Effects of Ceftriaxone on Ethanol Drinking and GLT-1 Expression in Ethanol Dependence and Relapse Drinking

Abstract

Repeated cycles of chronic intermittent ethanol (CIE) exposure increase voluntary consumption of alcohol (ethanol) in mice. Previous reports from our laboratory show that CIE increases extracellular glutamate in the nucleus accumbens (NAc) and that manipulating accumbal glutamate concentrations will alter ethanol drinking, indicating that glutamate homeostasis plays a crucial role in ethanol drinking in this model. A number of studies have shown that ceftriaxone increases GLT-1 expression, the major glutamate transporter, and that treatment with this antibiotic reduces ethanol drinking. The present studies examined the effects of ceftriaxone on ethanol drinking and GLT-1 in a mouse model of ethanol dependence and relapse drinking. The results show that ceftriaxone did not influence drinking at any dose in either ethanol dependent or non-dependent mice. Further, ceftriaxone did not increase GLT-1 expression in the accumbens core or shell, with the exception of the ethanol dependent mice receiving the highest dose of ceftriaxone. Interestingly, ethanol-dependent mice treated with only vehicle displayed reduced expression of GLT-1 in the accumbens shell and of the presynaptic mGlu2 receptor in the accumbens core. The reduced expression of the major glutamate transporter (GLT-1) as well as a receptor that regulates glutamate release (mGlu2) may help explain, at least in part, increased glutamatergic transmission in this model of ethanol dependence and relapse drinking.

INTRODUCTION

Excessive alcohol (ethanol) consumption can lead to dependence. Ethanol dependence is associated with many neuroadaptive changes (Hansson et al., 2008; Koob & Le Moal, 2008; Mulholland et al., 2016; Pandey et al., 2017; Spanagel, 2009) that, in turn, can precipitate withdrawal symptoms, increase propensity for relapse and lead to less controlled drinking (Becker, 2008; Heilig et al., 2010; Koob & Volkow, 2016; Vengeliene et al., 2009). One consistent, neuroadaptive feature of chronic ethanol exposure in animal studies is an up-regulation of glutamatergic activity (Gass & Olive, 2008). A variety of studies using microdialysis techniques (Baker et al., 2002; Dahchour & De Witte, 2003; Dahchour et al., 2000; Ding et al., 2013; Griffin et al., 2014; Melendez et al., 2005; Pati et al., 2016) and magnetic resonance spectroscopy (MRS) (Gu et al., 2014; Hermann et al., 2012; Zahr et al., 2009) in rodents show increased extracellular levels of glutamate following chronic ethanol treatment in numerous brain regions including dorsal striatum, NAc, and hippocampus and prefrontal cortex. Studies using other techniques support the idea that chronic ethanol exposure increases neuronal excitability in the brain (Cannady et al., 2018; Jeanes et al., 2011; Kircher et al., 2019; Marty & Spigelman, 2012a; Marty & Spigelman, 2012b; Padula et al., 2015; Roberto et al., 2004). Evidence from MRS studies in human alcoholics paints a more complex picture, showing that glutamate levels change dynamically over time, increasing about 48-72 hours after the last drink but can be lower at other times (Bauer et al., 2013; Hermann et al., 2012; Mon et al., 2012; Prisciandaro et al., 2016). Thus, across different model systems and procedures, chronic exposure to ethanol increases glutamatergic activity.

Studies from our laboratory using a well-characterized mouse model of ethanol dependence and relapse drinking show that chronic intermittent ethanol (CIE) exposure produces significant increases voluntary ethanol consumption (Becker & Lopez, 2004; Griffin, 2014; Griffin et al., 2009b; Griffin et al., 2014; Lopez & Becker, 2005) that has also been reported by a number of other laboratories (Bergeson et al., 2016; Dhaher et al., 2008; Gilpin et al., 2008; Huitron-Resendiz et al., 2018; Jeanes et al., 2011; Padula et al., 2015; Roberto et al., 2004; Roberts et al., 1996). Using in vivo microdialysis techniques, we previously reported that ethanol dependent mice have increased extracellular glutamate concentrations in the NAc (Griffin et al., 2015; Griffin et al., 2014) and pharmacologically increasing or decreasing accumbal glutamatergic concentrations increased or decreased, respectively, ethanol drinking in the model (Griffin et al., 2014). These findings indicate that accumbal glutamatergic transmission is important for regulating ethanol drinking and that the increased glutamate activity noted in the NAc may contribute the increased drinking by dependent mice. These findings are consistent with other reports demonstrating a relationship between glutamate activity in the nucleus accumbens and regulation of ethanol consumption (Kapasova & Szumlinski, 2008; Pati et al., 2016; Szumlinski et al., 2008). Together, these results provide evidence for a significant role for glutamatergic transmission in the addiction process as previously reviewed (Kalivas & O'Brien, 2008; Scofield et al., 2016).

