K. K. Szumlinski, Department of Psychology and Neuroscience Research Institute, University of California at Santa Barbara, Santa Barbara, CA 93106-9660, USA. E-mail: email@example.com
Alcohol-induced increases in nucleus accumbens glutamate actively regulate alcohol consumption, and the alcohol responsiveness of corticoaccumbens glutamate systems relates to genetic variance in alcohol reward. Here, we extend earlier data for inbred mouse strain differences in accumbens glutamate by examining for differences in basal and alcohol-induced changes in the striatal expression of glutamate-related signaling molecules between inbred C57BL/6J and DBA2/J mice. Repeated alcohol treatment (8 × 2 g/kg) increased the expression of Group1 metabotropic glutamate receptors, the NR2a/b subunits of the N-methyl-d-aspartate receptor, Homer2a/b, as well as the activated forms of protein kinase C (PKC) epsilon and phosphoinositol-3-kinase within ventral, but not dorsal, striatum. Regardless of prior alcohol experience, C57BL/6J mice exhibited higher accumbens levels of mGluR1/5, Homer2a/b, NR2a and activated kinases vs. DBA2/J mice, whereas an alcohol-induced rise in dorsal striatum mGluR1/5 expression was observed only in C57BL/6J mice. We next employed virus-mediated gene transfer approaches to ascertain the functional relevance of the observed strain difference in accumbens Homer2 expression for B6/D2 differences in alcohol-induced glutamate sensitization, as well as alcohol preference/intake. Manipulating nucleus accumbens shell Homer2b expression actively regulated these measures in C57BL/6J mice, whereas DBA2/J mice were relatively insensitive to the neurochemical and behavioral effects of virus-mediated changes in Homer2 expression. These data support the over-arching hypothesis that augmented accumbens Homer2-mediated glutamate signaling may be an endophenotype related to genetic variance in alcohol consumption. If relevant to humans, such data pose polymorphisms affecting glutamate receptor/Homer2 signaling in the etiology of alcoholism.
Furthering the link between NAC glutamate transmission and alcohol drinking, genetic vulnerability to high alcohol consumption relates to facilitated alcohol-induced glutamate sensitization within the NAC or its afferent structures. For example, selectively bred alcohol-preferring Indiana P rats exhibit approximately 50% higher evoked glutamate release in cortical slices, compared with selectively bred non-preferring NP rats. Moreover, this evoked glutamate release is enhanced by alcohol in P rats, but not in their NP counterparts (McBride et al. 1986). When compared with low alcohol-drinking Fischer 344 rats, the capacity of an acute alcohol injection (1 g/kg) to increase NAC glutamate levels is greater in alcohol-drinking Lewis rats (Selim & Bradberry 1996). More recently, our laboratory has observed marked differences between inbred B6 and the alcohol-avoiding inbred DBA2/J (D2) mouse strains regarding repeated alcohol-induced changes in NAC basal and alcohol-stimulated glutamate release, with B6 mice exhibiting indices of alcohol-sensitized glutamate transmission and D2 mice exhibiting neurochemical tolerance (Kapasova & Szumlinski 2008). Given our earlier data that NAC extracellular glutamate levels differentially regulate alcohol intake in these strains (Kapasova & Szumlinski 2008), this study extends our earlier B6/D2 data by examining for strain differences in the striatal protein expression of glutamate receptor signaling molecules. The functional relevance of the observed strain differences in NAC Homer2 expression for the divergent glutamate-sensitizing and alcohol-drinking phenotypes of B6 vs. D2 mice was also assessed using AAV approaches. The results obtained are consistent with the over-arching hypothesis of our laboratory that NAC glutamate signaling through Homer2 may be an endophenotype related to genetic variance in alcohol intake and support earlier indications for associations between polymorphisms in glutamate receptors and downstream kinases with risky drinking behavior in humans (Desrivieres et al. 2008; Schumann et al. 2008).
Materials and methods
Adult male (8 weeks of age) C57BL/6J (B6) and DBA2/J (D2) mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA). Mice in all treatment groups were individually housed in polyethylene cages (15 cm wide × 23 cm long × 16 cm high) in a colony room controlled for temperature (25°C) and humidity (71%), under a 12-h day/12-h night cycle (lights off at 1800 h). Animals were given ad libitum access to food and water. The animals were allowed to acclimate to the colony room for at least 7 days following arrival. All experimental protocols were consistent with the guidelines of the National Institutes of Health (NIH) Guide for Care and Use of Laboratory Animals (NIH Publication No. 80-23, revised 1996) and were reviewed and approved by the University of California, Santa Barbara, Institutional Animal Care and Use Committee.
Repeated alcohol treatment
For the immunoblotting and in vivo microdialysis experiments, B6 and D2 mice were subdivided into two treatment groups, the first of which received eight intraperitoneal (i.p.) injections of 2 g/kg alcohol, every other day, whereas the second group (control) received an identical regimen of saline injections (vol = 0.02 ml/g body weight). This repeated alcohol treatment regimen was selected for study as it produces behavioral and neurochemical sensitization in B6 mice, while inducing neurochemical tolerance in D2 mice (Kapasova & Szumlinski 2008; Szumlinski et al. 2008).
Behavioral pharmacological and genetic studies implicate members of the mGluR–Homer–PI3K and protein kinase C (PKCɛ) signaling cascades in regulating alcohol intake in rodents (Backstrom et al. 2004; Besheer et al. 2008; Blednov & Harris 2008; Cozzoli et al. 2009; Hodge et al. 2006; Lominac et al. 2006; Olive et al. 2005; Schroeder et al. 2005), and a recent study suggests that mGluR5-mediated activation of extracellular signal-regulated kinase (ERK) within the NAC shell subregion is involved in cue-induced reinstatement of alcohol-seeking behavior (Schroeder et al. 2008). Thus, immunoblotting was conducted on tissue from both the dorsal and ventral aspects of the striatum to assess strain differences in basal and alcohol-induced changes in the expression of the constitutively expressed Homer proteins Homer1b/c and Homer2a/b, their associated Group1 metabotropic glutamate receptor (mGluR) subtypes mGluR1 and mGluR5 (Tu et al. 1999), as well as their associated NR2a/b subunits of the N-methyl-d-aspartate (NMDA) glutamate receptor (Naisbitt et al. 1999). Moreover, we examined for strain differences in basal and alcohol-induced changes in the expression of the non-activated and activated forms of PKCɛ and ERK, as well as total PI3K expression and the levels of phospho-(Tyr)p85α as a read-out of PI3K activity (Zhang et al. 2006). Sample sizes ranged from 10 to 12 mice/group.
