Abstract

1. CACNA1C in mental disorders

While a number of medications are used for the treatment of psychiatric disorders, the majority were developed based upon efficacy to reduce symptoms rather than to eliminate pathological processes. Current treatments are inadequate for many patients as recurrence is common and the full clinical effects are often not obtained until after months of treatment. Despite these well-known inadequacies, improvements have been slow in coming largely due to a dearth of understanding regarding pathophysiology. This lack of knowledge has prevented the development of treatments specific for the underlying disease mechanisms of psychiatric illness.

Many neuropsychiatric disorders are highly heritable, indicating that genetics is certain to have a role in pathogenesis. For example, the individual heritability of bipolar disorder, major depression, and schizophrenia is estimated to be approximately 70–90%, 40–50%, and 80–85%, respectively (Cannon et al., 1998; Cardno and Gottesman, 2000; Kendler and Prescott, 1999; Kieseppä et al., 2004; McGuffin et al., 1996, 2003). There is also evidence for shared familial susceptibility between these three Diagnostic and Statistical Manual (DSM)-IV-defined disorders, and in particular between bipolar disorder and depression, and bipolar disorder and schizophrenia (Berrettini, 2003; Demjaha et al., 2011; Gottesman et al., 2010; Lichtenstein et al., 2009; McGuffin et al., 2003). These epidemiological data are supported by the more recent findings that specific genetic risk variants cross these diagnostic boundaries as well (Chubb et al., 2008; Maier, 2008; Williams et al., 2011). While there is a credible prediction that identifying susceptibility genes will eventually lead to targeted ‘cure therapeutics’ (Insel and Scolnick, 2006), there first remains a critical need to definitively identify such genes and to subsequently understand the functional consequences of the associated genetic variations (Fig. 1). However, identifying underlying susceptibility genes that are associated with psychiatric disorders in multiple populations has proven to be difficult due to a number of factors including less than adequate sample sizes, the polygenetic and multifactorial nature of the disorders, uncertainty or lack of precision in the diagnoses, and significant overlap both in terms of symptoms and likely pathophysiology as well.

Open in a separate window

Fig. 1

Psychiatric disorders and their relationship with susceptibility genes, physiological factors, current classes of drugs, and possible novel treatments

Current treatments for psychiatric disorders were developed to treat symptoms, and in many instances likely have limited effects on underlying pathophysiology. It is hoped that treatments targeted toward pathophysiology will provide greater and sustained rates of remission. However, currently knowledge is lacking regarding the pathophysiology of most psychiatric diseases. Treatments that target the underlying pathophysiology of psychiatric disorders such as mood disorders will derive from understanding the neurobiology underlying the disease. As diseases with high heritability, genetic underpinnings are a link to pathophysiology. The top portion of the figure shows current treatments. The bottom portion of the figure indicates examples of current top genetic findings in psychiatric disorders, including CACNA1C. The middle portion of the figure suggests levels of analysis that will be critical to understand the role of genetic variation on disease pathopysiology and to direct the subsequent development of future treatments. MAOI, Monoamine oxidase inhibitor; SSRI, Selective serotonin reuptake inhibitor; TCA, Tricyclic antidepressant; ANK3, ankyrin 3, node of Ranvier (ankyrin G); CACNA1C: voltage-dependent L type calcium channel, α1C subunit; MIR137, microRNA 137; ODZ4, odd Oz/ten-m homolog 4 (Drosophila); ZNF804A, zinc finger protein 804A.

However, encouraging results have emerged recently as the results of adequately powered genome-wide association studies (GWAS) have converged to implicate specific genetic polymorphisms. One gene in which such genetic polymorphisms have been identified is CACNA1C, which codes for the pore-forming α_1C subunit of the L-type voltage-gated calcium channel (LTCC), referred to as Ca_v1.2. Ca_v1.2 couples transient activation of inward calcium current to transcriptional regulation and plays an important role in dendritic development, neuronal survival, synaptic plasticity, memory formation, learning, and behavior (Barad, 2003; Dudkin et al., 1990; Kobrinsky et al., 2011; Moosmang et al., 2005; Shibasaki et al., 2010; West et al., 2001; Wheeler et al., 2008; White et al., 2008; Yoshii and Watabe, 1994).

The LTCC family consists of four distinct members referred to as Ca_v1.1–Ca_v1.4 (Catterall, 2011). Ca_v1.2 is the primary LTCC expressed in the mammalian brain. In the mouse brain, quantitative polymerase chain reaction (qPCR) of RNA transcripts revealed that Ca_v1.2 accounts for approximately 85% of the LTCCs with Ca_V1.3 accounting for most of the remainder (Sinnegger-Brauns et al., 2009). A functional Ca_v1.2 channel consists of three subunits: transmembrane α_1C (CACNA1C) and α₂δ (encoded by CACNA2D-1, 2 or 3) as well as intracellular β (encoded by CACNB1-4 genes) and calmodulin (CaM) (Dolphin, 2009). The major characteristics of the Ca_v1.2 channel such as voltage-sensing, ion selectivity, and pharmacological responses associated with binding of calcium channel blockers are encoded by CACNA1C. Auxiliary α₂δ, β and CaM are involved in regulation of expression and modulating select properties of Ca_v1.2.

