The NIEHS Environmental Genome Project: Enabling Studies of Gene-Environment Interaction

1
The NIEHS Environmental Genome Project Enabling
Studies of Gene-Environment Interaction

• Douglas A. Bell, Ph.D.
• Environmental Genomics Section
• National Institute of Environmental Health
  Sciences
• Professor, Dept of Epidemiology
• UNC School of Public Health

2
NIEHSs Environmental Genome ProjectResequencing
of 500 Candidate Genes Potentially Involved in
Environmental Disease

• Concept and rationale
• Examples of gene-environment interaction
• Resequencing studies, accomplishments, and
  accessing data.

3
Modulation of Response to Exposure
Exposure
Early Effects
Genetic Susceptibility
4
Genetic Modulation of Exposure, Damage, and
Biological Response
Exposure
Target tissue
Biological Response
Disease

• Genetic Variation in
• Metabolism, or distribution, affects dose to the
  tissue
• Detection and repair of damage
• Differences in growth and recovery from damage

5
Genetic Modulation of Exposure Risk
Resistant Genotype
Background Risk Level (low)
No Exposure
Sensitive Genotype
Resistant Genotype
2-Fold Risk
Exposure
Sensitive Genotype
4-Fold Risk
6
Benzoapyrene Metabolism
GST Glutathione
Inactive
CYP450
PAH-oxide
DNA Reactive
7
Benzoapyrene Metabolism
GST Glutathione
Inactive
GSTM1 Null
CYP450
PAH-oxide
DNA Reactive
8
Bladder Cancer Risk Associated with Smoking and
GSTM1 Null Genotype
GSTM1 ()
GSTM1 null
1 1.3 2.2
4.3 3.5 5.9
Nonsmokers 1- 50 Packyears Smoking gt50
Packyears Smoking
Plt0.001 Bell et al, JNCI 851559,1993
9
Examples of Gene-Environment Interaction (gene
modifies environmental effect)

• Malaria and Sickle Cell gene.
• HIV infection and CCR5 receptor variant.
• LPS sensitivity and Toll Receptor (TLR4)
• Adverse drug response and CYP2D6 poor metabolism.
• Alcohol intolerance and aldehyde dehydrogenase.
• Smoking, GSTM1 null, NAT2 slow genotypes, and
  bladder cancer risk .

10
Variation in Risk Estimates in Human Populations
Phenotypic variation in response due to
Physiology Metabolism Repair Growth Timing of
Exposure
Risk
Exposure
11
Example Metabolism Polymorphisms
Range of Enzyme Activity in Human Populations
No Phenotypic Polymorphism
frequency
Activity
12
Distribution of Polymorphic Enzyme Activity in a
Population
High
Low
High
Low
/
/
/-
/-
-/-
-/-
Activity
Examples N-Acetyltransferase 2, GSTM1, CYP2D6
13
How does frequency of a risk factor impact
exposure induced (G x E) risk in the population?
14
Effects of Exposure in High and Low Risk Human
Populations
Risk
100
High Risk
10
Average
Low Risk
0
Exposure
15
How will genetic data be used in public health
risk assessment?

• Given detailed information on the relationship
  between genotype and phenotype, more accurate
  risk assessments may be possible.

16
Risk Assessment Process
Hazard/Risk Assessment
Risk Management More/Less Control
Replace default assumptions about variability
17
Incorporating Human Genetic Polymorphism
Information Into Risk Assessment
Cancer - Yes/No Dose ? Extrapolate to Humans
Chemical X

• Biochemistry
• Mechanism of toxicity
• Genes, pathways
• Human genetics

Susceptible human subgroup?
18
Incorporating Genetics Into Risk Assessment
Issues

• A polymorphism may have different effects
  depending on the chemical, the target organ/
  disease, and the population being considered.
• Thus, a protective allele for one chemical may
  convey risk for a different chemical. Similarly
  one organ system may be protected at the risk of
  another e.g. immune system response could
  increase DNA damage or neurotoxicity.

19
GST Theta 1 (GSTT1) - One gene with 2 effects
Detoxication
Ethylene oxide
Inactive
GSTT1 Glutathione
Activation
(Unstable)
HCHO
DNA
Methylene chloride
GSTT1 Glutathione
DNA Reactive
DNA
(also Methyl chloride)
D.A.Bell NIEHS
20
Activation vs. Detoxication

• Effects of polymorphism dependent on chemical and
  toxicity pathway
• Activation - If the activation pathway is missing
  (null genotypes), some individuals may have zero
  risk even if they have exposure.
• Detoxication - Since this process will never be
  100 efficient, both functional and low activity
  genotypes will exhibit risk associated with
  exposure.

21
The Effect of GSTT1 Genotype on Metabolism of
Methyl Chloride
T1 Null No Metabolism
Measure exhaled methyl chloride
T1 Metabolism to DNA reactive forms
From Lof, A. et al, Pharmacogenetics 10645,
2000.
22
Smoking, GSTT1 Polymorphism, and Markers of
Genotoxicity in Erythrocytes

• Background Ethylene oxide hemoglobin adducts
  are a good measure of smoking exposure in blood.
• Experiment To test if GST genotypes modulated
  effects of smoking in erythrocytes, we measured
  ethylene oxide hemoglobin adducts in freshly
  collected human erythrocytes from nonsmokers and
  smokers.
• Results
• Ethylene oxide adducts (HEV) were 50 higher in
  GSTT1 null individuals.

