This interview took place on May 2, 2007, and was conducted by Norman Sussman, MD.


This interview is also available as an audio PsychCastTM at

Disclosure: Dr. Wray receives grant support from the Australian National Health and Medical Research Council, the National Institutes of Health, and Sequenom, Inc.


Dr. Wray is a statistical geneticist specializing in the genetics of complex diseases. She is senior research officer at Queensland Institute of Medical Research in Brisbane, Australia, where she leads the Anxiety and Depression Study of the Genetic Epidemiology Laboratory. The study represents a powerful design to identify a large cohort of individuals for genetic analysis for whom comorbidity between depression and anxiety subtypes is fully documented.


There are so many terminologies used in genetics that many may be unfamiliar with. Can you provide an overview of the jargon associated with genomes?

An individual’s genome is their complete deoxyribonucleic acid (DNA) sequence which is found in 22 pairs of chromosomes, plus the sex chromosomes. A high proportion of the DNA sequence is identical between people as well as between species. It is commonly noted that humans are related to apes in that they share a lot of the same DNA. However, there are sites throughout the genome called genetic polymorphisms which can vary. The variation at these sites can cause the differences between individuals. Similarly, they can cause similarities between relatives as a proportion of shared genetic material is passed from parent to offspring.

Different versions of polymorphisms are called alleles. A single-nucleotide polymorphism (SNP) is the simplest form of a genetic polymorphism, where a single nucleotide in the DNA strand can vary. Because each person has two sets of chromosomes, people carry two versions of each DNA strand. Thus, each SNP site has two versions of the nucleotides that may be either the same (homozygous) or different (heterozygous). The two copies that a person carries makes up their genotype. Within a population there are three possible types of genotypes at each SNP position that can exist. A gene is a chunk of DNA that codes for a particular protein. If an SNP within a section that codes for a gene is different, it might end up with placement of a different amino acid into that protein chain. However, the vast majority of these genetic polymorphisms do not alter the protein product, though it is believed that they are very likely to be involved in regulation of genes.

In a genetic association study, allele frequencies at a polymorphic site are examined in a set of cases and controls. If the allele frequency differs between the cases and controls, the polymorphism may be a causal risk factor for a disease. There are millions of genetic polymorphisms within the human genome, although some are very rare, and others almost always occur together. In fact, the majority of the variation in the human genome can be investigated by studying approximately 500,000 of these SNPs. The forefront of genetic research currently consists of genome-wide association studies where, in very large sets of cases and controls, half a million SNPs are genotyped. This should make groudbreaking progress in relation to complex diseases such as psychiatric disorders.

Semaphorins are molecular cues that have been implicated in the development of the nervous system and, in particular, in the guidance of axonal projections and neuronal migration. Semaphorins were only discovered approximately 13 years ago; thus, understanding their function is relatively new. The role of semaphorins and their receptors in the developing nervous system has been examined mostly in mice studies. As new axons made in the developing brain reach their target, they do not appear to be normal in mice who have defective genes-encoding semaphorins. These differences are very subtle, which makes them very interesting as potential candidate genes to study in relation to psychiatric disorders.1,2 Plexins are receptors for semaphorins. Interestingly, whereas the expression of semaphorins occur throughout the brain during development, in adults this expression is limited to the motor system and the olfactory-hippocampal pathway.


Are there certain areas of the brain where neurogenesis is more prevalent than others?

Current literature indicates that neurogenesis in adults is confined to the olfactory hippocampal pathway. There used to be a well-entrenched dogma that no new neurones were laid down post-puberty. Only recently was it recognized that neurogenesis is not restricted to the developing brain, and does occur in adults. Although the rate of neurogenesis is not high, it is significant when accumulated over time.


Are neurogenesis and neurodegeneration likely to be involved in the onset of mood disorders and anxiety?

At the moment there is a theory that something precipitates a reduction in neurogenesis in the adult brain which results in a depression that is alleviated when neurogenesis returns to normal levels. When this theory was first introduced the evidence simply came from the fact that brain volume in depressed patients is less common than in normal controls. This finding has been replicated in many studies. More recently, work with mouse and rat models show that the commonly prescribed mood-stabilizing drug lithium enhances hippocampal neurogenesis. Specifically, it was discovered that neurogenesis is a requirement for the behavioral responses in order for lithium to be effective. Further research suggests that all major pharmacologic treatments result in enhanced hippocampal neurogenesis.

The theory about adult neurogenesis complements other theories. Many factors associated with depression, such as changes in neurotransmitters, hormones, and physical exercise enhance neurogenesis, whereas factors such as age, stress, and other stimuli for the pituitary-adrenal axis reduce neurogenesis.Thus, the theory does not stand on its own. It is complementary. However, it is unclear whether changes in neurogenesis cause depression or whether another mechanism precipitates depression and, as a result, causes a reduction in neurogenesis.