Glutamate levels are tightly regulated in order to maintain low concentrations of glutamate in the extracellular space surrounding synapses. Several membrane bound transporters are important for this regulation, including five known members of the Exictatory Amino Acid Transporter (EAAT) family (Bridges & Esslinger, 2005; Danbolt, 2001). Of these transport proteins, GLT-1 (human EAAT2) is widely expressed in the brain and makes a large contribution to glutamate clearance from the extracellular space (Amara & Fontana, 2002). Because of this role, it is hypothesized that GLT-1 may be an important therapeutic target in a variety disease states that involve altered neuronal excitability (Pajarillo et al., 2019), including addictive disorders (Roberts-Wolfe & Kalivas, 2015). It has been reported that ethanol exposure can reduce GLT-1 expression in key brain areas such as prefrontal cortex, NAc and dorsal striatum (Das et al., 2016; Das et al., 2015; Mulholland et al., 2016) suggesting a mechanism for altered glutamatergic transmission in the context of ethanol dependence. Interestingly, it was also reported that treatment with the beta-lactam antibiotic, ceftriaxone, a third-generation cephalosporin, could increase expression of GLT-1 in vitro and in vivo (Rothstein et al., 2005). Other reports have shown that treatment with ceftriaxone and related compounds such cefazolin could also decrease ethanol drinking in rodent models (Alasmari et al., 2016; Alhaddad et al., 2014; Das et al., 2015; Sari et al., 2013; Stennett et al., 2017). Therefore, the current studies were designed to examine the effects of ceftriaxone in our model of ethanol dependence and relapse drinking. We hypothesized that ceftriaxone treatment would decrease drinking and this would be accompanied by increased GLT-1 expression in the NAc.

MATERIALS AND METHODS

Subjects

Male C57BL/6J mice (10-14 weeks) were obtained from Jackson Laboratories (Bar Harbor, ME) and maintained in a temperature- and humidity-controlled AAALAC accredited facility under a 12 hr light cycle (lights on 0000 hr). Mice were initially group housed during a 2 – 4 week period of acclimation to the vivarium, and then individually housed for the remainder of the experiments. Food and water were available ad libitum at all times. All experimental procedures were approved by the Institutional Animal Care and Use Committee at the Medical University of South Carolina and were consistent with the guidelines of the NIH Guide for the Care and Use of Laboratory Animals.

Chronic Intermittent Ethanol Exposure

Chronic intermittent ethanol (CIE) exposure was administered via inhalation using a well-established ethanol dependence model in mice (Becker & Lopez, 2004; Griffin, 2014; Lopez & Becker, 2005). Briefly, one group of mice (ethanol dependent; EtOH group) received CIE in ethanol vapor inhalation chambers (16 hr/day for 4 days) while the remaining mice (nondependent; CTL group) were similarly handled but maintained in control (air) inhalation chambers. This pattern of CIE (or air) vapor exposure was repeated over three weekly cycles, alternating with intervening weeks when mice resumed voluntary ethanol drinking in the limited access paradigm. Ethanol (95%) was volatilized, mixed with fresh air, and delivered to Plexiglas inhalation chambers at a rate set to yield blood ethanol levels in the range of 175-225 mg/dl. These values were verified by measuring chamber ethanol concentrations (daily) and blood ethanol concentrations (weekly) as previously described (Lopez & Becker, 2005). Prior to being placed in the ethanol vapor chambers, EtOH mice were administered a loading dose of ethanol (1.6 g/kg; 8% w/v; 20 ml/kg volume) and the alcohol dehydrogenase inhibitor pyrazole (1 mmol/kg) by intraperitoneal injection (Griffin et al., 2009a). CTL mice received injections of saline and pyrazole before being placed in control chambers. During the inhalation treatment, the housing conditions were identical to those in the colony room.

