Why are social factors important in genetics (and vice-versa)? Andrew Pickles University of Manchester Methodological challenges for the Twenty-First Century.

Why are social factors important in genetics

(and vice-versa)?Andrew Pickles

University of Manchester

Methodological challenges for the Twenty-First Century

RMP Meeting 22-23 Nov 2007

Rapid Change• Wholesale change of material since spoke to

this topic some 3-4 years ago

• Previously material still relevant– concerned to contrast environmental variation in

population means from variations in within population individual differences where estimates of additive genetic variance dominate and wider “shared” social factors often apparently negligible.

– How heritabilities need careful interpretation and can be entirely environmentally contingent

– Nature versus nurture contrast as unhelpful

Genetics versus environment: false dichotomy

DNA is very distal to the outcomes of interest - there are many intermediate stages at which the environment and the social will modulate genetic effects and vice-versa.

There are fewer genes than expected (e.g. too few for even blue-print of just the brain). Normal development requires environmental input.

The process of gene expression is proving to be much more complicated than expected.

What is the effect of a gene?

• Depends on who you are (other genes and proximal social factors)

• Depends on where you are (social context)

• Depends on your developmental stage

• Depends on what has happened to you

Gene-environment interaction

Environmentally contingent expression

Dunedin Cohort Study (Caspi, Moffitt et al)

MAOA, maltreatment and antisocial behaviour

ANTISOCIAL BEHAVIOR AS A FUNCTION OF MAOA ACTIVITY AND A CHILDHOOD HISTORY

OF MALTREATMENT (from Caspi et al., 2002)

-0.5

-0.25

0

0.25

0.5

0.75

1

1.25

None Probable Severe

Low MAOAactivity, n =163

High MAOAactivity, n =279

Childhood maltreatment

Composite index of antisocial behavior

(z scores)

Serotonin transporter gene, life events and depression

.00

.10

.20

.30

.40

.50

0 1 2 3 4 +

Pro

bab

ilit

y o

f m

ajo

r d

epre

ssio

n e

pis

od

e

Number of stressful life events

EFFECT OF LIFE STRESS ON DEPRESSION MODERATED BY 5-HTT

GENE (from Caspi et al., 2003)

s/s

s/l

l/l

s/s = short allele homozygousl/l = long allele homozygouss/l = heterozygous

COMT, cannabis and schizophrenia

0

5

10

15

20

No adolescent use Adolescent use

SCHIZOPHRENIA SPECTRUM DISORDER:CANNABIS USE INTERACTS WITH

GENOTYPE (Caspi et al., 2005)

%sc

hizo

phre

nifo

rm d

isor

der Met/Met

Met/Val

Val/Val

COMT genotype

Learning provides a mechanism for persisting effects

Biological systems can learn although this is often referred to as “programming”

Example Systems

Immunological response is largely programmed through exposure in early life

High risk genotypes (over responding) may benefit from early social exposure to risk agent

Neglect can be protective for asthmaDay-care can be protective for juvenile onset leukemia

What are the kinds of detailed mechanics that could give rise to these effects?

A proof of concept

Stress, Cortisol and the HPA system

Genotypes can confer advantage or disadvantage depending upon the

environment. How could disadvantage be moderated?

• Anxiety is disadvantageous when the environment is benign

Q: How does a mother rat “teach” a pup with a gene that predisposes for anxiety to be less anxious?

A: She stops the gene from being expressed.

How does she do that?By licking the pup a lot!

Licking and stress response• The pups with the gene predsiposing for a

more anxious response but who are licked more are less fearful, they produce less stress hormones when provoked and their heart rate doesn't go up as much.

• The licking had to have occurred in the first six days of life.

This is now known to be an example of epigenetics

Epigenetics

DNA does not just hang about in the cell with each bit ready for haphazard transcription and protein production.

The DNA strands are organised

Histones and methylation

Maternal behaviour, methylation and pup behaviour

• Some rat pups inherit a gene variant for cortisol that can make them anxious. At birth it is unmethylated.

• If the rat pup is licked a lot in the first six days, a single methyl group (CH3) becomes attached at one specific location on the DNA in brain cells.

• This makes it impossible for the gene to be transcribed. The gene is silenced by the mother rats behaviour and the rat pup is placid.

• Remove the methyl group (can be done chemically) and the rat pup reverts to being anxious, consistent with its genotype.

Many concepts, skills and fields of research

• Careful description of social behaviour and social development

• Inter-generational cultural transmission (in fact occurs beyond parent child to grand-parent child because some aspects of methylation can be heritable)

• Developmental biology• Detailed organ specific gene-expression, cell

biology, experimental design (cross-fostering), pathology etc

Genetics and inference for causal environmental effects

Discordant twins

Mendelian randomisation

MZ (Identical) co-twin controls

• In longitudinal studies we might control for omitted variables by examining how change in outcome (difference between time 1 and time 2) is associated with change in environmental exposure

• MZ differences in outcome can be similarly compared with differences in exposure

Path DiagramInstrumental Variable Model

V1 V2 V3

E2 E3

Note: Errors E2 and E3 are correlated (unmeasured confounder) V1 has no direct effect on V3. This is the exclusion restriction

Finding plausible IVs is the problem

Can we find genes that might meet the exclusion restriction

• Does alcohol causally increase risk of depression or do depressed people self-medicate with alcohol?

• There is a genotype that gives deficient alcholol dehydrogenase. Individuals with this genotype (10% UK, 50% in Japan) enjoy alcohol less and drink less.

• Using this genotype as an instrument (Mendelian randomisation) suggests alcohol increases risk of depression.

Conclusions• For social and many health outcomes effects

of genes are socially mediated and moderated• Time-restricted social inputs can have time-

persistent effects through programming. Epigenetics provides one such mechanism.

• Genetic designs provide another tool for unravelling social processes

• We need social science informed genetics and genetically informed social science.