INTRODUCTION

Although the prevalence of alcohol and drug use disorders has increased over the last several decades (Whiteford et al, 2013), few efficacious medications are available to treat alcohol and drug addiction. This is likely due to the complex interactions among genetic, epigenetic, and environmental factors that characterize addictive disorders. Based on results from meta-analyses and twin studies, an estimated 23–79% of the variance in alcohol and drug dependence is heritable (Agrawal et al, 2012), and matching variations in genes to therapeutics can improve treatment outcomes in individuals with alcohol use disorders (AUDs; Heilig et al, 2011; Kranzler and McKay, 2012; Sturgess et al, 2011). For example, clinical studies in individuals with AUDs have reported that carriers of the A118G polymorphism in the mu-opiate receptor gene OPRM1 show a better clinical response to treatment with naltrexone (Anton et al, 2008; Anton et al, 2012; Oslin et al, 2003; Ray and Hutchison, 2007; Schacht et al, 2013). Thus, identifying additional genetic factors that contribute to excessive use of alcohol or drugs is critical for developing more efficacious therapies for these disorders.

Small-conductance Ca2+-activated K+ (KCa2) channels link fluctuations in intracellular Ca2+ levels with functional changes in the membrane potential of neurons to regulate neuronal excitability, synaptic plasticity, and learning (Adelman et al, 2012). Cloning studies have identified three KCNN genes that encode KCa2 channel α subunits in the brain (KCa2.1: KCNN1 located on human Chr 19 and mouse Chr 8; KCa2.2: KCNN2, human Chr 2 and mouse Chr 18; and KCa2.3: KCNN3, human Chr 1 and mouse Chr 3; Kohler et al, 1996). Although KCa2 channel subunits are widely expressed throughout the brain, their expression pattern varies by brain region (Sailer et al, 2002; Sailer et al, 2004). KCa2.1 and KCa2.2 channels localize to the cerebral cortex, hippocampus, and lateral amygdala, and KCa2.3 channels are highly expressed in the thalamus, brain stem nuclei, and basal ganglia, including the nucleus accumbens (NAc) and ventral tegmental area (VTA). Although not extensively studied, there is evidence that KCa2.2 and KCa2.3 channel subtypes regulate different aspects of neuronal function. For example, only KCa2.2 channels affect stimulus-evoked excitatory postsynaptic potentials in hippocampal CA1 neurons (Bond et al, 2004; Hammond et al, 2006; Ngo-Anh et al, 2005). KCa2.3 channels influence the frequency of action potential (AP) firing and KCa2.2 and KCa2.3 channels regulate the precision of AP firing in dopamine neurons (Deignan et al, 2012; Wolfart et al, 2001).

In recent preclinical models, modulating KCa2 channel activity regulates alcohol and drug intake. In standard rodent drinking models administration of KCa2 channel-positive modulators reduced operant responding for alcohol and decreased voluntary intake levels (Hopf et al, 2010; Hopf et al, 2011; Padula et al, 2013). Positive modulation of KCa2 channels also attenuated the severity of handling-induced convulsions in alcohol-dependent mice and reduced acute alcohol withdrawal-induced hyperexcitability in cultured hippocampal slices (Mulholland, 2012; Mulholland et al, 2011). In addition, chronic alcohol exposure downregulated KCa2 channel function and expression in the hippocampus, NAc, and VTA (Hopf et al, 2010; Hopf et al, 2007; Mulholland et al, 2011). In contrast, repeated cocaine treatment and withdrawal enhanced the KCa2 channel-mediated component of the afterhyperpolarization (AHP) in NAc medium spiny neurons (MSNs; Ishikawa et al, 2009; Mu et al, 2010). Together, these findings provide compelling evidence that KCa2 channels are involved in alcohol and substance abuse.

To examine the role of the three genes that encode KCa2 channels in regulating alcohol-, nicotine- and illicit drug-related behaviors, the present study employed an integrative functional genomics approach and analyzed databases from clinical and preclinical studies. Results of that analysis demonstrated that the family of KCNN genes is present in multiple alcohol- and drug-related quantitative trait loci (QTL) and experimentally derived gene sets. Interestingly, KCNN2 and KCNN3 occurred in an alcohol preference QTL on chromosome 18 in genetically diverse BXD recombinant inbred (RI) strains of mice, as well as in an alcoholism susceptibility QTL on human chromosome 1. To better define the role of KCNN in voluntary alcohol consumption, we tested a panel of BXD RI strains along with the high-drinking C57BL/6J (B6) and low-drinking DBA/2J (D2) parental strains of mice in two well-established voluntary alcohol consumption models: the drinking in the dark (DID) procedure (Rhodes et al, 2005) and an alcohol dependence model of escalated drinking (Becker and Lopez, 2004; Lopez and Becker, 2005). We then used gene and protein expression analysis, electrophysiology, and pharmacological manipulation of KCa2 channel activity to correlate differences in drinking with the functional status of Kcnn.