There is evidence that functional mutations in the CACNA1C gene predispose to autism spectrum disorders. A rare human autosomal-dominant multisystem disorder, Timothy Syndrome, presenting with arrhythmias, webbed digits, congenital heart disease, and autism spectrum symptoms is associated with a mutation in the exon 8A splice variant of CACNA1C that comprises about 23% of Ca_v1.2 in heart and brain (Splawski et al., 2004). It has been demonstrated that Timothy Syndrome is associated with the G406R mutation in the cytoplasmic end of the exon 8A-coded transmembrane segment IS6. This mutation greatly reduces voltage-dependent inactivation of the calcium channel with little effect on calcium-dependent inactivation (Barrett and Tsien, 2008), and thus may potentially cause a pathogenic calcium overload in cardiac and neuronal cells. An analogous mutation in alternative exon 8 (mutually exclusive with exon 8A), G402S, and a corresponding reduction in voltage-dependent inactivation were observed in Ca_v1.2 of an individual with a severe Timothy Syndrome variant (Splawski et al., 2005, 2004). Taken together, these results show pathogenic effect of naturally occurring gain-of-function mutations probably affecting Ca_v1.2 molecular determinant(s) of slow inactivation (Depil et al., 2011; Shi and Soldatov, 2002).

Interestingly, prior to GWA approaches CACNA1C was considered a potential candidate gene for bipolar disorder. As reported in 2002, CACNA1C was included in a family based association study (470 parent-proband trios) of 76 candidate genes and was one of two genes (the other being brain derived neurotrophic factor (BDNF)) where a single nucleotide polymorphism (SNP) in CACNA1C was found to be nominally associated with a bipolar disorder diagnosis (Sklar et al., 2002). However, the association with CACNA1C did not survive the confirmation analysis in a separate cohort, though the association with BDNF did (Sklar et al., 2002). In hindsight this lack of replication was likely related to the genotyped synonymous (silent) SNP (rs72552065) being located in exon 44, approximately 450 kb downstream from the SNPs subsequently implicated by GWA studies conducted later on.

In a manner not biased by the previous results, in 2008 Sklar et al. reported the results of a GWA study of 1461 bipolar disorder and 2008 control subjects combined with the results of the Wellcome Trust Case–Control Consortium GWA study (Burton et al., 2007), and reported that the strongest signal (p = 1 × 10⁻⁴) of association from these combined samples were from a SNP within the CACNA1C gene (rs1006737) (Sklar et al., 2008). The accepted standard for GWA significance is 5 × 10⁻⁸ (Pe’er et al., 2008). Subsequent studies with increasingly larger sample sizes have since confirmed this finding (Ferreira et al., 2008; Keers et al., 2009). See Table 1, which details the genetic evidence associating CACNA1C with psychiatric disorders, and includes p values for all GWA studies, as well as odds ratios. The strongest implicated SNP in these early GWA studies was rs1006737. Most recently, the results of the Psychiatric Genetics Consortium (PGC) Bipolar Working Group comparing 11,977 bipolar disorder subjects and 51,672 controls identified another SNP (rs4765913) in CACNA1C as the most significant finding in that study, at a significance level of p = 1.52 × 10⁻⁸ (Sklar et al., 2011).

Many of the studies described above utilized overlapping patient and control populations, which increased with subsequent studies. As the population size increased the strength of association increased in these studies and the first strong association with bipolar disorder was observed at rs1006737. However, subsequent GWA studies with larger population size have identified significant associations with other SNPs. Therefore, while overlapping populations in successive studies possibly contributed to identifying rs1006737 as a candidate susceptibility SNP associated with bipolar disorder, they have not always resulted in consistent replication of this particular SNP. While the inability to fully replicate specific GWAS-identified SNPs in all the studies is a concern, the consistency with which GWAS signals are identified within the same general region (intron 3) of CACNA1C indicates the likely importance of the gene in bipolar disorder psychopathology (Fig. 2).

Open in a separate window

Fig. 2

Mental disorder significant findings projected on the genomic map of CACNA1C

Genomic map of CACNA1C beginning with exon 1 (Soldatov, 1994); exons are indicated with vertical bars separated by introns. Identified mental disorder associated SNPs are shown by colored circles at approximate locations on the gene. GWAS significant associations are: rs1006737 and rs4765905 (which are in complete LD)—bipolar disorder, schizophrenia, and major depression. rs4765913—bipolar disorder and schizophrenia/bipolar disorder combined analysis. rs7297582—bipolar disorder and major depression disorder combined analysis. rs2370411 and rs2370419: two SNPs showing sex specific association (females only). The red star (*) at exon 8/8a marks the Timothy syndrome loci. The image is of transcript variant 1, initially referred to as HFCC (Soldatov, 1992, GenBank accession #: NM_199460.2); not shown in the figure is alternative splicing of the gene.

