Transcript Protein Data Bank Advisory Committee

Polypharmacology: Drug
Discovery in the Era of Genomics
and Proteomics
Philip E. Bourne
University of California San Diego
[email protected]
Big Questions in the Lab
1.
2.
3.
4.
5.
August 14, 2009
Valas, Yang & Bourne 2009
Current Opinions in Structural Biology 19:1-6
Can we improve how
science is disseminated
and comprehended?
What is the ancestry of the
protein structure universe
and what can we learn
from it?
Are there alternative ways
to represent proteins from
which we can learn
something new?
What really happens when
we take a drug?
Can we contribute to the
treatment of neglected
{tropical} diseases?
Motivation
• The truth is we know very little about how the
major drugs we take work
• We know even less about what side effects
they might have
• Drug discovery seems not to have moved into
the omics era
• The cost of bringing a drug to market is huge
>$800M
• The cost of failure is even higher e.g., Vioxx
~ $5Bn
• Fatal diseases are neglected because they do
not make money
Motivation - Reasoning
• The truth is we know very little about how the
major drugs we take work – receptors/mechanism
is unknown
• We know even less about what side effects they
might have - receptors/mechanism is unknown
• Drug discovery seems not to have moved into the
omics era – systems biology can help but as yet
is unproven
• The cost of bringing a drug to market is huge
>$800M
• The cost of failure is even higher e.g., Vioxx ~
$5Bn - receptors/mechanism is unknown
• Fatal diseases are neglected because they do not
make money – there must be a workable
business model
Why Don’t we Do Better?
A Couple of Observations
• Gene knockouts only effect phenotype in 1020% of cases , why?
– redundant functions
– alternative network routes
– robustness of interaction networks
A.L. Hopkins Nat. Chem. Biol. 2008 4:682-690
• 35% of biologically active compounds bind to
more than one target
Paolini et al. Nat. Biotechnol. 2006 24:805–815
Why Don’t we Do Better?
A Couple of Observations
• Tykerb – Breast cancer
• Gleevac – Leukemia, GI
cancers
• Nexavar – Kidney and liver
cancer
• Staurosporine – natural product
– alkaloid – uses many e.g.,
antifungal antihypertensive
Collins and Workman 2006 Nature Chemical Biology 2 689-700
Implications
• Ehrlich’s philosophy of magic bullets
targeting individual chemoreceptors has
not been realized
• Stated another way – The notion of one
drug, one target, one disease is a little
naïve in a complex system
How Can we Begin to Address
the Problem?
• Systematic screening for multiple
targets by multiple drugs
• Integration of knowledge from multiple
sources
• Analyze the impact on the complete
living system
– Statically
– Dynamically
What if…
• We can characterize a proteinligand binding site from a 3D
structure (primary site) and search
for that site on a proteome wide
scale?
• We could perhaps find alternative
binding sites (off-targets) for
existing pharmaceuticals and
NCEs?
What Do These Off-targets Tell Us?
•
Potentially many things:
1. Nothing
2. How to optimize a NCE
3. A possible explanation for a side-effect of a
drug already on the market
4. A possible repositioning of a drug to treat a
completely different condition
5. The reason a drug failed
6. A multi-target strategy to attack a pathogen
Today I will give you brief vignettes of each of these
scenarios, but first the bioinformatics guts of the approach
Need to Start with a 3D Drug-Receptor
Complex - The PDB Contains Many
Examples
Generic Name
Other Name
Treatment
PDBid
Lipitor
Atorvastatin
High cholesterol
1HWK, 1HW8…
Testosterone
Testosterone
Osteoporosis
1AFS, 1I9J ..
Taxol
Paclitaxel
Cancer
1JFF, 2HXF, 2HXH
Viagra
Sildenafil citrate
ED, pulmonary
arterial
hypertension
1TBF, 1UDT,
1XOS..
Digoxin
Lanoxin
Congestive heart
failure
1IGJ
Computational Methodology
A Quick Aside – RCSB PDB
Pharmacology/Drug View Mid 2010
Asp
Drug Name
Aspirin
% Similarity to
Drug Molecule
Has Bound Drug
100
• Establish linkages to
drug resources (FDA,
PubChem, DrugBank,
etc.)
