Abstract
The most recent genome-wide association studies (GWAS) of schizophrenia (SCZ) identified hundreds of risk variants potentially implicated in the disease. Further, novel statistical methodology designed for polygenic architecture revealed more potential risk variants. This can provide a link between individual genetic factors and the mechanistic underpinnings of SCZ. Intriguingly, a large number of genes coding for ionotropic and metabotropic receptors for various neurotransmitters—glutamate, γ-aminobutyric acid (GABA), dopamine, serotonin, acetylcholine and opioids—and numerous ion channels were associated with SCZ. Here, we review these findings from the standpoint of classical neurobiological knowledge of neuronal synaptic transmission and regulation of electrical excitability. We show that a substantial proportion of the identified genes are involved in intracellular cascades known to integrate ‘slow’ (G-protein-coupled receptors) and ‘fast’ (ionotropic receptors) neurotransmission converging on the protein DARPP-32. Inspection of the Human Brain Transcriptome Project database confirms that that these genes are indeed expressed in the brain, with the expression profile following specific developmental trajectories, underscoring their relevance to brain organization and function. These findings extend the existing pathophysiology hypothesis by suggesting a unifying role of dysregulation in neuronal excitability and synaptic integration in SCZ. This emergent model supports the concept of SCZ as an ‘associative’ disorder—a breakdown in the communication across different slow and fast neurotransmitter systems through intracellular signaling pathways—and may unify a number of currently competing hypotheses of SCZ pathophysiology.
Similar content being viewed by others
Introduction
Schizophrenia (SCZ) is a complex human disease, with heritability (or disease risk due to genetic factors) estimates ranging from 0.6 to 0.8.1 Until recently, genome-wide association studies (GWAS) have identified only a small number of associated genes or loci accounting for a miniscule fraction of the heritability.2 The turning point was the establishment of the Psychiatric Genomic Consortium,3 which enabled the pooling of large numbers of independent studies, greatly increasing the power for identifying genes affecting disease risk, and confirming the polygenic nature of psychiatric disorders.2, 4, 5, 6 In addition to increasing the sample size, a key to improved yield from SCZ GWAS was refined statistical methods.7 Recent application of this approach in SCZ6 indicated an increase in power for gene discovery over standard methods,8 and extended the list of putative risk genes for SCZ.8 While these findings underscore the polygenic nature of SCZ, it is a challenge to provide a comprehensive explanation linking the gene associations with underlying pathophysiology.
The recently discovered loci include genes that may provide partial evidence for several previously proposed, competing hypotheses for the pathophysiological basis of SCZ. Historically, mechanistic investigations of SCZ have been driven by a number of seemingly mutually exclusive mechanistic hypotheses—dysregulation of dopaminergic,9 glutamatergic10 or GABAergic11 neurotransmission, disruption of calcium homeostasis or second messenger signaling,12, 13 and growth factor deficiency.14 The latest GWAS findings6 have not narrowed the set of possible mechanisms to a specific neurotransmission system or physiological process (e.g., synaptic release). Rather, they further revealed the wide spectrum of molecules involved in different aspects of brain structure and function including hits across neurotransmitter systems and intracellular machinery, seemingly supporting all the competing hypotheses at once.
These findings may indicate a number of different pathological mechanisms underlying the heterogeneous clinical manifestations indiscriminately diagnosed as SCZ. Alternatively, these seemingly independent hits may be tied together through a common, unifying neurobiological framework. Here, we examine this possibility from the standpoint of classical neurobiological knowledge. Specifically, we inspect the independently discovered genes within the identified loci and show that a substantial proportion of these genes are mapped onto a common network known to be involved in integration of signaling across different neurotransmitter systems. In this network, the signaling cascades triggered by activation of G-protein-coupled receptors (GPCRs) modulate ion channels and ionotropic receptors.
Approach
We examined the list of SCZ-associated single-nucleotide polymorphisms (SNPs) for the presence of genes related to neuronal electrical excitability and neurotransmission falling into the following categories or ‘gene sets’ (Table 1):
-
1
‘Ca2+ signaling toolkit’ gene set was derived from a published expert review unrelated to GWAS.15 This set included Ca2+ channels, pumps, exchangers and transporters located on the plasma membrane as well as internal membranes of Ca2+ stores (endoplasmatic reticulum and mitochondria), Ca2+-permeable receptors for neurotransmitters and an array of Ca2+-binding proteins.
-
2
‘Ion channels’ gene set included all plasma membrane Ca2+ channels from the ‘Ca2+ signaling toolkit' gene set and the following additional channels: voltage-gated Na+, voltage-gated K+, voltage-gated Cl−, Ca2+-dependent K+, Ca2+-dependent Cl−, inward-rectifying K+, K+ leak, cation leak and the hyperpolarization-activated current Ih. This set was compiled relying on gene naming conventions.
-
3
‘Ionotropic receptors for neurotransmitters’ gene set included AMPA, Kainate, Delta and N-methyl-D-aspartate (NMDA) receptors for glutamate, γ-aminobutyric acid A (GABAA) receptors, ionotropic serotonin (5HT) receptor 3A, nicotinic acetylcholine (ACh) receptors, and purinergic receptors P2X. This set was compiled relying on gene naming conventions.
-
4
‘Metabotropic receptors for neurotransmitters’ gene set included mGluR receptors for glutatame, GABAB receptors, dopamine receptors, 5HT receptors (excluding inotropic receptor 3A), muscarinic ACh receptors, noradrenalin (NA) receptors, opioid receptors, adenosine A1 and A2A receptors and purinergic P2Y receptors.
-
5
We derived two gene sets—‘DARPP-32’ and ‘Dopamine pathway’—from two review articles, unrelated to GWAS, focused on integration of fast and slow neurotransmission. The first article provides an overview of molecular players from the standpoint of the critical integrative role played by DARPP-32.16 The second is focused on the intracellular cascades downstream dopamine receptors and interaction with glutamate signaling.17
Finally, to confirm the relevance of the identified genes to brain function, we used data from the Human Brain Transcriptome Project18 to examine mRNA expression for each of the genes in Table 1.