The clinical genetic association of CACNA1C with mental disorders appears to be broader than for bipolar disorder only. Recent studies have established an association of CACNA1C genotype with schizophrenia (Bigos et al., 2010; Nyegaard et al., 2010) and major depression disorder (Casamassima et al., 2010b; Shi et al., 2011). Several studies have also implicated CACNA1C in combined analysis including distinct DSM-defined psychiatric disorders. For example, Green et al. (2010) conducted a GWAS with patients diagnosed with unipolar recurrent major depression (N = 1196), bipolar disorder (N = 1868) and schizophrenia (N = 479) compared to 15,316 nonpsychiatric control subjects. The authors found that the previously reported bipolar disorder risk SNP (rs1006737) conferred risk of schizophrenia and early recurrent major depression. The authors also noted that the association of rs1006737 with bipolar disorder was strengthened when they combined their data with that from the previous study by Ferreira et al. (2008). Recently, in a combined analysis of bipolar and schizophrenic (n = 16,373) patients compared to control (n = 14,044) subjects, rs4765905 located within CACNA1C was implicated as the top statistical hit (7.0 × 10⁻⁹) (Ripke et al., 2011). Another recent combined analysis of bipolar disorder and schizophrenia data sets revealed significant association with CACNA1C SNP rs4765913 (Sklar et al., 2011). Most recently Hamshere et al. (2012), reported that rs4765905 achieved genome wide significance for an association with schizophrenia after combining additional subjects with the analysis by Ripke et al. (2011). rs4765905 is in complete linkage disequilibrium (LD), defined as the occurrence of genotypes at two loci that is significantly different than by chance alone, with rs1006737 (D′= 1.0, r² = 1.0; LD is denoted by D′, or a correlation coefficient r², which both range from 0 to 1, with 1 indicating perfect LD) (Hill, 1974; Lewontin, 1964, 1988; Reich et al., 2001; Slatkin, 2008).

In terms of major depression, a meta analysis of two separate earlier GWA studies of bipolar and major depressive disorder patients, the CACNA1C SNPs rs1006737 and rs7297582 were found to reach a genome wide significance when both bipolar and unipolar mood disorders were combined (Liu et al., 2011). Further support for the involvement of CACNA1C in depression was provided by Dao et al. who reported that two SNPs in intron 3 (rs2370419 and rs2370411) showed significantly increased risk of illnesses in females, but not males, for bipolar and major depression disorder (Dao et al., 2010). This study used the GWAS data generated earlier from the National Institute of Mental Health-Genetics Initiative Bipolar Disorder (NIMH-GIBD) and Genetics of Recurrent Early-onset Major Depression (GenRED) consortium datasets, consisting of 2021 mood disorder cases and 1840 controls.

Thus, the results of a number of recently published studies in multiple populations have repeatedly linked polymorphisms in CACNA1C to multiple psychiatric disease diagnoses and indicate that associations between CACNA1C polymorphisms are not unique to one disorder (Table 1). The evidence implicates CACNA1C as currently the most statistically robust finding when combining multiple psychiatric disease categories. While the most statistically robust CACNA1C associations are in bipolar disorder, as compared to depression and schizophrenia, these data suggest that CACNA1C belongs to a class of shared susceptibility factors. These data support a conclusion that diagnostic categories are not appropriate for delineating the downstream effects of genes on behaviors.

Thus, CACNA1C is a common risk factor for which genetic evidence indicates that the identified intronic SNPs have significant correlation with pathological behaviors. The associated genetic changes in CACNA1C likely act through a common developmental and/or physiological process of central nervous system function in a manner that is common across current diagnostic categories, and at a level upstream from symptoms. Consistent with this hypothesis, as we discuss in more detail later, recent reports have associated the most consistently reported disease-associated SNP (rs1006737) with variations in brain structure and function in psychiatrically healthy human subjects. However, the mechanisms underlying how genetic changes in CACNA1C modify risk for developing psychiatric disorders are unclear. Indeed, defining the implications of such CACNA1C genetic change in a context relevant to psychiatric disorder pathophysiology is essential.

2. CACNA1C genetic variation

CACNA1C is located on the short arm of chromosome 12p13.3, spanning an approximately 6.45 Mb genomic region (Gene ID 775). The gene consists of at least 55 exons (Soldatov, 1994), which span an approximately 740 kb region. A number of SNPs in CACNA1C, concentrated within the large 328.5-kb intron 3, have been linked to psychiatric illness (Table 1), with limited information currently available regarding functional consequences. The locations of these SNPs within the gene are detailed in Fig. 2. Given their location in a non-coding region of the gene, they are not expected to directly interfere with the structure-functional properties of Ca_v1.2 (e.g., as in the case of G402S in Timothy syndrome). However, a likely consequence of intronic variation is regulation of expression of the channel. Beginning to address these questions Bigos et al. reported an association between total expression of CACNA1C and a SNP in complete LD with rs1006737 in an earlier microarray based expression analysis of postmortem brains (Bigos et al., 2010). It is currently not known whether this effect is constant throughout development, and in what specific tissues or brain regions.

A few studies have assessed for an imbalance in the relative expression of the two CACNA1C alleles, referred to as allelic expression imbalance (AEI) (Yan et al., 2002). Estimating AEI involves measuring the relative expression of the alleles of a given gene within the same cell or the same tissue sample. This approach minimizes the non-allelic factors, such as environmental exposures, that can modify gene expression and maximizes the identification of effects caused by variation in individual alleles. While AEI for CACNA1C was not identified in cardiovascular tissues from transplant recipients (Johnson et al., 2008), lymphoblastoid cell lines from normal controls did show AEI at some SNPs (Quinn et al., 2010). Interestingly, these other SNPs are not in LD with SNPs identified in GWA studies (Quinn et al., 2010).

It is known that the CACNA1C transcript is subject to extensive alternative splicing that potentially could generate thousands of splice variants. However, complementary DNA (cDNA) cloning experiments have yielded a much more limited number of partially overlapping sequences that comprises of, at a minimum, 40 known splice variants (Tang et al., 2004; Tiwari et al., 2006) of which at least 23 are protein coding (National center for Biotechnology information database). The high degree of variation is due to several different mechanisms of alternative splicing, including splicing of exons at alternate junctions, the presence of mutually exclusive exons, exons that either splice as alternate exons or at alternate donor sites, and through utilization of alternate promoters (Abernethy and Soldatov, 2002; Dai et al., 2002; Pang et al., 2003; Perez-Reyes et al., 1990a,b).

