Schizophrenia Research Forum - A Catalyst for Creative Thinking

Special K: Primate-specific Potassium Channel Variant Implicated in Schizophrenia

7 May 2009. Drawing on a plethora of techniques to build its case—genotyping and genomic meta-analysis, structural and functional MRI, analysis of postmortem brain tissue, and electrophysiological experiments—an international team of researchers reports that KCNH2, a gene that codes for a potassium channel best known for its role in heart function, is a candidate risk gene for schizophrenia. Moreover, the group proposes that a KCNH2 isoform specific to the primate brain is associated with both schizophrenia and with a “schizophrenia-like shift” in brain anatomy and in physiological and cognitive function in subjects unaffected by the disease.

The new research, led by Daniel Weinberger at the NIMH, Bethesda, Maryland, and published in the May 3 online edition of Nature Medicine, is the latest in a series of genetic studies by the group that follow a multi-faceted, hypothesis-driven strategy to identify risk alleles (see, e.g., SRF related news story; SRF news story). In this case, the team took its lead from an earlier study that reported differential expression of many genes in the prefrontal cortex (PFC) of patients with schizophrenia (Prabakaran et al., 2004). Choosing 10 of those genes for further exploration, the authors genotyped 170 families with offspring that had been diagnosed with schizophrenia and found significant association with a region of chromosome 7q36.1 near the NOS3 (nitric oxide synthase-3) gene. The most strongly associated SNP lay in the nearby gene KCNH2. Additional genotyping in several samples and a pooled meta-analysis revealed that four SNPs in a small (~3 kb) region of the second intron of KCNH2 were strongly associated with schizophrenia.

“Current” hypotheses

KCNH2, also known as hERG (human ether-à-go-go related gene), is usually associated with myocardial function, where KCNH2 mutations contribute to long QT syndrome. But the gene is also highly expressed in the brain, particularly in the PFC and hippocampus (Guasti et al., 2005), where the distinctive slow activation, fast inactivation, and slow and voltage-dependent deactivation of the potassium channel coded by the gene are thought to contribute to oscillations in cortical and hippocampal networks (see, e.g., Bazhenov et al., 2004).

Even though none of the KCNH2 SNPs identified in the new study group achieved a statistically significant genomewide association with schizophrenia, the researchers decided to look further for possible structural or functional brain phenotypes for the putative risk allele. The Weinberger group takes the position that “genes weakly associated with schizophrenia might show relatively robust effects on prefrontal cortex and hippocampal formation function in risk allele-carrying populations,” including healthy controls. Their related view that “the use of healthy controls for genetic association at the level of brain function avoids potential confounders related to chronic illness and medical treatment” (for a discussion, see Hariri and Weinberger, 2003) led the group to conduct structural and functional MRI comparisons of healthy non-carriers, heterozygous carriers, and homozygous carriers of KCNH2 risk alleles.

The researchers found that hippocampal volume decreased significantly in these healthy subjects with increasing “allelic load”; that is, homozygous carriers of KCNH2 risk alleles had significantly smaller hippocampi than heterozygous carriers or non-carriers. These structural correlations were reflected in cognitive and physiological fMRI measures: on tests of IQ and processing speed, carriers performed proportionally worse than non-carriers, but had significantly stronger BOLD (blood oxygenation level-dependent) signals in the hippocampus and prefrontal cortex during these tasks. The Weinberger team interprets these combined findings as an indication of overactive but far less efficient cognitive processing in carriers of KCNH2 risk alleles.

Finding a new isoform
To get at the possible mechanisms underlying these deficits, the team analyzed RNA in postmortem PFC samples from 10 individuals with schizophrenia. They identified a previously unknown KCNH2 isoform, KCNH2-3.1, in which the first 102 amino acids of the full-length transcript KCNH2-1A are replaced with six amino acids unique to KCNH2-3.1. Expression of the KCNH2-3.1 isoform was significantly greater than that of KCNH2-1A, and was also significantly associated with the same risk alleles tested in the group’s MRI experiments. However, expression of the isoform was most pronounced in those who had been diagnosed with schizophrenia.

