Designing a Fragment Process to Fit Your Needs - Fragment-Based Drug Discovery - page 21

Fragment library design. The challenge for all approaches to drug discovery is twofold:
one wants sufficient diversity to make sure one achieves adequate sampling of available
space and one wants the library to be small enough to screen quickly. Fragments have a
proven advantage over lead-like libraries in both of these areas. This is expertly presented
in Chapters 3, 8 and 9. Here we will briefly give an overview of the major components and
factors that need to be assessed in assembling this part of FBDD. The first component of the
library to address is the level of diversity that it will encompass. There is always going to be a
tradeoff between diversity and inherent SAR: one cannot reasonably have every compound
possible in a library and one wants to have as many as possible! It is also recommended to
have compounds which are closely related to each other so that SAR can be established.
The tension between these competing needs is what drives library composition. [ 10 ] Adiverse
fragment library is sufficiently diverse if it provides good chemical starting points for hit
optimization. Part of providing 'good' starting points is that the results lead to testable
hypotheses from the outset; this means that there must be SAR that results from the initial
screen. We believe that a sparse matrix approach with very limited SAR around each
populated node is the optimal approach. The actual composition of the library is dictated by
the results of target, assay and compound assessment. Another component of assembling
the library is that the fragment should be suitable for medicinal chemistry elaboration.
Many different approaches have been used in constructing fragment libraries to have the
follow-up chemistry built into the fragments, including capped chemistry handles, built-in
SAR and ready-to-use chemistry handles. [ 19, 58 ] If fragments do not have handles available
for immediate elaboration, it should be possible to elaborate off the scaffold should it hit.
This elaboration can use additional fragment libraries or newly synthesized (or purchased)
compounds that address the budding SARmade evident by the screen. It is also advisable to
have 'nested' libraries ready to follow upon the initial hit. This pyramidal approach has been
shown to be quite expedient. [ 59 ] Another approach similar to 'nested' libraries is 'cascaded'
libraries. In this approach, the initial library is set up to sample maximal fragment space
with the fewest possible compounds, which is sometimes called a 'sparse matrix' approach.
Follow-up libraries are prepared to sample a much tighter area of fragment space and are
screened if a hit is within a certain distance to a trigger point. The initial hit(s) may trigger
multiple follow-up libraries to be screened, but the results from the follow-up screening
will triangulate on a given region of fragment space and lead to obvious further testing.
At this point, the medicinal chemistry resources can be engaged having identified a very
robust SAR hypothesis.
2.3 Phase II Activities
2.3.1 The Screen
The variety of biophysical and biochemical techniques now available that can be used as
screens for fragment based studies is impressive. [ 9 ]
Biochemical techniques. Conceptually the most straightforward approach to identifying
fragments is through a functional (biochemical) screen; if the target activity is affected by
the compound, it is a 'hit'. However, since fragments usually have relatively low binding