Transcript Document

Genomic /Genetic Considerations in CNS Drug Development:
Current Status and Approaches
1. Basic Science Standpoint
Orest Hurko, MD
American Society for Experimental NeuroTherapeutics
11th Annual Meeting
Arlington Virginia
6 March 2009
AVP, Wyeth Research
Professor (Hon.), University of Dundee
Director: Translational Medicine Research Collaboration
ASENT.6 March 2009
Genomic /Genetic Considerations for CNS drugs
Genes & the pharmaceutical enterprise
Methods for studying genes
Genome-wide association studies
Opportunities
ASENT.6 March 2009
Decoding the human genome delivered lots of promise
… but the downstream challenges were underestimated
This enabled the shift from 500 pharmacologically
proven targets to 10,000’s of unproven targets,
changing the entire drug discovery process
Easy wins in rare single-gene disorders prompted
unrealistic optimism for common diseases with
complex genetics
Focus was on technology, not on analysis
Hundreds of millions were spent with minimal return
Millenium has completely exited genetics to become
a conventional drug company
Many companies have abandoned efforts in genetics
In favor of genomics
ASENT.6 March 2009
But now, ten years on, genetics has finally become a
practical translational tool for industry
• Understanding of statistical issues
• High-fidelity high-throughput genotyping
• Large repositories of population-based samples
• Consortia with standardized procedures
• Growing appreciation of the heuristic value of outliers
o
Target Validation
Target/Compound
Interaction
Pharmacodynamic
Activity
• Ever-growing number of robust validations
Disease Biomarker
& Disease
Modification
Patient Stratification
ASENT.6 March 2009
Rare diseases have simple genetics.
Common diseases have complex genetics.
ASENT.6 March 2009
Why should a big drug company be interested
in a rare disease?
Goldstein JL, Brown MS (1973) Familial hypercholesterolemia:
identification of a defect in the regulation of
3-hydroxy-3-methylglutaryl coenzyme A reductase activity
associated with overproduction of cholesterol.
Proc Natl Acad Sci U S A. 70: 2804-8.
“The homozygous form of the autosomal dominant disorder,
familial hypercholesterolemia, is characterized by the presence
in children of profound hypercholesterolemia,
cutaneous planar xanthomas, and rapidly progressive coronary
vascular disease that usually results in death before age 30
years….. “
.
ASENT.6 March 2009
Demonstrable value of studying extreme
phenotypes
Goldstein JL, Brown MS (1973) Familial hypercholesterolemia:
identification of a defect in the regulation of
3-hydroxy-3-methylglutaryl coenzyme A reductase activity
associated with overproduction of cholesterol.
Proc Natl Acad Sci U S A. 70: 2804-8.
“The homozygous form of the autosomal dominant disorder, familial
hypercholesterolemia, is characterized by the presence in
children of profound hypercholesterolemia, cutaneous planar
xanthomas, and rapidly progressive coronary vascular disease
that usually results in death before age 30 years. ….”
.
Abbott stands to gain as cholesterol-fighters cut risk in other heart issue
By Bruce Japsen | Tribune reporter November 13, 2008
ASENT.6 March 2009
But now, genome-wide association
studies allow efficient study of common
diseases as well.
ASENT.6 March 2009
Genomic /Genetic Considerations for CNS drugs
Genes & the pharmaceutical enterprise
Methods for studying genes
Genome-wide association studies
Opportunities
ASENT.6 March 2009
Three distinct methodologies for
three different questions
Genetics
Genomics
Molecular Biology
The study of variation and its inheritance
Intrinsically probabilistic
Heritability, segregation analysis, linkage, association
The study of expression of all genes of an organism
Deterministic
Transcriptional profiling , in situ hybdrization, difference
libraries
The study of molecules underlying genetics & genomics
Deterministic
Cloning; sequence and structural analyses; cross
hybridization; site-directed mutagenesis; si RNA
knockdowns; transgenics &b knockout abimal models
ASENT.6 March 2009
Genetic analyses should only be undertaken
if there is significant heritability
 Heritability estimates are always relative to the genetic and environmental factors
in the population
 Heritability describes the population, not individuals within that population
Heritability can be estimated in controlled experiments & in population studies
Phenotype (P) = Genotype (G) + Environment (E).
Var(P) = Var(G) + Var(E) + 2 Cov(G,E).
If Cov(G,E) = 0. then H2 = Var(G)/ Var(P)
ASENT.6 March 2009
The pattern of inheritance dictates the optimal genetic
approach
Rare sporadic childhood disorders
-- Chromosomal rearrangements (or recessives)
-- Responsible genes in breakpoints or duplications/deletions
Sequencing
Unusual high density families
-- dominant
-- point mutations (or microdeletions / duplications) in genes of major effect
Linkage
Common, adult disorders without pronounced familial grouping
-- multiple genes of additive effect
-- often major environmental interactions
Association
ASENT.6 March 2009
High-density families ideal for linkage analysis
