Molecular weight watchers

Most medicinal chemistry programmes start off with one or more lead compounds, usually picked from a list of hits in a high-throughput screen (HTS) of some sort. HTS will give you all kinds of things, of course - lists of compounds from previous programmes against similar targets, scattered oddities that no one can remember making or purchasing, compounds that on inspection turn out to hit over half the assays they see regardless of the target class (and should be thrown away before they go on wasting everyone’s time). Mixed in with these, if you’re lucky, will be a few chemically appealing compounds with reasonable affinity for the target, representing structural classes that you haven’t worked to death already.

Watch your weight 

So where to begin? One rule of experienced medicinal chemists is to start small. That’s because compound structures almost always expand as a project goes on. If you start with something with a reasonable molecular weight for a drug, you run a very real risk of finishing with something well over the line. (Where is that line, exactly? Opinions vary, but it’s safe to say that everyone’s very happy with a molecular weight in the 300s, well used to seeing compounds in the 400s, and will be increasingly nervous in the 500s). If you start small, on the other hand, with luck (and close attention) you can end up with a compound with more reasonable properties. 

This concept has been given a firmer footing in recent years with the growing popularity of ligand efficiency, basically a ratio of potency to molecular weight. Once you get used to thinking this way, you start to view your screening data differently. Compounds with nanomolar potency don’t look so appealing when they weigh 650; significant portions of a molecule that size aren’t really contributing to its binding, or it would be much more potent than that. The smaller ones whose potencies stretch out into the micromolar range, on the other hand, could be well worth investigating. Taken to its extreme, this leads you to fragment-based drug discovery, whose practitioners start with much smaller (and much less potent) compounds than many traditional med-chemists feel comfortable working with. In a few impressive examples, real drug candidates have been developed from starting molecules that formerly would have been categorized with pocket lint and carpet sweepings for their potencies.

Why do compounds put on weight as a project goes on? Well, other things being equal, size and general greasiness will buy you potency. A significant part of the energy that goes into binding comes from shoving water molecules around. Your compound has to like your protein target more than it likes the water that surrounds itself, and the cheap way to make that happen is to make it hate the water rather than love your protein. So if potency is all you’re after, and if you don’t care what devils you have to deal with to get it, you can make impressively tight-binding molecules. Just find a tolerant region of your starting structure and load it up with naphthalenes, polyfluorinated rings, adamantanes, or similar lumps of lard. The resulting binding affinities will impress the credulous. The resulting compounds will horrify the wise, though; they will rightly fear what will happen when the tests of metabolism, toxicology, and formulations have to be met.

Hit parade

One mechanism for compound bloat is the natural tendency to divide a lead molecule up into regions, with a different person or group assigned to each. Everyone will optimise their section for potency, and if they just add a bit of size and molecular weight while doing so, the resulting ’greatest hits’ structure will no longer look so reasonable when it’s assembled. A related problem is that many chemists have particular functional groups that have worked well for them in the past, which they’ll find a way to add to whatever molecule comes their way. (Counteracting this is the not-uncommon failure of the greatest hits approach, since many structure-activity relationships turn out not to be so simple and additive). 

Another problem is political, or at least psychological. Once a potent compound has been made in a project, it requires a great effort of will to walk away from it. Pressure comes from the other groups on a project, or from upper management: ’Why aren’t you people working on that nanomolar lead?’ Even after a compound (or a whole compound class) has demonstrated its failings in solubility, selectivity, or safety, it remains what a lawyer would call an ’attractive nuisance’. Those binding affinities! How can you pass them up? The best advice for long-term drug research happiness is: if you don’t want to hear about why you’re not following up on those horrible greasy compounds, don’t make any of them in the first place.

Derek Lowe is a medicinal chemist working on preclinical drug discovery in the US.