Summary of the article Gene–Environment Interactions in Severe Mental Illness - Uher, 2014

What is the role of genes and environment in severe mental illness?

We include the most disabling psychiatric disorders that require inpatient treatment in severe mental illness (SMI). Examples studied in this article are schizophrenia, bipolar disorder, and severe depression. These disorders share a common etiology and therefore it might be beneficial to study these disorders together.

Both genetic disposition and environmental exposures can increase the risk of SMI. There is more contribution of genes to SMI than to common mental disorders. The heritability of schizophrenia and bipolar disorder is estimated between 70 and 80%. The heritability of depression depends on its severity: in the general population we find a low heritability of around 38% and in hospital-ascertained severe depression the heritability lies between 48 and 75%. Most genetic variants increase the risk of more than one mental illness.

There are multiple environmental factors contributing to SMI, such as in utero infections, lack of nutrients, maternal stress, perinatal complications, social disadvantage, growing up in an urban environment, ethnic minority status, childhood maltreatment, bullying, traumas and cannabis usage. Multiple factors contribute to each case of schizophrenia. Sometimes risk factors are correlated or act in synergy. However, it seems many cases of SMI can be prevented by modifying the environment.

How do genes and the environment interact?

When there is a gene-environment interactions, one or more genetic variants and one or more environmental factors causally contribute to the condition in the same individuals, in which genetic factors can influence the sensitivity to environmental exposures. This is different from correlations, in which genetic factors only influence the probability of environmental exposure. There is an interaction, when this term is significant in a multiple regression. However, statistical interactions are different from actual biological interactions. Therefore, multiple methods are necessary.

Some people are resilient against the environmental risk factors, and there are many individual differences as to the impact of the environmental exposure. This might, partially, be due to genetic factors. Many cases of SMI arise because of a synergy between genetic and environmental causes. Twin studies are done to examine the heritability of a disorder. The genetic component is computed by gene-environment interactions, in which the environmental factors are shared within a family. However this is a misattribution, because the strongest known environmental factors are shared within families, but twin studies estimate almost no or little contribution of shared environment. The heritability based on twin studies does probably depict a larger part of the variance than the genetic variants actually account for. These ‘heritability gaps’ suggest that gene-environment interactions might explain large part of the variance.

What is genetic contribution and what are environmental factors?

There seems to be a stronger link between low birth weight and low educational achievement for children of biological parents with schizophrenia than for children with no family history of SMI. Gene-environment interactions have an influence in the early processes on the neurodevelopmental pathway of SMI. Low birth weight, however, is not only a reflection of environmental factors during pregnancy, but also has a genetic contribution. Therefore the interpretation is more complex.

Another result found was that excellent parenting and clear communication can reduce the risk of schizophrenia, but only for offspring of biological mothers with schizophrenia. Furthermore, a serious infection during pregnancy increases the risk of psychosis in offspring with a family history of psychotic illnesses. Economic disadvantages during upbringing can increase the risk of psychosis in adoptees with a family history of SMI in biological relatives. Between family history of psychosis and childhood maltreatment a correlation was found. In a twin study of depression it was found that genetic disposition interacted with environmental triggers to result in depression.

These studies indicate that the pathogenic effect of many environmental factors depends on the genetic disposition to SMI. Adoption and twin studies have been used to separate the genetic and environmental effects. However, even in adoption studies there is still some sharing of environment in early life and in twin studies monozygotic twins may share more environment than dizygotic twins. Therefore the interpretation of gene-environment studies by proxy is limited. Familal relatedness is not the same as genetic contribution and environmental factors can interact with genetic variants.

Molecular genetic variants

Molecular genetic variants may help the development of novel indications for therapeutics. The first discovered gene-environment interaction for a psychotic disorder was that of the COMT gene. This gene codes for an enzyme that metabolises dopamine, which can result in the positive symptoms of psychosis. When valine is replaced by methionine at position 158, which is called a single nucleotide polymorphism (SNP), the enzyme is less efficient. A study showed that the risk of developing psychotic symptoms and disorders is higher in individuals with the efficient Val alleles when using cannabis in adolescence, while the less efficient Met allele offered protection.

This finding had a use impact on personalised prevention of psychosis. However, the finding has been difficult to replicate. It might have been a false-positive finding. Recent data, however, suggest the interaction really exist, but depends on the proportion of tetrahydrocannabinol and cannabidiol. The findings were replicated for adolescent cannabis use with high tetrahydrocannabinol to cannabidiol ratio. Another finding was the three way interaction between COMT genotype Val alleles, childhood maltreatment and adolescent cannabis usage in the etiology of psychotic experiences.

There also seems to be an interaction between de AKT1 gene and the use of cannabis in the development of psychosis, namely for the carriers of the C/C genotype on rs2494732. The finding has been replicated multiple times. Therefore it is highly advisable for individuals with this genotype not to heavily use cannabis. Another finding was that two polymorphisms in the GRIN2B gene interacted with maternal positivity for the herpes simplex virus-2. This finding still has to be replicated.

The FKBP5 gene sensitised individuals to develop post-traumatic stress disorder after being maltreated in childhood. It also increased the risk of psychotic symptoms. The polymorphism 5-HTTLPR of the SLC6A4 gene moderates the effects of childhood maltreatment on depression. The short alleles of 5-HTTLPR increase risk of developing persistent depressive disorder after childhood maltreatment. Another study found that the short 5-HTTLPR alleles and a history of childhood maltreatment was associated with cognitive impairment in people with psychotic disorders. This finding was only true for physical abuse and physical neglect, not for all childhood maltreatment forms together.

In the development of depression a functional polymorphism (Val Met) in the brain derived neurotrophic factor gene interacts with stressful life events and childhood maltreatment. The people with the Met allele have a higher chance of developing depression after exposure to adversity. In a student sample it was also seen that the people with BDNF Met allele and a history of childhood abuse were more inclined to developing psychotic-like experiences.

For bipolar disorders gene-environment interactions have not often been studied. There has been one study stating that the people with bipolar disorder and the BDNF Met allele were more likely to have depressive episodes after a stressful life event than the people with the Val allele.

How do you search for gene-environment interactions across the genome?

When studying genes that are expected to be related to a disorder, a lot of genes can be overseen, since most genetic associations are found in genes that weren’t suspected to be involved. Gene-environment interaction searches across the genome might detect more genetic associations.

However, there has so far been only one such systematic search for gene-environment interactions in SMI. In this research they looked for genetic variants that sensitised individuals to developing schizophrenia when they were exposed to cytomegalovirus in utero. They looked at 29,000 polymorphisms that were significantly related to cytomegalovirus infection. The rs7902091 in the CTNNA3 gene significantly interacted with maternal cytomegalovirus infection in causing schizophrenia, but not with the accepted genome-wide level of significance. So it is not yet sure that this finding will be replicated.

What should future research investigate?

More genome-wide studies should be done in the future. However we need large samples for this, because the statistical power is lower in these studies. There has been even less gene-environment research for bipolar disorder, than schizophrenia and depression.

Before we can use genome-wide gene-environmental studies some challenge have to be overcome. The environmental variables have to be accurately assessed with a high quality. The environmental variables to be assessed should also be carefully selected, because with each environmental variable added, the number of potential gene-environment tests increases by the number of genetic variables and the sample size has to increase as well.

To do studies with larger samples requires some coordination between genetic and epidemiological studies. It would be good to obtain the genetic and environment data from the same sample. That way, the total costs of the research won’t increase.