Ethanol Drinking Procedures

Mice were trained to drink ethanol in the home cage under a limited access, 2-bottle choice situation as previously described (Becker & Lopez, 2004; Griffin, 2014; Lopez & Becker, 2005). A modified sucrose fading procedure was used, with 15% (v/v) ethanol as the final solution along with tap water as the alternative fluid. The 2-hr drinking sessions started at 1130 hr Monday through Friday. The amount consumed was recorded daily (± 0.1 ml) and body weights were recorded weekly. The position of ethanol and water bottles were alternated randomly to avoid side preferences. Although it is offered as a choice, water consumption was not measured in this study because our previous work indicates that this strain of mouse consumes very little water during this short, limited access procedure as indicated by sporadic and negligible bottle contacts when lickometers are used (Griffin et al. 2009b).

Drugs

Ceftriaxone and cefazolin were obtained from Sigma-Aldrich, Inc (St. Louis, MO). Solutions were made fresh daily by dissolving in normal saline (0.9% w/v) and given intraperitoneally in a volume of 10ml/kg once per day.

Western Blot Procedures

Brains were collected by rapid decapitation, positioned ventral side up in a plastic scintillation vial and rapidly frozen in 2-methylbutane that was super chilled with blocks of dry ice. Brains were stored at −80°C until the western blot assay that followed our routine procedures (Nimitvilai et al., 2017; Padula et al., 2015). At this time, brains were thawed and cut in 2mm serial coronal sections. For Experiment 1, 2mm punches were taken (Core and Shell combined) and, in Experiment 2, the NAc Core and Shell were collected separately using 1 mm punches. All punches were put into centrifuge tubes containing 2% lithium dodecyl sulfate (LDS). Individual samples were homogenized by sonication and an aliquot of each was taken to determine the protein concentration by the bicinchoninic acid assay (Pierce Biotechnology, Inc. Rockford IL). Samples were diluted in 4x LDS and reducing agent (NuPAGE, Invitrogen Corp., Carlsbad, CA) and denatured at 70 °C for 10min. Eight μg of each sample were loaded into 4-12% Bis-Tris MIDI gels with MOPS buffer in the electrophoresis chamber and protein was transferred to PVDF membranes using the Turbo-Transfer PVDF kit (Bio-Rad). The membranes were washed and treated with Swift Membrane Stain (G-Biosciences, St Louis, MO) for total protein visualization using a ChemiDoc MP Imaging system (Bio-Rad Laboratories, Hercules, CA). Membranes were destained, washed in PBST, and blocked with 5% nonfat dried milk (NFDM) for 1hr at room temp on a rocker plate. Primary antibody incubation included rabbit anti-GLT-1 primary antibody (1:1000; Cell Signaling Technology, Danvers, MA; Cat # 3838) or rabbit anti-mGlu2 primary antibody (1:1000; Abcam, Cambridge, MA; Cat # ab15672) diluted in PBST with 0.5% NFDM overnight at room temperature in roller tubes. We selected these two antibodies because of their extensive use in the addiction literature to measure expression of GLT-1 and mGlu2. The GLT-1 antibody was used to validated functional changes in glutamate uptake in chronically stressed rats that were reversed by ceftriaxone treatment (Garcia-Keller et al., 2016). The mGlu2 antibody was validated for western blotting in NP vs P rat lines (Zhou et al., 2013). The following day, membranes were washed and incubated in goat anti-rabbit HRP-conjugated secondary antibody (1:2000, ThermoFisher Scientific, Cat # G-21234) in PBST for 1hr at room temperature. Membranes were then washed and chemiluminescence was visualized after 1min incubation in Clarity Western ECL (Bio-Rad Laboratories, Hercules, CA). Blots with band intensities in the control samples that were within the middle portion of the linear range were selected for analysis. Mean band density for GLT-1 and mGlu2 bands eluting at roughly 65kDa and 100kDa, respectively, were quantified and normalized to the total protein stain using ImageJ-assisted densitometry (National Institute of Health, USA). Cropping of blot images was required in several cases so that treatment comparisons could be shown.

Data Analyses

Analyses were conducted with SPSS® (Version 24) using 2-way analysis of variance (ANOVA) with repeated measures, as appropriate. Significant interactions were followed up using pair-wise comparisons. For all analyses, statistical significance was established with α = 0.05.