MATERIALS AND METHODS

GeneWeaver Bioinformatics

The GeneWeaver software system (GeneWeaver.org) contains a curated database of functional genomic experimental results across nine species with many types of experimental data including gene expression analyses, QTL positional candidate sets, curated literature annotations for gene functions, chemical interactions, and mutation screens (Baker et al, 2012). This database was queried for ‘Kcnn1-3 alcohol’ and ‘Kcnn1-3 drugs’ to retrieve clinical and preclinical experiments that implicate these genes in alcohol-, nicotine-, and illicit drug-related experiments. Studies that did not explicitly focus on addiction were excluded. Using a conservative approach, we identified additional genes that were present in at least 70% of the gene sets. Genes were annotated using categories of the Gene Ontology (GO) Consortium (www.geneontology.org), and gene groups were analyzed with a Bonferroni adjusted over-representation test using the PANTHER (Protein ANalysis THrough Evolutionary Relationships) classification system (Mi et al, 2013).

Animals and Housing

Male and female C57BL/6J (B6) and DBA/2J (D2) parental strains were purchased from Jackson Laboratory (Bar Harbor, ME), and BXD RI strains were obtained from the University of Tennessee Health Science Center or Jackson Laboratory. Mice were individually housed in temperature- and humidity-controlled environments and kept on a 12-h light/dark cycle. Food and water were available ad libitum during all procedures. The Medical University of South Carolina Institutional Animal Care and Use Committee approved all procedures in accordance with NIH guidelines for the humane care and use of laboratory animals.

Chronic Intermittent Ethanol Exposure and Two-Bottle Choice Drinking

To establish baseline drinking, mice consumed alcohol in their home cage using a standard two-bottle choice (15% ethanol (v/v) vs water) limited-access (2 h) protocol for 6 weeks. Mice were then underwent four to five repeated weekly cycles of chronic intermittent ethanol (CIE) exposure in vapor inhalation chambers, alternated with weekly home cage drinking sessions (Becker and Lopez, 2004; Griffin et al, 2014; Lopez and Becker, 2005). Following the last vapor chamber exposure, mice remained abstinent from alcohol for 72 h prior to sacrifice and tissue collection for microarray analysis. Drinking data were analyzed by two-way ANOVA.

Microarray Data Generation and Analysis

Studies on correlations for Kcnn1-3 gene expression in NAc were completed using microarray datasets produced from BXD RI strains treated with either saline (Wolen et al, 2012) or the CIE protocol (Miles, Smith, Lopez, Becker and Williams, unpublished data). Tissue punch dissections, RNA isolation, and hybridization/scanning procedures with Affymetrix Mouse Genome 430 2.0 arrays (Affymetrix, Cat #900497) were performed as described previously (Wolen et al, 2012). Full presentation and discussion of the BXD CIE microarray data will be reported elsewhere (Miles, Smith, Lopez, Becker and Williams, manuscript in preparation). To minimize the risk of technical variation confounding the detection of an alcohol response, we simultaneously processed an air- and CIE-exposed mouse from an individual strain. Annotation data for Mouse Genome 430 2.0 probe sets were obtained from the GeneNetwork Data Sharing Zone (genenetwork.org/share/annotations). Expression data from the air and CIE treatment groups were background corrected, quantile normalized and summarized using the robust multi-array average (RMA) expression measure. All datasets generated for this paper are accessible on GeneNetwork (www.genenetwork.org) and from the Gene Expression Omnibus repository under accession number GSE28515. All data are MIAME compliant.

Phenotype Correlation Analysis

To avoid shared variance across traits inherent to within-subject designs that may artificially inflate correlations, we analyzed Kcnn1-3 transcript levels and drinking phenotypes using a between-subjects design across multiple cohorts of experimental mice. First, we conducted a linear regression analysis of Kcnn1 (probe set 1419617_at), Kcnn2 (1448927_at), and Kcnn3 (1421632_at) RMA levels in the NAc of alcohol-naive (saline treated) BXD RI mice from a previous study (Wolen et al, 2012) and baseline alcohol intake values from a separate cohort of corresponding BXD RI strains in the CIE-drinking model (as described above). Next, we conducted a linear regression analysis of NAc Kcnn1-3 RMA levels from air-exposed BXD RI mice and alcohol intake values from the 4-h session on day 4 of a binge alcohol drinking (DID) procedure (Jones, Lu and Williams, unpublished observation, GeneNetwork record ID 13565). Finally, drinking data in air- and CIE-exposed BXD RI mice and their associated NAc Kcnn3 RMA levels were assessed using a hierarchical linear model (HLM 7.0) with pre-post CIE exposure nested within subject and day further nested within pre-post CIE exposure. Pre-post CIE exposure was fit as a random effect.

