activity and abundance does not always correlate
with mRNA levels (Chenet al., 2002; Gygiet al.,
1999).
The ‘one-gene-one-protein’ hypothesis is now
well and truly dead. Proteins hugely outnumber
genes in all mammals. The term Proteomics has
been coined to describe the analogous study of
proteins within particular cells or tissues (Figeys,
2003; Petricoinet al., 2002; Tyers and Mann, 2003;
Zhuet al., 2003). Moreover, many proteins are
modified after translation in ways that are crucial
in regulating their function. Thus, the application
of proteomics also extends far beyond the target
identification stage in drug development.
Further exploitation of this genomic and protei-
nomic can be obtained by making comparisons of
these data with epidemiological observations in
human populations. Patterns of familial disease,
with mapping to differences between individuals
in terms of DNA or mRNA, can identify which of
many genetic variations is the etiology. This is
known as ‘Linkage Analysis’, and, ultimately, the
precise chromosomal location, relative to the loca-
tion of other known genes, can be found using a
technique known as ‘positional cloning’. An exam-
ple of new target identification using these methods
was the identification of ApoE as a causative factor
in Alzheimer’s disease (Pericak-Vance, 1991).
Mutations which cause disease can arise sponta-
neously. Genetic mapping methods utilizing posi-
tional cloning can help identify disease-causative
genes and their proteins in animals which have
spontaneously developed diseases similar to those
of humans. An example of this type of technology
is the ob/ob genetic mouse, which is obese and has
mutationsinageneforapeptidehormoneknownas
leptin. A similar mouse, the Agouti strain, is also
obese and has defects in melanocortin receptors,
which develops type II diabetes, and therefore can
be used as an animal model of that disease in
humans. Of course, human disease is rarely as
simple as a single genetic defect, so these models
must be used with some caution when testing drugs
or when identifying the causative genes. Pathophy-
siological studies of organisms that have been
engineered to contain (transgenic ‘splice in’), or
to be free from (‘knock out’) the identified gene is
an extension of this concept (see also below).
The sequencing of genes does not directly iden-
tify new molecular targets for disease. But what it
does do is to permit the rapid identification of target
proteins, because their codes are known. Usually,
only a few trial peptides need then be synthesized,
shaving months off of the discovery process. In
turn, this allows rapid identification and cloning of
new targets for assay development.4.3 Whole tissue studies
Pharmacologists are often able to develop tissue
and whole animal models of human disease. In
some instances, studies on isolated tissues, such
as blood vessels, heart muscle or brain slices, will
allow a tissue- or organ-specific understanding of
the effects of potential new drugs. Cardiovascular
pharmacologistsoftenstudyisolatedarterieswhich
are maintained in a physiological salt solution.
Electric stimulation can induce contraction of the
vascular smooth muscle, and the effects of hyper-
tensive drugs on vascular contraction can then be
measured. Historically, these systems were often
used as primary drug screening tools. Because
these methods are much less direct than molecular
screening, they are now relegated to secondary or
tertiary roles as validation of the targets or drugs
discovered, using assays that directly employ the
molecular or cellular targets. Wholeanimal models
areoftenseenascriticaldecision-makingpointsfor
a newly discovered drug.
Human pathology is inevitably more complex
than those of rats and mice. Thus, it is often neces-
sary to induce a pathological state by introduction
of a pathogen or stimulant directly into a healthy
animal. The development of new animal models is
a time-consuming process and must be overseen by
the appropriate ethics committees and expert veter-
inarian advice.
Why arein vivo(whole animal) studies still
important to drug discovery? All the new technol-
ogy, as well as mathematical modeling using com-
puters, has reduced but not eliminated the need for
animal experimentation. Computer models still
cannot accurately predict the effects of chemical
compounds on the cell, let alone in systems with
higher orders of complexity, that is whole tissues,48 CH4 DRUG DISCOVERY: DESIGN AND SERENDIPITY