Genetic Epidemiology of Oral Clefts

(1)

Genetic Epidemiology of Oral Clefts

Rolv T. Lie

Department of Global Public Health and Primary Care University of Bergen

Norway

(2)

Outline

1. Registry-based data on parent-offspring recurrence

– Birth defects have specific familial risks

– Recurrence risks are high for oral clefts (30-fold) – Multiple genes are assumed to be important

2. Hunt for genes for oral clefts

– GWAS have found 50+ associated SNPs – Estimates of associations

3. How much of the recurrence risk of oral clefts is

_

explained by the effect of these genetic variants?

(3)

Oral clefts

cleft lip only CLO

cleft lip with cleft palate CLP

cleft palate only CPO

O’Rahilly & Müller, 1992

(4)

1. Medical Birth Registry (MBR) of Norway

• All births in Norway since 1967 (~3 million)

• Medical information on delivery, child and mother

• Unique id-number for all persons i Norway

• Known id-number of child and both parents in MBR

• Possible to link a persons birth-record with birth- records of offspring

• Ascertainment vary for birth defects (90% for cleft

lip)

(5)

Risk of Similar and Dissimilar Birth Defects in Children According to Category of Birth Defect in the Mother.

From: A population-based study of survival and childbearing among female subjects with birth defects and the risk of recurrence in their children.

N Engl J Med. 1999 Apr 8;340(14):1057-62. Skjærven R, Wilcox AJ, Lie RT:

(Study of a total of 187 544 children of mothers born in 1967-82)

(6)

From: Survival and Reproduction Among Males With Birth Defects and Risk of Recurrence in Their Children.

Lie RT, Wilcox AJ, Skjærven R

JAMA. 2001;285(6):755-760. doi:10.1001/jama.285.6.755

(Study of a total of 110 427 children of fathers born in 1967-82)

(7)

From: Familial risk of oral clefts by morphological type and severity: population based cohort study of first degree relatives. BMJ. 2008 Feb 23;336(7641):432-4.

Sivertsen A, Wilcox AJ, Skjaerven R, Vindenes HA, Abyholm F, Harville E, Lie RT:

Data from surgical clinics were linked with MBR Study of 703 131 children and their parents

All estimates are for isolated oral clefts

(8)

From: Familial risk of oral clefts by morphological type and severity: population based cohort study of first degree relatives. BMJ. 2008 Feb 23;336(7641):432-4.

Sivertsen A, Wilcox AJ, Skjaerven R, Vindenes HA, Abyholm F, Harville E, Lie RT:

CLO and CLP have common etiology, CPO is a distinct category

(9)

Extensive data since 1954 on oral clefts and

parent-offspring recurrence from a case-registry in Denmark:

A cohort study of recurrence patterns among more than 54,000 relatives of oral cleft cases in Denmark: support for the multifactorial threshold model of inheritance.

J Med Genet. 2010 Mar;47(3):162-8.

Grosen D1, Chevrier C, Skytthe A, Bille C, Mølsted K, Sivertsen A, Murray JC, Christensen K.

RR 95% CI

CLO- CLO: ~46 (34-62)

CLP - CLP: ~29 (22-38)

CPO-CPO: ~28 (21-39)

(10)

Meta-analysis of published data from Denmark and Norway

Estimates of recurrence risk from parent to child of isolated oral clefts among Scandinavians:

_________

Absolute risks

^_________

RR (95% CI) Recurrence Reference

^____

CLO - CLO: 42 (31-56) 2.3 % 0.055 % CLP - CLP: 29 (22-37) 2.5 % 0.086 %

CPO -CPO: 32 (24-42) 2.3 % 0.073 %

^_____

(11)

2. The hunt for genes for Oral clefts

(12)

(13)

Sample size, Caucasians

Study CLO CPO CLP Controls Denmark 47 59 30 772

Iowa 40 61 50 254

NBDPS (USA) 245 323 248 1886 Norway 148 273 168 1139

Utah 85 137 97 470

Total 565 853 593 4521

(14)

Design of the study

2011 case-dyads (mother and child)

4521 control-dyads

(mother and child)

(15)

Is deviation form HW-equilibrium a problem?

Cont.