Genes coding for a wide range of iono- and metabotropic receptors, neurotransmitters and ion channels
We sought to provide a biological interpretation for the most recently published SCZ GWAS results,6 reanalyzed using a covariate modulated mixture model (CM3) statistical method.8 While the original GWAS report identified 108 gene loci implicating 348 genes in SCZ, based on standard GWAS methods,6 the CM3 methodology yielded 414 independent gene loci and 1285 potential gene targets with a replication rate of 80%8 (Supplementary Table 1). The recent pathway analysis of the same GWAS sample19 as well as studies of de novo mutations20, 21, 22 have implicated dysregulation of synaptic function and the balance between neuronal excitation and inhibition. Therefore, we prioritized genes within a general theme of synaptic communication and electrical neuronal excitability. We hypothesized that the increase in gene discovery offered by the CM3 method would provide further support for selective involvement of glutamatergic and GABAergic neurotransmission and voltage-gated calcium channels (VGCCs)—the players that have been implicated by prior studies.19, 20, 21, 22
In contrast to this expectation, our analyses revealed the presence of genes coding for a wide range of iono- and metabotropic receptors for neurotransmitters as well as various ion channels (Table 1). Given that SCZ is characterized by disabling and often chronic behavioral and cognitive abnormalities,23 the discovery of genes within a general theme of synaptic communication and electrical neuronal excitability is not unexpected (for a discussion of this topic, see McCarroll and Hyman24). More surprising, perhaps, is the presence of hits across many neurotransmitter systems—glutamatergic, dopaminergic, serotonergic, GABAergic, cholinergic and opioid—and ion transport machinery (channels, transporters and pumps) for all four major ions determining the membrane potential: Na+, K+, Ca2+ and Cl−.
This begs the question of whether these hits reflect independent mechanisms, or are tied together through common biological pathways and functions. To address this question, we turn to classical neuroscience knowledge of how neurons integrate signaling across different transmitter systems. The original evidence for coupling across neurotransmitter systems was documented by Greengard and co-workers16, 25 in the neostriatum, following the discovery of the protein named DARPP-32 and its role as a convergence node for integration of ‘slow’ (GPCRs) and ‘fast’ (ionotropic receptors) neurotransmission.26 The integration network described in these studies consists of GPCRs and the associated intracellular signaling cascades that—usually via a second messenger (cAMP, cGMP, Ca2+, diacylglycerol or inositol trisphosphate (IP3))—regulate a set of kinases and phosphatases. The kinases and phosphatases, in turn, control the sensitivity/permeability of ionotropic neurotransmitter receptors and ion channels, ensuring the balance across neurotransmission of different modalities (e.g., dopaminergic and glutamatergic, reviewed in Beaulieu and Gainetdinov17).
Motivated by this framework, we evaluated whether the genes identified by our analysis comprise putative elements of this neurotransmitter integration network: (1) Ca2+ signaling toolkit;15 (2) machinery for ion transport across plasma membrane including ion channels, transporters, pumps and ionotropic receptors; (3) typical GPCR signaling cascades involving GPCRs, G proteins and regulators of their activity, either adenylate cyclase (AC) or phospholipase C (PLC), and the downstream kinases and phosphatases. Table 1 lists the risk genes sorted into each of these categories. Functional interactions between the same genes are illustrated schematically in Figure 1. Below we describe each category; note that each gene has been discovered individually and independently of its function.
Ca2+ signaling toolkit
In excitable cells, Ca2+ is a main second messenger transmitting depolarization to the biochemical machinery of the cell. Therefore, intracellular concentration of free Ca2+ is tightly regulated by an array of channels, Ca2+-permeable ligand-operated receptors, pumps and exchangers. Taken together, these proteins are commonly referred to as the ‘Ca2+ signaling toolkit’.15, 27 Within the most significant detected loci (replication rate>0.8), we found 21 genes coding for the elements of this toolkit (Table 1 and Figure 1). These included VGCC subunits (CACNA1C, CACNA1D, CACNA1I and CACNB2), plasma membrane Ca2+ pump (ATP2B2), the endoplasmic reticulum (ER) Ca2+ pump SERCA (sarco/ER Ca2+-ATPase) (ATP2A2), Ca2+-permeable NMDA receptor (NMDAR, GRIN2A), IP3-regulated ER Ca2+ channel (ITPR3, ITPRIP), Ca2+-permeable transient receptor potential cation channel (TRPC4, TRPM6), Ca2+-binding proteins calneuron, secretagogin, neurogranin and nucleobindin (CALN1, SCGN, NRGN and NUCB2), ‘sperm-associated’ VGCC expressed in the brain (CATSPER2 and CATSPER2P1) and Ca2+-permeable nicotinic acetylcholine receptor (CHRNA2, CHRNA3, CHRNA5 and CHRNB4). These results extend the primary Psychiatric Genomic Consortium findings, based on the same data, that identified 9 out of these 21 genes (indicated by an asterisk in Table 1).6 In addition, we found α-centractin (ACTR1A) and three genes coding for the divalent metal cation transporter (CNNM2, CNNM3 and CNNM4). ACTR1A and CNNM2 were reported as Ca2+ homeostasis-related SCZ risk factors in an earlier GWAS.2, 6
To test statistically for enrichment of SNPs within ‘Ca2+ signaling toolkit’ genes, we derived a list of genes from a published expert review unrelated to GWAS15 (Supplementary Table 2). Ca2+ signaling toolkit genes were highly enriched compared with all genes, shown in the fold-enrichment plot (Figure 2).