The immunoblotting procedures employed for the detection of glutamate receptors/Homers/kinases were identical to those previously described by our group (Ary & Szumlinski 2007; Ary et al. 2007; Cozzoli et al. 2009; Szumlinski et al. 2008). In brief, at 24 h withdrawal from repeated alcohol/saline treatment, the shell and core subregions of the NAC (ventral striatum) and/or the dorsal aspect of the striatum were dissected over ice. To extend early preliminary data (n = 2) demonstrating significant and selective Homer2a/b knockdown in the NAC by our AAV–shRNA constructs, the NAC from control and shRNA-infused mice was dissected over ice at 2 months post-AAV infusion (see below). For both studies, brains were placed in an ice-cold mouse brain mold (Braintree Scientific, Braintree, MA, USA), sectioned along the horizontal plane (0.5 mm thick) and micropunches of tissue made with an 18-gauge needle for shell/core and a 1 mm biopsy punch for the dorsal striatum. Tissue punches were then homogenized in a solution consisting of 0.32 M sucrose, 2 mM ethylenediaminetetraacetic acid (EDTA), 1% wt/v sodium dodecyl sulfate, 50 µM phenyl methyl sulfonyl fluoride and 1 µg/ml leupeptin (pH = 7.2), and 50 mM sodium fluoride, 50 mM sodium pyrophosphate, 20 mM 2-glycerol phosphate, 1 mM p-nitrophenyl phosphate, and 2 µM microcystin LR was included to inhibit phosphatases. After centrifugation at 10 000 g for 20 min, the homogenates were quantified using the Bio-Rad DC protein assay (Bio-Rad, Hercules, CA, USA) and stored at −80°C. Protein samples (20 µg/lane) were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and Tris–Acetate gradient gels (3–8%) (Invitrogen, Carlsbad, CA, USA) were used to separate glutamate receptors, whereas Bis–Tris gradient gels (4–12%) (Invitrogen) or Tris–Acetate gels were employed to separate Homers and the kinases. Wet polyvinylidene difluoride (PVDF) (Bio-Rad) membrane transfer was employed, and membranes were pre-blocked with either phosphate-buffered saline containing 0.1% (v/v) Tween 20 and 5% (wt/v) non-fat dried milk powder (for non-phosphorylated proteins and kinases) or a 5% (wt/v) bovine serum albumin phosphate-buffered saline solution (for phosphorylated kinases) for a minimum of 2 h before overnight incubation with primary antibody. The following rabbit polyclonal antibodies were used: anti-Homer 1b/c and anti-Homer 2a/b (Dr. Paul F Worley, Johns Hopkins University School of Medicine; 1:1000 dilution), anti-mGluR5 (Millipore, Billerica, MA, USA; 1:1000 dilution), anti-NR2a and anti-NR2b (EMD Chemicals, Gibbstown, NJ, USA; 1:1000 dilution), anti-PI3K (Millipore; 1:1000 dilution), anti-p(Tyr)p85alpha binding motif (Cell Signaling Technology, Beverly, MA, USA; 1:500 dilution), anti-ERK1/2 (Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1:2000 dilution), anti-PKCɛ (Santa Cruz Biotechnology; 1:1000 dilution), and anti-p(Ser729)PKCɛ (Santa Cruz Biotechnology; 1:500 dilution). The following mouse primary polyclonal antibodies were also used: anti-mGluR1 (Millipore; 1:500 dilution) and anti-p(Tyr204)ERK1/2 (Santa Cruz Biotechnology; 1:1000 dilution). A rabbit primary anti-calnexin antibody (Stressgen, Ann Arbor, MI, USA; 1:1000 dilution) was used as a control to ensure even protein loading and transfer. Membranes were washed, incubated with horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Millipore; 1:5000 to 1:10 000 dilution) or anti-mouse secondary antibody (Jackson Immuno Research Laboratories, West Grove, PA, USA; 1:10 000) for 90 min, washed again, and immunoreactive bands were detected by enhanced chemiluminescence using either ECL Plus (Amersham Biosciences, Piscataway, NJ, USA) or Pierce SuperSignal West Femto (Thermo Fisher Scientific, Rockford, IL, USA). Primary antibodies were stripped off PVDF membranes with ReBlot Plus Strong Antibody Stripping Solution (Millipore). Image J (NIH, Bethesda, MD, USA) was used to quantify immunoreactivity levels. An analysis of the immunoreactivity for calnexin indicated even protein loading and transfer (Fig. 1). Thus, the density × area measurements for each protein band was averaged over the D2-saline control samples within a gel and all bands on that gel were expressed as percent of the average D2-saline value. To obtain an index of kinase activation, the density × area measurements for each phospho-protein was normalized to that of its corresponding non-phosphorylated protein prior to expressing the data as percent average D2-saline values.
Surgery and AAV infusion
As the immunoblotting studies indicated marked strain differences in NAC Homer2 expression, the effects of over- and under-expressing Homer2b within the NAC shell upon alcohol intake/preference and alcohol-induced glutamate release were assessed in B6 and D2 mice. The procedures to surgically implant bilateral indwelling guide cannulae and to infuse AAVs intra-NAC shell were identical to those previously implemented by our group (Cozzoli et al. 2009; Kapasova & Szumlinski 2008; Szumlinski et al. 2008). In short, under isoflurane gas anesthesia, mice were bilaterally implanted with 20-gauge stainless steel guide cannula (10 mm long) aimed 3 mm above the NAC shell (AP: +1.3 mm; ML: ± 0.5 mm; DV: −2.30 mm, relative to Bregma for B6 mice and AP: +1.0 mm; ML: ±0.5 mm; DV: −1.80 mm, relative to Bregma for D2 mice) (Franklin & Paxinos 1997). One-week post-surgery, 33-gauge injector cannula (12 mm long; threaded through a 24-gauge adaptor for stability) were bilaterally lowered into the NAC and AAVs carrying Homer2b cDNA, an shRNA against Homer2b (shH2b# 1; 5′-GUGUGAAUAUGUCUCUGAGTT-3′; Cozzoli et al. 2009), or a scrambled vector (control) was infused at a rate of 0.05 µl/min for 5 min (total vol = 0.25 µl/side) using a Harvard Apparatus PhD 2000 syringe pump (Holliston, MA, USA). Behavioral or neurochemical testing commenced 3 weeks post-transfection, a time when AAV-mediated neuronal transfection is maximal (Cozzoli et al. 2009; Klugmann & Szumlinski 2008; Klugmann et al. 2005; Lominac et al. 2005; Szumlinski et al. 2004, 2005, 2008).
As it was important to localize microinjector placement and the extent of AAV-mediated neuronal transfection within the NAC, as well as to verify transfection by both our AAV–cDNA and AAV–shRNA constructs in our experimental subjects, we employed our standard immunocytochemical approach (Cozzoli et al. 2009; Lominac et al. 2005; Szumlinski et al. 2005, 2007, 2008) to verify AAV-mediated neuronal transfection in animals following neurochemical and behavioral testing. As described previously (Cozzoli et al. 2009; Lominac et al. 2005; Szumlinski et al. 2005, 2007, 2008), B6 and D2 mice were transcardially perfused with saline, followed by 4% paraformaldehyde solution and as described previously, immunocytochemical staining for the HA tag was performed on 50 µm-thick serial tissue sections through the NAC (Fig. 2a–f). Mice failing to exhibit microinjector placements and/or HA staining within the boundaries of the NAC were excluded from the statistical analysis. By this criterion, one to two mice from each treatment group were excluded from the final statistical analysis of the data.
In Vivo microdialysis and HPLC analysis of glutamate
B6/D2 differences exist regarding both the acute and the sensitized NAC glutamate response to alcohol (Kapasova & Szumlinski 2008). As strain differences in NAC Homer2 expression were observed in the present study (see Results), we assessed the functional relevance of NAC Homer2 protein expression for strain differences in alcohol-induced changes in NAC extracellular glutamate. For this, AAV-infused B6 and D2 mice were subjected to conventional in vivo microdialysis procedures on days 1 and 8 of repeated alcohol/saline treatment. The procedures for microdialysis probe construction and dialysate collection were identical to those described previously (Kapasova & Szumlinski 2008; Lominac et al. 2006; Szumlinski et al. 2007, 2008). In short, a microdialysis probe (24-gauge, 13 mm in total length, including 0.7–1.0 mm active membrane) was unilaterally inserted into the NAC, fixed to the guide cannula with a drop of super glue, and artificial cerebral spinal fluid (146 nM NaCl, 1.2 mM CaCl2, 2.7 mM KCl, 1.0 mM MgCl2, pH = 7.4) was perfused through the probe at a rate of 2.0 µl/min using an automated syringe pump (kdScientific, Holliston, MA, USA). The side of probe insertion was counterbalanced across all treatment groups to reduce asymmetry confounds. Three hours were allowed for probe equilibration and then dialysate was collected in 10 µl of preservative (0.075 µM NaH2PO4, 25 µM EDTA, 0.0017 µlM 1-octansulfonic acid, 10% acetonitrile (v/v), pH = 3.0) at 20-min intervals for 1 h. Then, animals were i.p. injected with alcohol (2 g/kg) or saline and sample collection continued for an additional 3 h. At the end of the first microdialysis session, the probe was removed, the dummy cannula reinserted and mice were returned to the colony room. Mice then received injections 2–7 of their ethanol/saline treatment in the home cage, with injections administered every other day (Kapasova & Szumlinski 2008; Szumlinski et al. 2008). The morning of the eighth alcohol injection, the microdialysis procedures were repeated using the opposite side of the head. Sample sizes were 12 mice/group at the outset of experimentation, but because of attrition related to clogged guide cannulae, probe or transfection failure, final sample sizes ranged from 8 to 11 mice/group.
