validity of LDHA as a cancer target.