The Psychiatric Genomics Consortium’s most recent (wave 3) genome-wide association study found 256 loci significantly associated with schizophrenia. These findings derive from a comparison of 65,000 cases with 88,000 controls. As with the determination of height, each site individually makes only a small contribution – but, when taken together – they make schizophrenia a highly heritable disease, Robin Murray told the EPA meeting in a highly entertaining plenary address.

Prof Murray suggested that psychosis seems rather like blood pressure, with all of us somewhere on the continuum, and schizophrenia the equivalent of severe hypertension

Perhaps unsurprisingly, many of the risk loci relate to the dopamine D2 receptor or to glutamate and GABA – which themselves influence dopamine release. Other loci are on genes governing neurodevelopment, giving support to the idea of schizophrenia as a neurodevelopmental disorder, or the functioning of the immune system (including HLA), which links to evidence implicating response to infection and stress in the etiology of the disease.

Revealing the genetic architecture of psychosis

Putting together the information from single nucleotide polymorphisms (SNPs) enables a polygenic risk score to be calculated. People in the highest 10% are seven times more likely than those with the lowest scores to develop schizophrenia.

Importantly, the SNPs relating to schizophrenia overlap with those contributing to risk of major depression, post-traumatic stress disorder and – above all – bipolar disorder. One implication of this is that we may need to blur a little our distinctions between diagnoses. A wider implication of the multiplicity of mutations involved in determining risk is that psychosis is probably rather like blood pressure, with everyone on a continuum, and schizophrenia the equivalent of severe hypertension.

As well as the role of SNPs, an excess in certain copy number variants (CNVs) contributes to risk. CNVs affect far fewer patients but their role gives schizophrenia something in common with conditions such as learning disability and autism. So it seems that schizophrenia lies at the confluence of two different phenomena – adult psychiatric disorders on the one hand and neurodevelopmental impairments and the other.

SNPs related to schizophrenia risk also load for psychiatric illness more generally

Central role of dopamine

For twenty years, we have known that striatal dopamine (DA) synthesis is elevated in acute psychosis. It has recently become clear that this is also the case in the manic phase of bipolar disorder. So again there is commonality between schizophrenia and other disorders. And the reason we sometimes have difficulty distinguishing between schizophrenia and BPD is not because we are poor clinicians, said Professor Murray, but because there is overlap in genetic predisposition and a common tendency to synthesize too much dopamine.

No discussion of genetics should minimize the role of social and environmental factors, he emphasized. Childhood assault and neglect, being a migrant or from an ethnic minority, and stress in adulthood all contribute to risk. But striatal DA dysregulation again seems to be the final common pathway. That is also true with drug-induced psychosis.

No discussion of genetics should minimize the role of social and lifestyle factors

Professor Murray has made a major contribution to public health by amassing evidence of the role played by highly potent cannabis in the high levels of psychosis seen among urban populations such as south London. And this factor may explain the intriguing differences in psychosis incidence seen across Europe.

Why is psychosis incidence in London six times that in Palermo?

The recent EUGEI study gathered incidence data for the period 2011-2013 from 16 sites in five countries, each of which contributed information from one large urban center and at least one smaller city. Rates of psychosis were highest in London (61 cases per hundred thousand population per year), Amsterdam (47 cases) and Paris (45 cases). They were far lower in the southern European cities of Madrid (25), Barcelona (13) and Palermo (11).

Professor Murray speculated on possible causes of this remarkable difference in incidence. While not dismissing the potential protective role of a Mediterranean diet, sunlight (and hence vitamin D), relatively low migration, and absence of loneliness due to a strong role for the family, his belief is that the decisive factor is the widespread use of high potency cannabis (with a THC content greater than 10%) in northern European cities.

Compared with no use, Professor Murray suggested that such use of cannabis increases psychosis risk five-fold, and there is a strong correlation between lifetime use in different populations and psychosis incidence rates.