Intracranial Stereotaxic Surgery and Microinjections

Adult male B6 mice underwent bilateral cannula placement surgery in the NAc at least 2 weeks prior to the start of the drinking paradigm. Mice received microinjections during the third, fourth, and fifth test drinking sessions as previously described (Griffin et al, 2014). Apamin (0–0.8 ng; Sigma-Aldrich, St Louis, MO) was freshly dissolved in sterile saline and was microinfused 30 min prior to the limited access drinking sessions in a dose volume of 0.25 μl/min over 2 min. Cannula placements were verified at the end of the study by cresyl violet staining. Mice were excluded from analysis if the cannula tip was determined to be outside of the NAc as previously described (Griffin et al, 2014). Drinking data in mice that received microinjections were analyzed by two-way ANOVA.

Electrophysiology Recordings

We used standard whole-cell patch-clamp electrophysiology procedures and measured intrinsic excitability, AHP amplitude, and KCa2-mediated tail currents in MSNs from adult C57BL/6J mice. Acute slices (300 μm thick) were prepared from C57BL/6J mice following 3–7 days withdrawal from four cycles of CIE exposure to match the time point of the apamin microinjection study. In this cohort of CIE-exposed mice used for recordings, the BEC was 206.8±11.03 mg/dl (n=8 per mice group). Mice were decapitated and the brain was immediately removed and placed in an ice-cold dissection solution consisting of (in mM): 200 sucrose, 1.9 KCl, 1.2 NaH2PO4, 6.0 MgCl2, 0.5 CaCl2, 0.4 ascorbate, 10 glucose, and 25 NaHCO3. Slices were incubated at 34°C for 30 min and then kept at room temperature for at least 30 min in artificial cerebral spinal fluid (ACSF) solution containing (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH2PO4, 1.3 MgCl2, 2.0 CaCl2, 0.4 ascorbate, 10 glucose, and 25 NaHCO3. After incubation, slices were transferred to a submerged recording chamber bathed in 34°C ACSF. Incubation and recording buffers were continuously aerated with 5% carbon dioxide/95% oxygen. The pH of all solutions listed above was adjusted to 7.3 using NaOH, and osmolarity was measured to be ~300 mOsm. Recordings were made with a Multiclamp 700B amplifier (Axon Instruments, Union City, California) connected to a computer running Windows XP and Axograph X software. Patch pipettes (2–3 MΩ resistance) were pulled from thin-wall borosilicate glass (1.17 mm ID; Warner Instruments, Hamden, CT, USA) and filled with internal solution (pH 7.3 using KOH, osmolarity ~290 mOsm) containing (in mM): 140 K-methylsulfate, 8.0 NaCl, 1.0 MgCl2, 0.05 EGTA, 10 HEPES, and 5 Mg2 ATP. Currents were acquired at 10 kHz using an ITC-18 interface. To activate KCa2-mediated tail currents, neurons were held at −70 mV and depolarizing voltage steps (400 ms, from −30 to +10 mV in 10 mV increments) were applied with 2 s between each step. The protocol was repeated three times followed by a 10-min bath application of 100-nM apamin and three additional recordings. The data were averaged for each of the three protocols, and peak amplitude of the KCa2-mediated tail current that occurred up to 100 ms after each depolarizing pulse was measured. To determine the effects of withdrawal from CIE on AP spiking of neurons in the NAc core, current-clamp recordings were performed. Spike firing was induced by direct current injection (from 0 to 220 pA with 20 pA increments) through patch pipettes filled with K-methylsulfate internal solution (recipe as described above). Recordings were analyzed for number of spikes in response to each injected current, resting membrane potential (mV), frequency (Hz), interspike interval (ISI, msec), and AHP (mV). The first three AHP magnitudes were averaged, in which each AHP magnitude was calculated by subtracting the lowest potential during hyperpolarization from AP threshold. Averaged peak amplitude of the KCa2-mediated tail currents at each step before and after application of apamin was analyzed by two-way ANOVA. Differences in AP and AHP parameters between air-treated and CIE groups were analyzed by two-way ANOVA or unpaired two-tailed t-test.