Experimental Design

Two experiments were conducted. In both experiments, all mice were trained to drink ethanol in the limited access procedures and experienced CIE exposure procedures up through Test 3. The experiments were conducted differently beginning with Cycle 4 of the CIE exposure. In the first experiment (Figure 1A), the CTL and EtOH mice were divided into 3 treatment groups based on average individual average intakes for Test 3 and administered vehicle (0.9% saline), ceftriaxone (200 mg/kg) or cefazolin (200 mg/kg) intraperitoneally (ip.) once per day beginning on the first day of the 4^th CIE Cycle for 7 days. The initial choice of dose was based on the report from Rothstein and colleagues using 200 mg/kg (2005) showing effectiveness at increasing GLT-1 expression in vivo, though it should be noted that these earlier studies used different mouse strains than the current experiments. The rationale for drug treatment during the CIE procedures was that this procedure distinguished the two groups since both groups drink alcohol in the limited access procedure. Thus, it was expected that GLT-1 expression would be more strongly affected during CIE. The specific hypotheses were that beta-lactam treatment may attenuate or block a reduction in GLT-1 caused by CIE and also reduce drinking. In this experiment, some mice were sacrificed 24 hours after the last treatment without resuming drinking while other mice resumed drinking during Test 4, though the drug treatment was discontinued during Test 4. In this study, cefazolin, another beta-lactam antibiotic expected to have similar effects on GLT-1 was used for comparison (Alasmari et al., 2016). For the western blot analysis, a large area of tissue was collected from the NAc (2mm bilaterally) and both sides pooled.

Figure 1.

Experiment 1 ethanol drinking results. A) The experimental design involved repeated cycles of CIE exposure (or air) alternating with access to ethanol for voluntary drinking. Treatment with vehicle (VEH), cefazolin (CFZ) or ceftriaxone (CTX) occurred during the last CIE cycle but did not continue into the voluntary drinking period (gray box). B) Prior to drug treatment, ethanol dependent mice (EtOH) escalated voluntary ethanol consumption compared to their own baseline (*p<0.05) as well as compared to the non-dependent control mice (CTL; #p<0.05). C) Treatment with VEH, CFZ or CTX did not significantly alter ethanol drinking in either group and the EtOH mice continued to consume more than the CON mice (^p<0.05, indicating main effect of group). Data are means ± s.e.m.

Experiment 2 (Figure 3A) was conducted similarly but a higher dose of ceftriaxone was used in place of cefazolin because of the apparent ineffectiveness of the 200mg/kg ceftriaxone dose (and cefazolin) in Experiment 1. However, we also shifted the drug treatment to occur after CIE in order to more directly observe the effects of the beta-lactam treatment on drinking and so retained the lower ceftriaxone dose in the design. In this experiment, drug treatment began at the conclusion of the 4^th CIE cycle and continued into Test 4 after mice resumed voluntary access to ethanol (drug or vehicle administration occurred 30 min prior to ethanol access). This resulted in 3 treatment groups of 0 mg/kg, 200 mg/kg and 400 mg/kg ceftriaxone. Finally, because of the growing recognition of the functional differences between the NAc Shell and Core subregions, smaller tissue punches were dissected (1mm) and bilaterally pooled from the NAc so that expression of the target proteins could be evaluated in Shell and Core subregions.

Figure 3.

Experiment 2 ethanol drinking results. A) The experimental design was similar to that shown in Figure 1 but the drug treatment was extended 2 days into Test 4 (Gray box) and a higher dose of ceftriaxone was included). B) As in Experiment 1, the EtOH mice increased voluntary ethanol consumption compared to their own baseline (*p<0.05) and compared to CTL mice (#p<0.05). C) Continued treatment with ceftriaxone over two days, including a higher dose, did not influence ethanol drinking over those 2 days in either group and the EtOH mice continued to consume more ethanol than the CTL mice (^p<0.05). Abbreviations are the same as for Figure 1. Data are means ± s.e.m.

RESULTS

Experiment 1

In this study, ethanol dependent mice increased their amount of voluntary ethanol intake (g/kg) with repeated cycles of CIE relative to their baseline intake as well as the non-dependent control mice that experienced similar handling procedures but not ethanol exposure during CIE. The analysis of the data shown in Figure 1B using a 2(Group) X 4(Time) repeated measures ANOVA detected a significant interaction (F (3,192) = 10.259, p <0.0001) as well as main effects of Group (F (1, 64) = 9.426, p < 0.0001) and Time (F (1,64) = 27.632, p < 0.0001). Post-hoc analyses showed that ethanol dependent mice consumed more ethanol during Tests 1, 2 and 3 compared to their own baseline (*p<0.05) as well as the non-dependent control mice during Tests 2 and 3 (#p<0.05).