• Create query
capabilities for drug
information
• Provide superposed
views of ligand binding
sites
• Analyze and display
protein-ligand
interactions
Mockups of drug view features
RCSB PDB Ligand View
Peter Rose et al
A Reverse Engineering Approach to
Drug Discovery Across Gene Families
Characterize ligand binding
site of primary target
(Geometric Potential)
Identify off-targets by ligand
binding site similarity
(Sequence order independent
profile-profile alignment)
Extract known drugs
or inhibitors of the
primary and/or off-targets
Search for similar
small molecules
…
Dock molecules to both
primary and off-targets
Statistics analysis
of docking score
correlations
Computational Methodology
Xie and Bourne 2009
Bioinformatics 25(12) 305-312
Characterization of the Ligand Binding
Site - The Geometric Potential
 Conceptually similar to hydrophobicity
or electrostatic potential that is
dependant on both global and local
environments
• Initially assign Ca atom with
a value that is the distance
to the environmental
boundary
• Update the value with those
of surrounding Ca atoms
dependent on distances and
orientation – atoms within a
10A radius define i
GP  P 
Pi
cos(ai)  1.0

2.0
neighbors Di  1.0

Computational Methodology
Xie and Bourne 2007 BMC Bioinformatics, 8(Suppl 4):S9
Discrimination Power of the Geometric
Potential
4
binding site
non-binding site
3.5
• Geometric
potential can
distinguish
binding and
non-binding
sites
3
2.5
2
1.5
1
0.5
100
99
88
77
66
55
44
33
22
11
0
0
Geometric Potential
Computational Methodology
0
Geometric Potential Scale
Xie and Bourne 2007 BMC Bioinformatics, 8(Suppl 4):S9
Local Sequence-order Independent Alignment
with Maximum-Weight Sub-Graph Algorithm
Structure A
Structure B
LER
VKDL
LER
VKDL
• Build an associated graph from the graph representations of two
structures being compared. Each of the nodes is assigned with a
weight from the similarity matrix
• The maximum-weight clique corresponds to the optimum alignment
of the two structures
Xie and Bourne 2008 PNAS, 105(14) 5441
Similarity Matrix of Alignment
Chemical Similarity
• Amino acid grouping: (LVIMC), (AGSTP), (FYW), and
(EDNQKRH)
• Amino acid chemical similarity matrix
Evolutionary Correlation
• Amino acid substitution matrix such as BLOSUM45
• Similarity score between two sequence profiles
d   f a Sb   f b S a
i
i
i
i
i
i
fa, fb are the 20 amino acid target frequencies of profile a
and b, respectively
Sa, Sb are the PSSM of profile a and b, respectively
Computational Methodology
Xie and Bourne 2008 PNAS, 105(14) 5441
Nothing in Biology {Including
Drug Discovery} Makes Sense
Except in the Light of
Evolution
Theodosius Dobzhansky
(1900-1975)
What Do Off-targets Tell Us?
•
Potentially many things:
1. Nothing
2. How to optimize a NCE
3. A possible explanation for a side-effect of a
drug already on the market
4. A possible repositioning of a drug to treat a
completely different condition
5. The reason a drug failed
6. A multi-target strategy to attack a pathogen
Today I will give you brief vignettes of each of these
scenarios, but first the bioinformatics guts of the approach
How to Optimize a NCE
• African trypanosomiasis
(sleeping sickness)
• Carried by the tsetse fly
• Trypanosoma brucei is
the active agent
• Endemic to Africa
• 300,000 new cases
each year
• Sleep cycle disturbed
• Neurological phase
deadly
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
Optimize: Find Secondary
Targets of TbREL1
NCS45208
Aka Compound 1
TbREL1 – T. brucei RNA editing ligase I
IC50: 1.95 ± 0.33 μM
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
Workflow
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
Mitochondrial 2-enoyl Thioester
Reductase (HsETR1)
• Neither FATCAT nor CLUSTALW2
judged HsETR1 to be homologous to
the primary target.
