INTRODUCTION

Alcohol use can develop into addiction in a significant number of people, and yet the mechanism of transition between alcohol use and dependence is not sufficiently understood. The establishment of addiction remains largely unknown but it is thought to involve a number of distinct learning and memory processes. It is now widely believed that addiction and memory formation share molecular and anatomical pathways (Nestler, 2002; Kelley, 2004; Müller and Schumann, 2011).

Ca2+/calmodulin-dependent protein kinase II (CaMKII) plays a key role in the plasticity of glutamatergic postsynapses of the brain (Colbran and Brown, 2004; Irvine et al, 2006; Lamsa et al, 2007; Wayman et al, 2008) and appears crucial for learning and memory, with the α-subunit-composed heteromer (αCaMKII) being the most influential (Elgersma et al, 2004; Easton et al, 2012). There is considerable evidence for a role of CaMKII in drug addiction in humans (Li et al, 2008) and in animal models (Pierce et al, 1998; Tan, 2002; Licata and Pierce, 2003; Sakurai et al, 2007; Choe and Wang, 2002). In particular, CaMKII levels in the ventral tegmental area (VTA) and nucleus accumbens (NAcc) appear of crucial importance for psychostimulant sensitization, place preference, and self-administration (Licata et al, 2004; Anderson et al, 2008; Kourrich et al, 2012). Although little is known about the role in alcohol addiction, alcohol affects CaMKII levels in the brain (Smith and Navratilova, 1999; McBride et al, 2009; Mahadev et al, 2001; Lee et al, 2011) and increases phosphorylation rate at Thr286 (Garic et al, 2011; Wang et al, 2012). Both may control alcohol effects on cellular signaling (Liu et al, 2006; Xu et al, 2008), behavior (Kim et al, 2012; Wang et al, 2012), and neurotoxicity (Garic et al, 2011).

Following transient calcium/calmodulin activation, αCaMKII can switch to an autonomous mode of activity known as autophosphorylation. Mice with a point mutation in position 286 of the protein (αCaMKIIT286A) do not show autophosphorylation, but severe learning impairments (Giese et al, 1989; Irvine et al, 2005) and emotional dysregulation (Easton et al, 2011), suggesting that autophosphorylation controls, in particular, the speed of learning but not the capacity to learn or memory retrieval (Irvine et al, 2006).

Alcohol exerts its effects at the interface of the glutamatergic and monoaminergic systems, which in turn shape behavioral effects and addiction development (McBride, 2010; Spanagel, 2009). Here we asked whether αCaMKII autophosphorylation plays a role in the establishment of alcohol drinking behavior as a crucial component of (later) addiction development. We hypothesized that a genetically induced deficiency in αCaMKII autophosphorylation results in slower establishment of alcohol drinking in mice, which would make their drinking phenotype initially different from wild-type controls. With increasing learning experience, for example, after 5 learning trials (Irwine et al, 2005), which we considered as a subchronic exposition, they should approach the wild-type drinking phenotype. In parallel, different effects on neuronal mechanisms, such as dopamine (DA) and serotonin (5-HT) responses and cellular activation in the VTA, should occur after acute or subchronic alcohol exposure (Smith and Navratilova, 1999). Furthermore, we asked whether single-nucleotide polymorphisms (SNPs) related to αCaMKII function should change the risk for alcohol addiction in humans.

MATERIALS AND METHODS

Animals

Male and female αCaMKIIT286A mutant mice (Giese et al, 1989) were studied in gender-balanced designs in all experiments (MT, n=50; heterozygous (Het), n=43, wild-type (WT), n=52; for details, see Supplementary Information). This mutation blocks the autophosphorylation of CaMKII but does not affect the Ca2+-dependent activity. Animals were individually housed, provided with food and water ad libitum, and kept on a 12:12 h light/dark cycle (lights on at 0700 h). Behavioral tests were performed during the light cycle between 0900 and 1600 h. Room temperature was maintained between 19 and 22 °C at a humidity of 55% (±10%). All housing and experimental procedures were performed in accordance with the UK Home Office Animals (Experimental Procedures) Act 1986.