Beginning 24 hours after completing Cycle 4, dependent and non-dependent mice were sorted into drug treatment groups for once daily administration, balanced according to the individual average ethanol consumption from Test 3. Thus, there were 3 drug treatments per Group: vehicle, cefazolin (200 mg/kg) and ceftriaxone (200 mg/kg) with n= 5-6/group. The data from this phase of the study are summarized in Figure 1C and were analyzed using a 2(Group) X 3(Treatment) ANOVA. This analysis revealed a significant factor interaction (F (2,60) = 3.466, p = 0.008) and a main effect of Group (F (1,60) = 31.518, p < 0.0001) while the effect of Treatment was not significant (F (2,60) = 0.3, p = 0.694). Post-hoc analyses suggested that the factor interaction could be due to the presence of between group differences in the vehicle and cefazolin treatment groups (both p<0.05) but not the ceftriaxone group (p=0.165). However, the post-hoc analysis did not find any differences between vehicle and the drug treatments in either dependent or non-dependent mice (all p>0.117), indicating the drug treatments did not affect ethanol drinking. Finally, analysis of the GLT-1 western blot data summarized in Figure 2 did not reveal a change in total GLT-1 protein expression across both the NAc Core and Shell (2mm punches) in any of the groups (n=5-6/grp). The normalized OD values were analyzed similarly as described for Figure 1B and there were no significant factor interaction or main effects detected (all F’s <1.3, p > 0.05).

Figure 2.

Western blot analysis of total GLT-1 expression in the NAc (2mm punches, encompassing both Core and Shell) was unchanged by ethanol dependence or by drug treatment. Abbreviations are the same as for Figure 1. Data are means ± s.e.m.

Experiment 2

Figures 3, 4 and 5 summarize the results of this study. Again, ethanol dependent mice increased the amount of ethanol intake over time prior to initiating the drug treatment regimen, as expected. The data shown in Figure 3B were analyzed using a 2(Group) X 4(Time) ANOVA which revealed a significant factor interaction (F (3,138) = 10.648, p < 0.0001) as well as main effects of Group (F (1,46) = 10.104, p = 0.002) and Time (F (3,138) = 42.483, p < 0.0001). Post-hoc analyses showed that ethanol dependent mice drank more ethanol than their own baseline (*p < 0.05) as well as the controls (#p < 0.05) during Test 1, 2 and 3.

Figure 4.

Western blot data of total GLT-1 expression from the NAc Core and NAc Shell (1mm punches for each region). A) Neither ethanol dependence nor treatment with ceftriaxone affected GLT-1 expression in the NAc core from either group of mice. B) Interestingly, as noted under the vehicle treatment, ethanol dependence reduced GLT-1 expression in the NAc Shell (#p<0.05). With both doses of ceftriaxone treatment, there was reduced GLT-1 expression in the CTL mice (*p<0.05) while the EtOH mice showed an increase at the highest dose (#p<0.05) and compared to vehicle (*p<0.05). Abbreviations are the same as for Figure 1. Data are means ± s.e.m.

Figure 5.

Western blot data of total mGlu2 expression from the NAc Core and NAc Shell from the same samples shown in Figure 4. A) There was a significant (*p<0.05) reduction in mGlu2 expression in the EtOH mice compared to the CTL mice but ceftriaxone did not affect mGlu2 expression. B) No changes were found in mGlu2 expression in the Shell related to either ceftriaxone treatment or as a function of ethanol dependence. Abbreviations are the same as for Figure 1. Data are means ± s.e.m.

Similar to the first study, mice were again divided into 3 Treatment groups based on their individual average ethanol intake in Test 3 (n= 8-9/group). The 2(Group) X 3(Dose) ANOVA on the data summarized in Figure 3C did not detect a factor interaction (F (2,42) = 1.982, p = 0.151) or an effect of Dose (F (2,42) = 0.394, p = 0.677), though there was a main effect of Group (F (1,42) = 4.726, p = 0.035) consistent with the greater intake by the ethanol dependent mice.

