Structural Bioinformatics
in Drug Discovery
Melissa Passino
Structural Bioinformatics
• What is SBI?
“Structural bioinformatics is a subset of
bioinformatics concerned with the use
of biological structures – proteins, DNA,
RNA, ligands etc. and complexes thereof
to further our understanding of
biological systems.”
http://biology.sdsc.edu/strucb.html
SBI in Drug Design and Discovery
• SBI can be used to examine:
• drug targets (usually proteins)
• binding of ligands
↓
“rational” drug design
(benefits = saved time and $$$)
Traditional Methods of Drug Discovery
natural
(plant-derived)
treatment for
illness/ailments
↓
isolation of active
compound
(small, organic)
synthesis
of compound
↓
manipulation of
structure to get
better drug
Aspirin
(greater efficacy,
fewer side effects)
Modern Methods of Drug Discovery
What’s different?
• Drug discovery process begins
with a disease (rather than a treatment)
• Use disease model to pinpoint
relevant genetic/biological
components (i.e. possible drug targets)
Modern Drug Discovery
disease
→ genetic/biological target
↓
discovery of a “lead” molecule
- design assay to measure function of
target
- use assay to look for modulators of
target’s function
↓
high throughput screen (HTS)
- to identify “hits” (compounds with
binding in low nM to low μM range)
Modern Drug Discovery
small molecule hits
↓
manipulate structure to increase potency
i.e. decrease Ki to low nM affinity
↓
*optimization of lead molecule into candidate drug*
fulfillment of required pharmacological properties:
potency, absorption, bioavailability, metabolism, safety
↓
clinical trials
Interesting facts...
• Over 90% of drugs
entering clinical
trials fail to make it
to market
• The average cost
to bring a new
drug to market is
estimated at $770
million
Impact of Structural Bioinformatics
on Drug Discovery
Genome
Gene
Protein
HTS
Hit
Lead
Candidate Drug
Genomics
Fig 1 & 2
Bioinformatics
Structural Bioinformatics
Fauman et al.
Chemoinformatics
Structure-based Drug Design
ADMET Modelling
• Speeds up key steps in
DD process by combining
aspects of bioinformatics,
structural biology, and
structure-based drug
design
Identifying Targets:
The “Druggable Genome”
human genome
polysaccharides
lipids
nucleic acids
proteins
Problems with toxicity, specificity, and
difficulty in creating potent inhibitors
eliminate the first 3 categories...
human genome
polysaccharides
lipids
nucleic acids
proteins
proteins with
binding site
“druggable genome” = subset of genes which
express proteins capable of binding small drug-like
molecules
Relating druggable targets
to disease...
Analysis of pharm
industry reveals:
GPCR
• Over 400 proteins
used as drug targets
Other 110
families
STY kinases
Cys proteases
Gated ionchannel
Zinc peptidases
Ion channels
Nuclear PDE
receptor
Serine
proteases
P450 enzymes
Fig. 3, Fauman et al.
• Sequence analysis of
these proteins shows
that most targets fall
within a few major
gene families
(GPCRs, kinases,
proteases and
peptidases)
Assessing Target Druggability
• Once a target is defined for your
disease of interest, SBI can help
answer the question:
Is this a “druggable” target?
• Does it have sequence/domains similar to
known targets?
• Does the target have a site where a drug
can bind, and with appropriate affinity?
Other roles for SBI in drug discovery
• Binding pocket modeling
• Lead identification
• Similarity with known
proteins or ligands
• Chemical library design /
combinatorial chemistry
• Virtual screening
• *Lead optimization*
• Binding
• ADMET
SBI in cancer therapy:
MMPIs
• Inability to control metastasis is the
leading cause of death in patients
with cancer (Zucker et al. Oncogene. 2000, 19,
6642-6650.)
• Matrix metalloproteinase inhibitors
(MMPIs) are a newer class of cancer
therapeutics
• can prevent metastasis (but not cytotoxic);
may also play role in blocking tumor
angiogenesis (growth inhibition)
• Used to treat “major” cancers: lung,
GI, prostate
What is an MMP?
• Family of over 20 structurally related
proteinases
• Principal substrates:
• protein components of extracellular matrix
(collagen, fibronectin, laminin, proteoglycan
core protein)
• Functions:
• Breakdown of connective tissue; tissue
remodeling
• Role in cancer:
• Increased levels/activity of MMPs in area
surrounding tumor
Brown PD. Breast Cancer Res Treat 1998, 52, 125-136.
Whittaker et al. Chem. Rev. 1999, 99, 2735-2776
MMP-1,3,8
MMP-2
MMP-7
MMP-9
MMP-10 to 13,19,20
MMP-14
to 17
Whittaker et al. Chem. Rev. 1999, 99, 2735-2776
MMP catalysis
“metallo” in MMP = zinc
→ catalytic domain contains 2 zinc atoms
Whittaker et al. Chem. Rev. 1999, 99, 2735-2776
Peptidic inhibitors
• Structure based
design
– based on natural
substrate collagen
– zinc binding group
• Poor Ki values, not
very selective
(inhibit other MPs)
Brown PD. Breast Cancer Res Treat 1998, 52, 125-136.
Peptidic hydroxamate inhibitors
• Specificity for
MMPs over
other MPs
• Better binding
(low nM Ki)
• But poor oral
bioavailability
A (not very) long time ago,
in a town (not too) far away…
…lived a company
named Agouron…
…and this company
had a dream, a
dream to design a
nonpeptidic
hydroxamate
inhibitor of MMPs…
...so they made some special crystals…
used x-ray
crystallography/3D
structure of
recombinant human
MMPs bound to
various inhibitors
↓
to determine key a.a.
residues, ligand
substituents needed
for binding
Gelatinase A
http://www.rcsb.org/pdb/
…and used the magic of structural
bioinformatics to design many, many
nonpeptidic hydroxylates.
Ki
oral
bioavailabity
antigrowth
repeat…
antimetastasis
Results…
AG3340
“Prinomastat”
• Good oral
bioavailability
• Selective for
specific MMPs
– may implicate their
roles in certain
cancers
Prinomastat
• Evidence showing prevention of lung
cancer metastasis in rat and mice models
• Clinical trials
→ non small cell lung cancer
→ hormone refractory prostate cancer
…stopped at Phase 3 (Aug 2000) because
did not show effects against late stage
metastasis
Morals of the story…
• SBI can be used as basis for lead
discovery and optimization
• MMPs are good targets for chemotherapy
to help control metastasis…
…but MMPIs must be combined with other
cytotoxic drugs to get maximum benefits,
and used at earliest stage possible