Ion transport across plasma membrane
VGCCs constitute a hub for integration of the membrane potential, which reflects spiking and synaptic activity, with activation of GPCRs for neurotransmitters and neuromodulators. Indeed, VGCC properties (conductance, activation/inactivation, voltage dependence) are known to be modulated by the direct binding of G-protein Gβγ subunits.28 Therefore, alteration in other ion channels, ionotropic receptors, GPCRs or G proteins may lead to similar endophenotypes as functional deficits in VGCC themselves. Below, we evaluate this hypothesis, starting with mechanisms of ion transport followed by GPCR signaling cascades.
We considered known mechanisms of ion transport across the plasma membrane for the four major ions determining the membrane potential and neuronal electrical excitability: Na+, K+, Ca2+ and Cl−. These include ion channels and ionotropic receptors as well as pumps and transporters. In addition to VGCCs (CACNA1C, CACNA1D, CACNA1I and CACNB2), we found genes for K+-permeable channels of different kinds: voltage-gated KCNB1, KCNG2, KCNV1 and HCN1 (Ih current), inward-rectifying KCNJ3 and small conductance KCNN3. We also found voltage-gated Cl− channel (CLCN3), K+/Cl− cotransporter SLC12A4 and Na+/K+ ATPase-interacting NKAIN1.
For ionotropic receptors, in addition to the NMDAR (GRIN2A), and nAChRs (CHRNA2, CHRNA3, CHRNA5, CHRNB4) mentioned above, we found ionotropic glutamate receptors (AMPA subunit GRIA1 and Delta subunit GRID1) and GABAA receptor subunit GABRA2. Further, we identified a number of genes that have known direct functional interaction with NMDAR: SRR coding for astrocytic enzyme serine racemase known to produce a potent NMDAR modulator d-serine,29 MAGUK kinases DLG2 and DLG4 and synaptic Ras GTPase-activating protein 1 (SYNGAP1). Two additional genes coded for neurotransmitter transporters: GABA vesicular transporter SLC32A1 and choline transporter SLC44A4.
It is recognized that cellular membrane permeability to ions is collectively determined by channels, transporters, pumps and ionotropic receptors. However, a single list of all the molecular players is not available. To test for enrichment, we compiled two separate subsets of genes coding for ion channels and ionotropic receptors relying on gene naming conventions (Supplementary Table 2). The ion channels list included all plasma membrane Ca2+ channels from the ‘Ca2+ signaling toolkit' gene set and the following additional channels: voltage-gated Na+, voltage-gated K+, voltage-gated Cl−, Ca2+-dependent K+, Ca2+-dependent Cl−, inward-rectifying K+, K+ leak, cation leak and the hyperpolarization-activated current Ih. The ionotropic receptors list included AMPA, Kainate, Delta and NMDA receptors for glutamate, GABAA receptors, ionotropic 5HT receptor 3A, nicotinic ACh receptors and purinergic receptors P2X. The fold-enrichment plot in Figure 2 shows clear enrichment for the Ion channels and Ionotropic receptors for neurotransmitters gene sets.
GPCR signaling
Properties of ion channels and ionotropic receptors are regulated by downstream signaling cascade activation of GPCRs. For this reason, we examined genes within the identified loci for their association with the typical G-protein signaling cascades involving AC, PLC and the downstream kinases and phosphatases. We found eight hits on GPCRs (DRD2, GRM3, OPRD1, GABABR1, GABABR2, CHRM3, CHRM4 and HTR1A), opioid-binding protein (OPCML) having a role in opioid receptor function, G-protein β- and γ-subunits (GNB2, GNG7), two members of the family of regulators of G-protein signaling (RGS) controlling the duration of the G-protein-coupled signaling cascades (RGS6, RGS7BP), G-protein signaling modulator GPSM3 and three genes related to the β-arrestins/GPCR kinase pathway of GPCR regulation/internalization: AKT3, AKTIP and CLTA.
In addition, we found numerous genes coding for the classical AC- and PLC-mediated downstream signaling cascade of GPCRs. The AC cascade included phosphodiesterases that limit the duration of cAMP (and cGMP) signaling (PDE9A, PDE4B), SPHKAP, AKA6P and NBEA known for their role in anchoring of protein kinase A and cAMP-dependent CREB1. A hit on AC itself (ADCY1) was just below the chosen 0.8 replication threshold (Table 1).
PLC featured three hits: PLCB2, PLCH2 and PLCL1. G proteins that mediate activation of PLC result in the local depletion of a membrane phospholipid phosphatidylinositol-4, 5-biphosphate (PIP2), the presence of which is necessary for maintenance of the high-threshold VGCC currents.30 Indeed, a number of additional genes were specifically associated with the PIP2 metabolism: phospholipase A2 (PLA2G15), diacylglycerol kinase (DGKI, DGKZ), protein kinase C (PRKCB), protein kinase D (PRKD1, PRKD3) and a phosphatidylinositol transfer protein required for the synthesis of PIP2 (PITNM2).
We also found numerous hits on kinases and phosphatases that are known to have a central role in the integration of fast and slow neurotransmission:16, 26 the main integration hub DARPP-32 (PPP1R1B), other PP1 subunits (PPP1R3B, PPP1R10, PPP1R13B and PPP1R16B), PP2A (PPP2R2A, PPP2R3A, PPP2R2B, PPP2R3C and PPP2R5B), PKG (PRKG1), CLK2-stimulating PP2A assembly, PHACTR3-inhibiting PP1 and ANP32E-inhibiting PP2A. Other genes deserving specific comments are neuronal nitric oxide synthase (NOS1) and its interacting protein (NOSIP), which is an enzyme producing NO necessary for the NO→cGMP→PKG cascade featured in Greengard26 (see Figure 3). Finally, a neuron-specific activator of cyclin-dependent kinase 5 p35 (CDK5R1) acting on DARPP-32 was present just below the 0.8 replication rate threshold (Table 1).
