|
Leveraging
its expertise in biophysical chemistry for small-molecule
drug discovery, Surface Logix recently announced positive
data from a repeat-dose Phase I clinical trial to assess
the safety and tolerability of SLx-4090, an enterocyte-specific
microsomal triglyceride transfer protein (MTP) inhibitor
designed to treat dyslipidemia. Preclinical studies
demonstrated the ability of SLx-4090 to lower both
triglycerides and LDL cholesterol as well as reduce
weight. The company plans to initiate a Phase IIa study in
patients with dyslipidemia complicated by high
triglyceride levels later this year.
Surface
Logix applied its Pharmacomer Technology Platform to
design an MTP inhibitor that could be given orally and
would act selectively in the gastrointestinal tract to
block fat transport through the intestinal wall, without
causing fat accumulation in other organs such as the
liver, heart, or eye. The company is applying this core
platform to develop small-molecule therapeutics targeted
to specific physiological microenvironments in the body to
improve their pharmacokinetic and pharmacodynamic
properties, including solubility, permeability, and
metabolism. The company’s initial focus is on targets in
cardiovascular disease, oncology, and metabolic disorders.
Jim
Mahoney, president and CEO of Surface Logix, commenting on
the problems associated with traditional approaches to
structure-based design, states, “The most commonly used
rules are limited and not always accurate.” As an
example of the complexity of chemical design, Mahoney
explains that adding a piperazine ring to a pharmacophore,
the active regions of a drug, will yield a certain
activity, but if the ring is put on a different
pharmacophore, the resulting compound may behave
differently. “You have to understand the physicochemical
properties of the whole combination as well as of the
parts to get the answer,” he adds.
Overall,
current approaches to structure-based design are better
than traditional methods because there is a lot more
information available—better physicochemical and pathway
information and a broader range of assay and biological
data—to guide small-molecule design strategies. More and
better information is available on the structure of
targets such as enzymes and receptors, allowing chemists
to tweak the structure of the scaffold and rapidly assess
the impact on a specific target.
For
kinase inhibitors in particular, the ability to screen a
compound against a panel of kinases “is a valuable tool
for medicinal chemists and molecular biologists,” says
Tarak Mody, Ph.D., senior director of business development
and licensing at Pharmacyclics. “You can narrow down
what enzymes it is hitting,” identify which compounds
are inhibitors, and get a sense of what concentrations are
needed to inhibit a specific enzyme. “Then you can work
with the scaffold to change its functionality and see how
that impacts its effect on an enzyme and its druglike
properties.”
Kinase
targets present particular challenges because of the
structural similarities within kinase families and the
need in some cases to design inhibitors with exquisite
selectivity for one or a few key kinases involved in a
specific disease pathway. Pharmacyclics recently published
data demonstrating the selectivity of its orally active
small molecule designed to inhibit Bruton’s tyrosine
kinase (Btk), a signaling molecule expressed by immune
cells such as B cells, macrophages, and mast cells, which
is required for the B-cell activation implicated in
autoimmune diseases such as rheumatoid arthritis (ChemMedChem
2007;2(1):58–61). The drug candidates selected bind Btk but not binding
sites on other tyrosine kinases. Pharmacyclics is applying
structure-based design principles to target cancer and
autoimmune diseases.
Hitting the Target
As an
example of the trend toward more customized, target-biased
library synthesis based on “intricate knowledge of the
target or ligand(s),” Nikolay Savchuk, Ph.D., president
and CEO of ChemDiv, points to the development of kinase-targeted
compound sets. He describes a “growing demand for both
selective and ‘dual’ inhibitors of protein kinases
that are capable of shutting down two disease-relevant
targets representing different signaling cascades, such as
KDR/Raf1 or EGFR/Akt, with comparable EC50s.”
The company’s kinase-biased library comprises 30,000
compounds and about 700 templates. Similar focused
libraries target GPCRs, ligand- and voltage-gated ion
channels, nuclear hormone receptors, proteases, and
phosphatases.