(16)

Meta Analysis (post Haplin Strat)

R-syntax:

Run Haplin Strat

# Log-transformation of RRs log.RR <- log(esti$RR.est)

# Pull approximate SE for log RR from confidence intervals log.SE <- (log(esti$RR.upper) - log(esti$RR.lower))/(2*1.96)

# Weights for the meta-analysis are the inverse of the variances w <- 1/log.SE^2

# Common log RR

tot.log.RR <- sum(w*log.RR)/sum(w)

# SE for common log RR tot.log.SE <- 1/sqrt(sum(w))

# CI for common RR

tot.RR <- exp(c(est = tot.log.RR, lower = tot.log.RR - 1.96*tot.log.SE, upper = tot.log.RR + 1.96*tot.log.SE)) (heterogeneity test of RRs is a part of Haplin Strat)

(17)

Frequency of SNPs that had effect for one category of clefts

Gene SNP MAF

PAX7 rs742071 0.40

ABCA4_ARHGAP24 rs560426 0.46

IRF6 rs642961 0.20

THADA rs7590268 0.23

8q21.3 rs12543318 0.34

8q24 rs987525 0.22

FOXE1 rs3758249 0.39

KIAA1598-VAX1 rs4752028 0.17

SPRY2 rs8001641 0.49

TPM1 rs1873147 0.27

NOG1 rs227731 0.45

MAFB rs13041247 0.40

(18)

(19)

Effects of child’s own alleles

(20)

Poo-effects for iCLO

(21)

Poo-effects for iCLP

(22)

No maternal genetic effects for CLO etc.

(23)

Isolated vs. non-isolated CLO or CLP

(24)

3. How much of the recurrence risk is explained by the effect of the 12 SNPs?

RR % explained

observed by the 15 SNPs

CLO - CLO: 42 1.0% (1.42-1)/(42-1)

CLP - CLP: 29 1.0% (1.30-1)/(29-1)

CPO -CPO: 32 0.1% (1.03-1)/(32-1)

recurrence risk(genes) - population risk

% explained =

recurrence risk(observed) - population risk RR_genes - 1

= RR_obs -1

(25)

Assumptions and calculations,

RR of recurrence produced by each SNP

Assumptions:

• Hardy-Weinberg equilibrium

• Random mating

• Multiplicative effect used to combine RRs

) parent

| child (

P ) parent

| child (

RR

_Re_currence

= P + + + −

) parent (

P ) parent and

child (

) P parent

| child (

P + + +

=

+

(26)

Calculation of joint risk of both mother and child

Take sum across categories to obtain the probabilities on previous slide

(27)

Effects of child’s own alleles

(28)

If risk loci can be combined in a multiplicative model, then their total effect

on the recurrence risk ratio (RRR) is the product of the RRR from each locus

(29)

Tests of pairwise interactions

(beyond multiplicative)

(30)

Little evidence of higher order

interactions

(31)

Recurrence risk parent-to-offspring effected by each SNP

Gene SNP MAF CLO CLP CPO

Hypothetical if:

RRsd=3 RRdd=10

PAX7 rs742071 0.40 1.043 1.024 1.000 1.33

ABCA4_ARHGPA24 rs560426 0.46 1.003 1.013 1.001 1.30

IRF6 rs642961 0.20 1.057 1.015 1.000 1.36

THADA rs7590268 0.23 1.004 1.012 1.000 1.37

8q21.3 rs12543318 0.34 1.045 1.011 1.000 1.37

8q24 rs987525 0.22 1.116 1.097 1.002 1.36

FOXE1 rs3758249 0.39 1.011 1.013 1.002 1.34

KIAA1598-VAX1 rs4752028 0.17 1.020 1.019 1.001 1.35

SPRY2 rs8001641 0.49 1.009 1.013 1.002 1.29

TPM1 rs1873147 0.27 1.017 1.009 1.001 1.37

NOG1 rs227731 0.45 1.021 1.006 1.020 1.31

MAFB rs13041247 0.40 1.016 1.033 1.001 1.33

Total RR 1.42 1.30 1.03 33.3

(32)

What do we know about the contribution of common and rare genetic variants?

Weak effect Strong effect Common genetic

variants

Contributes

moderately Not found

Rare genetic variants

Probably not so important

Unknown, but

probably important!

(33)