Transcript Repair of DNA double-strand breaks and susceptibility to

Breast/Ovarian Family
† 57
54
22
43
† 49
55
48
51
32
+
+
PO
45
PO
45
B 32
O 59
39
Inherited predisposition
More BRCA-like genes
Rare, moderately strong variants
Common genetic variation
Role of normal genetic variation in
determining individual risk.
How useful is this information in
selection for screening and
prevention?
How do we find the genes?
Breast cancer as an example
Evidence that genetic variation
affects risk
Measure of variation = familial
clustering
Risk in close blood relative compared to
risk in population as a whole
= roughly 2-fold.
Is family clustering genetic?
MZ twin
DZ twin
Mother/sister
Incidence % per year
1.31
0.5
0.36
Patient’s contralateral breast
0.66
(Peto & Mack, Nat Genet 26, 411
(2000))
How much genetic predisposition is there?
How is it distributed?
Determines potential for
discriminating individual risks
risk
Breast/Ovarian Family
† 57
54
22
43
† 49
55
48
51
32
+
+
PO
45
PO
45
B 32
O 59
39
Familial clustering of breast cancer
OBS
EXP
Excess
177
106
71
Population
13
1.47
11.5
BRCA1/2
mutation
Fraction of excess familial clustering
attributable to BRCA1/2 = 15-20%
Familial clustering of breast
cancer
Risk to
1o relative
of case
Roughly 15-20%
due to BRCA1/2
2
Excess familial risk
1
ATM
Chk-2
Ha-ras
PTEN
What sort of genes may account
for familial risk apart from BRCA1/2?
Common low-penetrant
genes
BRCA3 etc
1.5
Allele freq.
1%
10%
30%
XsFRR
.25
2.3
5.3
Number
350
35
16
10
Allele freq.
0.2%
BRCA1, 2
Relative risk
XsFRR
Number
16
5
Patterns of breast cancer in
families
1500 cases, population based
BRCA1/2 excluded
What model fits best?
Best fit = combined result of several
factors,
individually of small
effect
= log-normal distribution of risk
in population.
Distribution of genotypes in
population and cases by genotype
risk
0.040
SD = 1.2
0.030
Population
Cases
0.020
0.010
0.000
0.01
0.10
1.00
Relative risk
10.00
100.00
Proportion of population and cases
above specified risk: SD = 1.2
Proportion above given risk (x)
100%
88%
Cases
Population
50%
46%
10%
0%
0% 3%
12%
20%
40%
60%
Risk of breast cancer by age 70
80%
Effects of normal genetic
variation on breast cancer risks
Population
10%
Cancers
Individual risk
by age 70
50%
46%
12%
>1:8
< 1 : 30
Proportion of population and cases
above specified risk: SD = 0.8
Proportion above given risk (x)
100%
80%
Cases
Population
50%
31%
10%
0%
0% 4% 11%
20%
40%
Risk of breast cancer by age 70
60%
80%
Proportion of population and cases
above specified risk: SD = 0.3
Proportion above given risk (x)
100%
Cases
Population
75%
50%
25%
0%
0%
20%
40%
Risk of breast cancer by age 70
60%
80%
Gail model of breast cancer
risk
Nurses Health Study Analysis
Excellent prediction of breast cancer incidence in
specified population.
Poor prediction of risk to individual.
2.8-fold between upper and lower deciles
cut-off for tamoxifen use defined 33% of
population with 44% of cases.
(Rockhill, JNCI 93, 358 (2001))
- find genes
- interactions
- validation
1/5
1/5
40x
risk
QuickTime™ and a
Photo - JPEG decompressor
are needed to see this picture.
How to find the genes?
Association studies
C
T
arg
cys
V
indirect
direct
linkage disequilibrium
Problems:
recombination
origins different time
multiple origins
Common variant : common disease
Marker
Disease allele
Rare variants
Candidate genes
Estrogen synthesis and degradation;
ER
Cell cycle checkpoints
DNA repair
TGFb pathway
IGF pathway
Carcinogen metabolism
Sample sets
Initial : 2000 cases, 2000 controls
Confirmatory : 2000 cases, 2000 controls
Cases -
Population based, East Anglia
simple epidemiology data, survival;
paraffin blocks
Controls - EPIC cohort, East Anglia
extensive epidemiological data, follow-up,
serum, mammography, bone density, etc
Power
Percentage polygenic variance explained.
90% power
p = 10-4
multiplicative
6000
Sample size
5000
4000
1%
2%
3000
5%
2000
10%
1000
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Allele Frequency
(Antoniou & Easton, submitted)
Provisional positive associations :
breast cancer
98 snps
47 candidate genes
Risk Br Ca Fraction
to age 70 of excess
(5.7%)
RR
Freq
OR
PAF
TGFb
BRCA2
XRCC3
ERa
14%
7%
15%
20%
1.25
1.31
1.34
1.27
2.9%
2.1%
4.4%
5%
6.8%
7.4%
7.4%
6.8%
0.2%
0.3%
0.5%
0.5%
Chk2
0.5%
2.4
0.6%
16%
0.5%
~2.0%
BRCA2 N372H association with breast cancer risk
Finns HH
HDB HH
UK set 3 HH
UK set 2 HH
UK set 1 HH
Joint HH
Joint NH
Joint NN
0.1
1
10
p=0.02
Tee et al. In prep.
3133
Fiegelson et al. 2001
OR breast
cancer
Haiman et al. 1999
1081
Mitrunen et al. 2000
744
Kristensen et al. 1999
CYP17 t -34 c
Spurdle et al. 2000
(cc Vs. tt)
Miyoshi et al. 2000
Kuligina et al. 2000
Hamajima et al. 2000
310
Huang et al. 1999
Conclusion:
This SNP has no main effect
on breast cancer risk!