• Both SOIPPA and AutoDock predicted it
was a secondary target.
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
Mitochondrial 2-enoyl Thioester
Reductase (HsETR1)
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
Mitochondrial 2-enoyl Thioester
Reductase (HsETR1)
• HsETR1 is thought to be essential for fatty
acid synthesis (FAS) type II.
• In the process of optimizing Compound 1
to make it more drug-like, modifications
that reduce binding to human HsETR1
may diminish unforeseen side effects.
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
T. brucei UDP-galactose 4epimerase (TbGalE)
• Neither FATCAT nor CLUSTALW2 judged
TbGalE to be homologous to the primary
target.
• AutoDock predicted it was a secondary
target, and it was homologous to a protein
that SOIPPA identified as a secondary target.
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
T. brucei UDP-galactose 4epimerase (TbGalE)
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
T. brucei UDP-galactose 4epimerase (TbGalE)
• Like TbREL1, TbGalE (galactose
metabolism) is essential for T. brucei
survival.
• Compound 1 inhibits two essential T.
brucei enzymes.
How to Optimize a NCE
Durrant et al 2009 PLoS Comp Biol in press
What Do Off-targets Tell Us?
•
Potentially many things:
1. Nothing
2. How to optimize a NCE
3. A possible explanation for a side-effect of a
drug already on the market
4. A possible repositioning of a drug to treat a
completely different condition
5. The reason a drug failed
6. A multi-target strategy to attack a pathogen
Today I will give you brief vignettes of each of these
scenarios, but first the bioinformatics guts of the approach
The Problem with Tuberculosis
•
•
•
•
One third of global population infected
1.7 million deaths per year
95% of deaths in developing countries
Anti-TB drugs hardly changed in 40
years
• MDR-TB and XDR-TB pose a threat to
human health worldwide
• Development of novel, effective, and
inexpensive drugs is an urgent priority
Found..
• Evolutionary linkage between:
– NAD-binding Rossmann fold
– S-adenosylmethionine (SAM)-binding domain of SAMdependent methyltransferases
• Catechol-O-methyl transferase (COMT) is SAMdependent methyltransferase
• Entacapone and tolcapone are used as COMT
inhibitors in Parkinson’s disease treatment
• Hypothesis:
– Further investigation of NAD-binding proteins may
uncover a potential new drug target for entacapone
and tolcapone
Repositioning - The TB Story
Kinnings et al. 2009 PLoS Comp Biol 5(7) e1000423
Functional Site Similarity between
COMT and InhA
• Entacapone and tolcapone docked onto 215 NADbinding proteins from different species
• M.tuberculosis Enoyl-acyl carrier protein reductase ENR
(InhA) discovered as potential new drug target
• InhA is the primary target of many existing anti-TB drugs
but all are very toxic
• InhA catalyses the final, rate-determining step in the fatty
acid elongation cycle
• Alignment of the COMT and InhA binding sites revealed
similarities ...
Repositioning - The TB Story
Kinnings et al. 2009 PLoS Comp Biol 5(7) e1000423
Binding Site Similarity between
COMT and InhA
COMT
SAM (cofactor)
BIE (inhibitor)
InhA
NAD (cofactor)
641 (inhibitor)
Repositioning - The TB Story
Kinnings et al. 2009 PLoS Comp Biol 5(7) e1000423
Summary of the TB Story
• Entacapone and tolcapone shown to have potential
for repositioning
• Direct mechanism of action avoids M. tuberculosis
resistance mechanisms
• Possess excellent safety profiles with few side effects
– already on the market
• In vivo support
• Assay of direct binding of entacapone and tolcapone
to InhA reveals a possible lead with no chemical
relationship to existing drugs
Repositioning - The TB Story
Kinnings et al. 2009 PLoS Comp Biol 5(7) e1000423
Summary from the TB Alliance
– Medicinal Chemistry
• The minimal inhibitory concentration
(MIC) of 260 uM is higher than usually
considered
• MIC is 65x the estimated plasma
concentration
• Have other InhA inhibitors in the
pipeline
Repositioning - The TB Story
Kinnings et al. 2009 PLoS Comp Biol 5(7) e1000423
What Do Off-targets Tell Us?