The apparatus and procedures for detecting dialysate glutamate levels using high-pressure liquid chromatography with electrochemical detection (HPLC–EC) were as previously described (Lominac et al. 2006; Szumlinski et al. 2007, 2008). The HPLC system consisted of a Coularray detector, a Model 542 autosampler and a Model 582 solvent delivery systems (ESA Inc., Bedford, MA, USA), with a detection limit of 0.01 fg/sample (20 µl/sample onto column). The mobile phase consisted of 100 mM NaH2PO4, 22% methanol (v/v), 3.5% acetonitrile (v/v) pH = 6.75 and glutamate was separated using a CAPCELL PAK C18 MG column (5 cm; Shiseido Company Ltd., Tokyo, Japan), eluting at 1.8 min. An ESA 5011A analytical cell with two electrodes (E1, +150 mV; E2, +550 mV) detected glutamate, following precolumn derivatization with o-phthalaldehyde (2.7 mg/ml) using the autosampler. The glutamate content in each sample was analyzed by peak height and was compared with an external standard curve for quantification using ESA Coularray for Windows software.
Alcohol preference and intake
To assess the functional relevance of B6/D2 differences in NAC Homer2 protein expression (see Results) for strain differences in alcohol intake and preference, a two-bottle-choice procedure was employed in which AAV-infused B6 and D2 mice were allowed 24 h-access to two identical 50 ml sipper tubes containing water and 3% alcohol (v/v) for five consecutive days. To generate a dose–response function, the concentration of alcohol available was increased to 6%, then to 12% and then to 18% with 5 days of access at each concentration. Bottles were weighed daily and the percent preference for each concentration and total alcohol intake (g/kg body weight basis) was determined. As described previously (Kapasova & Szumlinski 2008; Szumlinski et al. 2005), the weights of bottles on two dummy cages were recorded throughout testing and the average volume lost attributable to bottle handling/evaporation was subtracted from the daily record for each animal. Sample sizes were again 12 mice/group but because of attrition related to clogged guide cannulae and transfection failure, final sample sizes ranged from 9 to 11 mice/group.
With the exception of immunoblotting verification of Homer2 knockdown, where the data were compared between AAV-scrambled vs. -shRNA animals separately for each strain by t tests, all data presented in this report were analyzed using analyses of variance (ANOVAs). The immunoblotting data were analyzed using a Strain (B6 vs. D2) × ethanol (EtOH) (repeated saline vs. repeated alcohol) ANOVA. To more easily visualize strain and AAV effects upon alcohol preference/aversion, the relative preference for each concentration of alcohol vs. water was calculated (such that negative values reflect a preference for water over alcohol and positive values reflect preference for alcohol over water). The data for relative alcohol preference and intake (on a g/kg body weight basis) were then analyzed using a Strain × AAV (cDNA, shRNA and control) × Concentration (3%, 6%, 12% and 18% alcohol) ANOVA, with repeated measures on the concentration factor. Owing to technical difficulties with the HPLC, the sample sizes for all AAV groups were not identical on injections 1 and 8 of the microdialysis experiment. In order to include the data from all the animals in the statistical analyses, the Injection factor was treated as a between-subjects factor for the analysis of both basal glutamate levels and alcohol-induced changes in glutamate. Thus, the average baseline glutamate levels were analyzed using a Strain × EtOH × AAV × Injection ANOVA (all between-subjects factors). As group differences were not observed regarding baseline glutamate levels, the neurochemical data were expressed as a percentage of the average of the three baseline samples for each group to facilitate visualization of strain and alcohol effects upon the time–course of the glutamate response and to facilitate comparisons between the present data and that published previously by our group (Kapasova & Szumlinski 2008). These data were then analyzed using a Strain × EtOH × AAV × Injection × Time ANOVA, with repeated measures on the Time factor (12, 10-min bins). Planned comparisons between injections 1 and 8 of the repeated alcohol treatment were conducted independently for each AAV group to verify the presence of glutamate sensitization (B6 mice) and tolerance (D2 mice). Significant interactions were deconstructed for main effects and post hoc comparisons were conducted using Least Significant Difference tests when appropriate, α = 0.05.
Verification of AAV transfection
As previous attempts over the past 6 years to quantify by immunoblotting AAV–cDNA-induced increases in Homer2 protein expression have yielded inconsistent results (unpublished data), we did not attempt to quantify the extent of Homer2 over-expression post-AAV–cDNA infusion. However, preliminary quantification of AAV–shRNA-mediated knockdown of Homer2 expression (n = 2) has indicated that the shH2b# 1 construct employed in the present study reduces by approximately 50% NAC levels of Homer2 in B6 mice at 1–2 weeks post-infusion (Klugmann & Szumlinski 2009). To extend these earlier data to a later post-infusion time-point that is more relevant to the duration of the experiments described herein and to determine whether or not strain differences might exist in the efficiency of NAC Homer2 knockdown by our shRNA construct, we employed immunoblotting (see above) to assay the effects of a 0.5 µl/side, intra-NAC infusion of our AAV–shH2b# 1 shRNA construct upon the total protein expression of Homer2b in both D2 and B6 mice killed at 2 months post-infusion. Consistent with our previous in vivo data determined at 1–2 weeks post-infusion (Klugmann & Szumlinski 2009), our AAV–shRNA construct reduced NAC Homer2b levels by approximately 40–50% in both D2 and B6 mice [for D2: t(17) = 4.34, P < 0.0001; for B6: t(18) = 5.01, P < 0.0001], without influencing significantly the levels of Homer1b/c (Fig. 1; for both strains: t tests, P > 0.20). Also consistent with the stable and enduring nature of AAV-mediated neuronal transfection (see During et al. 2003; Klugmann & Szumlinski 2008 for discussion), as well as the immunocytochemical results of previous studies demonstrating AAV-mediated neuronal transfection within the NAC, hippocampus or prefrontal cortex at 1.5–3 months following AAV–cDNA infusion (Cozzoli et al. 2009; Klugmann et al. 2005; Lominac et al. 2005; Szumlinski et al. 2004, 2005, 2006, 2008), we could detect by immunoblotting significant Homer2 knockdown in the NAC of mice at 2 months post-infusion. Importantly for data interpretation, no obvious strain differences were found regarding either the extent of Homer2 protein knockdown or the extent to which our Homer2–shRNA construct influenced Homer1b/c expression between B6 and D2 mice (Fig. 1).
Strain differences in the expression and activational state of Homer-related proteins within the NAC shell
In contrast to alcohol-avoiding D2 mice, alcohol-preferring B6 mice exhibit glutamate sensitization within the NAC shell upon repeated alcohol treatment (Kapasova & Szumlinski 2008), which might relate to strain differences in the expression of Homer2 and associated receptor/intracellular signaling molecules (Lominac et al. 2006; Szumlinski et al. 2005, 2008). Thus, the first experiment assayed for strain differences in the total protein expression of the major CC-Homer isoforms within the NAC shell (Homer1b/c and Homer2a/b), their associated glutamate receptors (the Group1 mGluRs mGluR1/5 and the NR2a/b subunits of the NMDA receptor), as well as kinases activated downstream of glutamate receptor activation (PI3K, ERK and PKCɛ; Kumar et al. 2006; Messing et al. 1991; Olive et al. 2005; Schroeder et al. 2008; Rong et al. 2003). These assays were conducted in both alcohol-naÏve (saline-treated) mice and mice treated with a sensitizing injection regimen of repeated alcohol (8 × 2 g/kg, i.p., every other day; Kapasova & Szumlinski 2008; Szumlinski et al. 2008). The data for this experiment are summarized in Fig. 3.