Surface
Logix’s Pharmacomer platform is based on a method
developed by Professor George Whitesides at Harvard
University for producing self-assembling monolayers, which
can serve as inert environments for the attachment and
measurement of pharmacophores and Pharmacomers (monomers
designed by the company). These can be applied to existing
pharmacophores or to novel chemical scaffolds. By exposing
these small molecules to different types of biological
molecules, researchers can measure the various
interactions and begin to tease out the differences in the
biophysical and chemical attributes of small molecules
that affect tissue targeting and potential side reactions,
toxicities, or other liabilities.
“We
view the drug as a surface,” says Paul Sweetnam, Ph.D.,
CSO at Surface Logix. Using standard monomers found in
existing drugs, “we put those on the surfaces [of the
self-assembling monolayers] and found that they did not
have a lot of functionality in the physical-chemical
world. We then designed new monomers [Pharmacomers] that
exhibited unique functionality to address many
pharmacokinetic issues.”
The
company’s main focus is on “getting the drug to the
target,” says Mahoney. The Pharmacomers represent side
chains “designed to avoid the interactions a drug sees
on its way to the target,” he explains.
Building in Lead/Drug-Likeness
A study
evaluating the leadlikeness and diversity of commercially
available screening libraries identified ChemBridge
Research Laboratories’ (CRL) EXPRESS-Pick collection as
having the largest number of leadlike and unique
structures (Molecular Diversity 2006;10(3):377–388).
The company’s in-house NOVACore library boasts more than
65,800 unique structures and a high rate (about 90%) of
leadlike properties, according to the company.
Rongshi
Li, Ph.D., senior director of high-throughput medicinal
chemistry at CRL, which is a discovery chemistry CRO,
echoes the trend toward smaller, more focused libraries
that can be used to answer structure-activity relationship
(SAR) questions and that require shorter turn-around
times. CRL employs structure-based design based on either
x-ray crystal structures or homology modeling, combined
with pharmacophore (ligand)-based design and virtual
screening, together with leadlike/druglike property
filters developed based on physicochemical calculations.
Advances
in high-throughput purification technology together with
high-throughput organic synthesis have accelerated
parallel compound synthesis capabilities, but not
necessarily the creation of leadlike compounds. The most
critical goal in the early stages of library design is to
design novel compounds with low molecular weights and to
explore new areas of chemical space whenever possible, in
Li’s view. Then, “by applying physicochemical property
filters, we can cut a tremendous amount of design time,”
he says.
ChemDiv’s
modeling capabilities focus on the identification of
“druggable” areas on enzymes and proteins and, in
collaboration with MolSoft, is describing the scope of
ligand-binding pockets, or the “pocketome” (Genome
Informatics 2004;15(2):31–41).
This technology aids, for example, in the selection of
non-ATP competitive inhibitors of kinases and of unique
binders of nuclear hormone receptors.
Recognizing
a need in drug discovery for molecular probes that can be
used to identify novel biological targets, ChemDiv
developed Focused Diversity, an algorithm for assembling
focused libraries based on biologically driven diversity.
Their current library contains 2,500 molecules designed to
reflect “broad biological activity,” explains
Alexander Kiselyov, Ph.D., executive vice president of
R&D at ChemDiv. “Each molecule contains two to three
points of randomization for subsequent chemical
modification.” Active compounds can be identified in
high-content cell- or tissue-based screens.
ChemDiv’s
non-biased peptidomimetics library contains mainly beta-
and gamma-turn mimetics based on biaryl templates modified
with flexible and rigid substituents. “The geometry of
the designed biaryl fragments was compared using modern
computational methods with the dihedral angles reported
for several ‘natural’ beta- and gamma-turn motifs to
select the best match,” says Savchuk. Other components
of the library include spiro-bicyclic scaffolds, di- and
tri-peptide mimetics, SH2 domain mimetics based on the
company’s heterocyclic isosteres of phosphotyrosine, and
beta-sheet mimetics.
Copyright
2007, Cambridge Healthtech Institute. All Rights Reserved.
|