•
Potentially many things:
1. Nothing
2. How to optimize a NCE
3. A possible explanation for a side-effect of a
drug already on the market
4. A possible repositioning of a drug to treat a
completely different condition
5. The reason a drug failed
6. A multi-target strategy to attack a pathogen
Today I will give you brief vignettes of each {some } of these
scenarios, but first the bioinformatics guts of the approach
The TB Drugome
Existing Drugs
3. Protein-ligand
Docking
TB Structural
Proteome
…
TB Protein-drug
Interactome
2. Binding site
Similarity
Drugome/TB
1. Structural
Determination
& Modeling
TB Genome
4.2 Network
Integration
4.1 Network
Reconstruction
TB Metabolome
Target identification
Drug repurposing
Side effect prediction
New therapeutics
for MDR and XDR-TB
Drug resistance
mechanism
Bioinformatics 2009 25(12) 305-312
Multi-target strategy
Kinnings et al in Preparation
Structural coverage of the TB
proteome
3, 996
2, 266
284
1, 446
• High quality homology models from ModBase
(http://modbase.compbio.ucsf.edu) increase structural
coverage from 7.1% to 43.3%
Multi-target strategy
Kinnings et al in Preparation
Drug binding sites in the PDB
No. of drugs
• Searched the PDB for protein crystal structures
bound with FDA-approved drugs
• 268 drugs bound in a total of 931 binding sites
140
120
Acarbose
Darunavir
Alitretinoin
Conjugated
estrogens
Chenodiol
100
80
60
40
Methotrexate
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
No. of drug binding sites
Multi-target strategy
Kinnings et al in Preparation
SMAP p-value < 1e-5
drugs
TB
proteins
p < 1e-7
p < 1e-6
p < 1e-5
Multi-target drugs?
• Similarities between drug binding sites and
TB proteins are found for 61/268 drugs
• 41 of these drugs could potentially inhibit
more than one TB protein
20
18
16
chenodiol
No. of
drugs
14
12
testosterone
10
conjugated
estrogens &
methotrexate
raloxifene
ritonavir
8
levothyroxine
alitretinoin
6
4
2
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
No. of potential TB targets
Multi-target strategy
Kinnings et al in Preparation
Top 5 most highly connected
drugs
Drug
Intended targets
Indications
levothyroxine
transthyretin, thyroid
hormone receptor α & β-1,
thyroxine-binding globulin,
mu-crystallin homolog,
serum albumin
hypothyroidism, goiter,
chronic lymphocytic
thyroiditis, myxedema coma,
stupor
alitretinoin
conjugated
estrogens
retinoic acid receptor RXR-α,
β & γ, retinoic acid receptor cutaneous lesions in patients
α, β & γ-1&2, cellular retinoic with Kaposi's sarcoma
acid-binding protein 1&2
estrogen receptor
menopausal vasomotor
symptoms, osteoporosis,
hypoestrogenism, primary
ovarian failure
dihydrofolate reductase,
serum albumin
gestational choriocarcinoma,
chorioadenoma destruens,
hydatidiform mole, severe
psoriasis, rheumatoid arthritis
methotrexate
No. of
TB proteins
connections
14
adenylyl cyclase, argR, bioD,
CRP/FNR trans. reg., ethR,
glbN, glbO, kasB, lrpA, nusA,
prrA, secA1, thyX, trans. reg.
protein
13
adenylyl cyclase, aroG,
bioD, bpoC, CRP/FNR trans.
reg., cyp125, embR, glbN,
inhA, lppX, nusA, pknE, purN
10
acetylglutamate kinase,
adenylyl cyclase, bphD,
CRP/FNR trans. reg., cyp121,
cysM, inhA, mscL, pknB, sigC
10
acetylglutamate kinase, aroF,
cmaA2, CRP/FNR trans. reg.,
cyp121, cyp51, lpd, mmaA4,
panC, usp
raloxifene
estrogen receptor, estrogen
receptor β
osteoporosis in postmenopausal women
9
adenylyl cyclase, CRP/FNR
trans. reg., deoD, inhA, pknB,
pknE, Rv1347c, secA1, sigC
What Do Off-targets Tell Us?
