What can we learn from the high rates of schizophrenia in people with 22q11.2 deletion syndrome?

22q11.2 deletion syndrome (22q11.2DS) is a copy number variant (CNV) syndrome, resulting from a 1.5‐3Mb deletion on the long arm of chromosome 22. It occurs in at least 1 in 4,000 live births, making it one of the most common genetic deletion syndromes1. The physical phenotype is highly variable, involving multiple organ systems. The common manifestations include conotruncal cardiac malformations, cleft palate, renal abnormalities and immune dysfunction. Cognitive impairment is also common, with IQ distribution shifted about 30 points to the left of the typically developing population2.

In the 1990s, it was observed that a high proportion of adults with 22q11.2DS had schizophrenia3. Current estimates for lifetime prevalence of schizophrenia are approximately 25% in 22q11.2DS compared to about 1% in the general population. Furthermore, several studies have reported that the prevalence of the 22q11.2 deletion is 10‐20 times higher in patients with schizophrenia than the general population4. 22q11.2DS is therefore one of the strongest known risk factors for schizophrenia. One argument sometimes heard is that, since many people with 22q11.2DS have low IQ, and intellectual disability is a risk factor for schizophrenia, the psychosis seen in deleted cases is a consequence of low IQ and different nosologically from cases of schizophrenia in the non‐deleted population. However, 22q11.2DS increases risk of psychopathology independently of IQ2, 3, and the other clinical features of schizophrenia in deleted individuals do not differ from those seen in the wider population5.

Since carriers of the deletion are often identified as children, there has been much interest in 22q11.2DS as a high‐risk population in which the clinical, cognitive and neurobiological antecedents of schizophrenia can be studied. Interestingly, recent evidence suggests that the reciprocal duplication of chromosome 22 is associated with a lower risk of schizophrenia4. This further highlights the importance of studying this genetic locus to better understand resilience as well as risk, and to identify potential novel therapeutic targets.

The discovery of high rates of schizophrenia prompted many studies of psychiatric disorders in 22q11.2DS. The emerging picture is one of extensive pleiotropy. In childhood, neurodevelopmental disorders, such as attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD), are common2. In adolescence and early adulthood, the prevalence of mood disorders and psychosis increases, while anxiety disorders are common across the lifespan6. However, there is considerable variability between individuals in both the nature and severity of their psychiatric symptoms.

Recently, genomic technology has made it possible to detect CNVs across the genome, revealing that 22q11.2 is one among a number of CNVs associated with increased schizophrenia risk. Individuals with schizophrenia have an increased genome‐wide burden of large (>100kB), rare (<1%) CNVs generally7. Furthermore, a number of specific loci in addition to 22q11.2 are associated with risk of psychiatric disorders. Eleven such loci have been robustly associated with schizophrenia, with variable frequency and penetrance4. As with 22q11.2DS, these CNVs confer risk to a range of psychiatric disorders8.

The mechanisms by which 22q11.2 deletion increases risk of psychosis and other psychiatric disorders are still not well understood. Individuals with the deletion have only one copy of the genes within the deleted region (hemizygosity). The most likely mechanism is therefore that reduced dosage of one or more genes in the region leads to the clinical manifestations (haploinsufficiency), though other possibilities such as the unmasking of a recessive allele or position effects are also possible. While there are many interesting candidate genes in the region – including PRODH, COMT, DGCR8, RTN4R and TBX1 – none of these has been convincingly implicated as the single factor underlying the increased risk of schizophrenia or indeed other psychiatric phenotypes. It therefore remains possible that the increased risk of psychiatric disorders is conferred by the impact of the deletion on more than one and possibly several genes. This would not be surprising, since no single‐gene form of schizophrenia has been convincingly demonstrated.

Another possibility is that sets of functionally related genes (sometimes known as pathways) are particularly impacted by 22q11.2DS and other pathogenic CNVs. A recent pathway analysis across different CNVs has pointed to the role of synaptic genes influencing the balance of cortical excitation and inhibition in psychiatric risk9. This is an attractive model, since alterations in excitatory‐inhibitory balance have been reported in schizophrenia, ADHD and ASD, and therefore may help to explain the overlapping symptoms experienced by patients with 22q11.2DS, though whether a gene or genes on 22q11.2 are implicated in these pathways remains unclear.