Alcohol Drinking and Alcohol Deprivation Effect

Alcohol drinking was tested in naive αCaMKIIT286A (MT; n=10), WT (n=12) and Het (n=12) animals using a two-bottle free-choice drinking paradigm. Each cage was equipped with two bottles constantly available, one of which contained tap water and the other contained alcohol in various concentrations. After an acclimatization period to establish a drinking baseline, animals received alcohol at increasing concentrations of 2, 4, 8, and 12 vol.%. Mice were exposed to each concentration of alcohol for 4 days. Thereafter, alcohol concentration was switched to 16 vol.% and animals were allowed to drink for 2 weeks. In order to measure the alcohol deprivation effect (Spanagel and Hölter, 2000), baseline consumption of 16 vol.% alcohol was measured. Alcohol was removed for 3 weeks (both bottles now containing tap water) before it was re-introduced for 4 days. This procedure was repeated once more. Bottles were changed and weighed daily. The consumed amount of alcohol relative to body weight and the preference vs water were measured.

Taste Preference Test

Alcohol-experienced animals were used for this test (MT: n=10; WT: n=12; Het: n=12). Sucrose (0.45 and 5%) and quinine (10 and 20 mg/dl) preference was measured in a two-bottle free-choice test vs water. Each dose was offered for 3 days with the position of the bottles being changed and weighed daily (Spanagel et al, 2005).

Determination of Blood Alcohol Levels

Alcohol-naive animals were used for this test (MT: n=11; WT: n=12; Het: n=14). Animals were injected intraperitoneally (i.p.) with alcohol (3.5 g/kg). Mice were left undisturbed and then systematically culled at 15, 30, and 60 min after injection. Animals were immediately killed by cervical dislocation and trunk blood was collected. Blood was left at room temperature for 30–60 min to allow clotting. The blood was then placed in a centrifuge for 10–15 min at 3000–4000 r.p.m. at room temperature and the supernatant aspirated. Serum samples were then stored at −80 °C waiting further analysis. Analysis of the alcohol concentration in the serum samples was performed with quantitative enzymatic method (for details, see Supplementary Information).

Loss of Righting Reflex (LORR)

Alcohol-naive animals were used for this test (MT: n=5; WT: n=5; Het: n=5). Animals were administered with an alcohol injection of 3.5 g/kg (i.p.) (Spanagel et al, 2002) to induce LORR, and immediately placed in an empty cage. LORR was observed when the animal becomes ataxic and stopped moving for at least 30 s. The animal was then placed on its back. Recovery from alcohol administration was defined as the animal being able to right itself three times within a minute. A 2-h cutoff was used. Time taken for the animal to lose its righting reflex, and time to recovery from alcohol’s effect were recorded. LORR was carried out on day 1 of the experiment. Upon completion of the trial, the animal was returned to the home cage. Alcohol was then administered once daily on days 2–7, when animals were weighed and injected 2 g/kg (i.p.). LORR was again tested on day 8 of the experiment when animals received 3.5 g/kg (i.p.) alcohol.

In Vivo Microdialysis

Mice were deeply anesthetized and two guide cannulas were aimed at the prefrontal cortex (PFC) and the NAcc (Franklin and Paxinos, 1997). Animals were allowed at least 5 days for complete recovery before microdialysis testing (for details, see Supplementary Information).

On the day of the experiment, microdialysis probes with membrane lengths of 2 mm for the PFC (MAB 6.14.2.) and 1 mm (MAB 6.14.1.) for the NAcc were inserted into the guide cannulae under a short (3–5 min) isoflurane anesthesia and perfused with artificial cerebrospinal fluid (aCSF). After probe insertion, the animal was placed into an open field (21 × 21 × 30 cm) of a TruScan system (Coulbourn Instruments, Allentown, PA). Food and water were provided ad libitum and room temperature maintained between 19 and 22 °C. Samples were collected every 20 min. Three samples were taken as baseline and the neurotransmitters DA, 5-HT and NA were quantified by HPLC-EC as described previously (Müller et al, 2007; Pum et al, 2007). An injection of alcohol was then administered i.p. (2 g/kg) and further nine samples were collected. Locomotor activity was automatically measured by a TruScan system parallel to neurochemical activity. Locomotion was assessed as units in 20-min intervals.