Western Blotting

Tissue punches targeting the NAc core were extracted from a separate cohort of male B6 mice immediately following completion of their last drinking test session, flash frozen and stored at −80°C until processing. Samples were solubilized in 2% LDS and western blotting was performed using a primary antibody directed against KCa2.3 channels (1 : 4000; Alomone Labs, Jerusalem, Israel). An aliquot of each sample was taken for determination of protein concentration by the bicinchoninic acid assay (Pierce Biotechnology, Rockford, IL). Samples were diluted with NuPAGE 4X LDS sample loading buffer (Invitrogen Corp., Carlsbad, CA; pH 8.5) containing 50-mM dithiothreitol, and denatured for 10 min at 95°C. Approximately, 5 μg of each sample was separated using the Bis-Tris (375 mM resolving buffer and 125-mM stacking buffer, pH 6.4; 7.5% acrylamide) discontinuous buffer system with MOPS electrophoresis buffer (50-mM MOPS, 50-mM Tris, 0.1% SDS, and 1-mM EDTA, pH 7.7). Protein was then transferred to Immobilon P PVDF membranes (Millipore, Bedford, MA) using a semi-dry transfer apparatus (Bio-Rad Laboratories, Hercules, CA). After transfer, blots were washed with phosphate-buffered saline containing 0.1% Tween 20 (PBST) and then blocked with PBST containing 5% nonfat dried milk (NFDM) for 1 h at room temperature with agitation. The membranes were then incubated overnight at 4°C with the primary antibody diluted in PBST containing 0.5% NFDM and washed in PBST prior to 1-h incubation at room temperature with horseradish peroxidase-conjugated secondary antibodies diluted 1 : 2000 in PBST. Membranes received a final wash in PBST and the antigen-antibody complex was detected by enhanced chemiluminescence using a ChemiDoc MP Imaging system (Bio-Rad Laboratories, Hercules, CA). Bands were quantified by mean optical density using computer-assisted densitometry with ImageJ v1.41 (National Institutes of Health, USA). Because controls (eg, actin and GAPDH) used to normalize protein loading in western blot experiments can actually cause quantitation errors (Aldridge et al, 2008; Dittmer and Dittmer, 2006; Welinder and Ekblad, 2011), we used methods described by Aldridge et al, 2008 for normalizing to a total protein stain (Swift membrane stain; G Biosciences, St Louis, MO). Normalized protein expression data were analyzed with a two-tailed t-test. Analysis of the optical density values for the total protein stain demonstrated that the amount of protein loaded did not differ between the samples from the air and CIE-exposed mice (t(11)=0.203; p=0.843; n=6–7 per group).

RESULTS

KCNN1-3 and Alcohol, Nicotine, and Drug Addiction

The plausibility of a role for the family of KCNN genes in the behavioral effects of alcohol and other drugs of abuse was evaluated by the GeneWeaver database search of previous whole genome studies in which KCNN1-3 has been associated with alcohol- and drug-related behaviors. As shown in Table 1, multiple alcohol- and drug-related QTLs and gene sets from human and rodent studies contain KCNN1-3 genes. Moreover, this search showed that exposure to alcohol, opiates, or nicotine also significantly altered expression of KCNN1-3 in the NAc, superior frontal gyrus, or basolateral amygdala (Supplementary Table 1). Findings from this integrative genomics approach provide strong evidence that the family of KCNN genes are involved in alcohol, nicotine and illicit drug addiction and dependence and are consistent with evidence suggesting genetic overlap in substance use disorders (Agrawal et al, 2012). Although we identified KCNN1 and KCNN2 in some gene sets related to drugs of abuse, only KCNN3 is contained in gene sets for all of the drugs of abuse. Thus, we next asked whether other genes in these QTLs and gene sets were highly related to alcohol and drug addiction. As shown in Table 2 and Supplementary Figure 1, 70% or more of these gene sets also contain eight other genes (S100A6, S100A8, S100A9, SYT11, SHC1, JTB, GATAD2B, and RAB13). To identify whether specific signaling pathways were present in this network, we analyzed these genes with an over-representation test using PANTHER software (Mi et al, 2013). There were two PANTHER protein classes (intracellular Ca2+-sensing proteins and calmodulin) and one GO molecular function category (calmodulin binding) that were significantly over-represented (Bonferroni-adjusted over-representation test; p<0.05) in this gene network. These results support the idea that genes related to intracellular Ca2+ signaling and KCa2 channel function are positional candidates that may contribute to substance abuse.

Table 1 Gene Sets from GeneWeaver Related to Alcohol-, Nicotine-, and Illicit Drug-Related Behaviors that Contain the Family of KCNN Genes
Full size table
Table 2 A Conservative Network of Genes Found Within 70% or more of Alcohol-, Nicotine-, and Illicit Drug-Related GeneWeaver Gene Sets that also Contain KCNN3
Full size table

Kcnn3 Levels in the NAc Predict Drinking in Mice

Because KCNN2 occurred in an alcohol preference QTL on chromosome 18 in female BXD mice and KCNN3 occurred in an alcoholism susceptibility locus on human chromosome 1, we explored the relationship between KCNN genes and alcohol intake. First, we correlated NAc Kcnn1-3 RMA values obtained from alcohol-naive mice (Wolen et al, 2012) with baseline drinking data obtained from corresponding BXD RI strains of mice in the present study (n=14 strains; 106 total mice). Transcript levels of each KCNN gene in the NAc varied across the BXD strains (Kcnn1: 7.58±0.114 (mean transcript expression levels±SD), 1.53-fold range; Kcnn2: 9.59±0.099, 1.30-fold range; Kcnn3: 7.46±0.112, 1.44-fold range; Figure 1a). Similarly, average daily voluntary alcohol consumption by BXD mice during the 6-week baseline period varied across strains (Figure 1b). Linear regression analyses demonstrated that Kcnn3, but not Kcnn1 or Kcnn2 RMA values, significantly predicted average daily baseline drinking levels in these mice (Kcnn1: R2=0.033, p=0.54; Kcnn2: R2=0.024, p=0.57; Kcnn3: R2=0.373, p<0.05; Figure 1c) with high transcript levels of Kcnn3 in the NAc of BXD strains associated with low intake of alcohol. To further explore the negative relationship between Kcnn3 transcript levels and alcohol drinking, we analyzed Kcnn1-3 RMA values in air-exposed mice and drinking levels of BXD strains from the DID procedure (GeneNetwork record ID 13565; Supplementary Figure 2a). As shown in Supplementary Figure 2b, alcohol intake during the 4-h drinking session varied across BXD strains on day 4 of the DID procedure. Similar to results in the two-bottle choice limited-access model, RMA values for Kcnn3, but not Kcnn1 or Kcnn2, significantly predicted binge-like drinking in the DID model (Supplementary Figure 2c). In both the two-bottle choice and DID limited access models, Kcnn3 RMA values in the NAc were negatively associated with alcohol intake in BXD mice.