CACNA1C is expressed in many tissues in the human body across various developmental stages (Pang et al., 2003). Its expression has been observed in embryonic, fetal, infant, and adult stages. The gene is widely expressed in the brain (especially hippocampus and thalamus), cerebral arteries, whole heart, cardiac muscle and fibroblasts (Narayanan et al., 2010; Perez-Reyes et al., 1990b; Powers et al., 1991; Schultz et al., 1993; Soldatov, 1992). Expression of the transcripts are species- and tissue-dependent, developmentally regulated and is affected by pathophysiological conditions (Barad, 2003; Brillantes et al., 1994; Feron et al., 1994; Hofmann et al., 1994; Liao et al., 2007; Obermair et al., 2004; Tang et al., 2009; Tiwari et al., 2006). The majority of known splice variation, which predominates on the 3′ end of the gene, is not in proximity to genetic changes associated with mood disorders and schizophrenia (Fig. 2). However, the diversity in phenotypes of CACNA1C splice variants is relevantin studying the functioning of the gene in psychiatric disorders and may have importance in understanding disease mechanisms (Wang et al., 2006). It also provides avenues to interrogate potential targets for drug development and selective modification of physiological functioning.

3. Ca_v1.2 calcium channel

The Ca_v1.2 α1C subunit consists of four repeating domains (I–IV), each with six transmembrane α helical segments (S1–S6) that are connected by intra- and extra-cellular loops (Abernethy and Soldatov, 2002; Mikami et al., 1989; Tanabe et al., 1987) as well as a long C-terminus and shorter N-tail that both point to the cytoplasm. In neurons of the central nervous system, Ca_v1.2 channels are located primarily in the postsynaptic dendritic processes and somata, and are distributed throughout the dendritic tree (Leitch et al., 2009). This somatodendritic localization of Ca_v1.2 places it at a key position to couple neuronal excitation with Ca_v1.2-mediated Ca²⁺ signaling that modulates gene transcription making Ca_v1.2 an effective regulator of signaling pathways. The widespread downstream effects of Ca_v1.2 and other LTCCs are achieved through CaM associated with CaM-binding sites situated in the α_1C C-terminal tail (Halling et al., 2005) and some other signaling molecules localized probably close to the channel pore. CaM is an essential component of the mechanism of calcium-dependent inactivation guarding against calcium overloading during long action potentials (Zuhlke et al., 1999). Other data suggest that calcium binding protein-1 (CaBP1) may also be involved in regulation of Ca_v1.2 activity (Oz et al., 2011; Zhou et al., 2004). A key downstream mechanism of calcium signaling is calcium-induced intracellular calcium release (CICR) through ryanodine receptors (RyR), which has been widely demonstrated in cardiac and skeletal muscle LTCCs. CICR mediated by ryanodine-sensitive stores in hippocampal and other central neurons is not affected by calcium channel blockers (Sandler and Barbara, 1999), although there is some data indicating that neuronal LTCCs may in fact mediate CICR (Kim et al., 2007; Kolarow et al., 2007).

While the full details of LTCC signaling to the nucleus are yet to be elucidated, studies have extensively linked LTCC-mediated calcium influx to the Ca²⁺/CaM-dependent protein kinases (CaMK) cascade and the mitogen-activated protein kinase (MAPK) pathway. Immunoprecipitation studies have found that Ca_v1.2 is closely associated with CaMKII that is tethered to the channel by a binding site on the COOH terminus of the α_1C subunit in heart tissue (Hudmon et al., 2005) and to the Ca_vβ₁ and β₂ subunits in forebrain tissue (Abiria and Colbran, 2010). Such molecular association may be spatially and temporally flexible because Ca_vβ subunits tend to form homo- and hetero-oligomers at the functional Ca_v1.2 calcium channels (Lao et al., 2010). This localization of CaMKII close to the channel has been shown to be important in integrating and coupling local calcium increases to cyclic adenosine monophosphate (cAMP) response element binding protein (CREB)-dependent gene transcription (Wheeler et al., 2008). There is also evidence that CaM bound to LTCCs is particularly effective in activating the MAPK cascade leading to sustained CREB phosphorylation (Dolmetsch et al., 2001). For example, recent work has shown that inhibition of extracellular signal-regulated kinases (Erk) 1 and 2 increases synaptic vesicle release by increasing the number of surface-bound LTCCs suggesting a mechanism by which LTCCs may be involved in modulation of synaptic transmission (Subramanian and Morozov, 2011).

One of the most robust effects of postsynaptic LTCC calcium influx is activation of CREB-dependent transcription (Bito et al., 1996; Deisseroth et al., 1998; Dolmetsch et al., 2001; Weick et al., 2003; West et al., 2001). The investigation of CREB-dependent transcriptional regulation by the LTCC and cAMP established their role in activation of transient and stable signaling sub-microdomains of mixed and individual type (Kobrinsky et al., 2011). These discrete signaling microdomains might be differentially associated with such specialized processes as spontaneous contractions of muscle cells or the long-term changes underlying synaptic plasticity. While a number of genes are regulated by changes in free calcium generally, and CREB activity specifically, experiments show other levels of regulation. For example, BDNF is preferentially regulated by calcium influx through LTCCs when compared to calcium influx through N-methyl-D-aspartate (NMDA) receptors (Ghosh et al., 1994; Tabuchi et al., 2000). Specifically, calcium influx through LTCCs preferentially activates BDNF promoter I while BDNF promoter III is driven by LTCCs and NMDA receptors equally.