Alcohol-naive animals were used for this test (MT: n=12; WT: n=11; Het: n=12). The in vivo microdialysis was carried out on day 1 of the experiment when animals received the first of eight alcohol injections (2 g/kg, i.p.). A smaller locomotor stimulating dose of alcohol was used here and in the subsequent experiment to allow for behavioral sensitization/tolerance effects to be observed. After the microdialysis trial was completed, the animal was anesthetized using isoflurane. The probes were removed and guides re-inserted. Alcohol was then administered once daily on days 2–7 (2 g/kg, i.p.). The in vivo microdialysis was carried out again on day 8 of the experiment, providing a subchronic neurochemical and behavioral response to alcohol treatment. Once microdialysis experiments were complete, animals were killed by cervical dislocation. Brains were fixed in 4% formaldehyde solution and stored at 4 °C. Brains were sliced on a microtome and stained with cresyl violet for verification of probe placement.

c-Fos Activation after Acute and Subchronic Alcohol Treatment

Alcohol-naive animals were used for this test (MT: n=12; WT: n=12). Animals were transferred from the home cage to a temporary cage and injected either once (acute) or on seven consecutive days (subchronic) with alcohol at a dose of 2 g/kg (i.p.). Mice were left undisturbed for 70 min after injection. Thereafter, mice were culled under isoflurane narcosis and transcardially perfused. Brains were taken and c-Fos activation was measured in the rostral and caudal VTA (for details, see Supplementary Information).

Genetic Association in a Human Sample

Genotype data were extracted from a previously performed genome-wide association study (Treutlein et al, 2009; Frank et al, 2012). In brief, the sample comprised n=1333 male patients with severe DSM-IV alcohol dependence and n=939 male controls. Individuals were genotyped using Illumina infinium assays. From the autosomal SNPs available, we selected only those located within the transcript region of the CAMK2A gene, resulting in a set of 23 SNPs. We used Armitage trend tests to assess associations of CAMK2A SNPs with alcohol dependence. Analysis of linkage disequilibrium (LD) structure and haplotype association testing was performed using haploview version 4.2 (http://www.broad.mit.edu/mpg/haploview/). Haplotype blocks were defined according to the method of Gabriel et al (2002). All P-values were corrected using the false discovery rate (FDR) procedure (Benjamini and Hochberg, 1995).

Statistical Analysis

All quantitative data were expressed as mean±SEM. Data were analyzed using ANOVAs (for repeated measures where appropriate) followed by preplanned comparisons using Fisher’s LSD tests (Szumlinski et al, 2005). For experiment-specific details, see Supplementary Information. Although sex differences are well known in alcoholism-related behaviors (Lenz et al, 2012), we did not see significant sex differences in this study. Therefore, data were collapsed for analysis. The softwares SPSS 17.0, PLINK v1.07, and Statistica 9 were used. A significance level of P<0.05 was used.

RESULTS

Reduced Alcohol Preference

In order to test a potential involvement of αCaMKII autophosphorylation in the establishment of alcohol drinking behavior, we measured drinking in a two-bottle free-choice paradigm. We found that αCaMKIIT286A and Het animals drank significantly less alcohol than WT animals (Figure 1a; two-way ANOVA, genotype: F2, 124=11.17, P<0.0001; dose: F3, 124=55.10; P<0.0001; interaction: F6, 124=2.04, P=0.065). Pairwise comparisons showed significant differences between αCaMKIIT286A and WT animals at 8 vol.% (LSD, P=0.016) and 12 vol.% (P<0.0001). Also, Het alcohol consumption was reduced vs WT at 12 vol.% (P=0.0002). Alcohol preference vs water was significantly reduced in αCaMKIIT286A compared with WT animals (Figure 1b; two-way ANOVA, genotype: F2, 124=5.74, P=0.0041; dose: F3, 124=19.94, P<0.0001). Pairwise comparisons showed a significant difference between αCaMKIIT286A and WT animals at 4 vol.% (P=0.022) and a trend at 12 vol.% (P<0.058). These findings suggest that αCaMKII autophosphorylation is required for full establishment of alcohol drinking and alcohol preference over water in a standard drinking test.