Figure 1
figure 1

Kcnn3 RMA values in the NAc of saline-treated BXD RI mice predict average daily voluntary alcohol (15% v/v) consumption in BXD RI strains of mice. (a) Kcnn1-3 RMA values in the NAc of alcohol-naive saline-treated BXD RI mice. (b) Average alcohol consumption levels during the 6-week baseline period in air-exposed BXD strains in the two-bottle choice limited access model (n=14 strains, 106 total mice). (c) Regression analysis of Kcnn1-3 RMA values in the NAc of saline-treated mice and average baseline alcohol intake levels in non-dependent BXD RI strains of mice (Kcnn1: R2=0.033, p=0.54; Kcnn2: R2=0.024, p=0.57; Kcnn3: R2=0.373, *p<0.05).

PowerPoint slide

Full size image

Next, we used a HLM to analyze the relationship between Kcnn3 RMA values and alcohol consumption in BXD mice exposed to the CIE treatment. We selected HLM regression for this analysis because the predictor variables of the CIE dependence model vary at more than one level, and HLM can account for the shared variance in hierarchical data (Woltman et al, 2012). As shown in Figure 2a, a significant negative relationship occurred between Kcnn3 RMA values in the NAc and the amount of alcohol consumed during the 6-week baseline period of the study (air controls: b=−1.99, SE=0.76, t(52)=2.61, p=0.012; CIE mice: b=−2.61, SE=0.96, t(52)=2.72, p=0.009; n=28 strains, 114 total mice). These results are consistent with those depicted in Figure 1c and Supplementary Figure 2c, and suggest that Kcnn3 expression in the NAc predicts drinking among non-dependent BXD strains. In addition, the relationship between amount consumed and Kcnn3 RMA values in the NAc of dependent mice was even more robust (Kcnn3 slope x treatment interaction: b=−1.12, SE=0.55, t(51)=2.04, p<0.05). That is, the Kcnn3 slope was more extreme and declined more rapidly with increasing Kcnn3 RMA levels in CIE-exposed BXD mice compared to air-exposed controls (Figure 2b). Although only a modest effect, intercept values (drinking extrapolated to 0 Kcnn3) were significantly higher in CIE-exposed BXD mice than controls (intercept values: b=7.63, SE=3.79, t(51)=2.01, p<0.05).

Figure 2
figure 2

Kcnn3 transcript levels in the NAc predict baseline and CIE-induced escalation of drinking in BXD RI strains of mice. (a) A negative relationship occurred between Kcnn3 RMA values in the NAc and average alcohol consumption levels during the 6-week baseline period in air- and CIE-exposed BXD strains (air controls: b=−1.99, SE=0.76, t(52)=2.61, *p=0.012; CIE mice: b=−2.61, SE=0.96, t(52)=2.72, *p=0.009; n=28 strains, 114 total mice). (b) Graph of NAc Kcnn3 RMA levels in air- and CIE-exposed mice in relation to average alcohol drinking during the test drinking sessions (Kcnn3 slope (ie, slope by treatment interaction): b=−1.12, SE=0.55, t(51)=2.04, *p<0.05; intercept values: b=7.63, SE=3.79, t(51)=2.01, *p<0.05).

PowerPoint slide

Full size image

Blocking KCa2 Channels in the NAc Modulates Drinking in B6 Mice

The relationship between Kcnn3 expression levels in the NAc during baseline drinking and following CIE exposure suggests that KCa2.3 channels in the NAc may regulate consumption and escalation of drinking in CIE-exposed mice. To further test this hypothesis, we bilaterally microinjected the potent and selective KCa2 channel allosteric inhibitor apamin into the NAc of air control and CIE-exposed B6 mice. Micro-infusions began during the third drinking test cycle and continued through the fourth and fifth test cycles when higher drinking levels are typically observed in CIE-exposed mice (Griffin et al, 2014; Griffin et al, 2009). We microinjected vehicle (0.9% sterile saline) or apamin (0.2–0.8 ng/0.5 μl) 30 min prior to each drinking session in a quasi-random manner. As expected, CIE-exposed mice microinfused with vehicle consumed significantly more alcohol than air-exposed mice (two-way ANOVA, group x dose interaction, F(3,126)=4.481, p=0.005, SNK post-hoc, p<0.05, p<0.001, n=8–32 per group; Figure 3a). Microinfusion of a 0.4-ng dose of apamin significantly increased consumption in air-exposed mice, but not in dependent mice (Figure 3a). Although the high dose of apamin (0.8 ng) did not affect drinking in controls, it unexpectedly decreased voluntary consumption in CIE-exposed mice (Figure 3a). Histological confirmation of cannula placement is shown in Figure 3b and c. Monitoring behavior during and after the drinking sessions revealed an enhanced behavioral sensitivity to apamin in CIE-exposed mice that may have affected drinking. Thus, while only one of nine air-control mice displayed mild behavioral impairment (ie, involuntary twitching) after drug administration, 50% of the CIE-exposed mice developed spontaneous tonic/clonic seizures following the high-dose apamin infusion (χ2 for trend (1, n=15)=4.650, p<0.05; Figure 3d).

