Biology Reference
In-Depth Information
in clinical trials. So rather than use the technically correct
term 'failed drug' to describe nocodazole and similar
molecules, a better definition was needed. The term
'chemical probe' was introduced for this purpose and has
been widely accepted.
A chemical probe is a small molecule that targets
a specific protein or other cellular structure and, by binding
to its ligand, inhibits its activity. Nocodazole, and other
chemical probes that can selectively inhibit specific cellular
proteins, are powerful tools to elucidate gene function.
Furthermore, when used in combination with other agents
or together with genetic perturbations, they allow dissec-
tion of biological pathways at high resolution. Ultimately,
once developed and validated, some chemical probes may
graduate into lead candidates for new therapies, but this
progression involves issues of chemical tractability that are
often irrelevant for a probe to be a useful tool. Alterna-
tively, a chemical probe can serve to demonstrate that
a novel protein is druggable, thereby proving a potential for
therapeutic intervention. When applied to a cell, the effects
of chemical probes are complementary to genetic tools, yet
they have distinct advantages. Chemical probes are rapid,
reversible and tunable, and often directly affect the activity
of a target protein. Unlike drugs, chemical probes are not
constrained by pharmacokinetic and pharmacodynamic
considerations, and therefore the biological space they can
interrogate is potentially unlimited. Equally important, they
are not constrained by pharmacoeconomic considerations,
e.g., intellectual property is mostly irrelevant. An addi-
tional advantage of chemical probes is that, unlike genetic
approaches to perturb function, they can often be readily
transferred to other organisms (evolutionary conservation
permitting) that may lack genetic tools. The ability to
transfer chemical reagents across organisms increases their
utility, and by making them readily available to researchers
they promote interdisciplinary science. The concept of
a community process where promising chemical probes
are validated by community members as they use and
modify them
[36]
has received much attention. The initial
Molecular Libraries Initiative (MLI) component of the
NIH roadmap
[37]
was motivated, in part, by 'the need to
expand the availability, flexibility, and use of small-mole-
cule chemical probes for basic research'
[37]
. Our own
laboratory has adopted a strategy to provide resources for
this community-driven effort to reduce costs and increase
the efficiency of chemical probe identification and devel-
opment
[38]
.
The most important characteristic of a chemical probe
(desirable but not necessary for a drug) is that the mech-
anism of action be well understood. Because the purpose
of chemical probes is to act as molecular tools to elucidate
gene function, if, for example, polypharmacological
effects are not realized, the result may be mis-annotation
of gene function. As already discussed, understanding in
vivo MOA is challenging, particularly for compounds
that act by polypharmacological mechanisms, where
inhibiting one target may obscure the effects that result
from the inhibition of is a secondary targets. The tools
available to disentangle these different possibilities have
been most extensively developed in model organisms, and
several instructive examples are described in the following
sections.
CHEMOGENOMICS
A chemical
genetic assaymeasures the response of a single
gene to a small molecule, whereas a chemogenomic profile
is a combined set of measurements of the response of
each individual gene to a small molecule. Chemogenomic
profiles can be generated by nearly all genomic technologies
with the simple addition of a small molecule to the assay
condition. Responses measured are platform dependent and
can include, for example, RNA expression levels, protein
abundance and/or modifications, metabolite abundance, and
fitness in YKO and SGA screens.
e
Target Identification/Mechanism of Drug
Action
Many large-scale chemogenomic datasets have been used
to identify a small molecule's target. Knowledge of the
target of a small molecule is increasingly important in drug
development, providing the foundation for drug optimiza-
tion. Often the need for target identification arises from
a screen that identifies small molecules that induce
a phenotype of interest, yet does not reveal the mecha-
nism(s) that drive the phenotype. A common genomics
approach to target prediction uses the chemogenomic
profile of the small molecule of interest to query a large-
scale reference set of profiles derived from characterized
mutants (e.g., deletion strains) and/or small molecule with
known and unknown targets/MOAs
[26,39,40]
.The
assumption is that the small molecule of interest has
a MOA similar to the MOA of the best match in the
reference set, where matching is based on profile similarity.
All such guilt-by-association approaches are inherently
limited by the breadth and depth of the reference set.
Alternative approaches to target identification do not rely
on guilt-by-association. Target-oriented (or forward) drug
discovery dictates that, for lead small molecules that are
highly potent, in vitro target will be the in vivo target. This
approach (with important exceptions) has not generally been
successful. A reverse drug discovery approach requires
the measurement of the response of each gene to drug.
HaploInsufficiency Profiling (HIP) in yeast provides such a
reverse drug discovery approach (
Box 8.1
). Analogous
to fitness profiling, HIP results in a list of the most likely
drug targets, and often directly identifies
the target.