•
Potentially many things:
1. Nothing
2. How to optimize a NCE
3. A possible explanation for a side-effect of a
drug already on the market
4. A possible repositioning of a drug to treat a
completely different condition
5. The reason a drug failed
6. A multi-target strategy to attack a pathogen
Today I will give you brief vignettes of each {some } of these
scenarios, but first the bioinformatics guts of the approach
The Torcetrapib Story
PLoS Comp Biol 2009 5(5) e1000387
Cholesteryl Ester Transfer Protein (CETP)
CETP inhibitor
X
CETP
LDL
Bad Cholesterol
HDL
Good Cholesterol
• collects triglycerides from very low density or low density lipoproteins
(VLDL or LDL) and exchanges them for cholesteryl esters from high
density lipoproteins (and vice versa)
• A long tunnel with two major binding sites. Docking studies suggest
that it possible that torcetrapib binds to both of them.
• The torcetrapib binding site is unknown. Docking studies show that
both sites can bind to torcetrapib with the docking score around -8.0.
The Torcetrapib Story
PLoS Comp Biol 2009 5(5) e1000387
Docking Scores eHits/Autodock
Off-target
PDB Ids
Torcetrapib
Anacetrapib
JTT705
Complex ligand
CETP
2OBD
-11.675 / -5.72
-11.375 / -8.15
-7.563 / -6.65
-8.324 (PCW)
Retinoid X receptor
1YOW
1ZDT
-11.420 / -6.600
-6.74
-8.696 / -7.68
-7.35
-6.276 / -7.28
-6.95
-9.113 (POE)
PPAR delta
1Y0S
-10.203 / -8.22
-10.595 / -7.91
-7.581 / -8.36
-10.691(331)
PPAR alpha
2P54
-11.036 / -6.67
-0.835 / -7.27
-9.599 / -7.78
-11.404(735)
PPAR gamma
1ZEO
-9.515 / -7.31
> 0.0 / -8.25
-7.204 / -8.11
-8.075 (C01)
Vitamin D receptor
1IE8
>0.0/ -4.73
>0.0 / -6.25
-6.628 / -9.70
-8.354 (KH1) -7.35
Glucocorticoid
Receptor
1NHZ
1P93
Fatty acid
binding protein
2F73
2PY1
2NNQ
>0.0/ -4.33
>0.0/-6.13
/-6.40
>0.0/ -7.81
>0.0/ -6.98
/-7.64
-7.191 / -8.49
/-6.33
/6.35
???
T-Cell CD1B
1GZP
-8.815 / -7.02
-13.515 / -7.15
-7.590 / -8.02
-6.519 (GM2)
IL-10 receptor
1LQS
/ -4.59
/ -6.77
GM-2 activator
2AG9
-9.345 / -6.26
-9.674 / -6.98
(3CA2+) CARDIAC
TROPONIN C
1DTL
/-5.83
/-6.71
/-5.79
cytochrome bc1
complex
1PP9 (PEG)
/-6.97
/-9.07
/-6.64
1PP9 (HEM)
/-7.21
/8.79
/-8.94
1V5H
/-4.89
/-7.00
/-4.94
human cytoglobin
The Torcetrapib Story
/-4.43
/-5.63
/-7.08
/-0.58
/-7.09
/-9.42
/ -5.95
-8.617 / -6.17
???
??? (MYR) -4.16
PLoS Comp Biol 2009 5(5) e1000387
JTT705
Torcetrapib
Anacetrapib
JTT705
VDR
–
RAS
+
RXR
PPARα
PPARδ
FA
?
FABP
?
?