For this study, the tissue collected from the NAc during the dissection procedure was divided into Shell and Core and processed for western blot procedures (1mm punches). The 2(Group) X 3(Treatment) ANOVA on the band densities values shown in Figure 4A for the NAc Core tissue did not reveal any significant effects on GLT-1 expression either by Group or Treatment condition (all F’s <2.320, p > 0.373). On the other hand, the same analysis on the data from the NAc Shell found a significant interaction (Figure 4B; F (2,31) = 6.821, p = 0.004) and a main effect of Dose (F (2,31) = 3.398, p = 0.046) but not Group (F (1,31) = 0.045, p = 0.833). Post-hoc analyses revealed that both doses of ceftriaxone reduced GLT-1 expression in the non-dependent control mice compared to the vehicle condition (*p<0.05 versus vehicle). Interestingly, GLT-1 expression was reduced in the ethanol dependent mice compared to the control mice and that the highest dose of ceftriaxone reversed this effect (#p<0.05) though this increase in GLT-1 was not associated with changes in ethanol drinking.

Finally, enough tissue remained from Experiment 2 to examine total tissue expression of the presynaptic Group II metabotropic glutamate receptor, mGlu2. This receptor is of interest because of its role in regulating neuronal glutamate release (Cartmell & Schoepp, 2000; Lovinger & McCool, 1995). The data are summarized in Figure 5. Initial analyses of these western blot data were identical to that described for GLT-1. These ANOVAs did not reveal significant effects of group or treatment (all F’s <2.600, p > 0.090). However, there appeared to be a tendency for mGlu2 expression to be lower in dependent mice versus the non-dependent mice after vehicle treatment, an important a priori comparison considering the lack of effect of ceftriaxone on expression. Student’s T-test showed that total mGlu2 expression was reduced in the NAc Core in dependent mice compared to non-dependent mice (t = 2.832, p = 0.016), but not in the NAc Shell (t = 1.284, p > 0.2).

DISCUSSION

Consistent with our previous work these data indicate that there were robust increases in ethanol drinking in the ethanol dependent mice. However, treatment with the beta-lactam antibiotic ceftriaxone did not influence ethanol drinking in this model of ethanol dependence and relapse. Further, we found that neither ceftriaxone nor cefazolin systematically influenced GLT-1 expression in the NAc. However, analysis of GLT-1 expression in the NAc Shell subregion found a decrease in expression in ethanol dependent mice (~20%) and the highest dose of ceftriaxone reversed this decrease. Lastly, in these same samples, we examined the expression of mGlu2, a presynaptic glutamate receptor that regulates glutamate release (Cartmell & Schoepp, 2000; Lovinger & McCool, 1995) and is known to be important in regulating alcohol drinking (Meinhardt et al., 2013; Zhou et al., 2013). However, while mGlu2 was not influenced by ceftriaxone treatment, there was evidence for reduced expression of mGlu2 in the NAc Core of ethanol dependent mice.

The CIE model of ethanol dependence and relapse has been well characterized by our laboratory (Becker & Lopez, 2004; Griffin, 2014; Griffin et al., 2009b; Griffin et al., 2014; Lopez & Becker, 2005; Lopez & Becker, 2014; Lopez et al., 2012) and numerous other laboratories (Bergeson et al., 2016; Dhaher et al., 2008; Gilpin et al., 2008; Huitron-Resendiz et al., 2018; Jeanes et al., 2011; Roberto et al., 2004; Roberts et al., 1996). These reports reliably show increases in alcohol drinking over time as rodents are rendered dependent using the CIE procedures and given the opportunity to voluntarily drink ethanol. Our earlier work identified that the increased drinking in dependent mice was associated with increased extracellular glutamate in the NAc (Griffin et al., 2015; Griffin et al., 2014). Pharmacologically increasing glutamate levels in the NAc by microinjection of the non-selective glutamate transporter inhibitor TBOA could increase drinking in dependent mice as well as non-dependent mice, with the latter group increasing their intake enough so as to be indistinguishable from dependent mice in terms of intake (Griffin et al., 2014). Lowering accumbens glutamate levels by local injection of a mGlu2 agonist reduced drinking in both groups, with a more pronounced effect in the dependent mice (Griffin et al., 2014). These findings strongly implicate aberrant glutamatergic neurotransmission in the elevated drinking observed in ethanol dependent mice.