Intriguingly, while expression of KCNH2-3.1 was roughly equal across brain regions, its expression was three orders of magnitude greater in brain tissue than heart tissue compared to the full-length KCNH2 transcript. Furthermore, KCNH2-3.1 homologs were undetectable in the mouse brains and other mammalian brains, but abundant in rhesus monkey brain. A comparison of human prenatal and adult brain tissue indicated that KCNH2-3.1 expression is significantly higher before birth, which suggests that it could influence the trajectory of neural development and susceptibility to schizophrenia.

To complete the hypothetical circle, the authors conducted electrophysiological experiments on cells, including cortical cells, transfected with KCNH2-1A and -3.1, and found that transfection with the shorter isoform abolished the signature “tail current” that characterizes the slow deactivation of normal KCNH2 channels.

Based on its “convergent experiments,” the Weinberger team argues that, while the sustained neuronal firing pattern associated with the KCNH2-3.1 isoform may have been a beneficial evolutionary adaptation for primate cognition, the overexpression of the isoform seen in its postmortem studies of brains from schizophrenic individuals may perhaps contribute to psychosis. In another evocative aside, they propose that the cardiac side effects of atypical antipsychotics that bind to KCNH2 might be minimized—and the antipsychotic therapeutic benefits increased—if new drugs could be targeted to tamp down the expression of the brain-enriched KCNH2-3.1 isoform.—Peter Farley.

Huffaker SJ, Chen J, Nicodemus KK, Sambataro F, Yang F, Mattay V, Lipska BK, Hyde TM, Song J, Rujescu D, Giegling I, Mayilyan K, Proust MJ, Soghoyan A, Caforio G, Callicott JH, Bertolino A, Meyer-Lindenberg A, Chang J, Ji Y, Egan MF, Goldberg TE, Kleinman JE, Lu B, Weinberger DR. A primate-specific, brain isoform of KCNH2 affects cortical physiology, cognition, neuronal repolarization and risk of schizophrenia. Nat Med. 2009 May 3. Abstract

Comments on News and Primary Papers
Comment by:  Paul Shepard
Submitted 18 May 2009
Posted 19 May 2009
  I recommend the Primary Papers

The manuscript by Huffaker et al. extends the growing number of cardiac potassium channels that have found their way into the brain and onto the list of putative therapeutic targets for the treatment of neurological and psychiatric disease. In an extensive series of experiments, these investigators demonstrate an association between single nucleotide polymorphisms in a gene encoding an inwardly rectifying potassium channel (KCNH2), the expression of a previously unknown isoform (KCNH2-3.1), and schizophrenia. Named for the dance exhibited by ether-intoxicated fruit fly mutants in which the gene family was first identified, ether-a-go-go related gene or ERG K+ channels contribute to the repolarization of cardiac action potentials and the propensity of antipsychotic drugs to prolong the QT interval, a direct result of their ability to attenuate this current in the heart. The unique gating properties of ERG K+ channels (for review, see Shepard et al., 2007) give rise to a strong resurgent current that can profoundly alter both intermediate and slow components of neuronal signaling. Thus, ERG currents have been shown to alter spike timing (e.g., latency to first spike in a stimulus-evoked train, spike frequency adaptation) in cerebellar Purkinje (Sacco et al., 2003), medial vestibular nucleus (Pessia et al., 2008), and cultured cortical neurons, while in dopamine cells, they appear to underlie a slow afterhyperpolarization envisioned to contribute to the termination of plateau oscillations and the obligatory pause in firing after a burst of spikes (Canavier et al., 2007; Nedergaard, 2004).

Identification of a primate-specific KCNH2-3.1 isoform in hippocampus and cortex whose expression in brain alters the function of the channel begs a number of questions that will undoubtedly be the focus of subsequent research. Foremost among these is whether the therapeutic effects of antipsychotic drugs derive in some measure from their ability to block ERG channels containing the KCNH2-3.1 protein. Although the truncated KCNH2-3.1 isoform is unique to primates, phenotypic changes associated with expression of the protein result from loss of the PAS domain, a region of the protein responsible for the resurgent nature of the outward current. In addition to increasing the rate of ERG channel deactivation, expression of the truncated isoform may reduce the number of functional channels brought to the surface as suggested by the reported reduction in ERG current density in rat cortical neurons transfected with human KCNH2-3.1. The functional consequences associated with the loss of the PAS domain in individual cells can be characterized using dynamic clamp—a technique in which a computer simulation is used to introduce an artificial membrane conductance into individual neurons. However, the effects of the mutation on channel trafficking and assessment of the myriad of conductance states likely to result from heterologous expression with other ERG channel subunits will require a transgenic model, which if history serves, the Weinberger group has already begun constructing.