Figure 1. Pedigree structure of the two Chinese families with tooth agenesis.
ASENT.6 March 2009
Genomic /Genetic Considerations for CNS drugs
Genes & the pharmaceutical enterprise
Methods for studying genes
Genome-wide association studies
Opportunities
ASENT.6 March 2009
Linkage & Association – same general principle,
different time scales
Linkage
Association
20 generations
Crossovers increase with distance
ASENT.6 March 2009
Genome–wide
association studies are
G
e
based
on
a
very
simple
idea
n
o
m Score genomic DNA from a
e large sample of cases &
very
controls
for a very large number
of wsingle-nucleotide
i
polymorphisms
(SNPs)
d
e
1
 Compare the frequencies
among cases & controls
 Sites that differ significantly
between cases and controls
are then validated in
independent samples
ASENT.6 March 2009
Lessons learned from Genome-Wide Association
Studies (GWAS)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Genome-Wide Association Studies (GWAS) work
Effect sizes are usually small, so big samples needed
Rigorous quality control is paramount
GWAS may fail to detect certain susceptibility genes
Important to look well beyond the top few ‘hits’
Collaboration is important
Phenotype/selection is important
Validation is critical
Every SNP counts
“Low-hanging fruit” lead to more variants
ASENT.6 March 2009
Genome-Wide Association Studies (GWAS) work
Proof of Principle
 Complement factor H gene - -age-related macular degeneration
Reliable reproduction in follow-up studies:
PPARg & transcription factor TCF7L2 - Diabetes mellitus
IL23R, CARD15, NOD2 -- Crohn’s disease
chromosome region 8q24 - prostate cancer
GSTM1 null -- bladder cancer & acute leukemia
NAT2 slow acetylator -- bladder cancer
MTHFR C677T -- gastric cancer
Confident associations in other common diseases
coronary artery disease
atrial fibrillation
asthma
rheumatoid arthritis
obesity
breast cancer
coeliac disease
ASENT.6 March 2009
Effect sizes are usually small, so big
samples are needed
 Theory predicts that alleles of small effect are more frequent than alleles of
large effect
 Wellcome Trust Case Control Consortium (WTCCC) GWASs of seven
common diseases found per-allele odds ratios of 1.2–1.5
 Reasonable power to detect such loci requires 2000 cases and 2000 controls
 Failures to replicate findings in modestly sized samples do not constitute refutation
 Confidence attributable to a ‘significance level’ is influenced by sample size
 Rate of true positives increases with sample size because power to detect
true effects increases
ASENT.6 March 2009
Rigorous quality control is paramount
 Enormous data sets (samples & SNPs) in GWASs provide large
opportunities for spurious ‘associations’
 Data must be cleaned thoroughly to remove low-quality DNA
samples, genotype calls & individual samples
 Within WTCCC the best predictor of an SNP with poor QC was a
highly significant difference in genotype distributions between
cases and controls
-- validation is critical
ASENT.6 March 2009
GWAS fail to detect some susceptibility
genes
 Underpowered studies were a leading cause of failure
 Current technology surveys only a limited subset of potentially
relevant sequence variation
 Poor coverage of large genes
 Some mutations – such as copy number variations (CNVs)
from microduplications or microdeletions –are not detectable in many
SNP-based platforms used for GWAS
ASENT.6 March 2009
Phenotype & case-selection is important
Example: association of FTO gene with Type 2 Diabetes Mellitus
WTCCC (not matched for body mass)
-- 2000 cases & 3000 common controls
-- significant association @ P = 1.3 x 10-12
 DGI (matched for body mass)
-- 14,000 cases and controls
-- no association whatsoever
Subsequent work has shown that fat mass &
obesity-associated (FTO) influences risk of T2D
through a primary effect on body mass
ASENT.6 March 2009
Collaboration is important
 Benefits from collaborations that increase total sample sizes ,
test consistency & generalizability of findings
 Aggressive, very early, proactive data sharing key to
identification of several susceptibility loci not evident in any
single study alone
Standard phenotyping, threshholds for genome calls, raw data
sharing
 Diabetes Mellitus (Types 1 & 2), coronary artery disease,
ankylosing spondylitis benefited from collaborations
ASENT.6 March 2009
Genomic /Genetic Considerations for CNS drugs
Genes & the pharmaceuticals enterprise
Methods for studying genes
Association studies
Opportunities
ASENT.6 March 2009
Lessons for Experimental Neurotherapeutics

Learn from rare Mendelian variants of genes
encoding potential drug targets (OMIM)

Schizophrenia, autism, restless legs syndrome,
early onset depression, bipolar disease, multiple
sclerosis, Alzheimer disease, ADHD, & dyslexia are
heritable common diseases tractable for GWAS
 Do not succumb to the temptation of relaxing
diagnostic criteria to boost sample size
ASENT.6 March 2009
Special Thanks
Douglas Blackwood
Neil Craddock
Dan Crowther
Charles ffrench-Constant
Fred Immerman
Maha Karnoub
Robin Fears
Gino Miele
Ralph McGinnis
Victor McKusick
E.A.Murphy
Colin Palmer
David Porteous
Nigel Spurr
David St. Clair
Keith Vass
ASENT.6 March 2009