As illustrated in Fig. 3a, statistical analysis of the data for Homer2a/b showed both a main effect of Strain [F(1,45) = 14.02, P = 0.001] and Alcohol Treatment (EtOH) [F(1,45) = 4.60, P = 0.03], but no interaction between these variables (P = 0.78). Homer2a/b levels within the NAC shell were approximately 50% greater in both alcohol-naÏve and alcohol-treated B6 vs. their respective D2 mice, and repeated alcohol administration elevated Homer2 levels approximately 25% above that of saline-treated animals, irrespective of their strain. In contrast, Homer1b/c levels were not affected by either strain or alcohol treatment (no main effects or Strain × EtOH interaction, P > 0.05). A similar pattern of results were found for both subtypes of Group1 mGluRs as that observed for Homer2a/b [for mGluR1, Strain effect: F(1,41) = 3.02, P = 0.09; EtOH effect: F(1,41) = 11.05, P = 0.002; interaction: P = 0.81; for mGluR5, Strain effect: F(1,40) = 15.56, P < 0.0001; EtOH effect: F(1,40) = 5.00, P = 0.03; interaction: P = 0.22]. Neither strain nor alcohol treatment affected NAC shell levels of NR2a (no main effects or Strain × EtOH interaction, P > 0.05). However, the statistical analysis of the NR2b data showed main effects of both strain [F(1,48) = 7.28, P = 0.01] and alcohol treatment [F(1,48) = 6.21, P = 0.02], as well as a modest interaction between these two factors [F(1,48) = 3.12, P = 0.08]. As is clear from an inspection of Fig. 3a, both the main effects for NR2b can be attributed to the large rise in protein levels observed in alcohol-treated B6 mice.
As illustrated in Fig. 3b, statistical analysis of the total protein expression of non-activated forms of PI3K and PKCɛ did not show significant effects of, or an interaction between, the Strain and EtOH factors (P > 0.05), although a moderate Strain effect was observed for ERK [F(1,48) = 3.14, P = 0.08]. Analysis of the total protein expression of the phosphorylated forms of the kinases indicated no significant differences in p-ERK or p-PKCɛ expression. However, when the data for the phosphorylated forms of the kinases were normalized to their respective total protein expression, a moderate effect of alcohol treatment was observed for the p-ERK:ERK ratio [F(1,44) = 3.46, P = 0.07] and a clear strain effect emerged for the p-PKCɛ:PKCɛ ratio [F(1,48) = 5.11, P = 0.03]. Moreover, the analysis of the levels of p(Tyr)p85 PI3K binding motif showed an overall main effect of strain [F(1,39) = 13.33, P = 0.001], but no main effect of EtOH or Strain × EtOH interaction (P > 0.05). Together, these immunoblotting data for the NAC shell indicate that (1) repeated alcohol treatment can induce persistent (24 h) increases in the activity of PKCɛ and PI3K signaling, which are associated with increases in mGluR/Homer2/NR2 protein expression; (2) irrespective of prior alcohol history, strain differences exist regarding the expression of members of the mGluR/Homer2/NR2-mediated signaling cascades within the NAC shell, with high alcohol preference/intake associated with greater NAC shell protein expression and/or activation and (3) with the exception of NR2b, differences do not appear to exist regarding the alcohol responsiveness of these signaling cascades between B6 and D2 mice.
Strain differences in the expression and activational state of Homer-related proteins within the NAC core
The next experiment examined for the subregional specificity of the observed strain and alcohol effects upon Homer/mGluR/NR2/kinase expression within the NAC shell (Fig. 3) by conducting immunoblotting for our proteins of interest within the adjacent core subregion. The results of this study are summarized in Fig. 4. As can be observed by comparing Fig. 3a for the NAC shell and Fig. 4a for the NAC core, some of the data for Homer and glutamate receptor expression were consistent across the two NAC subregions. For example, B6 mice exhibited elevated NAC core levels of Homer2a/b, as well as mGluR1/5 expression, compared with D2 mice [Strain effect, for Homer2a/b: F(1,47) = 8.15, P = 0.007; for mGluR1: F(1,45) = 4.61, P = 0.04; for mGluR5: F(1,47) = 4.71, P = 0.04] and mGluR1 levels were elevated (albeit moderately) within the NAC core as a consequence of repeated alcohol injection [EtOH effect: F(1,45) = 4.24, P = 0.05]. Also, neither strain nor alcohol treatment affected NAC core levels of Homer1b/c (no main effects or EtOH × Strain interaction, P > 0.05). However, in contrast to the NAC shell (Fig. 3a), neither Homer2a/b nor mGluR5 levels were affected by repeated alcohol treatment within the NAC core (Fig. 4a; no EtOH effect or EtOH × Strain interaction, P > 0.05) and no strain- nor alcohol-dependent changes in NAC core NR2a/b levels were observed (no main effects or EtOH × Strain interaction, P > 0.05).
As illustrated in Fig. 4b, group differences were not observed for the total protein expression of PI3K and for either the non-activated or activated forms of ERK and PKCɛ within the NAC core (no main effects or EtOH × Strain interactions, P > 0.05). Although inspection of Fig. 4b suggested higher p-ERK:ERK ratios in B6 mice and a selective effect of alcohol treatment upon this measure in B6 mice, statistical analysis of the data failed to support this observation (no main effects or EtOH × Strain interaction, P > 0.05). As observed for the NAC shell (Fig. 3b), robust, alcohol-independent, strain differences in the NAC core p-PKCɛ:PKCɛ ratio were observed, with B6 mice exhibiting higher relative levels of the phosphorylated form of the kinase vs. D2 mice [Strain effect: F(1,45) = 11.39, P = 0.002; no EtOH effect or interaction, P > 0.05]. However, in contrast to the NAC shell (Fig. 3b), group differences were not observed for NAC core levels of p(Tyr)p85 PI3K binding motif (Fig. 4b; no main effects or EtOH × Strain interaction, P > 0.05). Taken together, these data indicate that (1) mGluR1 excepted, our repeated alcohol injection regimen does not up-regulate Homer-associated protein expression within the NAC core and (2) strain differences in basal mGluR/Homer2/PKCɛ signaling within the NAC core, with higher basal signaling associated with a high alcohol-preferring/consuming phenotype. As the core subregion of the NAC is critical for maintaining motivated behavior, idiopathic differences in basal mGluR/Homer/PKCɛ signaling within the NAC core may contribute to habitual alcohol drinking (Everitt et al. 2008).
Few strain differences in the expression and activational state of Homer-related proteins within the dorsal striatum
To further explore the subregional specificity of our observed strain and alcohol effects upon striatal mGluR/Homer/NMDA-related signaling, we also examined for differences in Homer/glutamate receptor/kinase expression within the dorsal striatum of alcohol-naÏve and alcohol-treated B6 and D2 mice. The results of this experiment are summarized in Fig. 5.
When compared with the data for the more ventral aspects of the striatum (Figs 3a and 4a), group differences in dorsal striatal Homer2a/b, mGluR1/5 and NR2a/b expression were either very modest or not at all apparent (Fig. 5a). Overall, B6 mice exhibited moderate, but non-significant, elevations in dorsal striatal levels of Homer2a/b and mGluR1/5 [Strain effect, for Homer2a/b: F(1,45) = 3.38, P = 0.07; for mGluR1: F(1,45) = 3.74, P = 0.06; for mGluR5: F(1,45) = 3.07, P = 0.08], but only mGluR5 expression was elevated by repeated alcohol treatment [F(1,45) = 5.64, P = 0.02; interaction: P > 0.05]. However, as observed for both NAC subregions, dorsal striatum Homer1b/c levels were unaffected by either strain or alcohol treatment, and similar to the NAC core, group differences in NR2a/b expression were also not observed (no main effects or EtOH × Strain interactions, P > 0.05).
As illustrated in Fig. 5b, analysis of the kinase data for the dorsal striatum failed to indicate group differences in the total expression of both the non-activated and activated forms of ERK and PKCɛ (no main effects or EtOH × Strain interaction, P > 0.05). Moreover, normalizing the levels of the phosphorylated forms of ERK and PKCɛ to their total protein expression still failed to show any significant group differences in kinase activity. Taken together, these data fail to indicate any short-term effects of repeated alcohol exposure upon mGluR/Homer/PI3K/PKCɛ signaling within the dorsal striatum and do not support a major contribution of striatal mGluR/Homer2/PKCɛ or PI3K signaling in mediating genetic variance in alcohol preference/consumption.