Canavier CC, Oprisan SA, Callaway JC, Ji H, Shepard PD. Computational model predicts a role for ERG current in repolarizing plateau potentials in dopamine neurons: implications for modulation of neuronal activity. J Neurophysiol . 2007 Nov 1 ; 98(5):3006-22. Abstract

Nedergaard S. A Ca2+-independent slow afterhyperpolarization in substantia nigra compacta neurons. Neuroscience . 2004 Jan 1 ; 125(4):841-52. Abstract

Pessia M, Servettini I, Panichi R, Guasti L, Grassi S, Arcangeli A, Wanke E, Pettorossi VE. ERG voltage-gated K+ channels regulate excitability and discharge dynamics of the medial vestibular nucleus neurones. J Physiol . 2008 Oct 15 ; 586(Pt 20):4877-90. Abstract

Sacco T, Bruno A, Wanke E, Tempia F. Functional roles of an ERG current isolated in cerebellar Purkinje neurons. J Neurophysiol . 2003 Sep 1 ; 90(3):1817-28. Abstract

Shepard PD, Canavier CC, Levitan ES. Ether-a-go-go-related gene potassium channels: what's all the buzz about? Schizophr Bull . 2007 Nov 1 ; 33(6):1263-9. Abstract

View all comments by Paul ShepardComment by:  Szatmar Horvath
Submitted 11 May 2009
Posted 1 June 2009
  I recommend the Primary Papers

Comments on Related News

Related News: DARPP-32 Haplotype Affects Frontostriatal Cognition and Schizophrenia Risk

Comment by:  Jonathan Burns
Submitted 14 February 2007
Posted 14 February 2007

This study provides hard empirical evidence for the hypothesis that psychosis (and schizophrenia in particular) represents a costly "byproduct" of complex human (social) brain evolution. Interestingly, the activation paradigms in the fMRI study (N-back and emotional face-matching tasks) are both testing social cognition. And the demonstrated changes in frontostriatal connectivity support the hypothesis that schizophrenia is a disorder of evolved intrahemispheric circuits comprising the Social Brain in our species.

I would suggest that further candidates (conferring vulnerability to psychosis) should be sought from amongst those genes known to have played a significant role in human brain evolution.


Burns J. (2007) The Descent of Madness: Evolutionary Origins of Psychosis and the Social Brain. Routledge Press: Hove, Sussex.

Burns J. The social brain hypothesis of schizophrenia. World Psychiatry. 2006 Jun;5(2):77-81. Abstract

Burns JK. Psychosis: a costly by-product of social brain evolution in Homo sapiens. Prog Neuropsychopharmacol Biol Psychiatry. 2006 Jul;30(5):797-814. Epub 2006 Mar 3. Review. Abstract

Burns JK. An evolutionary theory of schizophrenia: cortical connectivity, metarepresentation, and the social brain. Behav Brain Sci. 2004 Dec;27(6):831-55; discussion 855-85. Review. Abstract

View all comments by Jonathan Burns

Related News: DARPP-32 Haplotype Affects Frontostriatal Cognition and Schizophrenia Risk

Comment by:  Daniel Durstewitz
Submitted 8 June 2007
Posted 8 June 2007
  I recommend the Primary Papers

The phosphoprotein DARPP-32 occupies a central position in the dopamine-regulated intracellular cascades of cortical and striatal neurons (Greengard et al., 1999). It is a point of convergence for multiple signaling pathways, is differentially affected by D1- vs. D2-class receptor activation, and mainly through inhibition of protein-phosphotase-1 mediates or contributes to a number of the dopaminergic effects on voltage- and ligand-gated ion channels. These, in turn, by regulating intracellular Ca2+ levels, themselves influence phosphorylation of DARPP-32 and thereby interact with dopamine-induced processes.

