|Home | About | Journals | Submit | Contact Us | Français|
Genome-wide association studies allied with the identification of rare copy number variants have provided important insights into the genetic risk factors for schizophrenia. Recently, a meta-analysis of several genome-wide association studies found, in addition to several other markers, a single nucleotide polymorphism in intron 4 of the TCF4 gene that was associated with schizophrenia. TCF4 encodes a basic helix-loop-helix transcription factor that interacts with other transcription factors to activate or repress gene expression. TCF4 mutations also cause Pitt-Hopkins Syndrome, an autosomal-dominant neurodevelopmental disorder associated with severe mental retardation. Variants in the TCF4 gene may therefore be associated with a range of neuropsychiatric phenotypes, including schizophrenia. Recessive forms of Pitt-Hopkins syndrome are caused by mutations in NRXN1 and CNTNAP2. Interestingly, NRXN1 deletions have been reported in schizophrenia, whereas CNTNAP2 variants are associated with several neuropsychiatric phenotypes. These data suggest that TCF4, NRXN1, and CNTNAP2 may participate in a biological pathway that is altered in patients with schizophrenia and other neuropsychiatric disorders.
Although a small number of genetic loci have now been strongly implicated as risk factors for schizophrenia, most of these have yet to yield novel insights into the biology of the disease. A possible exception is the implication of neurexin-1 (NRXN1) and associated proteins in disease pathogenesis through the identification of copy number variants (CNVs) associated with an approximately 5-fold increased risk of schizophrenia.1 NRXN1 deletions are rare even in cases (0.19%), and this raises the question of the importance of NRXN1 for the disorder as a whole or at least a sizable proportion of it. Genome-wide association (GWA) studies have also led to the discovery of several new risk alleles for schizophrenia.2–5 Unlike NRXN1 deletions, these are common in the population, but the relative risks conferred are substantially lower (<1.5). For example, a synthesis of several recent GWA studies of schizophrenia with follow-up in additional samples identified a single-nucleotide polymorphism on chromosome 18q21.2 within transcription factor 4 (TCF4) that was associated with an increased risk of schizophrenia at a level of support (P = 4.1 × 10−9) that surpasses a widely held benchmark for genome-wide significance.5 TCF4 is a basic helix-loop-helix (bHLH) transcription factor that regulates gene expression in the immune system and in the brain during development. Haploinsufficiency of TCF4 causes a dominant form of Pitt-Hopkins Syndrome (PTHS), a developmental disorder associated with severe mental retardation. Remarkably, autosomal recessive forms of PTHS can also be caused by deletions and missense mutations in NRXN1 and another gene previously implicated by CNV analysis in schizophrenia and other neuropsychiatric diseases, contactin-associated protein like-2 (CNTNAP2).6 These findings potentially point to a functional link between TCF4 and both NRXN1 and CNTNAP2 and suggest that these proteins play a role in the pathogenic mechanisms of general relevance to schizophrenia and related disorders.
First, a word of caution concerning nomenclature is required. TCF4 (Gene ID: 6925) and TCF7L2 (Gene ID: 6934) loci are frequently confused because they share the TCF4 alias. TCF4 is the official symbol for TCF4, the protein discussed in this review, but it is also a widely used alternative name for T-cell–specific TCF4 (TCF7L2). In the present review, we have taken considerable care to include for consideration only those data which apply to the former. TCF4 is a member of the bHLH transcription factor family homologous to the Drosophila protein daughterless. bHLH transcription factors can be divided into several phylogenetic groups based upon their sequence composition, expression pattern, and ability to interact with other bHLH proteins.7 Briefly, group A, which includes TCF4, and group B bHLH proteins bind core DNA sequences referred to as E (Ephrussi)-boxes defined loosely by the sequence CANNTG. Group C bHLH proteins are also known as bHLH-Per-Arnt-Sim (PAS) because in addition to the bHLH domain, they also contain a PAS domain. Group D HLH proteins lack a basic domain and are hence unable to bind DNA. Group E-proteins are related to Drosophila hairy and enhancer of split proteins and bind preferentially to N-box sequences (CACGCG or CACGAG). Finally, Group F is characterized by the presence of the Collier/Olfactory-1/Early B-Cell Factor domain that is involved both in dimerization and in DNA binding.7
TCF4 is 1 of 4 mammalian E-proteins, the others being E12, E47, and HeLa E-box-binding factor. Basic amino acids in the bHLH domain of this family of transcription factors bind directly to DNA, recognizing the E-box consensus sequence in the regulatory regions of many genes. Although TCF4 can form homodimers, in common with other bHLH proteins, it appears to function as a transcriptional activator or repressor only by forming heterodimers with other group A or B bHLH proteins including atonal homolog 1 (ATOH1) and achaete-scute complex homolog 1 (ASCL1) (figure 1).9–11 These interacting proteins are often expressed in a tissue- or cell-type–specific manner. By contrast, heterodimerization of E-proteins with group D HLH proteins abrogates their transcriptional activity by sequestering them into inactive complexes that cannot bind DNA.