Strain differences in the effects of intra-NAC Homer2 AAV infusion upon NAC glutamate
Our earlier strain comparison indicated that B6 and D2 mice differed in their NAC glutamate response to both acute and repeated alcohol administration (Kapasova & Szumlinski 2008). Thus, to examine the functional relevance of the observed strain differences in NAC shell Homer2 expression (Fig. 3) for alcohol-induced glutamate release within the NAC, in vivo microdialysis was conducted on injections 1 and 8 of repeated alcohol treatment (2 g/kg, i.p., every other day) in groups of B6 and D2 mice infused intra-NAC with AAVs carrying Homer2b cDNA (to over-express), shRNA against Homer2b (to under-express) or a scrambled vector control. Representative pictographs of neuronal transfection by our three different AAV constructs are provided in Fig. 2a–f. As observed previously with our AAV–cDNA constructs (Cozzoli et al. 2009; Klugmann et al. 2005; Lominac et al. 2005; Szumlinski et al. 2004, 2005, 2006, 2008), transfection was apparent in both the cell bodies and processes of NAC neurons, with the good majority of neurons exhibiting membrane-localized immunostaining. Moreover, no obvious differences existed regarding the extent or pattern of immunostaining between our scrambled control, cDNA and shRNA constructs within the NAC (Fig. 2a–f). The results for basal NAC extracellular glutamate levels are presented in Table 1 and the data for alcohol-induced changes in NAC glutamate are summarized in Fig. 6.
Table 1. Summary of the average basal levels of glutamate (±SEM, in ng/20 µl sample) in the NAC of AAV-treated B6 and D2 mice prior to injections 1 and 8 of repeated alcohol treatment (2 g/kg)
The numbers in parentheses represent the sample sizes employed in the statistical analysis of the data.
1.07 ± 0.35 (8)
0.87 ± 0.14 (10)
1.27 ± 0.54 (9)
0.52 ± 0.09 (8)
0.73 ± 0.19 (9)
1.78 ± 0.56 (10)
0.73 ± 0.42 (9)
1.34 ± 0.54 (9)
0.71 ± 0.12 (9)
0.98 ± 0.23 (11)
0.51 ± 0.26 (10)
0.99 ± 0.46 (9)
Basal glutamate levels
As indicated in Table 1, an analysis of the average basal levels of NAC extracellular glutamate collected during the hour prior to alcohol injections 1 and 8 indicated no overall effects of Strain, AAV or Injection Number (P > 0.05). However, we did observe a modest, but significant, Strain × AAV interaction [F(2,110) = 3.2, P = 0.04], which reflected primarily a marginal increase in basal glutamate levels by AAV–cDNA in D2 mice (one-way ANOVA, P = 0.08) and a non-significant trend toward a reduction in basal glutamate by AAV–shRNA in B6 mice (one-way ANOVA, P = 0.21). As there were no consistent alcohol or AAV effects upon basal glutamate levels, the data post-alcohol injection for each group were normalized to their average baseline values to facilitate group comparisons of alcohol-induced glutamate release.
Alcohol-induced changes in glutamate
As observed previously (Kapasova & Szumlinski 2008; Szumlinski et al. 2008), eight intermittent injections with 2 g/kg alcohol elicited a rise in NAC glutamate that depended upon both the AAV treatment and strain of mouse (Fig. 6). This suggestion was supported by the results of the statistical analysis of the neurochemical time–courses indicating a significant four-way interaction [F(22,1089) = 1.84, P = 0.01].
To examine the interactions between AAV and strain upon the glutamate response to acute alcohol, the four-way interaction was deconstructed along the Injection Number factor and this analysis showed a significant three-way interaction for injection 1 [F(22,561) = 5.27, P < 0.0001]. As reported previously (Kapasova & Szumlinski 2008), further deconstruction along the AAV factor showed a significantly blunted glutamate response to acute alcohol in B6 control mice compared with D2 controls (gray symbols in Fig. 6a vs. 6c) [Strain × Time: F(11,176) = 7.21, P < 0.0001]. B6/D2 differences were also apparent between mice infused with cDNA [Strain × Time: F(11,187) = 2.07, P = 0.02], but in the case of the cDNA-infused animals, the converse relationship was observed (black symbols in Fig. 6a vs. 6c). Although cDNA infusion did not significantly affect the glutamate response to acute alcohol in D2 mice (Fig. 6c), consistent with our earlier report (Szumlinski et al. 2008) it facilitated a latent alcohol-induced increase in NAC glutamate levels in B6 mice [AAV × Time: F(22,253) = 3.94, P < 0.0001; post hoc tests]. In contrast, a significant strain difference in the glutamate response to acute alcohol was not observed between the two shRNA-infused groups (open symbols Fig in. 6a vs. 6c; Strain × Time, P = 0.09), which reflected a marked blunting of the alcohol-induced rise in D2–shRNA-infused mice relative to D2 controls [AAV × Time: F(22,308) = 5.74, P < 0.0001; post hoc tests], rather than an effect of shRNA infusion upon the acute glutamate response to alcohol in B6 mice (AAV × Time, P > 0.05).
An analysis of the glutamate response to injection 8 of repeated alcohol treatment also showed a significant three-way interaction (Fig. 6b,d) [F(22,528) = 2.33, P = 0.001] and thus the data were deconstructed along the AAV factor to examine for strain differences. As we have reported previously (Kapasova & Szumlinski 2008), the direction of the differences in alcohol-induced glutamate release between B6 and D2 control mice reversed upon repeated alcohol treatment, with B6 controls now exhibiting an alcohol-induced rise in neurotransmitter levels and D2 controls exhibiting a moderate reduction in glutamate levels below baseline (Fig. 6b vs. 6d) [Strain × Time: F(11,165) = 7.09, P < 0.0001]. Furthermore, and also consistent with our earlier data (Kapasova & Szumlinski 2008), a comparison of alcohol-induced glutamate release between injections 1 and 8 of repeated treatment showed neurochemical sensitization in control B6 mice (Fig. 6a vs. 6b) [Injection Number effect: F(1,15) = 18.39, P = 0.001; Injection Number × Time: F(11,165) = 5.49, P < 0.0001], whereas D2 control mice exhibited neurochemical tolerance (Fig. 6c vs. 6d) [Injection Number effect: F(1,16) = 22.43, P < 0.0001; Injection Number × Time: F(11,176) = 8.88, P < 0.0001]. A similar pattern of strain differences in the glutamate response to injection 8 [Strain × Time: F(11,176) = 3.46, P < 0.0001], as well as in the development of alcohol-induced glutamate plasticity [for B6-cDNA mice: Injection Number × Time: F(11,176) = 2.18, P = 0.03; for D2-cDNA: Injection Number effect: F(1,17) = 9.06, P = 0.008; Injection Number × Time: F(11,187) = 3.50, P < 0.0001] were observed inbetween the two groups of cDNA-infused mice. In contrast to the robust strain differences in repeated alcohol-induced glutamate release in control and cDNA-infused mice, the strain difference in the glutamate response to injection 8 for shRNA was considerably more moderate (Strain × Time: P = 0.05). As illustrated in Fig. 6b, the lack of a significant strain difference between shRNA-infused B6 and D2 mice reflected primarily a complete blockade of repeated alcohol's capacity to elevate NAC glutamate in B6-shRNA mice [AAV × Time: F(22,275) = 2.90, P < 0.0001]. Finally, the lack of a main effect of Injection Number and no significant Injection Number × Time interaction (both P > 0.05) confirmed that intra-NAC shRNA infusion completely prevented the development of alcohol-induced glutamate sensitization in B6 mice (Fig. 6a vs. 6b), as well the development of neurochemical tolerance in D2 mice (Fig. 6c vs. 6d). Thus, consistent with our earlier data for Homer2 KO mice (Szumlinski et al. 2005), knockdown of Homer2 selectively within the NAC is sufficient to prevent alcohol-induced glutamate plasticity in both high and low alcohol-preferring mouse strains.