Figure 3
figure 3

Pharmacological inhibition of KCa2 channels in the NAc modulates alcohol drinking and behavior differently in air- and CIE-exposed B6 mice. (a) Apamin (0–0.8 ng in 0.5 μl) was microinjected into the NAc of air controls and CIE-exposed mice (two-way ANOVA, group x dose interaction, F(3,126)=4.481, p=0.005, SNK post hoc, *p<0.05, **p<0.001, n=8–32 per group). (b) Representative cresyl violet stained tissue sections showing typical placements for the bilateral microinjector guides and tips. (c) A schematic of the plates from the Paxinos (2007) atlas demonstrates the locations of the tips of the microinjection cannula targeting the NAc. (d) Microinjection of 0.8 ng of apamin into the NAc of dependent C57BL/6J mice significantly increased the occurrence of seizure-like activity (0=normal behavior, 1=myoclonic jerks, 2=spontaneous tonic/clonic seizures; χ2 for trend (1, n=15)=4.650, *p<0.05).

PowerPoint slide

Full size image

CIE-Induced Functional Adaptations in Intrinsic Excitability and KCa2 Channels in the NAc

To determine if hypersensitivity of CIE mice to apamin was due to functional adaptations, we first examined CIE-induced changes in intrinsic excitability of MSNs in the NAc core using whole-cell slice electrophysiology. We performed the recordings at the same time point as the drinking studies to parallel the apamin microinjection experiment described above. CIE exposure did not alter the resting membrane potential (air: −69.6±1.36 mV, CIE: −70.85±1.37; t-test, t(18)=0.697, p=0.49, n=8–12 per group), but significantly increased basal firing rates of MSNs in the NAc core (two-way ANOVA, F(1,162)=35.34, p<0.0001; Figure 4a). CIE exposure also markedly enhanced AP frequency (t-test, t(18)=3.395, p=0.003, n=8–12 per group; Figure 4b) and decreased AP ISI (t-test, t(18)=3.618, p=0.002, n=8–12 per group; Figure 4c). Since KCa2 channels regulate intrinsic excitability by enhancing the AHP of NAc core neurons (Hopf et al, 2010), we examined the effects of CIE exposure on the AHP amplitude. CIE mice showed a significant reduction in AHP amplitude (t-test, t(18)=3.078, p=0.007, n=8–12 per group; Figure 4d), providing additional evidence that CIE exposure decreases KCa2 channel function.

Figure 4
figure 4

CIE exposure enhances intrinsic excitability and reduces the amplitude of the afterhyperpolarization (AHP) and KCa2 channel-mediated tail currents in MSNs in the NAc core. (a) Representative traces and quantitation show enhanced current-evoked AP firing in CIE-exposed mice (*p<0.05; n=8–12 neurons per group). (b, c) CIE exposure significantly alters the frequency and the interspike interval of MSNs. (d) CIE exposure reduces AHP amplitude (**p<0.01; n=8–12 neurons per group). (e) Representative trace showing a voltage step from−70 to +10 mV and the apamin-sensitive tail current that occurs after returning to –70 mV. (f) Examples of traces from tail currents recorded from NAc core MSNs in air- and CIE-exposed mice. Also shown are traces demonstrating the ability of 10 min bath application of 100 nM apamin to block these tail currents in air-exposed mice. (g–i) Quantitation of KCa2 channel-mediated tail currents in air- and CIE-treated mice before and after bath application of apamin (*p<0.05, **p<0.01; n=11–15 neurons per group).

PowerPoint slide

Full size image

We then applied depolarizing voltage steps in MSNs to activate KCa2 channel outward tail currents. As shown in Figure 4e–g, bath application of 100 nM apamin blocked the peak tail current in air-exposed mice (two-way ANOVA, F(1,65)=9.671, p=0.002, n=11 neurons). CIE exposure significantly reduced the peak tail current in comparison with currents recorded from controls (two-way ANOVA, F(1, 24)=7.925, p=0.009, n=11–15 neurons; Figure 4f and h), without altering cell capacitance values (two-tailed t-test, t(24)=1.55, p=0.134, data not shown). Bath application of apamin did not affect the peak tail currents in MSNs recorded from dependent mice, in contrast with control mice (two-way ANOVA, F(1, 115)=0.1883, p=0.67, n=15 neurons; Figure 4f and i). Altogether, these data suggest that CIE exposure produces profound adaptations in KCa2-mediated processes likely caused by a downregulation of functional KCa2.3 channels expressed in MSNs in the NAc core.

