Chemistry Reference
In-Depth Information
2. Hit rates from HTS are often low and the hits obtained fail to progress into lead
optimisation: HTS is the predominant technique for hit finding employed by the
majority of pharmaceutical companies and is central to modern drug discovery. How-
ever, many scientists are now regarding HTS as a costly necessity rather than a
method of choice.
[
9
]
3. HTS only samples a minute fraction of drug-like chemical space: A widely quoted
estimate
[
10
]
of the number of molecules containing up to 30 C, N, O or S atoms exceeds
10
60
(and the mass of a 1 pmol sample of each would of the same order as that of
the observable universe). An HTS screen of, say, 10
6
molecules would only sample a
minute portion of this space. By contrast, the number of synthetically feasible small
molecule fragments with masses up to 160 Da has been estimated
[
11
]
at 10
7
, hence a
typical fragment screen of 10
3
-10
4
molecules is sampling a much higher proportion of
this space.
4. Many companies have realised, to their significant cost, that screening non-proprietary
vendor compounds against non-proprietary targets can lead to difficulties in securing a
good intellectual property (IP) position: Increasingly, there is the possibility that another
company has screened similar compounds against a similar target and therefore it is
difficult to obtain strong patent protection against the chemical series. If competitor
patents do not appear for some time after the initiation of a drug discovery programme,
significant resources can be wasted on attempting to optimise compounds that another
organisation had already discovered and patented
[
12
]
.
Hence there is increased pressure to discover more suitable chemical starting points
and the term 'lead-like' has been used to describe these less complex molecules.
[
13
]
By its
very nature, fragment screening is ideally suited to finding low molecular weight bioact-
ive compounds. Because these compounds tend to be relatively low in potency (typically
in the 100-1000 M range) they are not identifiable by an HTS run at a typical com-
pound concentration of 10-30 M. Fragment screening collections, even those formed
from non-proprietary vendor compound collections, lie beyond the scope of HTS com-
pound collections, thus increasing the chances of identifying novel molecules that can be
optimised into patentable series.
Subsequent chapters will discuss in more detail and provide case studies of best practices
of using fragments in drug discovery programmes. It is useful, though, to start with an
understanding of what is meant by a fragment and how the use of fragments has reached
the current level of popularity.
1.2 What is a Fragment?
A quick search of the literature using the term 'drug-likeness' throws up thousands of
articles, reflecting the aspiration of the industry to be able to classify compounds readily
as drug-like or not. A wide variety of methodologies has been employed in the process of
classification, from simple filters based on physicochemical descriptors through to more
complex QSAR models. Although the concept of fragment-based drug discovery has been
around since the early 1980s, the application is relatively new and as such the volume
of literature around the subject can be measured in hundreds, not thousands. There are