What methodology was used in your study examining the possible association between genetic polymorphisms and anxiety and depression?

Our study3 was conducted after a study by Mah and colleagues4 in which 25,000 SNPs were genotyped in a set of schizophrenia cases and controls. This was the first study to use so many variants. From the discovery case-control set, they identified 62 SNPs that had a different allele frequency between the cases and controls. Certainly, when studying so many SNPs, some differences between cases and controls are expected to occur by chance. Thus, Mah and colleagues used these 62 SNPs and genotyped them on several additional case-control samples. One of the 62 SNPs—a SNP in the gene encoding plexin A2 (PLXNA2)—showed a difference in allele frequency between cases and controls across most of their replication samples. By replicating the result, it is quite unlikely that the difference in allele frequency was by chance.

One of the replication samples was genotyped in our lab at the Queensland Institute of Medical Research. We decided to study PLXNA2 in an anxiety-depression study sample. Although schizophrenia, anxiety, and depression are considered distinct disorders, there is a growing school of thought that there may be underlying risk variants in common. It did not seem too far fetched to think that something which was associated with schizophrenia might also be associated with anxiety and depression.

There is some background to the development of our study. In the early 1980s, Nick Martin, PhD, who heads the lab at the Queensland Institute of Medical Research, started to collect measurements on monozygotic and dizygotic twins as well as their siblings and parents. This classic twin family design can provide measurements to help understand the genetic basis of disease or, in fact, any phenotype. By comparing groups of identical and non-identical twins, we can tease apart the relative importance of common genetic factors versus common environmental factors. The collection of these records started way before the genomics revolution. The foresight of Martin has resulted in a collection of participants, which is a hugely valuable resource for the sorts of studies we can do today. Our laboratory has investigated genetic contribution to an amazing array of factors, including, for example, body mass index, sexual orientation, susceptibility to mosquito bites, and alcohol and nicotine addiction.

In a study conducted between 1980 and 1989, >18,000 people completed the Eysenck Personality Questionnaire (EPQ) to identify the personality trait of neuroticism.5 It has been frequently validated that people who score highly on the neuroticism scale are likely to have clinical disorders of anxiety and depression. The questions on the EPQ probe for anxious and depressive behaviors. We analyzed the results and identified sibling pairs who had extreme measures for neuroticism; siblings either both scored extremely high, both scored extremely low, or one scored extremely high and one scored extremely low. We invited these people to participate in another study on anxiety and depression.3 This kind of study design is called an extreme discordant and concordant design, and it is a very powerful way of trying to extract the maximum amount of information that was present in the original cohort of 18,000 people by investigating a much smaller subset. This large quantity of data and good consistency of quality is quite superior to what is usually available from clinical settings.

More than 3,000 participants in this anxiety depression study retook the EPQ and also completed the complete international diagnostic interview (CIDI) which was devised and validated by psychiatrists. It is possible to allocate Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition6 diagnoses of anxiety or depression based on the answers to this self-report questionnaire. High scorers on the EPQ were more often allocated diagnoses of anxiety and/or depression than low scorers. Six hundred and twenty four people met the criteria for diagnosis of anxiety or depression. Participants could be allocated more than one diagnosis and anxiety could be be broken down to the more specific diagnoses of phobias and obsessive compulsive disorder.

In the study sample, we genotyped variants of PLXNA2 and found an association with anxiety and depression. In our study samples we were able to probe to see exactly who was contributing to that association. We found they were people who showed anxiety either with or without depression, but not the people who had depression only. Our original result was really hypothesis testing, and this post-hoc analysis is really hypothesis generating. We hope others in the scientific community will replicate our results so that we can be truly confident.

The interesting thing was that the PLXNA2 gene lies within a homologous region, ie, the same region, from the mouse genome, which has been shown to be related to anxiety traits in mice. Large-scale mouse studies can produce powerful and much more certain results. Thus, the fact that we found the association with anxiety only in the CIDI study seemed to fit with the result from the mice study.


How do these results connect to the results of Caspi and Kendler?

Caspi and colleagues7 looked at the genetic polymorphism 5-HTTLPR which is in the serotonin transporter. (This study was later replicated by Kendler and colleagues.8) 5-HTTLPR is an insertion-deletion polymorphism, meaning  some people have a chunk of the DNA strand missing (deletion) and others have it present (insertion). The deletion form has been shown to reduce the transcription efficiency of the serotonin transporter gene. Caspi then examined stressful life events (SLE) and found that people with both SLE and the short form of 5-HTTLPR more likely to succumb to depression than others. This result for genotype-environment interaction is appealing because it fits with our knowledge about the relationship between stress and psychiatric disorders. A combination of the stress and genetic factors are more likely to predispose a person to depression.