Strain differences in the effects of intra-NAC Homer2 AAV infusion upon alcohol preference and intake
Our earlier studies conducted in B6 and B6-hybrid mice indicated a necessary and/or active role for NAC Homer2 expression in regulating alcohol intake (Cozzoli et al. 2009; Szumlinski et al. 2005, 2008). Given the genotypic differences in the effects of NAC Homer2 manipulations upon alcohol-induced glutamate release (Fig. 6), we next assessed the functional relevance of NAC Homer2 expression for strain differences in alcohol preference and consumption by infusing intra-NAC AAVs carrying either Homer2b cDNA to over-express Homer2 or Homer2b shRNA to knockdown Homer2 expression. These data are summarized in Fig. 7.
As expected (Belknap et al. 1993; McClearn & Rodgers 1959), D2 mice exhibited an overall lower preference for alcohol-containing solutions compared with B6 mice [Strain effect: F(1,55) = 552.35, P < 0.0001; Strain × Concentration: F(3,165) = 15.16, P < 0.0001], and all D2 groups showed alcohol avoidance at all concentrations in our two-bottle-choice procedure (Fig. 7a). Consistent with recent data indicating that our scrambled control vector does not inadvertently affect alcohol drinking in B6 mice (Cozzoli et al. 2009), the alcohol intake of the mice from both strains infused with our scrambled control vector was consistent with that reported previously by our group (Kapasova & Szumlinski 2008) and the literature (Belknap et al. 1993). Although AAV infusion did not alter the shape of the dose–response function for the relative alcohol preference in either strain [AAV × Concentration: P = 0.43; Strain × AAV × Concentration: P = 0.69], manipulating NAC Homer2 expression produced an overall effect upon alcohol preference that depended upon the strain tested [AAV effect: F(2,55) = 15.31, P < 0.0001; Strain × AAV: F(2,55) = 17.46, P < 0.0001]. In B6 mice, shRNA infusion produced a significant overall reduction in alcohol preference compared with the other two AAV treatment groups [AAV effect: F(2,27) = 19.13; AAV × Concentration: P = 0.48; post hoc tests]. However, despite an obvious rise in the preference for 18% alcohol in cDNA-infused B6 mice (Fig. 7a), post hoc analysis did not indicate an overall effect of Homer2 over-expression compared with scrambled controls. In contrast to the observed AAV effects in B6 mice, intra-NAC AAV infusion did not significantly affect alcohol preference in D2 mice (no AAV effect or AAV × Concentration interaction, P > 0.05).
We next compared the effects of intra-NAC AAV infusion upon alcohol intake (Fig. 7b), the analysis of which showed a significant three-way interaction [F(6,168) = 3.99, P = 0.0001]. Deconstruction of this interaction along the Strain factor indicated that the effects of AAV infusion upon the alcohol intake by B6 mice depended upon the concentration tested [AAV effect: F(2,28) = 15.48, P < 0.0001; AAV × Concentration: F(6,84) = 5.69, P < 0.0001]. As illustrated in Fig. 7b, shRNA infusion lowered the intake of 3%, 12% and 18% alcohol by B6 mice, whereas cDNA infusion elevated their alcohol intake but only at the 18% concentration (one-way ANOVAs, P < 0.05; post hoc tests). In contrast, neither AAV affected the low levels of alcohol intake exhibited by D2 mice (one-way ANOVAs, P > 0.05 for all alcohol concentrations).
In contrast to the above measures of alcohol reward in which AAV effects were observed only in B6 mice, an analysis of the effects of NAC Homer2 over- and under-expression upon water intake during the two-bottle-choice procedures failed to indicate strain differences in the effects of intra-NAC AAV infusion upon water drinking [AAV effect: F(2,56) = 7.66, P = 0.001; no interactions with the AAV factor, P > 0.05]. As illustrated in Fig. 7c, D2 mice exhibited greater water intake compared with B6 mice [Strain effect: F(1,56) = 383.64, P < 0.0001; Strain × Concentration: F(3,168) = 15.64, P < 0.0001] and an overall increase in water intake was observed in shRNA-infused animals of both strains compared with controls, whereas water drinking was not significantly affected in either strain by cDNA infusion (post hoc tests). Not shown, strain differences were not observed regarding the effects of AAV infusion upon total fluid consumption (no main effect of, or interactions with, the AAV factor, P > 0.05), although an overall strain difference in total fluid consumption was observed, with D2 mice drinking a larger volume of fluid than B6 mice, particularly when higher alcohol concentrations was presented [Strain effect: F(1,56) = 41.94, P < 0.0001; Strain × Concentration: F(3,168) = 4.51, P = 0.005]. These AAV data for alcohol reward are consistent with our previous demonstrations of an active role for NAC shell Homer2 expression in regulating alcohol preference and intake in mice with a genetic propensity to consume moderate to high amounts of alcohol (Cozzoli et al. 2009; Szumlinski et al. 2005, 2008) and provide further evidence that alcohol drinking by B6 mice exhibits greater responsiveness to glutamate manipulations of the NAC than that of D2 mice (Kapasova & Szumlinski 2008).
This report describes differences in NAC glutamate receptor-associated protein expression between two inbred mouse strains, C57BL/6J (B6) and DBA2/J (D2) mice, which are highly utilized for studying the neurogenetics of alcoholism (Crabbe 2002; Phillips 1997). Consistent with the existing immunoblotting data for alcohol drinking (Cozzoli et al. 2009; Szumlinski et al. 2008), repeated treatment of both strains with a sensitizing alcohol injection regimen elevated NAC expression of Group1 mGluR subtypes and their associated scaffolding protein Homer2 (c.f., Shiraishi-Yamaguichi & Furuichi 2007). Also consistent with the previous alcohol-drinking studies conducted in both mice (Cozzoli et al. 2009; Szumlinski et al. 2008) and P rats (Obara et al. 2009), no alcohol-induced changes in striatal Homer1b/c expression were observed and the majority of alcohol-induced changes in glutamate receptor/Homer2 levels were selective for the more ventral vs. dorsal striatum.
In the present injection study, the alcohol-induced changes in protein expression were more, rather than less, consistent between the two NAC subregions. These data extend earlier evidence for a hyper-responsiveness of NAC post-synaptic glutamate receptor signaling during early alcohol withdrawal (Cozzoli et al. 2009; Szumlinski et al. 2005, 2008) and indicate that repeated, behaviorally non-contingent, alcohol treatment augments Group1 mGluR/Homer2-mediated signaling within both the ‘motor’ (NAC core) and ‘motive’ (NAC shell) subcircuits of the limbic system (for recent reviews, Everitt et al. 2008; Meredith et al. 2008). Interestingly, in a recent study conducted in chronic (6 months) alcohol-drinking P rats, effects of alcohol intake were observed upon Group1 mGluR, NR2 and Homer2 protein expression in the NAC core subregion only (Obara et al. 2009). Such data are consistent with earlier immunoblotting data from cocaine studies, indicating that Homer-related signaling is activated within different ventral striatal subcircuits by self-administered vs. experimenter-administered drug, with the more dorsal core region exhibiting changes in self-administering, but not experimenter-injected animals (Ary & Szumlinski 2007; Ben-Shahar et al. 2009). Whether or not the apparent subregional differences in the effects of self-administered alcohol (Obara et al. 2009) vs. experimenter-injected alcohol (Cozzoli et al. 2009; Szumlinski et al. 2008) relate to the route of drug administration, the duration of alcohol exposure, other non-pharmacological factors associated with drug administration (e.g. stress of injection, predictability of alcohol effects/expectancy or conditioned factors) or the species employed cannot be discerned at the present time. Nevertheless, such data are consistent with current addiction theories that habitual drug-taking behavior is associated moreso with the activation of more dorsal striatal structures (Everitt et al. 2008; Kalivas & McFarland 2003; Meredith et al. 2008).
