Progress Educational Trust conference: What can we learn from twins?

Why do we think we can learn anything useful (other than about twins themselves) from twins? It might be thought that the most important thing about them is their 'twinliness', the one attribute denied to non-twins. Science thinks otherwise.

'Twin studies' have, for generations, been science's main weapon for attacking the 'nature-nurture' problem. Half a million have been published. To deny or even doubt their relevance and reliability is science's version of a sin against the Holy Ghost. Professor Marcus Pembrey did so, nevertheless, saying that in his view only a couple of twin studies had ever been of the slightest value. And remember that he was sitting next to one of the world's leading apologists for, and conductors of, twin studies, Professor Robert Plomin. While perfectly gentlemanly, the atmosphere was decidedly charged. Why did Marcus Pembrey say that? What was his evidence? In a word, 'epigenetics'.

Why twin studies in the first place, though? Because, it is argued, twins provide a naturally occurring perfect study design to tease out the relative effects of what on one hand, our genomes and, on the other, our surroundings (societies, families, schools, peer groups) contribute to our cognitive abilities and behaviour and our health, or lack of it. A simple model presumes the two contributions to be distinctive and, though interactive, discrete; the genome does not affect the environment nor does the environment influence the genome.

Pairs of twins share a combination of genes and environments to a greater degree than other pairs of children. 'Identical' twins share them extremely. Something that suggests they could be used as matched controls for any number of comparative studies. Not only do they have the 'same' genomes, but a significant part of their environment is an identical child, making identical demands on the available emotional and physical resources of mothers and others.

It would seem to follow that if 'identical' means what it seems and claims to mean it should be reflected in identical cognitive strategies and health histories as the twins grow. But it doesn't. Or at least it does not do it sufficiently to rule out some other influences. Third panel member, Dr Jordana Bell, painted a compelling picture of discordance between pairs of identical twins in the prevalence of several common diseases. The obvious conclusion is that the 'simple' model is wrong and that environment and genome do influence one another.

That genomes influence environments is uncontroversial, a point stressed by Robert Plomin. Human beings have consistently altered their surroundings to ones that suited them better than those they have been bequeathed. An individual genome can alter its immediate environment by creating its own differences or by exercising choice over what is already there. It becomes an active participant in its environment rather than merely responding passively to it, at least to some extent. Parents may choose a school for their child but it is hardly unknown for a child to persuade its parents to then change the school for another that it thinks it would prefer.

Genomes are not simply strings of genes. How and when genes are expressed are highly regulated. Robert Plomin referred to the idea of the Quantitative Trait Locus (QTL) around since the late 1980s. Traits that are not 'categorical' but quantitative (height is the traditional example but there is a huge number, including cognitive abilities) are influenced by several, perhaps many, genes simultaneously. QTLs are lengths of DNA, that, while not themselves encoding genes, regulate and co-ordinate transcription of groups of genes. It appears that the way QTLs work must include environmental as well as genetic influences.

The second option is this. Environments chemically alter the regulation of genomes by adding methyl groups both to selective sites on the DNA itself and on the 'histones' of the protein scaffolding around which the DNA is coiled in chromosomes.

These 'epigenetic' patterns play a part in the expression of genes including sometimes 'silencing' them, and therefore in the dynamics and fidelity of cellular differentiation. A huge part of their significance lies in the fact that the patterns are inheritable. In other words environmental influences in one generation can be passed to later ones. It is this fact above all, Marcus Pembrey argued, that underpinned his criticism of classical twin studies that are restricted to a single generation. That epigenetic patterns vary between identical twins as well as changing with their age is an attractive explanation for the observed departures from 100 percent concordance. If 'identical' twins are well-matched mutual controls, epigenetic variation provides a chance of identifying just which genes are implicated in particular diseases; though whether the epigenetic variation is cause or consequence still remains to be seen.

PET is grateful to the conference's sponsors Merck Serono, the London Women's Clinic, Ferring Pharmaceuticals and the Medical Research Council.

Please make a a donation to PET's 21st Anniversary Appeal, so that it can continue organising events and publishing BioNews throughout 2014 (and beyond) while keeping BioNews FREE for you to read.