Figure 1
figure 1

αCaMKII autophosphorylation-deficient mice initially drink less alcohol but reach wild-type (WT) level after repeated withdrawal. They show a largely preserved taste preference and avoidance and comparable alcohol bioavailability. (a) Alcohol consumption shown as mean consumption (±SEM) over 4 days of drinking for each dose of alcohol. (b) Alcohol preference (±SEM) vs water shown as mean preference (±SEM) over 4 days of drinking for each dose of alcohol. (c) Mean consumption (±SEM) of a 16 vol.% alcohol solution per day. Alcohol deprivation effect on consumption after two 3-week withdrawal periods (hatched bars; B1–B4—baseline). Alcohol consumption increased in the αCaMKIIT286A and Het animals over 4 days after the first 3-week withdrawal (T1–T4). αCaMKIIT286A but not Het consumption was still below WT levels. After the second 3-week withdrawal period, consumption increased in all animals, suggesting an ADE (T5–T8), with the αCaMKIIT286A then consuming at the same level as WT animals. (d) Mean preference vs water (±SEM) of a 16 vol.% alcohol solution per day. Alcohol deprivation effect on preference after two 3-week withdrawal periods (hatched bars). (e) Mean preference of sucrose and quinine solution over water measured over 3 consecutive days, respectively. (f) Blood alcohol levels (mean ±SEM) after a 3.5 g/kg (i.p.) injection of alcohol (*P<0.05, $P<0.001 vs WT).

PowerPoint slide

Repeated Withdrawal Eliminates Preference Differences

Initial alcohol preference was reduced in αCaMKIIT286A mice. We now investigated whether this initial difference could be overcome by repeated exposition/withdrawal from alcohol in a similar way at it was observed in nondrug-motivated learning tasks (Irvine et al, 2005). Voluntary alcohol consumption is known to escalate after repeated withdrawal periods, which is known as the alcohol deprivation effect (ADE; Spanagel and Hölter, 2000). We found that αCaMKIIT286A as well as Het animals consumed significantly less alcohol during a 4-day baseline when offered a 16 vol.% alcohol solution (Figure 1c; B1–B4, two-way ANOVA, genotype: F2, 124=11.67, P<0.0001). Pairwise comparisons vs WT on the factor genotype confirmed this (LSD; MT: P<0.0001; Het: P<0.0001). Also, alcohol preference vs water was altered in αCaMKIIT286A and Het animals during baseline (Figure 1d; B1–B4, two-way ANOVA, genotype: F2, 124=15.15, P<0.0001). Pairwise comparisons on the factor genotype showed a reduction for αCaMKIIT286A (P<0.0001) and Het animals (P<0.0001) vs WT. Alcohol consumption increased in the αCaMKIIT286A and Het animals over 4 days after the first 3-week withdrawal (T1–T4, two-way ANOVA, genotype: F2, 124=10.40, P<0.0001; time: F3, 124=2.59; P=0.056; interaction: F6, 124=2.20, P=0.048). However, in αCaMKIIT286A (P=0.0006) but not Het (P>0.05) animals, consumption was still below WT levels. Also, alcohol preference vs water increased during this interval (T1–T4, two-way ANOVA, genotype F2, 124=7.20, P<0.001; interaction: F6, 124=4.80, P=0.0002). Preference was still lower in αCaMKIIT286A (P=0.007) but not Het animals (P>0.05), compared with WT animals. After the second 3-week withdrawal period, consumption increased temporarily in all animals (T5–T8, two-way ANOVA, genotype: F2, 124=6.08, P=0.003; time: F3, 124=52.59; P<0.0001), suggesting a clear ADE with the αCaMKIIT286A then consuming at the same level as WT animals (P>0.05). Whereas the ADE was comparable between all genotypes on day 1 of reinstatement of alcohol drinking, it decreased stronger in the Het animals thereafter (vs WT, P=0.0018). This was also reflected in alcohol preference (T5–T8, two-way ANOVA; genotype: F2, 124=7.67, P=0.0007; time: F3, 124=10.83; P<0.0001), with αCaMKIIT286A preference at the same level as WT animals (P>0.05). These data suggest that the reduced alcohol preference in αCaMKIIT286A mice can be overcome by repeated withdrawal.

Sucrose Preference

We found no genotype difference in the preference of 0.45% sucrose (Figure 1e; P>0.05). However, a 5% sucrose solution was less preferred by the αCaMKIIT286A but not Het mice (genotype: F2, 93=6.20, P=0.003; LSD vs WT, P=0.002). Quinine was avoided by all mice. The 10 mg/dl concentration was avoided most by Het animals (two-way ANOVA, genotype: F2, 93=10.38, P<0.0001; time: F2, 93=5.25; P=0.007), which differed from WT (P<0.014). There was no genotype difference in the avoidance of a 20 mg/dl quinine solution (P>0.05). These data suggest that although taste sensitivity is not disturbed by αCaMKII autophosphorylation deficiency, it may, nevertheless, slightly reduce preference of sweet solutions at higher concentrations.

