Biomedical Engineering Reference
In-Depth Information
library being tested against a single target also increases the likelihood of discovering
a molecule capable of binding to that target in a novel manner [14].
The molecules that comprise these libraries, or indeed any compound collection,
may be obtained from either natural (natural products) or nonnatural (chemical syn-
thesis) sources. Nature has produced huge numbers of biologically relevant secondary
metabolites that have evolved to have specific and exquisite biological activity. These
compounds have been used medicinally for millennia, still provide many lead com-
pounds and drugs today [15,16], and almost certainly will continue to do so long
into the future [17]. Also, there can be no argument that natural products do not
represent a truly structurally diverse collection of compounds. Taking these factors
into account, the screening of natural products for biological activity is clearly an
extremely valuable thing to do. However, it is not realistic to suggest the production
of large libraries based solely on natural products, due predominantly to difficulties in
sourcing, isolating, and identifying the bioactive components, as well as in purifying
and chemically modifying these often extremely complex structures [18]. Therefore,
in terms of producing large numbers of compounds for screening, deliberate chemical
synthesis is generally considered to be the most efficient approach [13,19].
With the advent of combinatorial chemistry in the 1990s, it became possible for
chemists to produce very large numbers of compounds in an efficient manner. The
use of split-and-pool techniques and advances in automation made it possible to
synthesize literally millions of compounds in a short period of time [20]. These
libraries were, however, generally made up of broadly similar structures, resulting
in a fairly limited biological profile across the library. The molecules were usually
synthesized by combining a number of building blocks in different ways using the
same synthetic methods to generate distinct structures, so achieving a multiplicative
increase in the final number of compounds synthesized with an additive increase in
the number of building blocks used. This approach usually resulted in the variation
of substituents (R-groups) around a common scaffold. Libraries of this sort have
had limited success in the discovery of novel biologically active agents, a fact that
is attributed primarily to the relative lack of structural diversity within the libraries
[21]. It is therefore believed that the quality of the compounds that make up these
screening collections, in terms of structural complexity and diversity, is as, if not
more important than the total number of compounds present [2].
There are many commercially available or proprietary compound collections that
can be seen to represent a source of small molecules without the need for de novo
synthesis [12]. The compounds comprising these libraries are generally synthesized
in a combinatorial fashion and so suffer from the limitations described above in
terms of structural and hence functional diversity. These collections also suffer from
limitations that can be attributed to the desire to produce very large numbers of
compounds quickly: They are generally “flat” (based around aromatic core structuers
containing few stereocentres), structurally simple, and similar [22].
There is also a belief that these compound collections are too heavily biased toward
traits that are traditionally perceived to be desirable in drug-like molecules, such as
the Lipinski rules for oral bioavailability [12,14,23]. These collections are therefore
heavily weighted toward known bioactive chemical space (the region of chemi-
cal space spanned by known drug molecules and bioactive natural products). By