D.A.Bell NIEHS
23
GSTT1 null genotypes have higher levels of
smoking-induced hemoglobin adducts
Study Design 16 nonsmokers 32 smokers HEVal
hemoglobin adducts measure by mass
spectrometry P 0.001 for difference in
slopes Nonparametric analysis similar.
Fennel et al CEBP 9705,2000
24
Incorporating Genetics Into Risk Assessment
Needs

• Identify genes involved in toxicological
  response.
• Detailed population genetic information
  including
• Identify polymorphisms.
• Determine frequency in populations.
• Population-based risk estimates in large studies
  (n2000).
• Determine functional relationship between
  genotype and phenotype
• Biochemical
• In vitro, in vivo quantitative measurements of a
  cellular phenotype (tumors, adducts, mutation,
  cell death, gene expression).
• Consider role of multiple genes, multiple
  pathways, etc.
• Incorporate kinetic or other functional data into
  risk model.

25
Environmental Genomics
Discovery Phenotype-directed Genotype-directed
Functional Analysis
Disease Risk Characterization
CTTATGT A/C GGGTAT
Effects in Populations
Phenotype
Genotype
Altered Binding
26
Polymorphism and Function
Exon 1
Exon 2
Promoter
3 UTR
Gene Deletions, Duplications
e.g. GSTM1, CYP2D6

• Effects of Polymorphism
• Altered function
• Quantity of protein

27
PhenotypeDirected Approach to Find SNPs That
Alter Gene Expression Level

C TGGGCCCCGCCCCCTTATGTA
GGGTATAAAGCCC . CCCGTCACC ATG  
SP1/Oct
Liu, X. et al
28
Sequence-Directed Approaches to Catalogue All
Significant SNPs In The Human Population

• Resequencing Projects Describing candidate gene
  polymorphisms in diverse populations.
• 9 million SNPs in dbSNP now,
• by 2006, expect 20 million human SNPs.
• A SNP every 100 bases.
• Haplotype Map Describing which SNPs occur
  together on chromosomes in populations
  (haplotypes).

29
SNP Discovery Projects

• The SNP Consortium 1 million SNPs across
  genome
• NIEHS Environmental/toxicology genes
• NHLBI Heart disease genes, inflammation
• NIGMS Pharmacogenetic genes

SNP data is entered into the NCBI dbSNP database
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
(No Transcript)
37
UCSC
Hapmap
38
(No Transcript)
39

• U Wash EGP Website

40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
HapMap Website

• Characterize the large scale genetic structure
  across the genome.
• Genotyping SNPs at 1 kb interval across the
  genome in European, African, and Asian
  populations.

44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
(No Transcript)
48
Bioinformatic Tools Available For Picking
Haplotype Tagging SNPs

• HapMap Website
• Seattle SNPs or EGP website
• Many other freely available programs

49
NIEHS Environmental Genome Project

• Resequencing of candidate environmental disease
  genes
• Accomplishments
• Total genes sequenced 437
• Total kilobases sequenced 11,001 kb
• Total SNPs identified 59,475

50
NIEHSs Environmental Genome ProjectSummary

• Gene-environment interaction affects disease
  risk.
• Effects of G x E interactions can be complex.
• Resequencing projects are providing many new
  candidate gene polymorphism.
• Determining the important functional SNPs that
  affect disease risk is a difficult challenge.

51
Strategies For Incorporating SNPs Into
Epidemiology Studies

• Whole genome association studies
• Test 10,000-100,000 SNPs in case control studies.
  
• Identify candidate regions, genes, followup with
  candidate gene studies.
• 2. High resolution candidate gene studies.
• Test functional SNPs and additional haplotype
  tagging SNPs in case/control or other design.
• Bioinformatics to identify 1500 SNPs, 150 genes
  (10 SNPs/gene).
• Coding SNPs, regulatory SNPs, haplotype tag SNPs.
  

52
Bioinformatic Identification of SNPs That Affect
Gene Expression

• Application to p53 response elements
• Application to NRF2 response elements

53
p53 inducible genes contain p53 Response
Elements.
Following UV exposure p53 binds RE of target gene.
RNA Pol
SEI1 mRNA
p53
p53
RRRCWWGYYY
SEI1 gene
Using bioinformatic methods, identify SNPs that
disrupt p53 response elements.
54
Binding Site Consensus
NCBI/Ensemble Genome Data
dbSNP Data
Test SNPs Against p53 Response Element Consensus
RRRCWWGYYYRRRCWWGYYY
AAAGGACAAGTTGAAACTTGCACAAGCAGCCTCCATTCTG
Build Table of All Promoter SNPs
DNA ambiguity code R A or G Y C or T W A or
T
Access database
Filter Best Hits
Dan Tomso
55
CWWG motif
Dan Tomso
56
Do SNPs in putative p53 response elements affect
p53 induced expression in Saos2 cells?
Saos2 Osteosarcoma Cells (p53 null)
Strong
Weak
Mike Resnick, Alberto Inga, Daniel Menendez
57
Environmental Genomics Section

• Douglas A. Bell
• Gary S. Pittman
• Merrill Chip Miller, III
• Daniel J. Tomso
• Michelle R. Campbell
• Xuemei Liu
• Xuting Wang
• Monica Horvath

58
Phylogenetic Footprinting of NRF2/ARE Genes
4000 Mouse ARE containing genes
1000 human/mouse
4000 Human ARE containing genes
Human/ mouse/rat 380
2100 Rat ARE containing genes
59
(No Transcript)
60
Gene x Environment Interaction

• Pharmacogenetics
• Adverse drug reactions (toxicity)
• Reduced efficacy
• Environmental disease
• Modification of exposure-induced toxicity
• Modification of exposure-induced disease
• Can we generalize about risk associated with a
  specific gene?