No Effects on Alcohol Bioavailability

There was no significant difference in alcohol bioavailability between genotypes after a 3.5 g/kg alcohol i.p. injection (P>0.05; Figure 1f).

No Role in Acute Sedating Effects of Alcohol

There was no genotype difference in the sedating effects of alcohol between groups of alcohol-naive mice following acute or subchronic alcohol treatment as measured by the LORR (two-way ANOVA, factor: genotype, P>0.05; Figure 2a), but a significant tolerance development (factor: test time; duration: F1, 12=9.97, P=0.0083; time: F1, 12=12.19, P=0.0045).

Figure 2
figure 2

The sedating effects of alcohol are experienced similarly by αCaMKII autophosphorylation-deficient, heterozygous, and wild-type alcohol-naive mice. Loss of righting reflex (LORR) time to sedation (a) and LORR duration (b) after acute and subchronic (after seven prior alcohol treatments) alcohol (3.5 g/kg, i.p.) treatment. Both parameters suggest no genotype effects, but a significant tolerance development. (**P<0.01, two-way-ANOVA, factor: treatment time).

PowerPoint slide

Absence of Locomotor Activating Effects

Alcohol-naive mice showed no differences in locomotor activity between genotypes (P>0.05; Figure 3a). An acute alcohol injection increased locomotor activity in WT but not αCaMKIIT286A or Het mice (Figure 3b). Although two-way ANOVA failed to show significant genotype effects or interactions (P>0.05), preplanned comparisons revealed a significant difference between αCaMKIIT286A and WT mice 60 min after acute alcohol injection (LSD, P=0.03). Basal behavioral activation levels of αCaMKIIT286A but not WT or Het mice increased after subchronic alcohol treatment (LSD vs acute, P<0.05; Figure 3a). Following seven prior daily alcohol treatments, an alcohol injection produced a significant increase in locomotor activity in WT and Het but not αCaMKIIT286A mice (Figure 3c; two-way ANOVA; time: F11, 231= 2.97, P=0.001; interaction: F22, 231= 1.83, P=0.02). Preplanned comparisons showed a significant difference between groups at 20 min. (LSD, MT vs WT: P=0.05) after injection. These findings suggest that the locomotor activating effects of alcohol are absent in αCaMKIIT286A mice following both acute and subchronic exposure.

Figure 3
figure 3

The effects of an alcohol (2 g/kg, i.p.) injection on αCaMKII autophosphorylation-deficient and wild-type mice in a well-habituated (>2 h before baseline) open field. (a) Baseline behavioral activity is not different between genotypes (#P<0.05). Locomotor activity recorded after acute (b) and subchronic (seven) alcohol treatments (c). Arrow indicates time of alcohol injection (*P<0.05 vs MT).

PowerPoint slide

The Acute Dopaminergic Effects of Alcohol in NAcc Are Absent

Mesolimbic DA activation is a crucial mediator of the reinforcing effects of alcohol (Di Chiara and Imperato, 1988; Spanagel, 2009). There were no differences in basal DA levels in the NAcc or PFC between αCaMKIIT286A and WT mice (P>0.05, Table 1). Acute alcohol administration increased extracellular DA levels in the NAcc of WT mice (Figure 4a; two-way ANOVA, genotype: F2, 26=4.20, P=0.03; time: F11, 286=2.21, P=0.01). Preplanned comparisons revealed a significant increase in DA levels in the NAcc in WT mice at 40 min (LSD vs MT, P=0.028), 80 min (P=0.01), 100 min (P=0.01), and 120 min (P=0.04) after injection. This DA increase was absent in αCaMKIIT286A and Het mice. DA levels in the PFC of all mice did not change in response to an alcohol challenge (P>0.05, Supplementary Figure S1A). After subchronic alcohol exposure, an acute alcohol injection led to an increase in NAcc DA levels in WT mice, which was less pronounced than in naive animals (Figure 4b; two-way ANOVA, P>0.05). Preplanned comparisons revealed significant differences between WT and αCaMKIIT286A mice only 100 min (LSD vs MT, P=0.02) after alcohol administration. The DA response was unchanged in αCaMKIIT286A and Het mice, thus reducing response differences between genotype groups. Subchronic DA levels in the PFC were not significantly different between groups (P>0.05; Supplementary Figure S1B). These data suggest a functional link between αCaMKII autophosphorylation and the DA increase in the NAcc as a crucial marker for the (positive) reinforcing effects of alcohol.