The majority of functional studies on TCF4 concern its role in the immune system. Here TCF4 is required for the development of lymphoid progenitors to the B- and T-cell lineages and regulates plasmacytoid dendritic cell (PDC) differentiation.12–14 PDC cells secrete interferon in response to viral nucleic acids and form part of the innate immune response. Beyond the scope of this review, the role of TCF4 and associated transcription factors in the development of the immune system has been recently reviewed.15
During brain development, bHLH proteins modulate critical events in neuronal and glial progenitor cells, controlling the transition from proliferation to differentiation.16 Although TCF4 and the other E-proteins are highly expressed in neural progenitor cells, their role in brain development has not been studied in detail. Knockouts of the genes encoding each of the 4 E-proteins have been produced. In each case, Tcf4 included, mice that are homozygous null for any of the E-protein encoding genes die at birth, whereas heterozygous mice are viable.14 Clearly then, Tcf4 is required for postnatal survival. However, it appears to be dispensable, at least in mice for brain development because at the gross histological level brain morphology appears normal in Tcf4 null animals.9
Although little is known about the role of TCF4 in the brain, it is useful to consider the roles of the 2 proneural genes, ATOH1 (the mammalian orthologue of the Drosophila atonal gene, also known as MATH1) and ASCL1, that have been shown to form functional heterodimers with TCF4 in this review. The proneural genes, of which there are less than 10 in mammals, are key transcriptional regulators of neurogenesis that specify all the different neurons in the mammalian nervous system.16 During brain development in the mouse, Atoh1 is essential for the establishment of a neural progenitor population in the rhombic lip and external granule layer that gives rise to multiple hindbrain structures.17 Although Tcf4 is widely expressed, Atoh1 interacts with Tcf4 to form neurons specifically in the pontine nucleus, a region in the ventral pons that conveys information between the motor cortex and cerebellum.9 In spite of no obvious neurodevelopmental abnormality in Tcf4 null mice, Tcf4 is required for the differentiation of subsets of neurons in the developing brain. TCF4 also interacts with ASCL1 in SH-SY5Y neuroblastoma cells to form a heterodimer that drives transcription of E-box–containing reporter genes (figure 1).10 The formation of heterodimers between TCF4 and different proneural genes provides a mechanism to regulate gene expression in specific subsets of neuronal precursors. Temporal and spatial regulation of neurogenesis by TCF4 and other E-proteins can also be achieved by repressing the transcriptional activity in a dominant-negative manner.18 Again in a neuronal cell culture model, the transcriptional activity of TCF4 is attenuated through the formation of heterodimers with the class D bHLH protein ID2 (figure 1).19
Surprisingly, little is known about the genes regulated by TCF4 in the central nervous system. The rat tyrosine hydroxylase enhancer has been shown to have an E-box–binding site for Tcf4.20 That study also showed that TCF4 could act as a transactivator but only when co-expressed with the homeobox transcription factor CUX/CDP2 (CCAAT displacement protein-2). Interestingly, microarray detection of Atoh1-regulated genes in the developing cerebellum found that Atoh1 targets are E-box–regulated genes that cluster into a few functional categories that include transcriptional regulation, cell proliferation, and signal transduction.21 By comparison with Atoh1, the identification of genes regulated by TCF4 not only in the developing brain but also in adult brain is likely to be pivotal to understanding the role of TCF4 in schizophrenia.
Genetic studies of other common disorders such as type II diabetes have shown that genes carrying common risk alleles frequently contain rare, more highly penetrant variants that can be associated with more severe phenotypes.22 For example, loss of function mutations in WFS1 cause autosomal dominant Wolfram Syndrome whose symptoms include early-onset non-autoimmune diabetes, optic atrophy, and deafness.22 By contrast, common variants in WFS1 confer risk of type II diabetes.22,23
Heterozygous deletions of the TCF4 gene cause PTHS (OMIM: 610954); a neurodevelopmental disorder characterized by severe mental retardation, microcephaly, epilepsy, poor motor development, and breathing abnormalities.24 In these families, the disease is inherited in a dominant manner and is the result of haploinsufficiency of TCF4.25–27 In addition to chromosomal deletions, TCF4 nonsense and missense mutations also cause dominant forms of PTHS. In comparison to forms of PTHS caused by deletions or nonsense mutations, TCF4 missense mutations are associated with an increased incidence of seizures, suggesting subtle differences in the disease mechanisms by class of mutation.28 Importantly, most TCF4 missense mutations are located within the basic region of the bHLH domain and have been shown to abrogate transcriptional activity in cells co-expressing ASCL1.26,29 Heterodimerization of TCF4 with other bHLH transcription factors such as ATOH1 and ASCL1 may explain one of the cardinal features of PTHS; a respiratory abnormality associated with hyperventilation and apnea. Mice lacking either Atoh1 or Ascl1 die shortly after birth due to an apparent inability to initiate respiration.17,30 In these mice and patients with PTHS, the functional dependence of heterodimer formation between the products of these proneural genes and TCF4 may result in a shared deficit in formation or activity of subpopulations of neurons that control breathing.