Replication studies of the original Caspi study have had mixed results. However, these studies generally do not have the power to detect what they are trying to detect. The problem is that the necessary study samples are expensive and time-consuming to collect. Information on stressful life events, depression, or psychiatric disorders have to be collected from each participant. The, blood must be drawn in order to conduct a DNA study. Thus, I tend to look more toward animal models to really understand the interaction between genotype and environment.


Do your findings fit in with any studies on brain-derived neurotropic factor (BDNF)?

There are parallels between the studies. However, I do not think anyone has pieced them together. BDNF is a gene that is involved in neuronal survival and differentiations in synaptic plasticity. It also has a role in brain development and is expressed in the adult brain, particularly in the hippocampus, which is parallel to PLXNA2. There has been mixed success in the association studies regarding whether or not there are variants of the gene-encoding BDNF which directly have an effect on psychiatric disorders. However, I think the information coming from mouse studies is quite convincing that BDNF does play a role. In addition, BDNF levels are lower in patients with depression, and treatment with antidepressants increases BDNF levels in the adult hippocampus and increases adult neurogenesis. I would not be surprised if PLXNA2 and BDNF are related. However, studies are necessary to determine this.


If validated, what are the practical implications of your findings?

I think we are kidding ourselves if we think the brain is simple. There are many interacting factors. Part of the research will serve to help others to eventually understand more about the metabolic pathway. That is a long way from clinical input. However, it is interesting that our genetic research has found a small increase in risk for anxiety disorders. As we study more genes we are finding more variants, which each individually have a small effect on risk. I think we are coming to a point where we will realize that the people who are at higher risk of succumbing to a disease are those who actually harbor many of these risk variants. The sorts of studies which are currently underway will help us to produce interventions to help prevent disease from manifesting. Soon we will have a better understanding of the genetic architecture which underlies psychiatric disorders.

Not so long ago, psychiatric geneticists thought that for each psychiatric illness there was one or at most a few genetic variants. Now, the community is waking up to the fact that there are many variants. However, there is still debate in the literature. Two recent articles published in The British Journal of Psychiatry argued for very different spectrums of the genetic architecture.9,10 One argued for the fact that there would be very rare alleles or genetic loci underlying schizophrenia, so that families had private mutations that, if present, were very highly causal for the disease.9 Another article argued the complete opposite, that many genetic variants could be common, but individually they carried only a small increase in risk.10

The current era of genome-wide association studies, where very large, powerful studies look at many variants, will likely produce answers as to how many risk factors are involved. My personal thoughts are very much in line with the second article. Many simultaneous risk variants can cause the breakdown of metabolic pathways, thus causing disease to manifest itself. I think this a very exciting time for psychiatric genetics. PP



1. Morris D, Runker A, O’Tuathaigh CMP, et al. Animal knockout and human studies identify SEMA6A and PLXNA2 as schizophrenia candidate genes [abstract]. Am J Med Genet B Neuropsychiatr Genet. 2006;141B(7):737.
2. Suto F, Tsuboi M, Kamiya H, et al. Interactions between Plexin-A2, Plexin-A4, and semaphorin 6A control lamina-restricted projection of hippocampal mossy fibers. Neuron. 2007;53(4):535-547.
3. Wray NR, James MR, Mah SP, et al. Anxiety and comorbid measures associated with PLXNA2. Arch Gen Psychiatry. 2007;64(3):318-326.
4. Mah S, Nelson MR, Delisi LE, et al. Identification of the semaphorin receptor PLXNA2 as a candidate for susceptibility to schizophrenia. Mol Psychiatry. 2006;11(5):471-478.
5. Kirk KM, Birley AJ, Statham DJ, et al. Anxiety and depression in twin and sib pairs extremely discordant and concordant for neuroticism: prodromus to a linkage study. Twin Res. 2000;3(4):299-309.
6. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994.
7. Caspi A, Sugden K, Moffitt TE, et al. Influence of life stress on depression: moderation by a polymorphism in the 5HTT gene. Science. 2003;301(5631):386-389.
8. Kendler KS, Kuhn JW, Vittum J, Prescott CA, Riley B. The interaction of stressful life events and a serotonin transporter polymorphism in the prediction of episodes of major depression: a replication. Arch Gen Psychiatry. 2005;62(5):529-535.
9. McClellan JM, Susser E, King MC. Schizophrenia: a common disease caused by multiple rare alleles. Br J Psychiatry. 2007;190:194-199.
10. Craddock N, O’Donovan MC, Owen MJ. Phenotypic and genetic complexity of psychosis. Invited commentary on… schizophrenia: a common disease caused by multiple rare alleles. Br J Psychiatry. 2007;190:200-203. Erratum in: Br J Psychiatry. 2007;190:365.