Table 1 Baseline Monoamine Levels before Acute Treatment and after Subchronic Alcohol Exposure
Figure 4
figure 4

Acute and subchronic alcohol effects on monoamine activity represented as percent of baseline (mean±SEM). (a) Extracellular dopamine (DA) levels in the nucleus accumbens (NAcc) after acute alcohol treatment. (b) Effects of subchronic alcohol treatment on dopamine levels in the NAcc. (c) Extracellular serotonin (5-HT) levels in the prefrontal cortex (PFC) after acute alcohol treatment. (d) Effects of subchronic alcohol treatments on PFC 5-HT levels. Arrows indicate time of alcohol injection (preplanned comparison between groups at each time point: *P<0.05, MT vs WT).

PowerPoint slide

Increased 5-HT Response in the PFC after Acute Alcohol Treatment

Serotonergic mechanisms are well known to orchestrate the behavioral effects of alcohol (McBride, 2010). CaMKII was shown to control 5-HT activity at synaptic level (Liu et al, 2005). There were no differences in basal 5-HT levels in the NAcc or PFC between αCaMKIIT286A and WT mice (P>0.05, Table 1). After first alcohol treatment, 5-HT in the NAcc was not significantly different between genotypes in this study (Supplementary Figure S1C). In WT and Het animals, there was no effect of alcohol on 5-HT levels in the PFC. However, in αCaMKIIT286A mice, alcohol led to a 5-HT increase (Figure 4c; two-way ANOVA, genotype × time interaction: F22, 319=2.22, P=0.001). Preplanned comparisons showed a significant differences between αCaMKIIT286A and WT 20 min (LSD, P=0.008) and 40 min (P=0.001) after alcohol administration. After subchronic alcohol treatments, an acute alcohol challenge did not alter serotonergic responses in the NAcc (P>0.05; Supplementary Figure S1D). In the PFC, there was no effect of alcohol on 5-HT levels in WT or Het animals, but an increase in αCaMKIIT286A mice (Figure 4d; two-way ANOVA; P>0.05). Preplanned comparisons revealed significant differences between αCaMKIIT286A and WT 120 min (LSD, P=0.05) and 160 min (P=0.02) after alcohol injection. These findings suggest an inhibitory effect of αCaMKII autophosphorylation on 5-HT responses to alcohol in the PFC.

Enhanced c-Fos Activation in the Rostral But Not Caudal VTA after Alcohol

In order to determine the origin of altered DA responses to alcohol, we further investigated the effects of acute and subchronic alcohol administration on c-Fos activation in the rostral and caudal VTA (Figure 5). αCaMKIIT286A mice showed an increased c-Fos expression in the rostral part of the VTA after both acute (F1, 10=6.32, P<0.05) and subchronic (F1, 9=20.7, P<0.01) alcohol treatment. There was no significant difference between αCaMKIIT286A and WT animals in c-Fos expression in the caudal part of the VTA (P>0.05). The rostral VTA contains predominantly GABAergic neurons (Olson and Nestler, 2007). In order to estimate the relative number of GABAergic cells within the VTA that show c-Fos activation after alcohol treatment, we performed GAD67/c-Fos double labeling in representative animals. We found that 68 and 40% of the c-Fos-activated cells were also GAD67 positive in the rostral and caudal VTA, respectively (Supplementary Figure S2).

Figure 5
figure 5

αCaMKII autophosphorylation-deficient mice show an increased c-Fos expression in the rostral, but not caudal, ventral tegmental area (VTA) after both acute and subchronic alcohol treatment. c-Fos labeling of the caudal (a, c, e, g, P>0.05) and rostral VTA (b, d, f, h) after acute or 7 days of subchronic alcohol (2 g/kg, i.p.) treatment. C-Fos was determined 70 min after last alcohol injection (ANOVA, *P<0.05, **P<0.01). (e–h) c-Fos-labeled cells of the VTA after single acute alcohol treatment (bar=100 μm).