Little is known about the role of Ca_v1.2 during brain development, however, LTCCs in general are recognized to play a role in the development of ventral hippocampal parvalbumin containing interneurons (Jiang and Swann, 2005), as well as in calcium oscillations that occur in developing Purkinje neurons (Liljelund et al., 2000). It is worth noting that parvalbumin interneurons are reproducibly reported decreased in the brains of individual with schizophrenia (Lisman et al., 2008). Additionally, LTCCs are involved in gamma-aminobutyric acid (GABA) mediated calcium increases in embryonic hypothalamic neurons (Obrietan and van den Pol, 1995). LTCCs also mediate the increased amplitude of Ca²⁺ transients after activation of GABA-A receptors by muscimol in embryonic hypothalamic neurons after estradiol treatment (Perrot-Sinal et al., 2001), as well as the increased phosphorylation of cyclic AMP response element binding protein found in the hippocampus and hypothalamus of newborn male rats (Perrot-Sinal et al., 2003).

Ca_v1.2 has been associated with changes in calcium signaling and its effects in aging. For example, phosphorylation of Ca_v1.2 is increased in the hippocampus of aged rats, possibly accounting for the age-related increase in neuronal Ca2+ influx (Davare and Hell, 2003). Additionally, Ca_v1.2 channels have been shown in vitro to interact with amyloid precursor protein (APP), and APP deletion in mice leads to both increased Ca_v1.2 levels and GABAergic calcium currents in the striatum and hippocampus (Yang et al., 2009). Much work is needed to define how alterations in Ca_v1.2 levels and function in development and aging may progress into psychiatric dysfunctions.

4. Function of Ca_v1.2 in the brain circuits and behaviors of rodents

The importance of Ca_v1.2 channels in regulating the functions of brain circuits and behaviors has been shown through administration of LTCC agonists and antagonists, as well as using several genetic knockout approaches. A germline knockout of Cacna1c is embryonically lethal (Seisenberger et al., 2000), however conditional knockout and heterozygous mouse models have been successfully used. LTCCs have been found to play a role in behaviors mediated by the mesolimbic pathway. This pathway traditionally defines the ventral tegmental area (VTA) to nucleus accumbens (NAc) circuit, modulated by interactions with the hippocampus, medial prefrontal cortex, and amygdala (Pierce and Kumaresan, 2006), and is thought to be critical for reward and motivation that can be altered in mood disorders and schizophrenia (Juckel et al., 2003; Nestler and Carlezon, 2006). In rats, administration of the hydrophobic dihydropyridine LTCC antagonist isradipine dose-dependently attenuates intake of sweetened drinking water, a rewarding liquid (Calcagnetti and Schechter, 1992), as well as the reinforcing effects of amphetamine (Pucilowski et al., 1995), while nimodipine suppresses the effect of nicotine (Biala, 2003) in conditioned place preference. Additionally, there is evidence that the LTCC antagonist D-cis-diltiazem enhances the rewarding effects of cocaine in the conditioned place preference test when injected in the ventral NAc shell (Chartoff et al., 2006). In mice sensitized to psychostimulants, increases in Ca_v1.2 protein levels in the frontal cortex and limbic forebrain were found (Ford et al., 2009; Shibasaki et al., 2010), as well as increased surface expression of Ca_v1.2 channels in pyramidal neurons of the medial prefrontal cortex (Ford et al., 2009).

Ca_v1.2 channels have also been shown to be important during sensitization to stimulants, a behavior mediated in part by VTA to NAc pathway. In rats sensitized to amphetamine, there is an up regulation of the Ca_v1.2 channel mRNA and protein in dopaminergic neurons in the VTA (Rajadhyaksha, 2004). Repeated micro-injections of LTCC dihydropyridine agonist BayK 8644 in the VTA of rats resulted in an increased behavioral response to cocaine (Licata et al., 2000), and BayK 8644-induced stimulation of LTCCs in the caudate putamen increases extracelluar dopamine levels as measured by microdialysis (Okita et al., 2000). Additionally, it has been shown that peripheral administration of dihydropyridine antagonists attenuates sensitization to stimulants (Biala and Weglinska, 2004; Mills et al., 2007). Some data suggest that Ca_v1.2 channels are important for the behavioral expression of sensitization, but not the development of sensitization (Giordano et al., 2010; Schierberl et al., 2011).

Dysfunction of the hippocampus has been implicated in the pathogenesis of both mood disorders and schizophrenia, and may be relevant to emotional instability and hippocampal memory dysfunction observed in both disorders (Small et al., 2011). A region specific knockout of Cacna1c in the hippocampus and neocortex of mice leads to a loss of Shaffer collateral/CA1 late-phase long-term potentiation (LTP) (Moosmang et al., 2005). Ca_v1.2 channels regulate excitability, and have been linked to decreased activation of the MAPK pathway and reduced CREB-dependent transcription, in CA1 pyramidal neurons (Lacinova et al., 2008; Moosmang et al., 2005). Conditional deletion of Cacna1c in the hippocampus and cortex results in hippocampal-dependent behavioral changes, such as impaired spatial memory in the Morris water maze and in a labyrinth maze test (Moosmang et al., 2005; White et al., 2008).