A second, related, question is what factor or factors influence the very different psychiatric and cognitive outcomes seen in 22q11.2DS. The majority of patients with 22q11.2DS (90%) have a 3Mb deletion, which involves around 60 genes, whilst approximately 10% have a nested 1.5Mb deletion involving about 30 genes. It could be that the size of the deletion influences outcome, although this seems unlikely, since no study has yet identified a significant difference in the psychiatric phenotype between those with a 1.5 or a 3Mb deletion, though this issue will not be definitely resolved until larger studies are completed.

Another possibility is that alleles on the intact chromosome are responsible for the pleiotropic outcomes. If this were the case, these would have to be very common to account for the high prevalence rates and there is no current evidence for such variants of large effect. One more possibility is that “second‐hit” CNVs or point mutations contribute to increased risk, but again these would need to be common to explain all the increased risk.

Genomic studies have indicated clearly the polygenic nature of psychiatric disorders. In schizophrenia, at least a third of the genetic variance is captured by the combined effects of many hundreds of common SNPs, and it is now possible to assay this using the polygenic score approach9. However, as yet, sample sizes have not been sufficiently large to test whether polygenic background can account for the differences in psychiatric outcomes in 22q11.2DS. The effects of environmental factors have not yet been the focus of much research in 22q11.2DS, but the study of the interaction between genes and environment in this at‐risk group will be crucial to our understanding of pleiotropy and also of environmental risk factors for schizophrenia more generally. A large international collaboration6 is currently underway to investigate these questions in a large sample of patients with 22q11.2DS.

There has been much recent research aimed at identifying those who are most likely to develop psychosis within the 22q11.2 population. Other psychiatric disorders, particularly anxiety disorders, have been found to be associated with the development of schizophrenia in 22q11.2DS and, although cognitive impairment per se does not mediate the association between 22q11.2DS and psychopathology, cognitive decline may be associated with the onset of psychosis10. Whether these represent independent markers of increased risk or prodromal symptoms is not yet clear. Much more longitudinal data are needed to explore predictors of psychosis, including clinical symptoms and intermediate phenotypes, for example neuroimaging abnormalities.

As one of the strongest known risk factors for schizophrenia, 22q11.2DS offers a relatively homogenous high‐risk population for exploring precursors and predictors in longitudinal study designs. Furthermore, given the recent advances in genome engineering, it is increasingly feasible to develop both cellular and animal models of the deletion and to relate findings in these to those from human studies11. The genetic and mechanistic overlap between neurodevelopmental disorders is a particularly interesting avenue to explore and is timely given recent drives to improve psychiatric classification systems (e.g., Research Domain Criteria12). The variability in the presentation of CNV syndromes suggests that defining phenotypes categorically may not be the most appropriate approach to capture the spectrum of symptoms experienced by patients, and a more dimensional system may be more appropriate.

Comparing data across risk CNVs will be an important area of research and may help to identify final common pathways to psychosis and related disorders. Of additional interest will be advancing understanding of not only how CNVs act to increase risk, but also how CNVs such as the 22q11.2 duplication may exert protective effects. This would have clear implications for developing novel treatment approaches.

Although animal models are likely to yield important mechanistic insights, they have their limitations. For example, while there is synteny between the deleted region in humans and mice, there are differences in genomic organization. Moreover, it is likely that human psychiatric phenotypes reflect additional genetic and environmental factors. Furthermore, there are difficulties in modelling complex psychiatric phenotypes in animals. Animal work will nevertheless be an important component of a multi‐level approach to the study of 22q11.2DS, which should integrate other approaches such as cellular models and human brain imaging.

In conclusion, 22q11.2DS and other pathogenic CNVs offer new approaches to studying schizophrenia risk and the relationships between neurodevelopmental disorders. Longitudinal studies of high‐risk populations will be crucial and these will inform, and be informed by, animal and cellular studies of CNVs.

Michael J. Owen, Joanne L. Doherty
MRC Centre for Neuropsychiatric Genetics and Genomics, Neuroscience and Mental Health Research Institute, Cardiff University, Cardiff, UK