"We wonder whether terms like druggable and undruggable are misguided, and whether we might be better served considering terms such as assayable and unassayable when judging protein targets." - Zhang and Cravatt, 2024

I enjoyed this comprehensive review on covalent ligand discovery for cancer targets from the Cravatt lab which has been a pioneer in activity-based protein profiling (ABPP) for a long time.

The review notes many examples of how ABPP was used to locate and target unusual cancer targets: KEAP1, the RNA-binding protein NONO, JAK1, E3 ligases, adaptor proteins, protein-DNA interactions and protein-protein interactions. In all cases cysteines were targeted, so the universe of lysines, tyrosines and other residues is going to be even larger.

Of all these the E3 ligase and JAK1 examples are particularly interesting, the E3 ligase example because it shows that even fractional engagement of cysteines can trigger ubiquitination of the target (the whole business of TPD is highly non-linear), and the JAK1 example because it shows the value of cell-based screening in finding cryptic pockets.

There's a marvelous quote toward the end that stems from the JAK1 case that I have quoted above. The cryptic pocket in the protein was seen only in cell-based and not in biochemical assays, pointing to the value of chemoproteomics in live cells:

“These screens can be performed in living cancer cells, enabling the discovery of cryptic druggable pockets that may have been overlooked by more traditional studies using purified proteins. One compelling example is the recent discovery of the allosteric, isotype-restricted JAK1 inhibitor VVD-118313 (Kavanagh et al. 2022). Chemical proteomics data suggest that this site on JAK1 is actually quite druggable in that it reacts with a range of electrophilic fragments and more elaborated covalent compounds and, in some cases, with excellent potency (low-nanomolar engagement in cells) and proteome-wide selectivity (Kavanagh et al. 2022). Given the extensive efforts within the pharmaceutical industry to develop JAK inhibitors (Schwartz et al. 2017), it merits asking, Why wasn't this allosteric site discovered previously? We would posit that such an allosteric, druggable site may be paradoxically much easier to discover in cells compared with purified proteins, which are often studied as simplified domains (e.g., the kinase domain in the case of JAKs) that may obscure allosteric modes of regulation.”

The context-specific nature of the cryptic pocket makes the authors wonder whether "druggable" and "undruggable" should be replaced by "assayable" and "unassayable".

Personally I love this change in terminology and am heartily on board. After all when we say "undruggable", what we really mean is "undruggable as measured by current assays", so why not skip the abstract theoretical definition and use an instrumental definition anyway, especially when the former might be misguided and context-dependent as is the case with JAK1? Some have pointed out that “undruggable” was always understood by any reasonable-minded person to mean “undruggable until further notice”. But I still don’t like the label because it has a ring of finality. “Unassayable” seems much better at giving the definition a provisional air, a feeling that it’s, well, undruggable only until the next killer assay comes along. There’s also the point that for medicinal chemists in particular, “undruggable” even now largely means “not targeted by a small or large molecule”. But as the Cravatt review notes, of 628 cancer targets identified by CRISPR screens, only 40 have been targeted by small molecules. So are the other 588 druggable or undruggable? Having a more operational definition of a protein target would make the distinctions much clearer.

It’s worth noting how so many targets recently have turned from undruggable to druggable not because of some intrinsic change in our philosophy of picking these targets but because of different technologies to target them and detect target engagement; in other words, through better assays. Take the classic example of KRAS which now has a dozen or more clinical candidates targeting it just in the last decade or so; that barrier fell quickly and precipitously. The turning point in case of KRAS G12C was a fragment-based tethering approach from UCSF in 2013 that opened the gates to covalent discovery, culminating in the discovery of Sotorasib, the first direct KRAS-targeting covalent drug from Amgen. The key innovation here was a new platform of covalent targeting, again translating to a different assay. A target that previously wasn’t assayable through non-covalent means became assayable through covalent ones.

This discussion reminds me of Sydney Brenner’s evergreen quote:

“Progress in science is driven by new techniques, new discoveries and new ideas, probably in that order.”

It’s sometimes easy to underestimate the impact of new techniques and assays, not just in uncovering new targets but in uncovering new paradigms. Chemists and biologists should especially be familiar with this idea of tool-driven revolutions: think of how NMR spectroscopy or x-ray crystallography or gene sequencing have revolutionized drug discovery. Scientists like big ideas. The next paradigm shift is often seen to come from a fundamental change in thinking. But more often than not it comes from a new way to measure and track things. And in drug discovery that can count for a hell of a lot.