Library Synthesis or Rational Drug Design?

A question that comes up once in a while ponders whether library synthesis (sometimes called parallel synthesis) has any advantage over traditional medicinal chemistry. And if so, what is it?

Not too long ago, there wouldn’t be a question, just a declaration of the awesome power of rapidly producing thousands of compounds, picking out the hits in assays and going on from there. However, that potential never really turned into a real paradigm shift in how we do drug discovery and thus it took its place in the medicinal chemist’s tool box.

By library synthesis I mean where one or more scaffolds are built on using straightforward and standardized chemistry to decorate the molecule, using automation and plates rather than hands and flasks. The initial chemistry is worked out on a small set of compounds then this is applied to a larger set, in which you have worked out methods for reacting, quenching and purifying many reactions in a parallel manner. Plus, you need some method (usually an LC/MS) to show you have made what you say you made. The problems are mostly ones of logistics as you try to do many things in parallel, plus the general issue that standard conditions just don’t work for all the reagents you selected, leaving holes in your final matrix. Purity can also be an issue where either your default or automated purification method is unable to purify sufficiently, making any result from a biological assay of questionable value (it might be good because of the impurity or you might miss a good compound due to the impurity). Compound identity can be incorrect as well, depending on the method of analysis – LC/MS may see the right mass but you have an isomer, for example.

The power of the approach comes when you have made and tested 100 or 1000 compounds and have enzyme inhibition data for all those compounds, giving a broad view of the side-chains which are tolerated, which are bad. It can be a fast way to run the quick pass of methyl, ethyl, butyl, futyl to explore an area of activity. It can also give serendipity a chance, as you stumble upon something unexpected. However, it is not so likely to give you that final compound – you are still likely to do have to do that final optimization. Plus, you can only be rational to a limited extent – the changes you make to the final product will be somewhat dictated by the chemistry you are using, the generic methodology, plus reagent availability.

With rational drug design, you take a more cautious one step at a time approach. In effect, you decide what you are going to make, make it, then see what that tells you in order to decide on the next step. In series with laborious syntheses, it is literally that way, with the length of time to make each compound high, you cannot just blast wildly and make anything. It is too wasteful. You have to take a studied approach and draw as much information as you can before you commit to the next iteration. But realistically, most med chem projects will involve making and submitting a small set, getting them tested while the next small set is in progress, then deciding on future targets when the new data is in hand.

Ultimately, this approach is the way that a final optimization will be done. For making more specialized side chain or reagents, you are more likely to be working on each molecule individually. Also, as you are giving each compound attention, you can adapt your chemistry to suit the characteristics of each molecule. So you can make that one particular compound as the parallel synthesis is completing the synthesis of many compounds but somehow did not make the one you most wanted.

This is not really a boxing match or a debate in which one approach is better, both methods have their place. Taking hesitant steps into new chemical space is painfully slow – it is often much preferable to give yourself a rough idea of the lay of the land before trying to find the best place to dig. Other data may help direct the selection of targets as well, in particular x-ray or computational models. As the structure-activity relationships emerge, there is less value in making many compounds and much greater return in making the one compound you want.


One comment on “Library Synthesis or Rational Drug Design?

  1. CMCguy says:

    Good post. The problem with both Rational Drug Design and Combinatorial Libraries was they were each promoted in some quarters as “quick and easy” means to totally replace traditional med chem. The approaches typically have different starting points. Rationale based will have some data (x-ray or other structural info) that can “test” theoretical molecules against before choosing what compound(s) to make. Libraries are physical compounds that can be used to generate initial hits in a bioassay or if have a hit can build a series to provide SAR but here the product is really the data that again should aid selection of (a few) compounds to make. Both are tools that can be helpful to traditional method but designing a drug typically is very iterative, to which experience of knowing what to try or avoid greatly enhances outputs. Even when design and select the “best” candidate the complexities of human systems create an adventure for success.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s