PPARγ
High blood
pressure
+
Anti-inflammatory
function
JNK/IKK pathway
JNK/NF-KB pathway
Immune response
to infection
The Torcetrapib Story
PLoS Comp Biol 2009 5(5) e1000387
Chang et al. 2009 Mol Sys Biol Submitted
Some Limitations
• Structural coverage of the given
proteome
• False hits / poor docking scores
• Literature searching
• It’s a hypothesis – need experimental
validation
• Money 
Limitations
What Do Off-targets Tell Us?
•
Potentially many things:
1. Nothing
2. How to optimize a NCE
3. A possible explanation for a side-effect of a
drug already on the market
4. A possible repositioning of a drug to treat a
completely different condition
5. The reason a drug failed
6. A multi-target strategy to attack a pathogen
Today I will give you brief vignettes of each of these
scenarios, but first the bioinformatics guts of the approach
Acknowledgements
Lei Xie
Li Xie
Roger Chang
Bernhard Palsson
Jacob Durant
Andy McCammon
Sarah Kinnings
http://funsite.sdsc.edu
43,738
Human Proteins
map human proteins to
drug targets with BLAST
e-value < 0.001
map human proteins to
PDB structures with >95%
sequence identity
13,865
Human Proteins
(2,002 Drug Targets)
3,158
Human Proteins
(10,730 PDB Structures)
map drug targets to
PDB structures
1,585
PDB Structures
(929 Drug Targets)
cover 929/2,002 = 46.4%
drug targets structurally
remove redundant
structures with 30%
sequence identity
2,586
PDB Structures
remove redundant structures
with 30% sequence identity,
825
PDB Structures
(druggable)
What we Search Against
The Human Target List
Computational Methodology
Selective Estrogen Receptor
Modulators (SERM)
• One of the largest
classes of drugs
• Breast cancer,
osteoporosis, birth
control etc.
• Amine and benzine
moiety
PLoS Comp. Biol., 2007 3(11) e217
Side Effects - The Tamoxifen Story
Adverse Effects of SERMs
cardiac abnormalities
thromboembolic
disorders
loss of calcium
homeostatis
?????
ocular toxicities
PLoS Comp. Biol., 3(11) e217
Side Effects - The Tamoxifen Story
Structure and Function of SERCA
Sacroplasmic Reticulum (SR) Ca2+ ion channel
ATPase
• Regulating cytosolic
calcium levels in
cardiac and skeletal
muscle
• Cytosolic and
transmembrane
domains
PLoS Comp. Biol., 3(11) e217
• Predicted SERM
binding site locates in
the TM, inhibiting Ca2+
uptake
Side Effects - The Tamoxifen Story
Binding Poses of SERMs in
SERCA from Docking Studies
• Salt bridge
interaction between
amine group and
GLU
• Aromatic
interactions for both
N-, and C-moiety
6 SERMS A-F (red)
PLoS Comp. Biol., 3(11) e217
Side Effects - The Tamoxifen Story
The Challenge
• Design modified SERMs that bind as
strongly to estrogen receptors but do
not have strong binding to SERCA, yet
maintain other characteristics of the
activity profile
PLoS Comp. Biol., 3(11) e217
Side Effects - The Tamoxifen Story
What Do Off-targets Tell Us?
•
Potentially many things:
1. Nothing
2. How to optimize a NCE
3. A possible explanation for a side-effect of a
drug already on the market
4. A possible repositioning of a drug to treat a
completely different condition
5. The reason a drug failed
6. A multi-target strategy to attack a pathogen
Today I will give you brief vignettes of each of these
scenarios, but first the bioinformatics guts of the approach
Bioinformatics Final
Examples..
• Donepezil for treating Alzheimer’s
shows positive effects against other
neurological disorders
• Orlistat used to treat obesity has proven
effective against certain cancer types
• Ritonavir used to treat AIDS effective
against TB
• Nelfinavir used to treat AIDS effective
against different types of cancers
Lots of Opportunities
Summary
• We have established a protocol to look for offtargets for existing therapeutics and NCEs
• Understanding these in the context of
pathways would seem to be the next step
towards a new understanding –
cheminfomatics meets systems biology
• Lots of other opportunities to examine
existing drugs – DrugX and the Recovery Act