Alcohol exposure or re-exposure to alcohol-associated cues is reported to elevate the activational state, translocation or total protein expression of PKCɛ, PI3K and ERK (Cozzoli et al. 2009; Messing et al. 1991; Olive et al. 2005; Schroeder et al. 2008; but see Kumar et al. 2006). In contrast to these earlier results, but consistent with a report indicating a tolerance to the effects of alcohol-induced PKCɛ translocation with repeated alcohol treatment (Kumar et al. 2006), we failed to detect the effects of repeated alcohol treatment upon either total kinase expression or indices of kinase activation at 24 h following the last alcohol injection. Thus, in contrast to repeated bouts of excessive alcohol consumption (∼1.5 g/kg alcohol in 30 min/day for 7 days; Cozzoli et al. 2009) or alcohol-associated cues in rats with a history of response-contingent oral alcohol intake (Schroeder et al. 2008), repeated, experimenter-administered injections of 2 mg/kg alcohol do not elicit an enduring increase in the constitutive activity of these downstream mediators of glutamate receptor signaling, at least when assessed by total protein expression. These discrepant results might relate to the very distinct alcohol delivery procedures employed between the injection and self-administration studies, implicating differences in alcohol pharmacokinetics and/or non-pharmacological factors associated with alcohol delivery (i.e. motivation, stress, control over intake, anticipation and associative learning) in the pattern or duration of alcohol-induced kinase activation within the NAC.
A positive relation exists between a genetic propensity to consume high amounts of alcohol and basal NAC mGluR/Homer2/kinase activity, as assessed by comparing between selectively bred mouse lines and between wild-type and mGluR5 mutant mice that differ in alcohol intake under scheduled, limited-access conditions (Cozzoli et al. 2009). Consistent with these ealier data, alcohol-preferring B6 mice exhibit higher indices of activated PI3K and PKCɛ in the NAC compared with alcohol-avoiding D2 mice. Moreover, the strain differences in basal kinase activity were positively associated with a parallel divergence in basal Group1 mGluR and Homer2 expression. As (1) Group1 mGluRs and Homer2 regulate alcohol-induced glutamate release within the NAC (Lominac et al. 2006; Szumlinski et al. 2005, 2008) and (2) pronounced B6/D2 differences exist in the effects of repeated alcohol administration upon NAC extracellular glutamate (Fig. 6; Kapasova & Szumlinski 2008), one might also expect that repeated alcohol treatment would sensitize glutamate receptor/Homer2-mediated signaling in B6 mice, whereas D2 mice would exhibit either no change in, or a tolerance to, alcohol-induced signaling upon repeated alcohol treatment. Although inspection of the NAC shell data for mGluR5, NR2b and the p-PKCɛ:PKCɛ ratio indicated trends that support for this hypothesis, significant alcohol-dependent strain differences in protein expression were not observed for the NAC shell. Such data indicate that the differential behavioral sensitivity to injected alcohol observed between B6 and D2 mice (Cunningham et al. 1992; Phillips 1997; Phillips et al. 1994) might relate moreso to strain differences in basal post-synpatic glutamate signaling, rather than in the responsiveness of this signaling to repeated treatment with moderate doses of alcohol (i.e. 2 g/kg). Although D2 mice exhibit a shift to the left relative to B6 mice in the dose–response functions for both alcohol-induced locomotor sensitization and conditioned place-preference (Camarini et al. 2000; Cunningham et al. 1992; Lessov et al. 2001; Metten et al. 1998; Phillips 1997; Phillips et al. 1994), repeated treatment with 2 g/kg alcohol is capable of eliciting locomotor sensitization and a conditioned place-preference in B6 mice (Lominac et al. 2006; McGeehan & Olive 2003; Szumlinski et al. 2008, but see Camarini & Hodge 2004). NAC Homer2 over-expression in B6 mice can promote the locomotor-sensitizing and conditioned rewarding properties of injected alcohol (Szumlinski et al. 2008). In this study, the magnitude of the rise in NAC mGluR/NR2/Homer2 expression and kinase activity elicited by repeated treatment with 2 g/kg alcohol was more rather than less strain-independent. From such observations, it is suggested that certain forms of alcohol-induced behavioral plasticity may be mediated by a relative increase in NAC glutamate signaling with repeated alcohol experience, rather than the absolute amount of Homer-mediated signaling through Group1 mGluRs and NMDA receptors following alcohol experience.
Although the average baseline glutamate values obtained in this study are lower than those reported in some of our previous reports (a finding that very likely reflects variance in probe recovery/construction), consistent with the results of our previous neurochemical comparison of B6 vs. D2 mice (Kapasova & Szumlinski 2008), strain differences in NAC baseline glutamate were not observed in alcohol-naÏve mice, suggesting that basal glutamate tone within the NAC is not probably a major contributing factor to the divergent alcohol phenotypes of these two inbred mouse strains. Repeated treatment with 1 g/kg alcohol is reported to elevate NAC extracellular glutamate levels in rats (Melendez et al. 2005) and we reported previously that strain differences exist with respect to the capacity of a repeated 2 g/kg alcohol injection regimen (identical to that employed in the present study) to elevate NAC baseline glutamate levels, with B6, but not D2, mice exhibiting this form of neurochemical plasticity (Kapasova & Szumlinski 2008). However, in the present study, we failed to detect an alcohol-induced rise in NAC glutamate content in B6 mice (Table 1), a finding that is consistent with the results of two other reports from our laboratory demonstrating no significant effect of repeated alcohol injection (at either 2 or 3 g/kg alcohol) upon indices of basal NAC glutamate content in either inbred B6 or B6–129 hybrid mice (Szumlinski et al. 2005, 2008). Interestingly, the two studies in the literature indicating an effect of repeated alcohol injection upon NAC basal glutamate content were conducted in virus-naÏve animals (Kapasova & Szumlinski 2008; Melendez et al. 2005), whereas the negative results in the present study (Fig. 6), as well as those previous from our group (Szumlinski et al. 2005, 2008), were derived from AAV-infused mice. This raises the possibility that our AAVs may be interfering somehow with the development of this specific alcohol-induced neurochemical adaptation in B6 animals. However, arguing against this possibility, repeated bouts of binge alcohol drinking (∼1.5 g/kg/day, every third day for a total of 6 days) do not elevate basal NAC glutamate levels in AAV-naÏve B6 mice (Szumlinski et al. 2007). Alternatively, the discrepancies in findings across studies might be attributable to differences in probe recovery; however, our earlier studies and that conducted by Melendez et al. (2005) all confirmed the presence or absence of alcohol effects in follow-up experiments employing the no net-flux technique, which is independent of probe recovery (Parsons & Justice, 1994 for discussion). Thus, inconsistencies across data sets do not appear to be reflect any obvious issues associated with differential probe recovery. Alternatively, localization of the microdialysis probes within the NAC may be a contributing factor to account for the presence vs. absence of an alcohol effect upon basal NAC glutamate content. In this study, we included data from all animals exhibiting neuronal transfection within the boundaries of the NAC, irrespective of their subregional localization, whereas in our previous strain difference study, data were included only for animals exhibiting placements within the NAC shell subregion (Kapasova & Szumlinski 2008). However, arguing against this case, our probe placements were localized primarily to the NAC shell in our previous AAV studies of B6 and B6–129 hybrid mice (Szumlinski et al. 2005, 2008), whereas those in the rat study by Melendez et al. (2005) did not discriminate between NAC subregions. Thus, at the present time, it is difficult to discern precisely why the effects of repeated alcohol injection upon basal NAC glutamate levels are inconsistent across studies. However, the fact that such inconsistencies exist supports a lesser role for this particular neurochemical adaptation in mediating not only strain differences in alcohol behavioral sensitivity, but also the effects of repeated alcohol administration upon the subsequent behavioral and neurochemical responsiveness to alcohol within a particular strain.
