More Than Genes IV: Epigenetics, the Womb, and Mental Illness

11/17/2011 09:02 am ET

It's natural that during the past two hundred years we've been involved in a roaring debate about whether mental illness is caused by inherited genes or by our various social environments. It's also a puzzle to an observer that the first nine months of life, from conception to birth, have been almost totally ignored in the debate--although not ignored by those who directly do research about fetal development. The common mantra for so long has been that it's either before you're born or after you're born--and if it's before you're born it must be genes.

After the previous essays in this series, the common mantra does sound completely silly. Any impact that affects the developing fetal brain has the potential to be involved in later mental illness--and that's not genes, and it's not social environments.

But we need to recognize that the term "mental illness" is about as ambiguous as you can get, a label that covers everything from mild mood problems to complete derangement in violent paranoid schizophrenia. My focus here is on the relation between fetal impacts and mental illness and not on debates about diagnostic criteria or the usefulness of any particular diagnosis.

Scientific and clinical data suggest there may be important connections between the fetal environment and later mental disorders. The pile of data, in fact, is large enough so that all we can do in this essay is sketch the current situation.

The most significant mental disorder, not in terms of prevalence but in terms of debilitation, is psychosis. In plain language, psychosis is madness, the working of the mind gone severely awry--and to a biologist, that means the working of the brain gone severely awry.

If psychosis is madness, then schizophrenia is the Queen of Madness, complex and always terrifying. With the flowering of medical psychiatry in the late 19th century, the character of the disorder was described and named, and the modern description is not much different than the description popular 100 years ago: behavior that includes delusions, hallucinations, disorganized thinking and speech--bizarre and inappropriate behavior. Its age of onset is usually in late adolescence or early adulthood, although precursor ("prodromal") signs of the disorder may begin to manifest themselves some years prior to the first psychotic episode. After that, if untreated, the illness in most cases produces a progressive deterioration in the individual's ability to function.

As one schizophrenic patient once told me, his condition was like a dark solitary confinement without walls. Things that made sense to him apparently made no sense to other people. He was always alone. That any schizophrenic patient is among three million other Americans who will be diagnosed with schizophrenia during their lives (and among 60 million worldwide)--an estimate representing a lifetime prevalence (probability of the diagnosis at any time during a lifespan) of about 1 percent of the population --would be cold comfort to him.

It's unfortunate that the overemphasis by some people on genetic determinism in psychosis has led to labeling various mental disorders as "genetic diseases" when there is no evidence that their cause is merely the inheritance of a genetic mutation. This is certainly true for schizophrenia. Many biologists are unhappy with such simplistic gene-mongering. So-called "genetic" psychiatric disorders usually have many apparent genetic involvements. In addition, the clinical outcome always depends on environmental circumstances that begin with conception. Thus the rationale for the emphasis on chromosome "markers" is evidently flawed. Chromosome markers for psychiatric disorders, including schizophrenia, routinely turn up on different chromosomes in different cases.

Neurodevelopmental models of schizophrenia have for the most part emphasized prenatal brain development, focusing on effects that may alter gene expression. For example, according to one idea, individuals with schizophrenia inherit genes that cause structural brain deviations that may be compounded by early environmental impacts. Much of the focus of this hypothesis has been on postnatal impacts, such as stress, that change brain chemistry, but the idea can also accommodate a focus on prenatal impacts.

Another idea is that the genes involved in schizophrenia influence synaptic plasticity and the development and stabilization of neuronal microcircuitry in the brain. According to this approach, influences on synaptic plasticity would involve specific synaptic transmitters such as dopamine and gamma-amino butyric acid (GABA).

Studies of twins clearly demonstrate that schizophrenia is a complex syndrome involving both genetic and environmental influences. But the genetics of schizophrenia is certainly not Mendelian genetics. in which discrete traits are always linked to discrete genes with links always obeying the various Mendelian laws of heredity.

The term "concordance" applied to twins refers to a condition common to both twins. The term "discordance" is the opposite--a condition in one twin but not in the other twin.

Twin research in schizophrenia presents us with concordance3 puzzles. For example, although the usual quoted figure for concordance of schizophrenia in monozygotic (identical) twins is approximately 50 percent, the classical modern study on the subject shows a variation of concordance of 31 to 58 percent depending on which Western country provides the statistics. The concordance rate in the United States, for example, is only 31 percent, while in the United Kingdom it's 58 percent. That's not a minor difference.

So the discordance for schizophrenia between some pairs of monozygotic twins (i.e., one twin has schizophrenia, the other doesn't) can be as high as 69 percent (the figure for the United States). I think this is a more important fact than any estimates of "heritability" (usually about 80 percent for schizophrenia) based on the customary questionable assumptions in the behavioral genetics of twin research. One such assumption is that there is complete linearity of the heritability equation--ignoring, especially, gene-environment interactions during gestation. In general, the so-called "heritability" of schizophrenia is as useless a calculation as the heritability of IQ. It's based on problematic statistical assumptions about correlations and the origins of variance. Current calculated heritability is a function of the populations studied, the postnatal environment, the phenotypic criteria used, the statistical model used, and many other factors. The heritability measure does not get us very far in understanding schizophrenia.

In contrast, the high discordances in some monozygotic twins, particularly for schizophrenia, are telling us something about origins--that something is happening during fetal development that can trump genome identity. This suggests that avoiding general consideration of the fetal environment in estimating the contribution of heredity to schizophrenia is a serious mistake; the discordance evidence is telling us that the fetal environment is important.

The best view we have from twin studies is that schizophrenia results from both genetic and shared-environment etiological influences, with "shared environment" meaning classical postnatal environmental factors plus prenatal environmental factors such as exposure to infectious agents, macro- or micronutrient dietary characteristics, and exposure to environmental toxic chemicals, teratogens, and other assaults to the prenatal environment. Which factors and assaults are important can differ from one individual to another for reasons still unknown. Schizophrenia is not a simple puzzle.

If discordances in monozygotic twins for psychosis imply that environmental differences are involved, the discordance may be caused by epigenetic differences. One idea is that epigenetic differences in MZ twins are stochastic, due to random effects altering gene expression. Another idea is that differing intra-uterine environments in MZ twins can lead to discordant consequences. The effects of differences in placentas of MZ twins illustrate how uterine environment can affect development in schizophrenia.

Given the complexity of the human brain and behavior, there's much value even in a suggestion about how things work and about the origins, prenatal and otherwise, of dysfunctional childhood or adult behavior. For that suggestion may ultimately lead toward the prevention of psychiatric disorders that--no matter how they're labeled--bring untold suffering to countless people.

In my next and last essay in this series, I'll discuss how culture can impact the fetal environment to shape later behavior. Previous articles in the series are available from my Huffington Post page at

[Parts of the above text are adapted from the new book More Than Genes: What Science Can Tell Us About Toxic Chemicals, Development, and the Risk to Our Children. Author: Dan Agin. Oxford University Press, 2009.]