Email Icon Facebook icon Twitter Icon GooglePlus Icon Contact

User Top Menu

New Study Finds Schizophrenia Risk Gene TCF4 Regulates Neuronal Excitability

15 Mar 2016

March 16, 2016. Thanks to a novel technique that allows researchers to probe gene expression in a subpopulation of genetically altered cells, a new study provides insights into how a schizophrenia risk gene regulates the excitability of neurons in the rodent prefrontal cortex. The study of transcription factor 4 (TCF4), which has been repeatedly associated with schizophrenia, was led by Brady Maher of the Lieber Institute for Brain Development and was published online on March 10 in Neuron.

This study provides "intriguing data that the schizophrenia risk gene, Transcription Factor 4 (TCF4), normally functions to repress the expression of two ion channels: KCNQ1 and Nav1.8 (SCN10a) that are not naturally expressed in brain," wrote Amy Arnsten of Yale University, who was not involved in the study, in a comment to SRF.

TCF4 has been linked to several neurodevelopment disorders, including a type of autism called Pitt-Hopkins syndrome (PTHS), 18q syndrome, and schizophrenia. The genomic region containing TCF4 was one of the earliest identified in genomewide association studies (GWAS) as containing single nucleotide polymorphisms (SNPs) associated with schizophrenia. And unlike other previous GWAS hits, TCF4 SNPs were found to be strongly significant also in the most recent GWAS conducted by the Schizophrenia Working Group of the Psychiatric Genomics Consortium (see SRF related news report).

Specific TCF4 polymorphisms have been associated with differences in cognitive function and sensorimotor gating in people with schizophrenia and controls (see SRF related news report and Zhu et al., 2013).

TCF4 is a transcription factor that can either repress or activate the transcription of genes and is highly expressed in the human central nervous system during development. However, little is known about the genes regulated by TCF4 and how changes in expression of TCF4 could lead to its associated disorders.

In this study, Maher and colleagues explored these questions by using two rodent models of PTHS, which is caused by TCF4 mutations that lead to protein deficiency. For their first model, the researchers used in-utero electroporation (IUE) to insert shRNAs or CRISPR-Cas9 constructs into the rat prefrontal cortex to knock down expression of TCF4 just before neurogenesis. The researchers chose this time point after determining that TCF4 mRNA expression peaks in the late prenatal period in both rats and humans. The IUE technique targeted neurons in the 2/3 cortical layer.

When Maher and colleagues then used whole-cell electrophysiology to record action potentials in IUE-transfected neurons, they found that knocking down TCF4 protein severely decreased the frequency of action potentials. This result suggests that TCF4 regulates the intrinsic excitability of these neurons. The researchers determined that the decrease in firing was due to an increase in the after-hyperpolarization current, which makes it harder for neurons to repeatedly fire.

"If such changes occurred in neurons in humans with mutations to TCF4, it could have devastating effects on prefrontal cortical function," wrote Arnsten.

These findings suggested that knocking down TCF4 changed the expression of one or more ion channels. To determine which ion channels were affected, Maher and colleagues needed to look at the expression of ion channels in only the cells that had been electroporated with the TCF4 knockdown constructs. In order to do this, they developed a novel technique that they named iTRAP, where they used translating ribosome affinity purification (TRAP) in IUE-transfected neurons.

"It allowed us to purify RNA only from the cells that we transfected, thus allowing us to compare the RNA between TCF4 knockdown cells and control cells," Maher told SRF.

The team determined that the expression of two ion channels—KCNQ1 and SCN10a—was significantly upregulated in the cells where TCF4 had been knocked down. Follow-up experiments showed that TCF4 binds directly to the KCNQ1 and SCN10a genes, suggesting that TCF4 could be directly repressing the expression of these genes.

"Both of these ion channels, KCNQ1 and SCN10a, are primarily thought to be expressed in the peripheral nervous system," said Maher. "We think what's happening is that normally, TCF4 represses the expression of these ion channels in the central nervous system. And in people who are lacking TCF4, they're having ectopic expression of these peripheral ion channels in the central nervous system."

Next, the researchers looked at action potential frequency and SCN10a and KCNQ1 levels in a transgenic mouse model of PTHS, which has a truncated copy of TCF4 on one allele. Whole-cell recordings from neurons in the 2/3 layer of the prefrontal cortex also showed a decrease in action potential frequency. Intriguingly, there was increased expression of SCN10a RNA in the frontal cortex in these mice; however, levels of KCNQ1 were actually lower. These findings were mirrored in a pharmacological experiment that found that SCN10a antagonists could rescue action potential output, but KCNQ1 antagonists could not.

TCF4 and schizophrenia

The relevance of these findings for increasing our understanding of the biology of schizophrenia is unclear at this point—primarily because it is not known how, or even if, the schizophrenia-risk SNPs in TCF4 change the expression or the function of the protein.