PowerPoint slide

CAMK2A Gene SNPs Predict Alcohol Dependence in Male Humans

Although there is no naturally occurring gene mutation affecting the autophosphorylation site in humans, there are SNPs in the CAMK2A gene that may affect general activity and indirectly autophosphorylation. The genotype distribution of all investigated SNPs was in accordance with Hardy–Weinberg expectations (P>0.1). Of the 23 SNPs typed from the CAMK2A gene, 7 were found to be significantly associated with alcohol dependence: rs3822607 (P=0.015, FDR=0.0496), rs3776825 (P=0.012, FDR=0.0460), rs7711562 (P=0.0034, FDR=0.0391), rs3756577 (P=0.012, FDR=0.0460), rs10463293 (P=0.0028, FDR=0.0391), rs4958445 (P=0.0091, FDR=0.0460), and rs4958902 (P=0.012, FDR=0.0460; Table 2). Haplotype-based analysis of human genotype data did not reveal additional insights (Supplementary Table S1). These findings may suggest a functional effect of genetic mutations in the CAMK2A gene on alcohol consumption in humans also.

Table 2 Association of CAMK2A Single-Nucleotide Polymorphisms with Alcohol Dependence in a Sample of 1333 Male DSM-IV Alcohol-Dependent Patients and 939 Male Controls

DISCUSSION

Here we report that αCaMKIIT286A mice prefer alcohol significantly less than WT animals—especially at higher percent alcohol solutions. This effect was persistent until animals were withdrawn twice from alcohol. Taste sensitivity remains largely unaffected, although αCaMKII autophosphorylation deficiency may slightly reduce the incentive properties of a sucrose solution. Alcohol bioavailability was found to be comparable across genotypes. Whereas the locomotor activating effects of a low dose of alcohol were absent in αCaMKIIT286A mice after acute and subchronic exposure, sedating effects of a high dose were preserved. The lack of αCaMKII autophosphorylation did not change DA or 5-HT basal activity, but coincided with a lack of the acute DA response to alcohol in the NAcc. Interestingly, it provoked a 5-HT increase in the PFC. Both effects suggest a strong link between αCaMKII autophosphorylation and monoaminergic responses to alcohol. An analysis of the cellular activation of the VTA, as an origin of the mesolimbic DA projections, revealed an enhanced activation after acute and subchronic alcohol in the rostral but not caudal region in the αCaMKIIT286A mice, which appeared to be driven predominantly by GABAergic neurons. Animal data suggest that αCaMKII autophosphorylation may control the speed at which alcohol preference is established, but not the capacity to consume alcohol. A comparative approach, using a human sample, confirmed the link between genetic mechanisms controlling αCaMKII activity and alcohol consumption by means of an association between naturally occurring SNPs and alcohol dependence.

The DA system is critical for the establishment of the acute reinforcing effects of alcohol (Koob et al, 1998; McBride et al, 1999). We have shown that an acute alcohol challenge induces a DA increase in the NAcc of WT mice. This response was entirely absent in Het and αCaMKIIT286A mice. This finding suggests that αCaMKII autophosphorylation is required for the alcohol-induced DA increase in the NAcc. The lack of a DA response in the αCaMKIIT286A mice is therefore a likely mechanism to explain the initially reduced alcohol preference in a free access condition. Activation of the dopaminergic neurons in the VTA is believed to be critically involved in the development of addiction (Robbins and Everitt, 1996; Koob et al, 1998). Projections from the VTA extend to regions implicated in the rewarding potential of drugs of abuse, including the NAcc (Van Bockstaele and Pickel, 1995; Olson et al, 2005) and the PFC (Carr and Sesack, 2000). An activation of the caudal VTA dopaminergic neurons, which is mediated by a reduced GABAergic signaling (Theile et al, 2011), is one of the cellular bases for alcohol reinforcement (Brodie et al, 1999; Xiao et al, 2009). Several studies have demonstrated that as the GABAergic activity in rostral VTA increases, the dopaminergic activity decreases, resulting in a diminished experience of reward and reinforcement (Ikemoto et al, 1997; Ding et al, 2009). We observed an increased rostral VTA c-Fos activation after alcohol in the αCaMKIIT286A mice, which is in line with the observation of reduced alcohol consumption and the lack of locomotor activating effects. The majority of the c-Fos-activated cells appeared to be GABAergic. Activation of GABAergic neurons in the VTA may induce an aversive state (Tan et al, 2012) and disrupt reward consumption. This may occur by inhibiting DA neurons and DA release, or via ascending GABAergic projections (van Zessen et al, 2012). We did not find an altered activation after alcohol in the caudal VTA of αCaMKIIT286A mice. Together, this may suggest that the lack of an alcohol-induced DA increase in the NAcc was not mediated by altered responsiveness of DA neurons in the VTA, but possibly by direct VTA GABAergic inputs to the NAcc (van Zessen et al, 2012).