Helzlouler et al. 1998
230
Weston et al. 1998
Bergman-Jungestrom et al. 1999
226
Young et al. 1999
Weston et al. 1998
N
0.1
1
10
100
Ye & Parry, 2002
Mutagenesis 17:119-126
Why a p value of p = 0.01 is not persuasive
True
association
False
association
Prior probability of result
(snp causing 1% of FRR,
100,000 snps in genome)
1/1000
999/1000
Probability given result
has p = 0.01
99/100
1/100
99/100,000
999/100,000
Assuming random choice of ‘candidate’ gene
only ~ 10% results at p = 0.01 are true
(~50%, at p = 0.001)
Summary of results
96 snps, 47 genes
~2000 cases, 2000 controls
p-value
0.001
p = 0.01/0.0004 for comparison
of distributions
0.01
0.05
observed
0.10
chance
1.00
0
10
20
30
40
50
SNP
60
70
80
90
100
% of excess FRR
explained
0.5
1
1.3
2
relative risk
Some reasons why human
association studies may be
difficult
Inappropriate genetic models eg rare/multiple alleles
Regulatory vs coding polymorphisms
Numbers : inadequate statistical power
Genetic background effects; interactions
weak ‘main effect’, high-order interactions
‘null’ result = balance of susceptible and resistant on
different BG
Phenotypic heterogeneity eg ER+/ER-; histology
Cancer/no cancer endpoint lacks power
Intermediate phenotypes
Serum estradiol and
CYP19
Exon
10 t>c 3’UTR
20
18
Serum SHBG and SHBG
Exon 8 g>a or D356N
60
50
16
40
14
30
12
10
tt
tc
P homogeneity =
0.0005
P trend <0.0001
cc
20
gg
ga
aa
P homogeneity = 0.006
P trend = 0.006
(Ponder, Dowsett labs;
EPIC; unpublished)
Implications for breast cancer
risk
2 fold increase in estradiol  30% increase
in risk of breast cancer
tt genotype of CYP19 c>t associated with
14% increase in estradiol: equivalent to
1.04 fold increase in breast cancer risk
Where next?
Empirical vs candidate approaches
Snp genotyping now ~17c/genotype :
? screen 600 “enriched” cases/600
controls
vs 1150 coding snps
~$240,000
Candidate gene approaches
Candidates from cell biology
Epidemiology
Regulatory variants
Quantitative phenotypes
Leads from mouse models
Mouse/human collaborations
1. Candidate susceptibility genes/regions
mapped in susceptible/resistant crosses
refined by amplicons/deletions in tumours
allele-specific differences in expression/somatic change
(easier in mouse because extended haplotypes)
loci involved in control of gene regulation
loci influencing intermediate phenotypes
set up large cross and score multiple phenotypes
How tightly should the
region be defined?
300 kb
Say 5 genes
First pass = find all coding region snps at >5%
Construct haplotypes, select minimum snp set = ? 30 snps
Genotype 30 snps in 2000 cases/2000 controls = 120,000
genotypes
Genotyping cost ~$20,000 @ 17c/genotype
BUT : currently requires ~1000 snps at a time
Mouse/human collaborations
2. Interactions
Identification of interacting loci
potentially approachable in mouse
Develop and evaluate programmes to
search for higher order
interactions;
? applicability to man
Mouse/human collaborations
3. Stages of cancer development
? Distinguish loci that influence
multiplicity
latency; progression
invasion
metastasis
and resistance to these
? Loci that affect treatment response
Mouse/human collaborations
4. “End game” - which is the active gene,
snp?
strain comparisons of variants
dissection of complex QTLs
transgenic models
A new horizon in medicine?
“‘Risk factor’ analysis will facilitate environmental modification,
screening and therapeutic management of people before they
develop symptoms”
(Bell, BMJ 1998)
“Differences in social structure, lifestyle and environment account
for much larger proportions of disease than genetic differences
…… Those who make medical and scientific policies ….. would do
well to see beyond the hype”
(Holtzman & Marteau, NEJM 2000)
Strangeways Research Laboratories University of Cambridge
Bruce Ponder
Paul Pharoah
Alison Dunning
Fabienne Lesueur
Bettina Kuschel
Annika Auranen
Katie Healey
Craig Luccarini
Jenny He
Louise Tee
Gary Dew
Doug Easton
Antonis Antoniou
Mitul Shah
Julian Lipscombe
Nick Day; EPIC
UCSF
Allan Balmain
Mandy Toland
Joe Gray
Mark Sternlicht
NCI
Kent Hunter
Biochemistry, Cambridge
Jim Metcalfe
Cancer Research UK; MRC
TGFb
t/c
-509
t
0.25
Pro/Leu
10
P
tt vs cc
c
P
c
L
0.11
0.60
PP vs LL OR 1.25 (1.1 - 1.4)
p = 0.01
OR 1.30 (1.1 - 1.5)
p = 0.01
Which SNP is the functional variant?
tt ProPro
Pro10 homozygotes
have increased risk
regardless of c-509t
genotype
ct ProPro
cc ProPro
ct LeuPro
cc LeuPro
cc LeuLeu
0.1
1.0
Odds Ratio 10
TGFb in vitro secretion
Time Course
End Point
4
Pro10
3
Ratio P:L
TGFb1
ng/ml
2
Leu10
1
0
0
6
12
18
hours
(Metcalfe, Ponder labs, 2002)
Funnel Plot For TGFb L10P
O R (PP Vs. LL)
N
* Cohort study
4517
ABC
875
HDB
939
Finn
404
Hishido et al.
*
3075
Ziv et al.
238
Frei
146 cases 2929 controls
0.1
1
10