To test for enrichment, we compiled a separate set of genes coding for metabotropic receptors (for glutamate, GABA, dopamine, 5HT, ACh, NA, opioids, adenosine and ATP), relying on gene naming conventions (Supplementary Table 2). In addition, we compiled two separate curated gene lists based on two review articles: The first article provides an overview of molecular players from the standpoint of the critical integrative role played by DARPP-32.16 The second one focused on the intracellular cascades downstream dopamine receptors and interaction with glutamate signaling.17 Although metabotropic receptors were not enriched, we found clear and strong enrichment for both the DARPP-32 gene list, and the Dopamine pathway gene list (Figure 2).
Gene expression profile
To play the proposed role in dysregulation of neuronal excitability and synaptic integration, the genes identified by our analysis (Table 1) should be, at the very least, expressed in the brain tissue. To this end, we used data from the Human Brain Transcriptome Project18 to examine mRNA expression for each of these genes in six brain regions of the developing and adult human brain. With the exception of CATSPER2, RGS7BP and SPHKAP, for which expression data were not available, our analysis confirmed that these genes are indeed expressed in the brain in a region-specific manner and with the expression profile following specific developmental trajectories, underscoring their relevance to brain organization and function (Supplementary Figure 1).
Comprehensive model of SCZ pathophysiology
SCZ is known as a complex polygenic disorder and is sometimes discussed as a heterogeneous clinical diagnosis of what could be unrelated dysfunctions on the level of cell biology.31 In isolation, individual genes identified in our study can be viewed as supportive of the dopaminergic,17 glutamatergic32 or calcium homeostasis13 hypotheses of the disease. However, considered together, the present results suggest an alternative model in which the existence of the SCZ phenotype with multiple genetic risk factors can be explained by the convergence and tight communication between seemingly independent neurotransmitter pathways controlling neuronal excitability and, ultimately, neuronal network function/dysfunction. The overlap between the SCZ risk genes discovered using our analysis and the previously described neurotransmitter integration network is illustrated in Figure 3, which was generated by populating a figure from Greengard’s 2001 review26 with our identified genes (only a subset of those from Table 1/Figure 1 is shown).
The idea that multiple factors of a polygenic disease are not distributed randomly across the genome but involve SNPs that share a common biological function has been discussed previously for brain diseases,33, 34, 35 and has been demonstrated for non-neuropsychiatric disorders.36, 37, 38, 39 In SCZ, the evaluation of functional relationships between individual risk variants finally became possible because of the availability of large data sets and powerful methods for statistical analysis that enable identification of a sufficient number of genome-wide molecular risk factors. Our results corroborate earlier reports on the association of SCZ with polymorphisms in genes coding for VGCCs,4 and genes related to synaptic function and neuronal excitability.4, 5, 19, 20, 21, 33, 40, 41, 42, 43, 44 Importantly, we show that many of these genes are mapped onto a common network, where the signaling cascades triggered by activation of GPCRs modulate ion channels and ionotropic receptors. This is achieved by activation of the intracellular kinases and phosphatases ensuring the appropriate level of excitability, and thus neuronal output, given the intensity and modality of the input. This is in line with the notion that specific therapeutic effects of antipsychotic agents involve mechanisms converging on normalizing synaptic transmission.45 An effect on DARPP-32 of antipsychotics has also been reported in neuronal cell cultures.46
While each of the SNPs involved have only minor effects by themselves on neuronal function, a critical number of synergistic variants may result in impaired neuronal function and thus SCZ symptoms. For example, alteration in Ca2+ signaling would affect such fundamental cellular level mechanisms as communication between the apical dendrite and the soma in cortical pyramidal cells leading to altered association of inputs at the cellular level and disturbed synaptic communication in neuronal networks.47 This could be collectively driven by the structural variation in the subunits of VGCCs, their post-translational modification and change in the electrical membrane excitability because of the up/downregulation of K+ conductances.
There is compelling evidence that SCZ is a neurodevelopmental disorder involving abnormal synaptic formation and connectivity (for review see Insel23 and Marenco and Weinberger48). With this in mind, it is tempting to speculate that any of the factors that alter the integration network depicted in Figure 1 are likely to evoke a series of compensatory mechanisms guided by neuronal plasticity rules, such as achieving the desirable end-point level of synaptic transmission.49 These mechanisms of adjustment have been documented on the single-neuron and neuronal network level in the valproic acid rat model of autism.50, 51 In this model, a decrease in the intrinsic (single-cell) excitability was accompanied by a compensatory increase both in NMDA receptor expression and synaptic connectivity.52 Thus, compensatory mechanisms set off by a deviation in the developmental course may attempt but not achieve the normal phenotype capable of supporting typical brain function. As such, these regulatory mechanisms may push the neuronal network to a new ‘state’, with dysfunctional information processing that it is associated with the SCZ behavioral phenotype.53 Thus, the common denominator among the genetic risk factors for SCZ might be the neurodevelopmental convergence upon a similar end-point functional network organization. This convergence is not unexpected taking into account the tight interactions across neurotransmitter systems.16
While the present analysis is focused on neuronal signal transduction and membrane excitability, one can imagine that similar compensation/plasticity can be triggered by deficits in the establishment, function, and modification of synaptic connections related to abnormal surface recognition, axonal guidance, vesicular transport or exocytosis machinery. Indeed, regulated cytoskeleton-associated protein complexes, important for synaptic structure, function and plasticity, have been implicated in SCZ by recent de novo mutation studies.20, 22 In addition, impairment of cellular energetics may compromise neuronal repolarization, Ca2+ clearance and ionic homeostasis contributing to the deficit in excitability/intracellular signaling. Careful consideration of these mechanisms is beyond the scope of the current perspective article.
Our current approach has a number of limitations. As with all GWASs, identification of disease-associated genes in the current study relies on the presence of common variants in the sample population. Therefore, the absence of known risk genes, such as DISC1, among our identified genes does not contradict their established relationship to SCZ,54 but rather suggest that common inherited variants in these genes are not significant contributors to disease risk. Moreover, a lack of association for a particular gene does not imply that the molecule does not have a significant role in the relevant signaling cascades, but rather that the specific common variants or SNPs of the gene do not alter the relevant functional properties. In some cases significant functional variants may not be tolerated.55 Another limitation shared between all GWASs is that each of the identified gene loci may contain multiple genes in linkage disequilibrium (i.e., inherited together) where individual contributions cannot be disentangled with current methods. Therefore, variation in some of the identified genes may not, in fact, contribute to SCZ pathophysiology.