The amygdala plays a central role in regulating fear circuits (Moosmang et al., 2005) and overall limbic system function that is implicated in neuropsychiatric illness (Krishnan and Nestler, 2010). LTCCs are known to modulate amygdala function, as well as behaviors mediated by the amygdala. For example, they have been shown to be important for long-term fear memory formation and LTP in the lateral amygdala (Bauer et al., 2002; Langwieser et al., 2010), as well as for induction of LTP in the basolateral amygdala to dentate gyrus pathway after stimulation (Niikura et al., 2004). Pharmacological inhibitors have demonstrated a role for LTCCs in presynaptic LTP at corticoamygdala synapses (Fourcaudot et al., 2009). In behavioral tests of fear memory, peripheral administration of LTCC blockers interfere with extinction, but not acquisition of fear memory (Cain et al., 2002; Suzuki et al., 2004). Administration of LTCC blockers into the ventricles or directly to the amygdala interferes with acquisition, but not extinction of fear learning (Bauer et al., 2002; Langwieser et al., 2010). Using a conditional knockout mouse model, as well as pharmacological methods, it has been shown that Ca_v1.2 but not Ca_v1.3 is the relevant LTCC genetic isoform in the thalamus-amygdala pathway (Langwieser et al., 2010). Genetically modified mice expressing dihydropyridine-insensitive Ca_v1.2 (Ca_v1.2DHP−/−) were resistant to the effects of the dihydropyridine antagonist nifedipine to impair fear extinction, suggesting a crucial role for Ca_v1.2 channels in fear conditioning (Busquet et al., 2008). Supporting this conclusion, the same group reported that Ca_v1.3-deficient mice manifested extinction identical to the respective wild type mice (Busquet et al., 2008). Additionally, it has been found that amygdala levels of the Ca_v1.2 protein is increased in response to fear conditioning in rats (Shinnick-Gallagher et al., 2003). Ca_v1.2 channels also play a role in observational fear learning, as shown by impaired observational fear learning in a mouse model with a selective deletion of Ca_v1.2 channels in the anterior cingulate cortex (Jeon et al., 2010).

Cacna1c heterozygous knockout mice, expressing decreased Ca_v1.2 protein levels as well as decreased LTCC current as adults, were protected against depression-like phenotypes (Dao et al., 2010). This finding, although reported in both sexes, was more robust in females. This is consistent with the finding that some SNPs in CACNA1C are associated with a mood disorder diagnosis only in females (Dao et al., 2010). Of potential relevance, the gonadal steroid estrogen has been reported to induce calcium influx via LTCCs in cultured hippocampal neurons in a non-genomic manner (not involving estrogen receptors) (Sarkar et al., 2008). A variety of Ca_v1.2 antagonists have been shown to reduce despair-like behavior similar to existing antidepressant medications (Cohen et al., 1997; Czyrak et al., 1989, 1990; Mogilnicka et al., 1987; Sinnegger-Brauns et al., 2004). Genetically modified dihydropyridine-insensitive Ca_v1.2DHP−/− mice were resilient to the antidepressant-like effect of nifedipine in the forced swim test, suggesting that Ca_v1.2 mediates the antidepressant-like effects of LTCC antagonists (Sinnegger-Brauns et al., 2004).

Moosmang et al. (2005) reported that a conditional forebrain knockout of Cacna1c did not alter anxiety-like or exploratory behavior. However, while Cacna1c heterozygous knockout mice males showed normal performance in anxiety-related procedures, females showed increased anxiety or fewer risk taking behaviors (Dao et al., 2010). In a recent study Lee et al. (2012) showed that mice constitutively heterozygous for Cacna1c manifested anxiety-like behavior in the elevated plus maze test (adult females as well as adolescent males). The authors also generated forebrain specific conditional knockout mice and a prefrontal cortex specific cacna1c knockdown which both also resulted in increased anxiety like behavior (Lee et al., 2012). LTCC antagonists were reported to increase anxiety-like behavior in rats (Viveros et al., 1996), but the effect is apparently species-dependent, as anxiety-like behavior in mice was not affected (Biala and Kruk, 2007). Finally, it has recently been reported that mice expressing a constitutively active G406R mutation of Ca_v1.2 α_1C subunit, causative of Timothy syndrome in humans, manifested a number of behavioral changes reminiscent of autism including repetitive/perseverative behaviors, altered social interaction and ultrasonic vocalizations, and enhanced tone cued and contextual memory following fear conditioning (Bader et al., 2011).