"We've been trying to find an association—we're still working on it—between the schizophrenia GWAS positive SNPs and TCF4 expression. The TCF4 locus has been associated with risk for schizophrenia, and the overlapping symptomatology between autism and schizophrenia adds to the link, but we don't know the specific mechanism in the case of schizophrenia," said Maher. "The mechanism could be anything at this point. It could be too much TCF4, or too little, or involve a previously uncharacterized isoform or function of TCF4. We also don't know when during development it could matter." Maher says his team is exploring this question using RNA sequencing from postmortem human brains across development and in various clinical samples.

However, expression of KCNQ1 and/or SCN10a in the prefrontal cortex could hypothetically explain some of the cognitive deficits seen in schizophrenia. "[I]f KCNQ1 is expressed on dlPFC [dorsal lateral prefrontal cortex] neurons in subjects with TCF4 mutations, it could result in weaker dlPFC network connectivity, reduced firing, and impaired cognitive abilities," wrote Arnsten. "It is also interesting to speculate that increased expression of Nav1.8 [SCN10a] in brain could result in nonspecific increases in excitability under conditions of inflammation, as research on the schizophrenia prodrome indicates that inflammation heralds the descent into illness."

Arnsten also notes that "a larger picture is emerging whereby schizophrenia is increasingly associated with genetic insults to ion channels (e.g., Cav1.2, KCNH2, Nicotinic-_7, NMDAR, HCN), or to regulators of cAMP-PKA or calcium signaling that regulate the open state of those channels (e.g., mGluR3, DISC1 anchoring of PDE4A, VIPR2). These genetic alterations would change neuronal excitability, particularly under conditions such as stress, where calcium-cAMP signaling is increased and defects may be most evident. Such genetic alterations to neuronal excitability may be most devastating in the newly evolved layer III dlPFC circuits that must generate precisely patterned and precisely timed representations needed for healthy cognitive function."—Summer E. Allen.

Reference:

Rannals MD, et al. Psychiatric Risk Gene Transcription Factor 4 Regulates Intrinsic Excitability of Prefrontal Neurons via Repression of SCN10a and KCNQ1. Neuron. Published online 2016 March 10. Abstract

Comments

Submitted by Amy Arnsten on

Rannals et al. provide intriguing data that the schizophrenia risk gene, transcription factor 4 (TCF4), normally functions to repress the expression of two ion channels, KCNQ1 and Nav1.8 (SCN10a), that are not naturally expressed in brain. KCNQ1 is normally found in the heart and intestine, where it interacts with KCNE1 to form the slow, depolarization-activated potassium current, I<sub>Ks</sub>. This current is powerfully increased by cAMP-PKA (protein kinase A) phosphorylation, which increases chloride secretion in the intestine and shortens ventricular action potential duration in the heart (and where mutations lead to the fatal, long QT syndrome). Similarly, Nav1.8 is a voltage-gated sodium channel that is normally not found in brain but rather in dorsal root ganglion cells, where it mediates nociception. Similar to I<sub>Ks</sub>, the Nav1.8 current is increased by inflammatory factors through cAMP-PKA phosphorylation of Nav1.8, leading to hyperalgesia. Rannals et al. found that knockdown of TCF4 led to expression of KCNQ1 and Nav1.8 in neurons of the rodent medial prefrontal cortex (PFC). The expression of KCNQ1 and Nav1.8 in PFC neurons greatly reduced their firing rate by 1) KCNQ1 increasing the after-hyperpolarization of the action potential, and 2) Nav1.8 increasing the resting membrane potential. If such changes occurred in neurons in humans with mutations to TCF4, it could have devastating effects on prefrontal cortical function.

Schizophrenia is associated with marked reductions in the activity of the dorsolateral PFC (dlPFC) during working memory—deficits that correlate with thought disorder (Perlstein et al., 2001). Postmortem studies have shown marked dendritic spine loss (Glantz and Lewis, 2000) and metabolic hypoactivity (Arion et al., 2015) in deep layer III, the circuits that represent information in the absence of sensory stimulation, the foundation of abstract thought (Goldman-Rakic, 1995). Our research in monkeys has shown that deep layer III dlPFC circuits are uniquely regulated at the molecular level, whereby feedforward calcium-cAMP-PKA signaling in spines, for example, as occurs during stress exposure, opens nearby KCNQ and HCN channels to weaken connectivity (Arnsten, 2015). KCNQ2, KCNQ3, and KCNQ5 are normally localized on dendrites and dendritic spines in deep layer III of the primate dlPFC, where channel opening by cAMP-PKA signaling weakens dlPFC network connectivity and reduces neuronal firing during working memory. Thus, if KCNQ1 is expressed on dlPFC neurons in subjects with TCF4 mutations, it could result in weaker dlPFC network connectivity, reduced firing, and impaired cognitive abilities. It is also interesting to speculate that increased expression of Nav1.8 in brain could result in nonspecific increases in excitability under conditions of inflammation, as research on the schizophrenia prodrome indicates that inflammation heralds the descent into illness.