Further investigation will be required to address the significance of specific SNP variants for the function of the respective genes (e.g., conductance, gating or voltage sensitivity of VGCC), using an approach that takes into account the polygenic context. Similarly, dedicated mechanistic studies will be needed to test the hypothesis of SCZ etiology as a dysfunction of integration across neurotransmitter systems. These studies would leverage methods and technologies beyond genetic association, such as measurement, manipulation and genetic engineering in neurons derived from human-induced pluripotent stem cells (hiPSCs). With small effects of individual risk variants, the difference in function due to, for example, alteration in a single type of ion channel may be too small to detect. Therefore, we propose using functional readouts at the level of convergence, such as neuronal excitability of single-neuron or/and neuronal network, while choosing ‘cases’ and ‘controls’ based on a predefined set of high-risk variants. Specifically, we envision the following lines of inquiry:
-
1
Establishing the cellular and small network-level functional endophenotype characteristic of SCZ using hiPSC-derived neurons. These experiments would take advantage of novel gene-editing methods,56 which now allow replacement of large DNA segments, as well as the growing arsenal of optogenetic and chemogenetic manipulation tools57, 58, 59 and optical reporters,60 being developed under the BRAIN Initiative.61 In this approach, we would rank patient-derived hiPSCs according to the number of high-risk variants, weighted by their strength of disease association, within a predefined set of genes, such as that comprising Table 1. Next, we would choose hiPSCs with the highest risk scores, and generate the corresponding negative controls by substituting high-risk variants with those carrying low risk. Conversely, we would choose hiPSCs with the lowest risk scores, and generate the corresponding positive controls by substituting low-risk variants with those carrying high risk. Then, we would subject the resulting cell lines to a series of well-controlled opto- and/or chemogenetic stimuli using specific optical reporters for high-throughput phenotyping of induced responses.62 These experiments would establish a mapping between groups of risk variants and the cellular/network behavior, and answer the question of whether accumulation of risk variants along seemingly independent intracellular signaling pathways can result in a similar functional endophenotype.
-
2
Direct opto- or chemogenetic manipulation of the key integration hubs such as DARPP-32. The question is whether directly interfering with the activity of converging nodes integrating fast and slow neurotransmission in control neurons (and networks) can evoke functional endophenotype characteristic for neurons bearing a significant load of SCZ risk variants.
-
3
Extension of (1) and (2) to organized cell cultures, such as ‘corticoids’,63, 64 and ‘chimeric’ mice.65 These model systems promote neuronal maturation and may better approximate a natural neuronal circuit environment.
-
4
Investigation in transgenic mouse models of SCZ.66 Although the validity of these models remains controversial,67 some of them, such as DISC1L100P,54, 68 exhibit neuroanatomical and behavioral changes seen in SCZ patients and respond to antipsychotic drugs.69 The question is whether DISC1 dysfunction results in the same functional endophenotype as in (1).
-
5
Investigation of the effects of antipsychotic drugs (e.g., haloperidol and clozapine) on normalizing the SCZ functional endophenotype in hiPSC-derived neurons and in responsive mouse models.
-
6
Combination of the above-mentioned experimental approaches (1–5) with ‘computational psychiatry’.70, 71 The basic idea is to use experimentally obtained mapping between SCZ risk variants and the cellular/network behavior to constrain computation models of single neurons and neuronal networks providing mechanistic biophysical basis for the SCZ functional endophenotype. A detailed ‘realistic’ model of cortical column72, 73 has recently become available because of a decade-long effort by the Blue Brain Project currently continuing under the umbrella of the Human Brain Project. These models of synaptically connected spiking neurons with realistic morphologies can be combined with the physical principles governing generation of the extracelluar potentials74, 75 for prediction of macroscopic electro- or magnetoencephalography signals resulting from complex circuit phenomena. These models may help bridge the cellular level biophysics to well-established neuronal deficits measured noninvasively in SCZ patients such as the lack of prepulse inhibition.76
Although the focus of the current analysis is limited to SCZ, it is quite possible that the biological mechanisms are more closely associated with specific symptom dimensions, such as psychosis, across psychiatric diseases. In fact, recent cross-disorder GWAS analyses, leveraging novel statistical methodologies,77 suggest extensive genetic pleiotropy among psychiatric diseases,78, 79, 80, 81 and other phenotypes.82, 83, 84 The relationship between specific symptom dimensions and biological mechanisms can be further investigated by focusing on substudies with more comprehensive assessments of symptom dimensions, rather than case–control information alone.
To summarize, the spectrum of molecular risk factors identified in our study supports the concept of SCZ as an ‘associative’ disorder: a breakdown in the communication across different slow and fast neurotransmitter systems through intracellular signaling pathways. This emergent model may unify a number of currently competing hypotheses, for example, dopaminergic, glutamatergic, calcium, second Messenger (reviewed in Lidow13), and the recently put forward Balance between Excitation and Inhibition.20 Future investigations, leveraging novel technologies for measurement, manipulation and gene editing, as well as novel computational tools, will be required to bridge the gap between the correlation results of GWAS and the underlying biological mechanisms, enabling ‘brain-based’ diagnoses and, ultimately, treatment of mental disease.
References
Cardno AG, Marshall EJ, Coid B, Macdonald AM, Ribchester TR, Davies NJ et al. Heritability estimates for psychotic disorders: the Maudsley twin psychosis series. Arch Gen Psychiatry 1999; 56: 162–168.
Ripke S, O'Dushlaine C, Chambert K, Moran JL, Kahler AK, Akterin S et al. Genome-wide association analysis identifies 13 new risk loci for schizophrenia. Nat Genet 2013; 45: 1150–1159.