In a separate cohort of B6 mice (n=17–19/group), baseline alcohol consumption was established and then the mice were treated with repeated weekly cycles of CIE exposure alternated with test drinking sessions. As shown in Figure 5a, CIE exposure significantly increased drinking in comparison with air-exposed controls during test cycles 1 and 3 (two-way ANOVA, group x test session interaction, F(4,136)=3.190, p=0.015, SNK post-hoc, p<0.05 vs air). We extracted tissue punches targeting the NAc core from these air- and CIE-exposed mice after their last drinking session and used western blotting to examine protein expression. Consistent with the electrophysiology results, CIE exposure significantly reduced KCa2.3 channel protein levels in the NAc core (t(11)=2.511, p<0.05, n=6–7 per group; Figure 5b and c).

Figure 5
figure 5

Dependence-induced escalation of drinking reduces KCa2.3 channel expression in the NAc core. (a) CIE-induced escalation of alcohol drinking in dependent B6 mice (two-way ANOVA, group x test session interaction, F(4,136)=3.190, p=0.015, SNK post hoc, *p<0.05 vs air). (b) Representative images of the total protein stain (5 μg per lane) and blot of KCa2.3 channels in air- and CIE-exposed B6 mice. The dashed box in lane 1 represents the quantified area of the total protein stain. (c) Normalized protein expression levels of KCa2.3 channels in air- and CIE-exposed B6 mice that have been drinking alcohol in the CIE-induced escalation of drinking model (two-tailed t-test, t(11)=2.511, *p<0.05, n=6–7 per group).

PowerPoint slide

Full size image

DISCUSSION

A major finding of this study is that KCNN genes exist in multiple alcohol- and drug-related QTLs and gene sets from both clinical and preclinical studies. Moreover, alcohol, opiates, and nicotine alter expression of KCNN genes in critical brain regions, further implicating these genes in substance abuse. Our study also showed that KCNN3 is in a gene set for an alcoholism susceptibility locus located on human chromosome 1. In support of this clinical observation, Kcnn3 transcript levels in the NAc of genetically diverse BXD strains predicted baseline alcohol intake in two limited-access drinking models (two-bottle choice and DID). Alcohol-dependent mice with low Kcnn3 RMA levels consumed more alcohol than controls with equivalent Kcnn3, and higher Kcnn3 levels protected against CIE-induced escalation of drinking in alcohol-dependent mice. These findings suggest that low KCNN3 expression and function may be a risk factor for developing AUD.

Previous studies reported that chronic alcohol exposure reduced the function and expression of KCa2 channels in the NAc, VTA, and hippocampus of rodents (Hopf et al, 2010; Hopf et al, 2007; Mulholland et al, 2011). Extending these preclinical findings, the integrative functional genomic analysis identified a gene set showing a decrease of KCNN2 in the superior frontal gyrus and the basolateral amygdala of alcoholics (Ponomarev et al, 2012). Notably, in opposition to the effect of chronic alcohol exposure on KCa2 channels, we found that psychostimulants and heroin increased KCNN3 expression or function of KCa2 channels in the NAc (Albertson et al, 2006; Ishikawa et al, 2009; Mu et al, 2010; Wang et al, 2008; Supplementary Table 2). Chronic morphine treatment also altered KCNN1 transcript levels in the mouse NAc (Grice et al, 2007), although the authors did not report the direction of the change. The findings of the present study add to a growing literature showing that alcohol, nicotine, and illicit drugs significantly alter KCa2 channel expression and function in key regions within the addiction neurocircuitry.

Congruent with our findings that low Kcnn3 transcript levels in the NAc are associated with higher alcohol intake, we demonstrated that pharmacological inhibition of KCa2 channels in the NAc increased alcohol consumption in B6 mice. Previous evidence has demonstrated that positive modulators of the KCa2 channel decrease operant responding for alcohol and voluntary drinking in non-dependent rats and mice (Hopf et al, 2010; Hopf et al, 2011; Padula et al, 2013). Our data show that apamin can increase drinking, provide strong evidence that KCa2 channels in the NAc bi-directionally modulate alcohol intake. In addition to demonstrating the genetic relationship between Kcnn3 and alcohol intake in mice, the results of the present study also show that CIE exposure produces neuroadaptations in accumbal KCa2 channels. Microinfusions of a higher dose of apamin into the NAc produced spontaneous seizures only in dependent mice and this heightened susceptibility was associated with enhanced intrinsic excitability of NAc MSNs, a reduction in the amplitude of their AHP and KCa2-mediated tail currents and downregulation of KCa2.3 channel protein levels. These findings are consistent with the known role for the KCa2.3 channel as a low-pass filter that regulates firing frequency (Deignan et al, 2012; Wolfart et al, 2001). Since blocking KCa2 channels in control B6 mice increases drinking to levels comparable to those observed in dependent mice, we speculate that the neuroadaptations in KCa2.3 channels produced by CIE exposure may contribute to escalation of drinking associated with dependence.