In general, a larger picture is emerging whereby schizophrenia is increasingly associated with genetic insults to ion channels (e.g., Cav1.2, KCNH2, Nicotinic-α7, NMDAR, HCN), or to regulators of cAMP-PKA or calcium signaling that regulate the open state of those channels (e.g., mGluR3, DISC1 anchoring of PDE4A, VIPR2) (reviewed in Arnsten et al., 2012). These genetic alterations would change neuronal excitability, particularly under conditions such as stress, where calcium-cAMP signaling is increased and defects may be most evident. Such genetic alterations to neuronal excitability may be most devastating in the newly evolved layer III dlPFC circuits that must generate precisely patterned and precisely timed representations needed for healthy cognitive function.

Submitted by Courtney Thaxton on

This comment was co-written by Courtney Thaxton and Benjamin Philpot.

This paper provides invaluable data uncovering potential gene regulation changes within neurons in the absence of transcription factor 4 (TCF4). To date, we know very little of the gene(s) regulated by TCF4 in the brain. It was only in 2007 that TCF4 was identified as the causative factor of Pitt-Hopkins syndrome, an autism-associated neurodevelopmental disorder characterized by severe intellectual disability, absent speech, epileptic seizures, and breathing abnormalities (Amiel et al., 2007; de Pontual et al., 2009; Peippo et al., 2006; Zweier et al., 2007).

Rannals and colleagues found that two ion channels, KCNQ1 and SCN10a, which are not normally expressed at high levels within most brain regions, are upregulated in the absence of TCF4. The three-pronged approach used by Rannals et al., including shRNA knockdown, CRISPR/Cas9 mediated knockdown, and the Tcf4tr/- mouse model, solidify SCN10a as a bona fide target for TCF4-mediated transcriptional regulation. It will be interesting to see through future experimentation whether the SCN10a protein is in fact upregulated, or mislocalized, in mPFC neurons or a subset of neurons throughout the brain. The identification of E-Box sites by ChIP analysis within both KCNQ1 and SCN10a suggests a direct regulation of these genes by TCF4, but future studies to verify that TCF4 can regulate these genes will be essential.

Interestingly, SCN10a has been shown to be upregulated in response to inflammation and nerve injury in dorsal root ganglia neurons (Belkouch et al., 2014; Qu et al. 2013). TCF4 has long been studied for its role in the immune response in dendritic cells (Zhuang et al., 1996; Bergqvist et al., 2000; Cisse et al., 2008; Ghosh et al., 2010). Thus, it is intriguing to ask whether the increase in SCN10a could be due, at least in part, to inflammation occurring in the absence of TCF4 in the brain. Overall, this paper has uncovered key elements of TCF4 function in neurons and has opened the door for future experimentation to parse the effects of TCF4 haplo-insufficiency in the pathophysiology of neurodevelopmental and neuropsychiatric disorders.

Submitted by Brady Maher on

Authors' Reply

We appreciate the positive comments by Amy Arnsten, Courtney Thaxton, and Ben Philpot about our recent article describing how the psychiatric risk gene TCF4 can regulate neuronal excitability. Both groups discuss an interesting connection between Scn10a and its sensitivity to inflammatory signals and the fact that TCF4 regulates B-lymphocyte and plasmacytoid dendritic cell development. Although our paper primarily focuses on TCF4 with regard to Pitt Hopkins syndrome, which is a rare autism spectrum disorder, it will be important to determine if defects in human TCF4 function actually contribute to risk for schizophrenia. It is tempting to speculate that TCF4's role in immune function is a possible link, given that much of the MHC region is so strongly associated with schizophrenia.

Single nucleotide polymorphisms (SNPs) within introns of TCF4 achieve GWAS-significant association with schizophrenia. Unfortunately, to date we don't know if these variants are regulating TCF4 expression and/or splicing, and if so, how this would lead to increased risk. As mentioned in this SRF news report about our article, SNPs near TCF4 have been associated with differences in cognitive function and sensorimotor gating, but these may have to do more with these particular phenotypes in general than their role in schizophrenia risk. The required next step in establishing a risk mechanism for TCF4 will be to determine if risk SNPs near TCF4 are expression quantitative trait loci (eQTLs) by correlating them to TCF4 expression. Michael O'Donovan's group initially performed this type of analysis in adult postmortem brain and observed no evidence for the schizophrenia risk SNP rs9960767 being a cis-eQTL for TCF4 (Williams et al., 2011). Although this study was negative, cis regulation by risk SNPs may still occur, perhaps only during a particular developmental stage and/or in a particular cell type. In addition, there is substantial complexity in the splicing of the 5' exons of TCF4 (Sepp et al., 2011), and the expression of some of these unique transcripts could be regulated by different local risk SNPs. Progress in answering these fundamental questions needs to be made before it is safe to classify TCF4 as a schizophrenia risk gene.