Sullivan PF . The psychiatric GWAS consortium: big science comes to psychiatry. Neuron 2010; 68: 182–186.
Cross-Disorder Group of the Psychiatric Genomics C. Identification of risk loci with shared effects on five major psychiatric disorders: a genome-wide analysis. Lancet 2013; 381: 1371–1379.
Schizophrenia Psychiatric Genome-Wide Association Study C. Genome-wide association study identifies five new schizophrenia loci. Nat Genet 2011; 43: 969–976.
Schizophrenia Working Group of the Psychiatric Genomics C. Biological insights from 108 schizophrenia-associated genetic loci. Nature 2014; 511: 421–427.
Schork AJ, Thompson WK, Pham P, Torkamani A, Roddey JC, Sullivan PF et al. All SNPs are not created equal: genome-wide association studies reveal a consistent pattern of enrichment among functionally annotated SNPs. PLoS Genet 2013; 9: e1003449.
Wang Y, Thompson WK, Schork AJ, Holland D, Chen CH, Bettella F et al. Leveraging genomic annotations and pleiotropic enrichment for improved replication rates in schizophrenia GWAS. PLoS Genet 2016; 12: e1005803.
van Os J, Kapur S . Schizophrenia. Lancet 2009; 374: 635–645.
Moghaddam B, Javitt D . From revolution to evolution: the glutamate hypothesis of schizophrenia and its implication for treatment. Neuropsychopharmacology 2012; 37: 4–15.
Nakazawa K, Zsiros V, Jiang Z, Nakao K, Kolata S, Zhang S et al. GABAergic interneuron origin of schizophrenia pathophysiology. Neuropharmacology 2012; 62: 1574–1583.
Berridge MJ . Calcium signalling and psychiatric disease: bipolar disorder and schizophrenia. Cell Tissue Res 2014; 357: 477–492.
Lidow MS . Calcium signaling dysfunction in schizophrenia: a unifying approach. Brain Res Brain Res Rev 2003; 43: 70–84.
Moises HW, Zoega T, Gottesman II . The glial growth factors deficiency and synaptic destabilization hypothesis of schizophrenia. BMC Psychiatry 2002; 2: 8.
Berridge MJ . Calcium signalling remodelling and disease. Biochem Soc Trans 2012; 40: 297–309.
Svenningsson P, Nishi A, Fisone G, Girault JA, Nairn AC, Greengard P . DARPP-32: an integrator of neurotransmission. Annu Rev Pharmacol Toxicol 2004; 44: 269–296.
Beaulieu JM, Gainetdinov RR . The physiology, signaling, and pharmacology of dopamine receptors. Pharmacol Rev 2011; 63: 182–217.
Kang HJ, Kawasawa YI, Cheng F, Zhu Y, Xu X, Li M et al. Spatio-temporal transcriptome of the human brain. Nature 2011; 478: 483–489.
Network, Pathway Analysis Subgroup of Psychiatric Genomics C. Psychiatric genome-wide association study analyses implicate neuronal, immune and histone pathways. Nat Neurosci 2015; 18: 199–209.
Pocklington AJ, Rees E, Walters JT, Han J, Kavanagh DH, Chambert KD et al. Novel findings from CNVs implicate inhibitory and excitatory signaling complexes in schizophrenia. Neuron 2015; 86: 1203–1214.
Purcell SM, Moran JL, Fromer M, Ruderfer D, Solovieff N, Roussos P et al. A polygenic burden of rare disruptive mutations in schizophrenia. Nature 2014; 506: 185–190.
Fromer M, Pocklington AJ, Kavanagh DH, Williams HJ, Dwyer S, Gormley P et al. De novo mutations in schizophrenia implicate synaptic networks. Nature 2014; 506: 179–184.
Insel TR . Rethinking schizophrenia. Nature 2010; 468: 187–193.
McCarroll SA, Hyman SE . Progress in the genetics of polygenic brain disorders: significant new challenges for neurobiology. Neuron 2013; 80: 578–587.
Walaas SI, Aswad DW, Greengard P . A dopamine- and cyclic AMP-regulated phosphoprotein enriched in dopamine-innervated brain regions. Nature 1983; 301: 69–71.
Greengard P . The neurobiology of slow synaptic transmission. Science 2001; 294: 1024–1030.
Berridge MJ, Bootman MD, Roderick HL . Calcium signalling: dynamics, homeostasis and remodelling. Nat Rev Mol Cell Biol 2003; 4: 517–529.
Zamponi GW, Currie KP . Regulation of Ca(V)2 calcium channels by G protein coupled receptors. Biochim Biophys Acta 2013; 1828: 1629–1643.
Mothet JP, Parent AT, Wolosker H, Brady RO Jr., Linden DJ, Ferris CD et al. d-Serine is an endogenous ligand for the glycine site of the N-methyl-D-aspartate receptor. Proc Natl Acad Sci USA 2000; 97: 4926–4931.
Buraei Z, Yang J . Structure and function of the beta subunit of voltage-gated Ca(2)(+) channels. Biochim Biophys Acta 2013; 1828: 1530–1540.
Brennand KJ, Landek-Salgado MA, Sawa A . Modeling heterogeneous patients with a clinical diagnosis of schizophrenia with induced pluripotent stem cells. Biol Psychiatry 2014; 75: 936–944.
Kantrowitz JT, Javitt DC . N-methyl-d-aspartate (NMDA) receptor dysfunction or dysregulation: the final common pathway on the road to schizophrenia? Brain Res Bull 2010; 83: 108–121.
Lips ES, Cornelisse LN, Toonen RF, Min JL, Hultman CM et al, International Schizophrenia C. Functional gene group analysis identifies synaptic gene groups as risk factor for schizophrenia. Mol Psychiatry 2012; 17: 996–1006.
Torkamani A, Topol EJ, Schork NJ . Pathway analysis of seven common diseases assessed by genome-wide association. Genomics 2008; 92: 265–272.