Alcohol administration increases 5-HT activity in the ventral hippocampus (McBride et al, 1999; Thielen et al, 2002). In our study, alcohol did not affect 5-HT activity in the NAcc or PFC in WT or Het animals. In αCaMKIIT286A animals, however, alcohol induced a significant 5-HT increase in the PFC. Interestingly, Het mice showed neither a strong DA nor 5-HT response, suggesting an absence of the reinforcing effects of alcohol. This is in line with Het behavior in that they drop back to low consumption after two withdrawals (Figure 1c). Deficiencies in brain 5-HT levels and turnover have previously been associated with high alcohol drinking (Murphy et al, 1982; Virkkunen and Linnoila, 1997; Smith et al, 2008). Conversely, artificially induced increases in extracellular 5-HT levels can reduce alcohol drinking (Boyce-Rustay et al, 2006). The alcohol-induced 5-HT increases in the PFC of αCaMKIIT286A mice may, therefore, contribute to the reduced alcohol drinking seen in these mice. There is a strong link between CaMKII and serotonergic function in the brain. CaMKII is required for the phosphorylation and activation of tryptophan hydroxylase (TPH), the rate-limiting enzyme in the biosynthesis of 5-HT (Kuhn et al, 1982, 2007; Ehret et al, 1989). The administration of the CaMKII inhibitor, KN-62, has been shown to increase the firing rate of 5-HT neurons (Liu et al, 2005). This finding is in line with the present data, whereby a reduced CaMKII function may increase 5-HT neuronal firing and terminal 5-HT release, which inhibits alcohol consumption. As the 5-HT response is still observed after subchronic administration in the αCaMKIIT286A mice, one may speculate that the ADE-induced increase in alcohol drinking is mediated by nonmonoaminergic mechanisms (Spanagel, 2009).

At a low dose (2 g/kg, i.p.), alcohol has locomotor stimulant effects upon acute and subchronic treatment in the WT animals. This effect was absent in αCaMKIIT286A and Het mice. These findings may suggest that αCaMKII autophosphorylation is required for the locomotor stimulant effects of a low dose of alcohol. Subchronic treatment with alcohol increased baseline locomotor activity in αCaMKIIT286A but not WT or Het mice. This may have been caused by a parallel increase in basal DA levels in the NAcc in αCaMKIIT286A but not WT or Het mice (Gong et al, 1999).

A recent study showed that alcohol-induced LORR is controlled by αCaMKII-bearing neurons in the dorsal striatum (Kim et al, 2012). Here we report that the sedative properties of a high dose of alcohol were not affected by a deficit in αCaMKII autophosphorylation. This may suggest that alcohol LORR may depend on αCaMKII but not its autophosphorylation.

In agreement with the animal data, we found several associations of polymorphisms in the human CAMK2A gene with alcohol dependence. Of the 23 SNPs typed from the CAMK2A gene, 7 were found to be significant, one of which (rs10463293) has previously been associated with working memory performance (Easton et al, 2012) suggesting that CAMK2A may play a role in both. The nonsynonymous autophosphorylation SNP described in the animal literature (Giese et al, 1989) is a coding SNP located in exon 11 of the human CAMK2A gene (rs35495165). One of the significantly associated SNPs (rs3756577) in this study, located in intron 11, immediately flanks the functional polymorphism of interest and may therefore be acting as a tagging SNP for the autophosphorylation SNP, which was not typed in this sample.

Altogether, the present study suggests that αCaMKII autophosphorylation might play an important role in the establishment of alcohol drinking behavior in mice and humans. A lack of αCaMKII autophosphorylation delays the establishment of this behavior significantly. αCaMKII autophosphorylation is required to express a DA increase in the NAcc as a functional marker for the positive reinforcing effects of alcohol. Its lack leads to a DA-5–HT response imbalance in the mesocorticolimbic system.

Figure 6
figure 6

Exon–intron structure of the CAMK2A gene (isoforms 1 and 2), according to UCSC genome browser (https://genome.ucsc.edu). Included are chromosomal position and haplotype block structure of assessed markers. Haplotype blocks are framed in black. Numbers in gray-shaded squares denote pairwise D′ between single markers. Gray scheme encodes LD between markers using D′ confidence intervals (uninformative: gray; strong evidence for LD: dark gray; strong evidence for recombination: white).

PowerPoint slide