In addition to TCF4, autosomal recessive forms of PTHS have recently been found to be caused by mutations in NRXN1 and CNTNAP2 (Caspr2).6 Neurexins are synaptic cell adhesion molecules found on axon terminals that together with their cognate neuroligins connect the pre- and postsynaptic membranes of synapses.31 Three neurexin genes exist in humans, each one encodes 2 major isoforms (in the case of NRXN1, NRXN1α, and NRXN1β) that are transcribed from different promoters. There is now strong evidence that deletions of NRXN1, or parts of that gene, increase risk of schizophrenia, implicating one or both of the major NRXN1 isoforms in the etiology of that disorder.1
The second autosomal recessive PTHS gene CNTNAP2 has been implicated in the genetic etiology of several diseases, again including schizophrenia (table 1). Truncating mutations in CNTNAP2 cause autosomal recessive cortical dysplasia-focal epilepsy (CDFE). CDFE is a rare congenital epilepsy syndrome with neuropsychiatric comorbidities that include mental retardation, autism, and attention-deficit hyperactivity disorder.35 By contrast, heterozygous genomic deletions of CNTNAP2 have been described in patients with mental retardation, epilepsy, and schizophrenia.33 The expression of CNTNAP2 is regulated in part by FOXP2; a forkhead transcription factor that has an important role in the neurobiology of speech and language acquisition.36 CNTNAP2 is also a member of the neurexin superfamily and is encoded by one of the largest genes in the human genome spanning in excess of 2 Mb. CNTNAP2 and its ligand contactin form a receptor signaling complex that mediates neuron-glial interactions and neuronal migration in the developing cortex and regulates the clustering of Kv1 channels at nodes of Ranvier.37 Histological examination of mice lacking CNTNAP2 revealed no gross abnormalities in the brain of homozygous animals contrasting with the defect in cortical lamination found in patients with CDFE.35,38
In a recent report, a small proportion of patients diagnosed with Angelman Syndrome, a neurodevelopmental disorder affecting the epigenetic regulation of the ubiquitin ligase UBE3A, were in fact found to have TCF4 mutations.39 The phenotypic similarities shared by these disorders include mental retardation and motor dysfunction. This study suggests that TCF4 mutations may be found in other mental retardation syndromes including Rett Syndrome that have a similar spectrum of phenotypes.
There is increasing evidence that schizophrenia results from a combination of rare mutations of relatively large effect and common variants that confer low risk. For example, common variants in TCF4 and rare CNVs in NRXN1 and, though the evidence is weaker, also CNTNAP2 are associated with schizophrenia. It can be argued that rare, highly penetrant variants are useful for defining biological pathways whose disruption can lead to schizophrenia, but their presence in only a limited number of cases means that further evidence is required to determine their relevance to schizophrenia more generally. The observation that rare mutations in TCF4, NRXN1, and CNTNAP2 can result in similar neurodevelopmental phenotypes suggests a functional link between the proteins they encode. The implication of a common variant in TCF4 in schizophrenia is therefore evident that the functions of NRXN1 and CNTNAP2 might also be of general importance in this disorder. In doing so, it demonstrates the utility of seeking both common and rare variants in complex genetic disorders.
The identification of biological pathways that may be altered in schizophrenia is a fundamental aim in deciphering the complex genetic factors that contribute to disease susceptibility. Several of the CNVs associated with schizophrenia and other neuropsychiatric disorders are found in genes that encode synaptic and neurodevelopmental genes.40 Moreover, some of the CNVs associated with schizophrenia are also found in patients with mental retardation, autism spectrum disorder, or bipolar disorder, suggesting that these may represent a continuum of overlapping phenotypes that range in severity of neurodevelopmental insult and age of onset.41,42 In common with many other GWA candidates in complex diseases, schizophrenia-associated mutations or alterations in gene expression have yet to be described for TCF4. Alterations in transcript levels, alternative splicing, or alternative promoters may generate subtle functional variants of TCF4 that could be altered in schizophrenia. While it is tempting to speculate that the expression of NRXN1 and CNTNAP2 may be regulated by TCF4, more work is required to further delineate the genetic role of TCF4 in schizophrenia.
Wellcome Trust (WT088866); Medical Research Council.