Comments on News and Primary Papers
Comment by: Ben Pickard
Submitted 23 May 2012
Posted 24 May 2012
The paper by Talkowski and colleagues describes the application of cutting edge genomics techniques to the molecular characterisation of multiple balanced chromosomal abnormalities (BCAs) linked to autism, autism spectrum disorders, and general neurodevelopmental disorders. In a single publication it has probably assigned more candidate genes than the entire conventional cytogenetic output from schizophrenia and autism in the preceding 15 years.
The authors carry out a great deal of complementary genomic analyses which add to the strength of their argument that these genes are indeed causally involved in illness. Without these additional data there would be one potential criticism of the paper in that the same power of analysis was not applied to BCAs in healthy controls. This is an important ascertainment issue because previous studies have not only identified disrupted genes in the healthy population (Baptista et al., 2005) but also shown that CNVs deregulating specific genes may only show an increased—as opposed to exclusive—representation in the ill population.
The observed overlaps between some of the identified BCA genes in ASD/neurodevelopmental disorders and those identified in GWAS studies of schizophrenia and bipolar disorder is fascinating but may be a double-edged sword. On the one hand, support for rare genetic contributors (CNVs/sequence variants/BCAs) to complex genetic disorders has often been drawn from those that are co-incident between studies. In that respect, this study is remarkable for highlighting the same genes from methods that detect very different mutation types. I’m genuinely surprised that there appears to be a convergence of ancient (read subject to evolutionary selection/population effects) and recent (meaning random) mutations. On the other hand, there is the disconcerting possibility that schizophrenia GWAS are only powered to detect the causes of blunt neurodevelopmental disturbances (which are perhaps less sensitive to issues of diagnostic categorisation) and not the fine-grained genetic hits that determine a precise clinical endpoint. If this is the case then we could end up with a situation where the genotypic distance between disorders is apparently much less than the phenotypic distance. This is most likely an extreme outcome that will be remedied once the genomic analysis of complex genetic disorders is able to factor in the composite effects of BCAs, CNVs, rare SNPs, and common SNPs—at the level of the single individual, and perhaps conditioned on the presence of big neurodevelopmental hits.
Quite logically, the presence of genes spanning diagnoses has been explained in terms of shared predisposition derived from early neurodevelopmental insults that are subsequently pushed down diagnostic pathways by other genetic or environmental factors. However, this assumption needs formal testing. The problem is reminiscent of the debate that circled the early use of constitutive mouse knockouts: how is it possible to disentangle developmental from adult functional phenotypes in a null? The advent of inducible Cre-LoxP technologies allowed that question to be directly addressed and may be the means to test the neurodevelopmental contribution of diagnosis-spanning candidate genes such as TCF4.
Could the approach detailed in this paper be applied directly to schizophrenia? It would certainly add substantially to the ‘confirmed’ gene list and would detect any reciprocal relationships with ASD/neurodevelopmental disorders. One issue is that ASD appears to have a higher overall incidence of chromosomal and genomic structural rearrangements than schizophrenia, but perhaps the greater question is availability of an appropriate sample set. The concerted cytogenetic screening that took place in Scotland coupled with an ability to cross-reference these findings with incidence of psychiatric disorder was instrumental in the discovery of DISC1 and other genes in Scotland (Muir et al., 2008)
but this resource is now largely exhausted of relevant BCAs. To my knowledge, the Danish registry represents the best bet for such an approach to succeed for schizophrenia (Bache et al., 2006).
Baptista J, Prigmore E, Gribble SM, Jacobs PA, Carter NP, Crolla JA. Molecular cytogenetic analyses of breakpoints in apparently balanced reciprocal translocations carried by phenotypically normal individuals. Eur J Hum Genet. 2005 Nov;13(11):1205-12. Abstract
Muir WJ, Pickard BS, Blackwood DH. Disrupted-in-Schizophrenia-1. Curr Psychiatry Rep. 2008 Apr;10(2):140-7. Abstract
Bache I, Hjorth M, Bugge M, Holstebroe S, Hilden J, Schmidt L, Brondum-Nielsen K, Bruun-Petersen G, Jensen PK, Lundsteen C, Niebuhr E, Rasmussen K, Tommerup N. Systematic re-examination of carriers of balanced reciprocal translocations: a strategy to search for candidate regions for common and complex diseases. Eur J Hum Genet. 2006 Apr;14(4):410-7. Abstract
View all comments by Ben PickardComment by: Patrick Sullivan, SRF Advisor, Jin Szatkiewicz
Submitted 29 May 2012
Posted 29 May 2012
I recommend the Primary Papers
In this exceptional paper, the authors combined new technology with old-school genomics to deliver convergent data about the genomic regions that predispose to neuropsychiatric disorders. The first goal of psychiatric genetics is to identify the “parts list,” an enumeration of the genes and genetic loci whose alteration clearly and unequivocally alters risk. The results of this intriguing paper connect rare and powerful genomic disruptions with loci identified via common variant genomewide association screens.