Given its central, vital role in dopamine-regulated signaling pathways, it is quite surprising that (to my knowledge) only a few studies exist on the implications of DARPP-32 variations for cognitive functions and brain activity. Therefore, this comprehensive series of studies by Meyer-Lindenberg et al. combining human genetics, structural and functional MRI, and behavioral testing represents an important milestone. Meyer-Lindenberg et al. identified different functionally relevant DARPP haplotypes, associated with differential DARPP mRNA activity in postmortem studies, and found that these were linked to significant differences on a number of cognitive tests probing “executive functions,” as well as to differences in putamen volume and activity, and structural and functional covariation between striatal and prefrontal cortical areas. Thereby, they paved the way for detailed investigations of the role of DARPP-32 in human cognition.

Since DARPP-32 is so intricately interwoven into so many intracellular and physiological feedback loops, as with dopamine itself (Durstewitz and Seamans, 2002), mechanistic accounts for the functional involvement of DARPP-32 variations in neural network dynamics may be hard to obtain. “Linear” causal thinking usually breaks down in such complex functional networks constituted of so many interacting positive and negative feedback loops on different time scales. Thus it may still be a while until we gain a deeper, biophysically based understanding of the neural processes that mediate the influence of DARPP variations on cognition, and integrative computational approaches may be required to help resolving these issues. Given the complexity of DARPP-regulated networks, I also would expect that fine-grained behavioral testing and analysis of error types of human subjects on different cognitive tasks may ultimately reveal quite subtle and differential effects of DARPP polymorphisms. Moreover, the effects on neural network dynamics may be such (e.g., changing the temporal organization of spiking patterns) that they may not always be detectable by current neuroimaging methods, meaning that while the most dramatic effects were found on activation and volume of striatum, where DARPP-32 is most abundantly expressed, a significant contribution of other brain areas in DARPP-associated cognitive differences may not be ruled out. Regardless of these difficulties in unraveling the underlying neural mechanisms, the work by Meyer-Lindenberg et al. allows us to tackle the question of how the balance in dopamine-regulated intracellular networks relates to cognition in humans, and points toward the neural structures and interactions most interesting to look at.

View all comments by Daniel Durstewitz

Related News: Genetics, Expression Profiling Support GABA Deficits in Schizophrenia

Comment by:  Karoly Mirnics, SRF Advisor
Submitted 26 June 2007
Posted 26 June 2007

The evidence is becoming overwhelming that the GABA system disturbances are a critical hallmark of schizophrenia. The data indicate that these processes are present across different brain regions, albeit with some notable differences in the exact genes affected. Synthesizing the observations from the recent scientific reports strongly suggest that the observed GABA system disturbances arise as a result of complex genetic-epigenetic-environmental-adaptational events. While we currently do not understand the nature of these interactions, it is clear that this will become a major focus of translational neuroscience over the next several years, including dissecting the pathophysiology of these events using in vitro and in vivo experimental models.

View all comments by Karoly Mirnics

Related News: Genetics, Expression Profiling Support GABA Deficits in Schizophrenia

Comment by:  Schahram Akbarian
Submitted 26 June 2007
Posted 26 June 2007
  I recommend the Primary Papers

The three papers discussed in the above News article are the most recent to imply dysregulation of the cortical GABAergic system in schizophrenia and related disease. Each paper adds a new twist to the story—molecular changes in the hippocampus of schizophrenia and bipolar subjects show striking differences dependent on layer and subregion (Benes et al), and in prefrontal cortex, there is mounting evidence that changes in the "GABA-transcriptome" affect certain subtypes of inhibitory interneurons (Hashimoto et al). The polymorphisms in the GAD1/GAD67 (GABA synthesis) gene which Straub el al. identified as genetic modifiers for cognitive performance and as schizophrenia risk factors will undoubtedly spur further interest in the field; it will be interesting to find out in future studies whether these genetic variants determine the longitudinal course/outcome of the disease, treatment response etc etc.

View all comments by Schahram Akbarian