Holmans P, Green EK, Pahwa JS, Ferreira MA, Purcell SM, Sklar P et al. Gene ontology analysis of GWA study data sets provides insights into the biology of bipolar disorder. Am J Hum Genet 2009; 85: 13–24.
Solovieff N, Cotsapas C, Lee PH, Purcell SM, Smoller JW . Pleiotropy in complex traits: challenges and strategies. Nat Rev Genet 2013; 14: 483–495.
Cotsapas C, Voight BF, Rossin E, Lage K, Neale BM, Wallace C et al. Pervasive sharing of genetic effects in autoimmune disease. PLoS Genet 2011; 7: e1002254.
Wagner GP, Zhang J . The pleiotropic structure of the genotype-phenotype map: the evolvability of complex organisms. Nat Rev Genet 2011; 12: 204–213.
Sivakumaran S, Agakov F, Theodoratou E, Prendergast JG, Zgaga L, Manolio T et al. Abundant pleiotropy in human complex diseases and traits. Am J Hum Genet 2011; 89: 607–618.
Mirnics K, Middleton FA, Lewis DA, Levitt P . Delineating novel signature patterns of altered gene expression in schizophrenia using gene microarrays. Scientific World J 2001; 1: 114–116.
Mirnics K, Middleton FA, Lewis DA, Levitt P . The human genome: gene expression profiling and schizophrenia. Am J Psychiatry 2001; 158: 1384.
Mirnics K, Middleton FA, Lewis DA, Levitt P . Analysis of complex brain disorders with gene expression microarrays: schizophrenia as a disease of the synapse. Trends Neurosci 2001; 24: 479–486.
Mirnics K, Middleton FA, Marquez A, Lewis DA, Levitt P . Molecular characterization of schizophrenia viewed by microarray analysis of gene expression in prefrontal cortex. Neuron 2000; 28: 53–67.
Vawter MP, Barrett T, Cheadle C, Sokolov BP, Wood WH III, Donovan DM et al. Application of cDNA microarrays to examine gene expression differences in schizophrenia. Brain Res Bull 2001; 55: 641–650.
Thomas EA . Molecular profiling of antipsychotic drug function: convergent mechanisms in the pathology and treatment of psychiatric disorders. Mol Neurobiol 2006; 34: 109–128.
Valjent E, Bertran-Gonzalez J, Bowling H, Lopez S, Santini E, Matamales M et al. Haloperidol regulates the state of phosphorylation of ribosomal protein S6 via activation of PKA and phosphorylation of DARPP-32. Neuropsychopharmacology 2011; 36: 2561–2570.
Larkum M . A cellular mechanism for cortical associations: an organizing principle for the cerebral cortex. Trends Neurosci 2013; 36: 141–151.
Marenco S, Weinberger DR . The neurodevelopmental hypothesis of schizophrenia: following a trail of evidence from cradle to grave. Dev Psychopathol 2000; 12: 501–527.
Spitzer NC . Electrical activity in early neuronal development. Nature 2006; 444: 707–712.
Rinaldi T, Kulangara K, Antoniello K, Markram H . Elevated NMDA receptor levels and enhanced postsynaptic long-term potentiation induced by prenatal exposure to valproic acid. Proc Natl Acad Sci USA 2007; 104: 13501–13506.
Rinaldi T, Silberberg G, Markram H . Hyperconnectivity of local neocortical microcircuitry induced by prenatal exposure to valproic acid. Cereb Cortex 2008; 18: 763–770.
Markram H, Rinaldi T, Markram K . The intense world syndrome—an alternative hypothesis for autism. Front Neurosci 2007; 1: 77–96.
Adams RA, Stephan KE, Brown HR, Frith CD, Friston KJ . The computational anatomy of psychosis. Front Psychiatry/Front Res Found 2013; 4: 47.
Brandon NJ, Sawa A . Linking neurodevelopmental and synaptic theories of mental illness through DISC1. Nat Rev Neurosci 2011; 12: 707–722.
Visscher PM, Brown MA, McCarthy MI, Yang J . Five years of GWAS discovery. Am J Hum Genet 2012; 90: 7–24.
Gaj T, Sirk SJ, Shui SL, Liu J . Genome-editing technologies: principles and applications. Cold Spring Harb Perspect Biol 2016; 8: pii: a023754.
Kramer RH, Mourot A, Adesnik H . Optogenetic pharmacology for control of native neuronal signaling proteins. Nat Neurosci 2013; 16: 816–823.
Spangler SM, Bruchas MR . Optogenetic approaches for dissecting neuromodulation and GPCR signaling in neural circuits. Curr Opin Pharmacol 2016; 32: 56–70.
Roth BL . DREADDs for neuroscientists. Neuron 2016; 89: 683–694.
Tenner B, Mehta S, Zhang J . Optical sensors to gain mechanistic insights into signaling assemblies. Curr Opin Struct Biol 2016; 41: 203–210.
Devor A, Bandettini PA, Boas DA, Bower JM, Buxton RB, Cohen LB et al. The challenge of connecting the dots in the B.R.A.I.N. Neuron 2013; 80: 270–274.
Zhang H, Reichert E, Cohen AE . Optical electrophysiology for probing function and pharmacology of voltage-gated ion channels. Elife 2016; 5: e15202.
Mariani J, Simonini MV, Palejev D, Tomasini L, Coppola G, Szekely AM et al. Modeling human cortical development in vitro using induced pluripotent stem cells. Proc Natl Acad Sci USA 2012; 109: 12770–12775.
Pasca AM, Sloan SA, Clarke LE, Tian Y, Makinson CD, Huber N et al. Functional cortical neurons and astrocytes from human pluripotent stem cells in 3D culture. Nat Methods 2015; 12: 671–678.
Espuny-Camacho I, Michelsen KA, Gall D, Linaro D, Hasche A, Bonnefont J et al. Pyramidal neurons derived from human pluripotent stem cells integrate efficiently into mouse brain circuits in vivo. Neuron 2013; 77: 440–456.