A classical approach in human genetics is to study affected individuals with balanced translocations. Using next-generation sequencing, these authors identified the precise locations of 38 rare balanced chromosomal abnormalities in subjects with neurodevelopmental disorders. They identified 33 disrupted genes, of which 22 were novel risk loci for autism and neurodevelopmental disorders. The other disrupted genes included many that had previously been identified by genomic searches for rare variation and common variation (e.g., AUTS2, CHD8, TCF4, and ZNF804A).
The authors then sought secondary genomic support for disease association with these 33 risk loci by analyzing a large collection of psychiatric GWAS data. They found an increased burden of copy number variants (CNVs) among cases as well as a significant enrichment of common risk alleles among both autism and schizophrenia cases. This research suggests that autism and neurodevelopmental disorders may have commonalities with psychiatric disorders such as schizophrenia at the molecular level, underscoring the complexity of genetic contribution to these conditions.
CNVs discovered from microarrays are mainly large, rare CNVs spanning multiple genes. Exome sequencing is limited to coding regions of the genome. In contrast, as illustrated in Talkowski et al. (2012), it is possible to identify individual lesions with nucleotide resolution in both coding and non-coding regions. Thus, this research suggests that sequencing individuals with pathogenic balanced translocations could provide a complementary strategy for mutation identification and gene discovery.
The experimental procedures were technically well done; all BCA breakpoints were confirmed by PCR and capillary sequencing. In seeking the secondary genomic support, the authors were keen on evaluating and eliminating the possibility for any confounding factors that may cause spurious association. For example, CNV burden analysis was conducted with respect to differential sensitivities from microarrays, and all results remained robust to various subset analyses and to one million simulations designed to establish empirical significance. To examine the potential for spurious enrichment of common risk alleles, the authors additionally conducted identical analysis in phenotype-permuted datasets from well-powered GWAS data for schizophrenia and autism as well as in well-powered GWAS data for eight unrelated traits, and therefore eliminated unforeseen confounders.
Impressively, many of the loci identified here now have convergent genomic results with support across multiple different samples and technical approaches. For example, TCF4 harbors common variation identified via GWAS, a Mendelian disorder, and now a gene disruption. These convergent genomic results markedly increase confidence that TCF4 is truly in the “parts list” for neurodevelopmental disorders. In contrast, there remain multiple questions about the genomic evidence for DISC1, where such convergence has not been achieved.
This paper also provides important results relevant to resolving the rare “versus” common variation debate. This appears to be a false dichotomy where, often, both rare and common variations contribute to the parts lists for these disorders.
View all comments by Patrick Sullivan
View all comments by Jin SzatkiewiczComment by: Bernard Crespi
Submitted 29 May 2012
Posted 29 May 2012
I recommend the Primary Papers
Balanced chromosomal abnormalities (BCAs) provide extremely useful alterations for linking of specific loci with psychiatric conditions, because they exert penetrant effects and localize to specific genes. The recent study by Talkowski et al. (2012) used direct sequencing of breakpoints, based on 38 subjects, to generate a set of genes with putative links to different neurodevelopmental disorders, broadly construed as including autism spectrum disorders, intellectual disability, and/or developmental and other delays.
One of the most striking results from their study was the presence, in their set of breakpoint-altered genes, of five genes that have been associated from other work with schizophrenia and related psychotic-affective spectrum disorders (such as bipolar disorder and major depression), including TCF4, ZNF804A, PDE10A, GRIN2B, and ANK3. These results suggest, according to the authors, the presence of shared genetic etiology for ASD, schizophrenia, and other neurodevelopmental disorders (mainly developmental delays). The authors also show overlap of their gene list with results from CNV and GWAS of autism and schizophrenia, further suggesting genetic links between these two conditions.
