A new study sheds further light on the genetic basis of the group of psychiatric disorders known collectively as schizophrenia. Further, the study (actually a collection of four studies) takes a new approach that might prove generally useful in associating genetic variation with disease risk, even beyond psychiatry.

Schizophrenia

In popular culture the term “schizophrenic” is often used to mean split personality or multiple personality, but this has never been the actual definition of the term. I’m not sure what the origin of this misconception is. The word “schizophrenia” does mean “split mind” but refers to mental illness characterized by disordered or delusional thinking. The “split” is between reality and mental function.

For at least several decades it has been clear that schizophrenia is not one discrete disorder, but rather it is a set of similar disorders. Symptoms include hallucinations, delusions (persistent false beliefs that do not have a cultural cause), impaired reality testing, bizarre thoughts and behaviors, often but not always paranoid in nature, a disconnection between thoughts and emotions, and lack of motivation or activity.

Part of the challenge of studying schizophrenia is that it is a clinically defined set of disorders, meaning that the category is based upon the signs and symptoms displayed, not any knowledge about underlying cause or biology. The brain, as you might suspect, is an incredibly complex organ with many interacting parts, and so there is likely to be a complex relationship between the underlying mechanisms of schizophrenia and the clinical manifestations.

This caused decades of debate about how to classify the different schizophrenias, and how many underlying disorders do they actually represent. Some researchers came to suspect that the clinical classifications might be fundamentally flawed, holding back basic research into genetics or other underlying mechanisms.

There has been a great deal of research into the genetics of schizophrenia. Researchers tried to find specific genetic mutations that would predict the development of schizophrenia. A recent review of twin studies found:

They yielded probandwise concordance rates of 41-65% in monozygotic (MZ) pairs and 0-28% in dizygotic (DZ) pairs, and heritability estimates of approximately 80-85%.

This means identical twins have about a 50% concordance (both either have or do not have schizophrenia), while non-identical twins have much lower concordance. This means that genetics are very important to the risk of schizophrenia, but not everything. There must be environmental factors as well. Overall, genes contribute 80-85% to the risk of developing schizophrenia.

Attempts to identify specific genetic variants that predict schizophrenia have been mixed. Specific gene variants seem to correlate with schizophrenia in some studies, but not others, or only in a subset of schizophrenic patients.

Such results are partly what lead to the hypothesis that schizophrenia is a set of related but distinct disorders. Certain genetic variants are found in some schizophrenics but not others because they actually are different disorders. The challenge then became identifying clinical subtypes of schizophrenia that were likely to reflect underlying genetic contributions. If our clinical categories did not cleave in the same places as the underlying genetics, attempts to correlate the two would be forever frustrated.

A New Approach

This brings us to the newly published paper. Arnedo et al, researchers at the Washington University of St. Louis, did a series of studies in which they looked at single nucleotide polymorphisms (SNP) in 4,200 people with schizophrenia and 3,800 healthy controls. SNPs are essentially point mutations in a single gene – one altered nucleotide. Instead of looking at the relationship between individual SNPs and schizophrenia, they looked at clusters of SNPs. The idea is that a cluster of SNPs would reflect a set of interacting genes that control the development and organization of the brain.

They also looked at the specific symptoms of schizophrenia in their patient population and identified 8 clusters of symptoms. When they looked at the SNPs in these patients they found 42 clusters of SNPs that were associated with a 70% or greater risk of having schizophrenia. Some of the clusters had higher risks, including some with 100% correlation with schizophrenia. They report:

These disjoint genotypic networks were associated with distinct gene products and clinical syndromes (i.e., the schizophrenias) varying in symptoms and severity. Associations between genotypic networks and clinical syndromes were complex, showing multifinality and equifinality. The interactive networks explained the risk of schizophrenia more than the average effects of all SNPs (24%).

Equifinality means that multiple pathways can lead to the same outcome, or in this case, multiple gene clusters could correlate with the same clinical syndrome. Multifinality means that one cause can have multiple effects, or here that one gene cluster could correlate with multiple clinical syndromes. This is not surprising. We see this in may multigene disorders, even outside of psychiatry. Specific clinical manifestations often have a probabilistic relationship to markers of underlying cause.

Finally, the clusters of SNPs were more predictive than looking at cumulative predictive value of the individual SNPs. This highly suggests that it is the interaction among the various genes that is important, and not just the adding up of individual gene effects.

One of the authors, Robert Cloninger, is quoted as saying:

“What we’ve done here, after a decade of frustration in the field of psychiatric genetics, is identify the way genes interact with each other, how the ‘orchestra’ is either harmonious and leads to health, or disorganized in ways that lead to distinct classes of schizophrenia.”

Conclusion

This is an exciting study for two reasons. The first is that it sheds significant light on the genetics of schizophrenia. This may lead to a new consensus about the clinical subtypes of schizophrenia. It will be interesting to see if the 8 clusters of symptoms identified in this study become officially recognized as the clinical subtypes of schizophrenia.

Second, and I think more importantly, this study shows the power of this approach to correlating genetics and clinical syndromes – looking specifically at the interaction of clusters of genes, rather than the contribution of individual genes. Applying this technique to many psychiatric and even non-psychiatric conditions may prove highly fruitful.

What this reflects is that, not only is the brain extremely complex, but genetics is extremely complex as well. The genetics of brain development is therefore complex squared. We need to take an equally complex approach to asking questions about the relationship between specific gene variants and resulting neurological outcomes. This approach perhaps gets us to the next level, looking at gene clusters, although I would not be surprised if deeper levels of complexity await our discovery.

I would like to see the common public conception of genetics get away from the simplistic notion of “the gene for X,” or the false dichotomy between genes and environment. Genes are not blueprints. They do not work from the top down. Genes are complex dynamic instructions that interact with other genes and a variety of environmental factors. The field of epigenetics is essentially the study of those factors that affect genetic expression. But really genetics and epigenetics are part of the same big picture of how genes work.

The other take-home message is that many diseases and disorders are really categories of similar diseases and disorders. This is particularly true in psychiatry, although not unique to psychiatry. For all the above reasons we may never identify a pure schizophrenia subtype. Such categories will always be fuzzy around the edges and blur one into the other. That is simply the nature of a complex dynamic system.