As we reported previously (Kapasova & Szumlinski 2008), D2 mice exhibit a large rise in NAC glutamate in response to an acute injection of 2 g/kg alcohol, whereas an acute alcohol injection or an acute bout of alcohol drinking elicits a very modest or no rise in B6 animals (Fig. 6a; Kapasova & Szumlinski 2008; Szumlinski et al. 2007, 2008). Our earlier data indicate a necessary role for both Group1 mGluRs and Homer2 for acute alcohol-induced glutamate release within the NAC (Lominac et al. 2006; Szumlinski et al. 2005), although the mechanism(s) through which this occurs is elusive, given the post-synpatic localization of Homer proteins (Shiraishi-Yamaguchi & Furuichi 2007). It is interesting to note that inbred mouse strain differences in the NAC glutamate response to acute alcohol appear to be inversely related to NAC glutamate receptor/Homer2/kinase expression (Table 2). Such discrepancies in findings point to genetic variance in other regulators of NAC extracellular glutamate/glutamate release (e.g. Group II mGluR autoreceptors, excitatory amino acid transporters and the cystine–glutamate exchanger; see Baker et al. 2002; Melendez et al. 2005) to account for the divergent acute alcohol glutamate phenotypes between B6 and D2 mice. Although it remains to be determined whether or not the large glutamate response to an acute injection of alcohol observed in D2 mice reflects some general hypersensitivity of NAC glutamatergic terminals, the fact that this strain exhibits lower expression of glutamate receptors/Homer2 suggests that this may be the case. That is to say, the lower glutamate receptor/Homer2 expression exhibited by D2 mice reflects a compensatory downregulation in post-synaptic aspects of glutamate transmission within the NAC of this strain. This hypothesis and the relevance of such neuroadaptations to the increased sensitivity of D2 mice to the acute psychomotor-activating properties of alcohol (Phillips 1997) will be topics of future studies.
Table 2. Comparison of strain differences in NAC protein expression, alcohol-stimulated glutamate release and measures of alcohol reward between control (Scrambled) B6 and D2 mice
Basal and alcohol-induced changes in protein expression
B6 > D2
B6 > D2
B6 > D2
NR2b (only following alcohol Rx)
B6 > D2
B6 > D2
B6 > D2
NAC extracellular glutamate
B6 = D2
Acute alcohol-induced release
B6 < D2
Repeated alcohol-induced release
B6 > D2
B6 > D2
B6 > D2
B6 < D2
However, consistent with an active and necessary role for NAC Homer2 expression in the development of alcohol-induced NAC glutamate sensitization (Szumlinski et al. 2005, 2008), a positive relation exists between B6/D2 strain differences in repeated alcohol-induced NAC glutamate sensitization (Kapasova & Szumlinski 2008) and basal NAC total protein expression of Group1 mGluRs/NR2b/Homer2/activated PI3K and PKCɛ (Table 2). Not precluding a role for the other major molecular regulators of extracellular glutamate mentioned above, such data compliment earlier correlative evidence obtained from studies of binge alcohol-drinking mice (Cozzoli et al. 2009; Szumlinski et al. 2007), as well as significant associations between polymorphisms in the p85α subunit of PI3K with risky alcohol-drinking behavior in human adolescents (Desrivieres et al. 2008), and further one of the working hypotheses of our laboratory that idiopathic increases in mGluR/Homer2-mediated signaling within the NAC facilitates the development of glutamate sensitization, which in turn increases the propensity for subsequent alcohol consumption.
Indeed, activity within glutamate receptor/Homer2/PKCɛ/ PI3K signaling cascades is necessary for the full expression of alcohol intake in rodents with high or moderate alcohol-consuming phenotypes (Backstrom & Hyytia 2004; Besheer et al. 2006; Choi et al. 2002; Cozzoli et al. 2009; Hodge et al. 2006; Lominac et al. 2006; Newton & Messing 2007; Olive et al. 2000, 2005; Schroeder et al. 2005; Szumlinski et al. 2005, 2008), a finding supported further by the present behavioral and neurochemical data for shRNA- and cDNA-treated B6 mice (Table 3). Moreover, immunoblotting data obtained from inbred, selectively bred and transgenic mouse models of genetic variability in alcohol intake provide congruent, albeit correlative, evidence for a link between variance in basal NAC glutamate receptor/Homer2/PKCɛ/PI3K signaling and alcohol preference/intake (Figs 3 and 4; Cozzoli et al. 2009). Although strain differences in alcohol intake/preference between B6 and D2 mice have been attributed to strain differences in taste reactivity (Blizard 2007; McClearn & Rodgers 1959), D2 mice exhibit less operant behavior for intravenous alcohol reinforcement than do B6 mice (Grahame & Cunningham, 1997), indicating that strain differences in mechanisms other than taste reactivity are also contributing to the alcohol-avoiding/low preference phenotype of D2 mice. In further support of this suggestion, an intermittent alcohol injection regimen, akin to that demonstrated presently to elevate NAC mGluR/NR2/Homer2 expression and kinase activation, can promote alcohol consumption in D2 mice (Camarini & Hodge 2004). These latter data suggest that repeated experience with injected alcohol can elicit some form of neuroplasticity that promotes drinking behavior that does not relate directly to orosensory experience. That being said, the rostrocaudal portions of the NAC regulate taste hedonics in rodents and glutamate transmission appears to play a role in this regard (Dourish et al. 1986; Kelley & Swanson 1997; Maldonado-Irizarry et al. 1995; Reynolds & Berridge 2002). As NAC Homer2 over-expression augments free-access alcohol intake in both high alcohol-preferring B6 mice (Szumlinski et al. 2008), as well as more moderate alcohol-drinking B6–129 hybrid mice (Szumlinski et al. 2005) and shRNA-mediated knockdown of NAC Homer2 expression attenuates binge alcohol drinking in B6 mice (Cozzoil et al. 2009), it was hypothesized that if reduced basal NAC shell Homer2 expression actively contributed to the alcohol-avoiding profile of D2 mice (by affecting taste hedonics, general reward mechanisms or both), then Homer2 over-expression in this region (i.e. mimicking the effects of injected alcohol) should promote higher alcohol intake/preference in D2 mice, whereas Homer2 knockdown might further reduce these indices of alcohol reward in this strain. However, our data clearly show that neither AAV construct significantly modified alcohol drinking by D2 mice, despite robust neuronal transfection (Figs 1 and 2) and despite producing the predicted effects in B6 animals (Fig. 7). Interestingly, the strain differences in sensitivity to AAV-mediated changes in NAC Homer2 expression are paralleled by the present data for alcohol-induced glutamate sensitization (Table 3) and are more or less consistent with the results of a recent pharmacological study in which manipulating NAC shell extracellular glutamate levels was found to actively regulate alcohol drinking in B6 mice, but not in D2 mice (Kapasova & Szumlinski 2008). One very plausible explanation for the apparent behavioral and neurochemical insensitivity of D2 mice to NAC glutamate/Homer2 manipulations could relate to their relatively low glutamate receptor expression, the functional consequences of which (i.e. taste hedonics or reward/reinforcement) may not be overcome by simply increasing the amount of a downstream mediator-like Homer2. Unfortunately, AAV or lentiviral vectors carrying cDNA for Homer2-associated glutamate receptors (mGluR1/5, NR2) are not currently available to test this hypothesis directly. However, an alternate route of investigation might involve extending the present results to include other inbred mouse strains that exhibit not only high and low, but also intermediate alcohol-drinking phenotypes, to better correlate genetic variance in intake with basal/alcohol-stimulated changes in mGluR/NR2/Homer2/kinase signaling, as well as the effects of NAC Homer2 manipulations upon measures of alcohol reward.
Table 3. Comparison of the effects of AAV-mediated increases and decreases in NAC Homer2 expression upon strain differences in alcohol-induced NAC glutamate release and alcohol-drinking behavior
+, AAV-mediated facilitation of phenotype expression relative to respective control; −, AAV-mediated reduction in phenotype expression relative to respective control; nc, no change in phenotype expression, relative to respective control.
Acute alcohol-induced glutamate release
Repeated alcohol-induced glutamate release
Alcoholism is considered a complex genetic trait, the precise etiology of which is not known. The present data for B6 and D2 mice compliment earlier findings indicating an important role for glutamate signaling through Homer2-associated pathways within the NAC in mediating genetic vulnerability to alcoholism-related behaviors. If relevant to the human condition, these data provide a potential role for polymorphisms in genes affecting Group1 mGluR/NR2/Homer2/PI3K and/or PKCɛ signaling cascades in the etiology of alcoholism.
Funding for this work was provided by NIH grants AA015351 and AA016650 to K.K.S. The authors wish to thank the laboratory of Dr. Paul F. Worley (Johns Hopkins University School of Medicine) for providing the Homer antibodies employed in this study.