Do these results mean that autism and schizophrenia share genetic risk factors? Perhaps, but also perhaps not. Two important caveats apply.
First, schizophrenia involves well-documented premorbidity, in a substantial proportion of cases, that centers on developmental, social, and language deficits and delays (e.g., Saracco-Alvarez et al., 2009; Gibson et al., 2010). In children, premorbidity to schizophrenia most commonly involves "negative" symptoms, including deficits in social interaction (Remschmidt et al., 1994; Tandon et al., 2009), which can overlap with symptoms of autism spectrum disorders (Goldstein et al., 2002; Sheitman et al., 2004; Tjordman, 2008; King and Lord, 2011). Males are more severely affected, as in autism (Sobin et al., 2001; Rapoport et al., 2009; Tandon et al., 2009). Schizophrenia mediated by CNVs, or BCAs, is likely to exhibit relatively high levels of premorbidity, due to the penetrant, syndromic, and deleterious nature of these alterations (Bassett et al., 2010; Vassos et al., 2010). A recent study by Sahoo et al. (Sahoo et al., 2011) provides evidence consistent with such premorbidity, in that of over 38,000 individuals (predominantly children) referred for developmental delay, intellectual disability, autism spectrum disorders, or multiple congenital anomalies, 704 exhibited one of seven CNVs (del 1q21.1, dup 1q21.1, del 15q11.2, del 15q13.3, dup 16p11.2, dup 16p13.11, and del 22q11.2) that have been statistically associated with schizophrenia in studies of adults (Levinson et al., 2011).
These findings suggest that the subjects in Talkowski et al. (Talkowski et al., 2012), (most of them children, for individuals with age data given) who harbor alterations to schizophrenia-associated genes may, in fact, be severely premorbid for schizophrenia. Diagnoses of ASD (commonly PDD-NOS) in such individuals may represent either false positives (Eliez, 2007
; Feinstein and Singh, 2007), or true positives, with ASD as a developmental stage followed, in some individuals, by schizophrenia. This latter conceptualization considers autism as akin to childhood schizophrenia, a view which contrasts sharply with the classic criteria derived from Kanner (Kanner, 1943), Asperger (1991) and Rutter (Rutter, 2000, Rutter, 1972, Rutter, 1978), who consider autism as a lifelong condition present from early childhood. Of course, diagnosing premorbidity to schizophrenia as such is challenging, but if any data can help, it is data from highly penetrant alterations such as CNVs and BCAs, as well as from biological and neurological (rather than just behavioral) phenotypes.
Second, association to an overlapping set of genes need not make two disorders similar, or similar in their genetic etiology. For example, as noted by Talkowski et al. (Talkowski et al., 2012), variation in TCF4 has been associated with both Pitt-Hopkins syndrome and schizophrenia, but these conditions show essentially no overlap in phenotypes. Similar considerations apply to CACNA1C, linked to the autism-associated Timothy syndrome (via an apparent gain of function) as well as to schizophrenia and bipolar disorder. A key to sorting out the huge clinical and genetic heterogeneity in autism, and in schizophrenia, is subsetting of cases by similarity in alterations to pathways and phenotypes. Lumping of autism with schizophrenia, based on overlap in risk loci without consideration of the nature of the overlap, will make such subsetting all the more difficult.
Data on genes disrupted by balanced translocations are tremendously useful, but their usefulness will, as for other data such as CNVs, be circumscribed by diagnostic considerations, especially when the subjects are children. Bearing in mind the possibility that some childhood diagnoses may represent false positives, and that overlap in genes need not mean overlap in causation, should help in moving the study of both autism and schizophrenia forward.
Asperger H; translated and annotated by Frith U (1991) . Autistic psychopathy' in childhood. In Frith, U. Autism and Asperger syndrome. Cambridge University Press. pp. 37-92.