5. Effects of CACNA1C on human brain function

Supporting the rodent data implicating Ca_v1.2 in basic brain function, a number of recent reports have associated a primary disease-associated SNP in CACNA1C, rs1006737, with variation in human brain function and structure in subjects who have no diagnosable psychiatric illness (Table 2). While polymorphisms in CACNA1C are significantly associated with bipolar disorder, depression, and schizophrenia, such genetic changes only increase probability of disease, and are not deterministic. Thus, a large percentage of the population who do not have disease also carry disease associated SNPs in CACNA1C, which allows for studies to dissect the effects of SNPs on brain function and structure. Specifically, multiple reports have associated rs1006737 genotype with differences in performance on a number of neuropsychological tasks in healthy control subjects. Krug et al. (2010) reported that the risk allele was associated with decreased performance on a test of semantic verbal fluency (Krug et al., 2010). The risk allele has also been associated with decreased attention-related performance in alerting and orienting as assessed with a scanner-adapted version of the Attention Network Test (Thimm et al., 2011), lower extraversion and higher harm avoidance, trait anxiety, and paranoid ideation scores as assessed via personality questionnaires, and increased startle reactivity (Roussos et al., 2011). Risk allele carriers were also reported to have manifested higher psychopathology scores for depression, anxiety, obsessive-compulsive thoughts, interpersonal sensitivity, and neuroticism (Erk et al., 2010). Strohmaier et al. provided an interesting report of sex-specific effect of CACNA1C SNP rs1006737 in a large population based cohort (Strohmaier et al., 2012). In their investigation of personality traits relevant to affective disorders and schizophrenia the authors found that the risk allele of rs1006737 was significantly associated with higher sense of coherence in women. On the other hand, this allele in men was associated with lower sense of coherence, albeit statistically not significant. Overall, Strohmaier et al. observed opposite direction of effect of the genotype of rs1006737 in males and females for most of the personality traits studied in the cohort. Zhang et al. (2011) reported that the risk allele was associated with impaired working memory in schizophrenia patients and healthy controls, but not bipolar patients. Soeiro-de-Souza et al. (2012) have also recently reported that the risk allele was associated with impairment in facial emotion recognition in bipolar patients, but not in healthy controls.

Consistent with these findings, a large number of functional neuroimaging studies have revealed an association between rs1006737 genotype and brain activity in healthy controls while undertaking specific tasks. To date these findings include changes in brain activity while performing an attention network task (Thimm et al., 2011), amygdala activity during emotional processing (Jogia et al., 2011; Wessa et al., 2010), bilateral hippocampal activation during episodic memory recall and diminished functional coupling between left and right hippocampal regions (Erk et al., 2010), increased activationof the leftinferior frontalgyrus aswellas the left precuneus during a separate semantic verbal fluency task (Krug et al., 2010), mediotemporal emotional processing and prefronto-cortical working memory processing (Bigos et al., 2010), and reduced corticolimbic/frontotemporal functional connectivity during emotional face-processing (Wang et al., 2011). Radua et al. (2012) found that this same SNP was associated with significant modulation of effective connectivityfrommedial frontalgyrus to left putamensignificantlyreduced during perception of fearfulfaces,but that the effect was primarily in BPD subjects (vs. unaffected relatives and healthy controls).

Three published reports have now indicated that the rs1006737 risk allele is associated with increased brain gray matter, either total gray matter volume or in specific brain regions in healthy controls (Kempton et al., 2009; Perrier et al., 2011; Wang et al., 2011). Kempton et al. (2009) reported an increase in total gray matter volume. In a larger study with 585 individuals, Franke et al. (2010), specifically attempted to replicate these findings, but could not indicating discordance among reports. It is possible that differences in total gray matter may be reflective of changes in select brain regions, as reported by Perrier et al. (2011) where increases in right amygdala and right hypothalamus were noted. Wang et al. (2011) reported an increase in gray matter volume in select brain regions including bilateral ventral, rostral, and dorsolateral prefrontal cortex, and anterior cingulate and temporal cortices. Perrier et al. (2011) also reported that there was a significant interaction between genotype and a bipolar disorder diagnosis whereby the left putamen was smaller in bipolar patients carrying the risk allele. Additionally, it has been reported that the minor allele for three CACNA1C SNPs (rs2051992, rs2239050, rs7959938) that are all in high LD with each other but distant from rs1006737 are associated with increased brainstem volume in healthy controls (Franke et al., 2010). In this study there was also an association between rs7959938 and white matter volume in the brainstem, largely overlapping in areas of the midbrain and pons, and also in cerebellum. It is unclear if these effects of CACNA1C SNPs on structural and volumetric outcomes arose during early development, or are only present in adults. However, it is clear that such changes on their own are not pathological as they were identified mostly in psychiatrically normal subjects (Table 2).

6. L-type calcium channel antagonists in the treatment of psychiatric disease

LTCC antagonists are mainly used clinically for treating high blood pressure, angina, and abnormal heart rhythms. Chemically, they are represented by three different structural classes: dihydropyridines, benzothiazepines, and phenylalkylamines, which all act by binding to different sites on Ca_v1.2 and blocking the calcium current (Triggle, 1992). LTCC antagonists within the same structural class may vary in their affinity for the same calcium channel subtype, as well as for their penetration of the blood-brain barrier. For example, nimodipine and nicardipine, unlike some other dihydropyridines, are considered to have good CNS penetration (Kaplan, 1989). LTCC antagonists have been assessed for the treatment of bipolar disorder and also unipolar depression (see Casamassima et al., 2010a and Levy and Janicak, 2000 for comprehensive reviews). However, studies have yielded mixed results suggesting response in only a subset of patients (most notably rapid cycling where patients transition between mood episodes occurs at a rapid rate) or that these treatments have not yet been optimized (e.g. selectivity for Ca_v1.2 or blood brain barrier penetrance (Casamassima et al., 2010a; Levy and Janicak, 2000; Mallinger et al., 2008; Pazzaglia et al., 1998)). Many of the early studies were performed with the phenylalkamine verapamil, which has low blood-brain barrier permeability. More recent double-blind placebo studies with dihydropyridines including nimodipine and nicardipine that more efficiently cross blood-brain barrier have suggested that this class of calcium channel medication has efficacy as an adjunctive antidepressant treatment in combination with selective serotonin reuptake inhibitors (SSRIs) or electroconvulsive therapy (Dubovsky et al., 2001; Taragano et al., 2001, 2005). To the best of our knowledge, LTCC antagonists have not been assessed for efficacy in the treatment of schizophrenia. More extensive investigation with LTCC antagonists that readily cross the blood brain barrier are clearly necessary to define the specific profile of this class of medications in bipolar disorder, as well as depression and schizophrenia.