Evidence suggests a genetic overlap in individuals addicted to alcohol, nicotine, or illicit drugs (Agrawal et al, 2012), and our functional genomic analysis implicates KCNN1-3 in QTLs and behaviors directly related to alcohol and drug intake. Along with KCNN3, results from our bioinformatic analysis in alcohol- and drug-related QTLs and gene sets identified a conservative subset of genes related to intracellular calcium signaling and calmodulin binding. These data suggest that these gene networks may contribute to individual differences in vulnerability to alcohol and drug abuse and dependence. Importantly, there is a genetic basis underlying the ability to remain abstinent during withdrawal from harmful drug intake (Crettol et al, 2006; Crettol et al, 2005; Fonseca et al, 2011; Uhl et al, 2008), and findings from genomic studies support the use of personalized medicines for treating addictions (Deschaux et al, 1997; Heilig et al, 2011; Kranzler and McKay, 2012; Sturgess et al, 2011). Indeed, matching SNPs in the OPRM1 gene to naltrexone treatment leads to more successful clinical outcomes in individuals with AUDs (Anton et al, 2008; Anton et al, 2012; Oslin et al, 2003; Ray and Hutchison, 2007; Schacht et al, 2013). In addition, polymorphisms in genes that encode methadone-metabolizing enzymes and transporter proteins contribute to low methadone plasma concentrations and ineffective methadone maintenance treatment programs (Crettol et al, 2006; Crettol et al, 2005; Fonseca et al, 2011). Studies also provide evidence that variants related to dopamine transmission can predict successful smoking cessation (Johnstone et al, 2004; Lerman et al, 2006; Swan et al, 2005; Uhl et al, 2008). Interestingly, KCNN3 is found in convergent genome-wide association studies of successful smoking abstainers (Uhl et al, 2008), and KCa2 channels can modulate evoked DA release and dopaminergic neurotransmission (Deignan et al, 2012; Herrik et al, 2012; Jacobsen et al, 2008; Soden et al, 2013). A number of polymorphisms in KCNN3 have been reported (Bowen et al, 2001; Chandy et al, 1998; Grube et al, 2011), and these polymorphisms reduce surface trafficking and currents from KCa2 channels (Chandy et al, 1998; Grube et al, 2011; Miller et al, 2001; Soden et al, 2013). However, whether individuals with KCNN3 polymorphisms that reduce KCa2 channel function have a higher risk for developing an AUD is still unknown. Thus, it will be important to determine whether KCa2 channel ligands are a suitable treatment for alcohol, nicotine, and illicit drug abuse, and whether KCNN polymorphisms predict effective treatment outcomes in substance abusers.

In the present study, blocking KCa2 channels in the NAc increased drinking presumably by enhancing neuronal activity of MSNs. MSNs in the dorsal and ventral striatum segregate into two distinct populations that are intermixed and indistinguishable in morphology. However, MSN subtypes express different genes, project to distinct brain regions, and have opposing roles in motivated and goal-directed behaviors (Heiman et al, 2008; Keeler et al, 2014). D1-dopamine receptor (D1DR)-expressing MSNs in the direct pathway project to the ventral pallidum, globus pallidus interna, substantia nigra pars reticulata, and VTA, while D2DR-expressing MSNs in the indirect pathway project to the globus pallidus externa. Although these pathways work together to balance ongoing normal behavior, mounting evidence suggests that drugs of abuse and chronic stress alters the balance between D1DR and D2DR MSN output (Francis et al, 2014; Kim et al, 2011; Lobo et al, 2010; MacAskill et al, 2014). Indeed, a recent report showed that CIE exposure and withdrawal prevented the induction of low frequency long-term depression (LTD) in D1DR-expressing MSNs and instead led to the occurrence of LTD in D1DR-non-expressing MSNs (Jeanes et al, 2014). Although we did not observe any apparent bimodal effects of CIE exposure on KCa2 channel function in the NAc core, it is possible that CIE exposure produces cell-type-specific effects on KCa2 channel function in D1DR and D2DR-expressing MSNs that could drive escalation of drinking in dependent mice. Further studies are needed to address this possibility.

In summary, results of the present study clearly demonstrate that KCNN1-3 gene expression in the brain is an important factor in alcohol, nicotine, and drug addiction. We validated these clinical and preclinical observations using genetically diverse BXD strains and targeted pharmacological inhibition of KCa2 channels in limited-access and dependence-induced escalation of drinking models. Based on our findings, we propose that human laboratory studies and clinical trials coupled with genetic profiling of heavy drinking alcoholics should be initiated to determine if KCa2 channels are a promising pharmacogenetic target for treating AUDs. Additional studies should also explore the suggestion that KCa2 channels are a possible biomarker for substance abuse or a pharmacogenetic target for reducing the intake of nicotine and illicit drugs.

FUNDING AND DISCLOSURE

The authors declare no conflict of interest.