Bassett AS, Scherer SW, Brzustowicz LM. Copy number variations in schizophrenia: critical review and new perspectives on concepts of genetics and disease. Am J Psychiatry. 2010;167(8):899-914. Abstract
Eliez S. Autism in children with 22q11.2 deletion syndrome. J Am Acad Child Adolesc Psychiatry. 2007;46(4):433-4. Abstract
Feinstein C, Singh S. Social phenotypes in neurogenetic syndromes. Child Adolesc Psychiatr Clin N Am. 2007;16(3):631-47. Abstract
Gibson CM, Penn DL, Prinstein MJ, Perkins DO, Belger A. Social skill and social cognition in adolescents at genetic risk for psychosis. Schizophr Res. 2010;122(1-3):179-84. Abstract
Kanner L. 1968;2:217–250. Abstract
King BH, Lord C. Is schizophrenia on the autism spectrum? Brain Res. 2011;1380:34-41. Abstract
Levinson DF, Duan J, Oh S, Wang K, Sanders AR, Shi J, et al., Copy number variants in schizophrenia: confirmation of five previous findings and new evidence for 3q29 microdeletions and VIPR2 duplications. Am J Psychiatry. 2011;168(3):302-16. Abstract
Rapoport J, Chavez A, Greenstein D, Addington A, Gogtay N. Autism spectrum disorders and childhood-onset schizophrenia: clinical and biological contributions to a relation revisited. J Am Acad Child Adolesc Psychiatry. 2009;48(1):10-8. Abstract
Remschmidt HE, Schulz E, Martin M, Warnke A, Trott GE. Childhood-onset schizophrenia: history of the concept and recent studies. Schizophr Bull. 1994;20(4):727-45. Abstract
Rutter ML. Relationships between child and adult psychiatric disorders. Some research considerations. Acta Psychiatr Scand. 1972;48(1):3-21. Abstract
Rutter M. Diagnosis and definition of childhood autism. J Autism Child Schizophr. 1978;8(2):139-61. Abstract
Rutter M. Genetic studies of autism: from the 1970s into the millennium. J Abnorm Child Psychol. 2000;28(1):3-14. Abstract
Saracco-Alvarez R, Rodríguez-Verdugo S, García-Anaya M, Fresán A. Premorbid adjustment in schizophrenia and schizoaffective disorder. Psychiatry Res. 2009;165(3):234-40. Abstract
Sahoo T, Theisen A, Rosenfeld JA, Lamb AN, Ravnan JB, Schultz RA, et al., Copy number variants of schizophrenia susceptibility loci are associated with a spectrum of speech and developmental delays and behavior problems. Genet Med. 2011; 13(10):868-80. Abstract
Sheitman BB, Kraus JE, Bodfish JW, Carmel H. Are the negative symptoms of schizophrenia consistent with an autistic spectrum illness? Schizophr Res. 2004;69(1):119-20. Abstract
Sobin C, Blundell ML, Conry A, Weiller F, Gavigan C, Haiman C, et al., Early, non-psychotic deviant behavior in schizophrenia: a possible endophenotypic marker for genetic studies. Psychiatry Res. 2001;101(2):101-13. Abstract
Talkowski ME, Rosenfeld JA, Blumenthal I, Pillalamarri V, Chiang C, Heilbut A, Ernst C, Hanscom C, Rossin E, Lindgren AM, Pereira S, Ruderfer D, Kirby A, Ripke S, Harris DJ, Lee JH, Ha K, Kim HG, Solomon BD, Gropman AL, Lucente D, Sims K, Ohsumi TK, Borowsky ML, Loranger S, Quade B, Lage K, Miles J, Wu BL, Shen Y, Neale B, Shaffer LG, Daly MJ, Morton CC, Gusella JF. Sequencing Chromosomal Abnormalities Reveals Neurodevelopmental Loci that Confer Risk across Diagnostic Boundaries. Cell. 2012;149(3):525-37. Abstract
Tandon R, Nasrallah HA, Keshavan MS. Schizophrenia, "just the facts" 4. Clinical features and conceptualization. Schizophr Res. 2009;110(1-3):1-23. Abstract
Tjordman S Reunifying autism and early-onset schizophrenia in terms of social communication disorders. Behav Brain Sci. 2008;31(3):278-9.
Vassos E, Collier DA, Holden S, Patch C, Rujescu D, St Clair D, et al., Penetrance for copy number variants associated with schizophrenia. Hum Mol Genet. 2010;19(17):3477-81. Abstract
View all comments by Bernard Crespi