Robertson HR, Feng G . Annual Research Review: transgenic mouse models of childhood-onset psychiatric disorders. J Child Psychol Psychiatry 2011; 52: 442–475.
Geyer MA, Olivier B, Joels M, Kahn RS . From antipsychotic to anti-schizophrenia drugs: role of animal models. Trends Pharmacol Sci 2012; 33: 515–521.
Porteous DJ, Millar JK, Brandon NJ, Sawa A . DISC1 at 10: connecting psychiatric genetics and neuroscience. Trends Mol Med 2011; 17: 699–706.
Clapcote SJ, Lipina TV, Millar JK, Mackie S, Christie S, Ogawa F et al. Behavioral phenotypes of Disc1 missense mutations in mice. Neuron 2007; 54: 387–402.
Wang XJ, Krystal JH . Computational psychiatry. Neuron 2014; 84: 638–654.
Maki-Marttunen T, Halnes G, Devor A, Witoelar A, Bettella F, Djurovic S et al. Functional effects of schizophrenia-linked genetic variants on intrinsic single-neuron excitability: a modeling study. Biol Psychiatry Cogn Neurosci Neuroimag 2016; 1: 49–59.
Hill SL, Wang Y, Riachi I, Schurmann F, Markram H . Statistical connectivity provides a sufficient foundation for specific functional connectivity in neocortical neural microcircuits. Proc Natl Acad Sci USA 2012; 109: E2885–E2894.
Markram H, Muller E, Ramaswamy S, Reimann MW, Abdellah M, Sanchez CA et al. Reconstruction and simulation of neocortical microcircuitry. Cell 2015; 163: 456–492.
Reimann MW, Anastassiou CA, Perin R, Hill SL, Markram H, Koch C . A biophysically detailed model of neocortical local field potentials predicts the critical role of active membrane currents. Neuron 2013; 79: 375–390.
Hagen E, Dahmen D, Stavrinou ML, Linden H, Tetzlaff T, van Albada SJ et al. Hybrid scheme for modeling local field potentials from point-neuron networks. Cereb Cortex 2016; 26: 4461–4496.
Swerdlow NR, Braff DL, Geyer MA . Sensorimotor gating of the startle reflex: what we said 25 years ago, what has happened since then, and what comes next. J Psychopharmacol 2016; 30: 1072–1081.
Schork AJ, Wang Y, Thompson WK, Dale AM, Andreassen OA . New statistical approaches exploit the polygenic architecture of schizophrenia—implications for the underlying neurobiology. Curr Opin Neurobiol 2016; 36: 89–98.
Andreassen OA, Thompson WK, Schork AJ, Ripke S, Mattingsdal M, Kelsoe JR et al. Improved detection of common variants associated with schizophrenia and bipolar disorder using pleiotropy-informed conditional false discovery rate. PLoS Genet 2013; 9: e1003455.
Le Hellard S, Wang Y, Witoelar A, Zuber V, Bettella F, Hugdahl K et al. Identification of gene loci that overlap between schizophrenia and educational attainment. Schizophr Bull 2016: pii: sbw085 (e-pub ahead of print).
Desikan RS, Schork AJ, Wang Y, Witoelar A, Sharma M, McEvoy LK et al. Genetic overlap between Alzheimer's disease and Parkinson's disease at the MAPT locus. Mol Psychiatry 2015; 20: 1588–1595.
Andreassen OA, Harbo HF, Wang Y, Thompson WK, Schork AJ, Mattingsdal M et al. Genetic pleiotropy between multiple sclerosis and schizophrenia but not bipolar disorder: differential involvement of immune-related gene loci. Mol Psychiatry 2015; 20: 207–214.
LeBlanc M, Zuber V, Andreassen BK, Witoelar A, Zeng L, Bettella F et al. Identifying novel gene variants in coronary artery disease and shared genes with several cardiovascular risk factors. Circulation research 2016; 118: 83–94.
Andreassen OA, Zuber V, Thompson WK, Schork AJ, Bettella F et alConsortium P. Shared common variants in prostate cancer and blood lipids. Int J Epidemiol 2014; 43: 1205–1214.
Andreassen OA, Djurovic S, Thompson WK, Schork AJ, Kendler KS, O'Donovan MC et al. Improved detection of common variants associated with schizophrenia by leveraging pleiotropy with cardiovascular-disease risk factors. Am J Hum Genet 2013; 92: 197–209.
Acknowledgements
We gratefully acknowledge support from the NIH (NS057198, EB00790, R01MH111359), the Research Council of Norway (229129, 213837, 223273), the South-East Norway Regional Health Authority (2017-112, 2016-064) and KG Jebsen Stiftelsen (SKGJ-MED‐008).
Author information
Authors and Affiliations
Corresponding authors
Ethics declarations
Competing interests
The authors declare no conflict of interest.
Additional information
Supplementary Information accompanies the paper on the Molecular Psychiatry website
Rights and permissions
About this article
Cite this article
Devor, A., Andreassen, O., Wang, Y. et al. Genetic evidence for role of integration of fast and slow neurotransmission in schizophrenia. Mol Psychiatry 22, 792–801 (2017). https://doi.org/10.1038/mp.2017.33
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/mp.2017.33
- Springer Nature Limited
This article is cited by
-
Association of neurotransmitter pathway polygenic risk with specific symptom profiles in psychosis
Molecular Psychiatry (2024)
-
Genetic overlap between multivariate measures of human functional brain connectivity and psychiatric disorders
Nature Mental Health (2024)
-
Shared genetic architecture between mental health and the brain functional connectome in the UK Biobank
BMC Psychiatry (2023)
-
The effect of alterations of schizophrenia-associated genes on gamma band oscillations
Schizophrenia (2022)
-
Zebrafish Larvae Carrying a Splice Variant Mutation in cacna1d: A New Model for Schizophrenia-Like Behaviours?
Molecular Neurobiology (2021)