7. Conclusions

Underlying genetics are among the strongest risk factors for the development of mood disorders and schizophrenia, and there is shared genetic overlap between DSM-defined disorders including bipolar disorder, unipolar depression, and schizophrenia. This shared genetic overlap is coincident with shared environmental risk factors, as well as extensive overlap in neurobiology–as evidenced by biological markers and endophenotypes–and in efficacious treatments (Demjaha et al., 2011; Gottesman and Gould, 2003; Insel and Cuthbert, 2009). This has contributed to the view that future classification systems for psychiatric illness should be agnostic to current criteria (Cuthbert and Insel, 2010; Insel et al., 2010) and that newtreatments that are basedupon pathophysiology will eventually emerge through a thorough understanding of genetic susceptibility factors (Insel and Scolnick, 2006).

SNPs in CACNA1C are associated with risk of developing diverse DSM-defined disorders including bipolar disorder, depression, and schizophrenia. In addition to being a current statistical “top hit” resultant from bipolar disorder GWA studies, it is also the top finding when combining results of depression or schizophrenia with bipolar disorder sample sets, reinforcing the well established notion that diagnostic categories are not appropriate demarcation points for the effects of genes on behaviors (Gottesman and Gould, 2003; Meyer-Lindenberg and Weinberger, 2006). Available data suggest that CACNA1C belongs to a class of shared susceptibility factors, which cross DSM diagnostic boundaries to influence susceptibility (Table 1).

We have reviewed these human genetic and clinical findings, as well as the biological functions of Ca_v1.2 that are likely most relevant to psychiatric disease. While involved in numerous cellular functions, Ca_v1.2 is strongly implicated in the coupling of cell membrane depolarization to changes in calcium permeability and resultant intracellular signaling pathway activity and ultimately gene transcription and synaptic plasticity. Mouse genetic knockout experiments and studies with LTCC antagonist and agonists indicate that Ca_v1.2 is involved in proper functioning of many brain regions, including the mesolimbic reward system, hippocampus, and amygdala. Much of the work exploring the effects of LTCC antagonists on activation of signaling pathways and rodent behavior have not been able to distinguish the effects of inhibiting all LTCCs, most notably Ca_v1.2 and Ca_v1.3 in the brain, though genetically modified mice expressing dihydropyridine-insensitive Ca_v1.2 (Ca_v1.2DHP−/−) have been generated and used to begin to address this critical question (Busquet et al., 2008; Giordano et al., 2010; Sinnegger-Brauns et al., 2004). Existing preclinical and some clinical pharmacological data support inhibiting LTCCs as a possible mechanism of mood stabilization.

Prior to the recent genetic findings, there was a priori evidence suggesting that calcium signaling may play a role in bipolar disorder. Studies of bipolar patients have consistently reported elevated basal and stimulated intracellular calcium levels in peripheral blood cells (Casamassima et al., 2010a; Dubovsky et al., 1989; Emamghoreishi et al., 1997; Perova et al., 2008). However, the relevance of these reports to the current genetic association with CACNA1C is presently unclear. Additional investigation of the role of Ca_v1.2 in brain function inboth humans and animal models will likely provide novel insights into the pathophysiology of psychiatric disorders and elucidate new approaches for their treatment.

It is evident from the increasing size of cases and controls that are being assessed and the odds ratios reported that the identified polymorphisms in CACNA1C impart only a small contribution to the overall heritability of psychiatric illness (Table 1). However, a number of reports now clearly indicate that genetic variant at SNP rs1006737 is associated with changes in brain structure and function in normal control subjects with a larger effect size (Table 2 and (Rose and Donohoe, 2012)). There is evidence indicating that genetic variation at rs1006737 may be involved in modulating gene expression, with the minor (risk) allele associated with increased expression levels of CACNA1C (Bigos et al., 2010). However, the mechanisms regarding how this gene is regulated by these SNPs or other SNPs in LD is not clear, and the resultant changes in Ca_v1.2 function and how it is related to disease have not been characterized. The mechanism by which genetic variation in CACNA1C influences risk is likely on the function of particular neuronal circuits, at a level far upstream from the development of disease or presentation of symptoms. Noting that CACNA1C is a susceptibility gene for diverse psychiatric diseases, these changes in circuit function likely converge in a manner that is common across diagnostic categories. It is, however, unclear if this effect of CACNA1C genetic variation is during development, in adults, or perhaps more likely both. A number of additional molecular genetic questions remain to be answered. These include the identification of CACNA1C splice variants associated with particular SNPs, and whether such an association is dependent on brain regions. It also remains critically important to fully address the consequences of changes in splice variation of CACNA1C on neurophysiology and whether these structural changes may lead to an identification of α_1C molecular determinant(s) to be used in high-throughput screening for research and development leading to new specific drugs to combat disease (Zuccotti et al., 2011). Defining the answers to these and other critical questions will likely provide important insight into neuropsychiatric disease susceptibility and potentially lead to novel therapeutic targets and approaches.

Acknowledgments

Abbreviations

References