In P rats, alcohol consistently reduces GLT-1 expression and ceftriaxone reverses this effect (Das et al., 2016; Das et al., 2015). Importantly, ceftriaxone treatment also reduces alcohol intake in both inbred and outbred rat strains (Alhaddad et al., 2014; Das et al., 2015; Sari et al., 2013; Stennett et al., 2017). Further, the increased GLT-1 caused by treatment with ceftriaxone restores extracellular glutamate concentrations in a rat model of ethanol drinking (Das et al., 2015) as well as in a model of cocaine self-administration (Trantham-Davidson et al., 2012). However, in the two experiments reported here, GLT-1 expression was not significantly affected in either group of mice by ceftriaxone, though there was a specific increase in the NAc Shell in the ethanol dependent mice at the highest dose. Despite this specific increase in GLT-1, there was no effect of ceftriaxone on ethanol drinking in either experiment or at either dose. We also tested another beta-lactam antibiotic, cefazolin, which has been shown to increase GLT-1 expression and reduce alcohol drinking and the reinstatement of alcohol-seeking in a manner similar to ceftriaxone (Alasmari et al., 2016; Weiland et al., 2015). In the current study, treatment with cefazolin did not affect drinking and also did not alter total GLT-1 expression. Thus, counter to our hypothesis, we did not find that ceftriaxone (or cefazolin) altered GLT-1 expression or alcohol drinking in our model of ethanol dependence and relapse. Interestingly, a recent study using outbred rats found that, while ceftriaxone treatment did reduce ethanol drinking, it did not affect nucleus accumbens core GLT-1 expression (Stennett et al., 2017). Lastly, although we did not find that ceftriaxone altered mGlu2 expression, there is new data indicating that ceftriaxone treatment can increase expression of mGlu2 receptors in the nucleus accumbens core (Logan et al., 2020).

The reasons why ceftriaxone did not influence GLT-1 expression or alcohol drinking in the present experiments are unknown. Of course, one consideration is species differences. The relevant ethanol studies showing effects on ethanol drinking and expression of GLT-1 cited above were conducted using the P strain of high drinking rats. However, studies using ceftriaxone to increase GLT-1 have reported been in mice [e.g. (Rothstein et al., 2005)] and include a study showing that ceftriaxone normalized GLT-1 expression compared to wild-type controls in a genetic mutant with naturally low GLT-1 (Lee et al., 2013). However, across multiple rodent models of neurological and substance use disorders, ceftriaxone is most effective at increasing GLT-1 expression when it has been decreased by disease/insult (Smaga et al., 2020). Considered in the context of the data presented here using C57BL/6J mice, in addition to the role of species, it is possible that ceftriaxone was unable to alter GLT-1 expression because it is not reduced in the nucleus accumbens core in this model. In support of this idea, where GLT-1 expression was reduced by ethanol in the accumbens shell, ceftriaxone was able to restore expression. Additionally, Stennett and colleagues found that an outbred rat strain did not show increased GLT-1 following treatment with ceftriaxone, nor did they show reduced GLT-1 expression after alcohol (Stennett et al., 2017).

We did find increases in GLT-1 expression in ethanol dependent mice at the highest ceftriaxone dose in the NAc Shell, although it was without effect on alcohol drinking, suggesting that future experiments could possibly examine higher doses of ceftriaxone in our mouse model. Following from this observation, another consideration regarding differences in the present study compared to other reports is the brain penetration of these compounds. As a group, cephalosporin antibiotics generally have low blood-brain-barrier penetration. Microdialysis studies in rats show that only 2% of ceftriaxone (Granero et al., 1995) and only 6% of cefazolin (Tsai & Chen, 2000) reach the brain compartment. While it is clear from the other reports cited here that ceftriaxone and cefazolin do access the brain and modulate GLT-1 expression, it is possible that there are as yet undefined characteristics of the blood-brain-barrier that vary across species, strains, or experimental conditions that reduce the brain penetration of these compounds. Intriguingly, there is growing evidence that ethanol exposure compromises the integrity of the blood-brain-barrier (Rubio-Araiz et al., 2017), offering a potential explanation for why ceftriaxone treatment lead to a selective increase in GLT-1 in the ethanol dependent mice compared to the non-dependent control mice in our study. Further, while ceftriaxone has received much attention for its ability to modulate GLT-1 expression, it should be noted that other beta-lactam containing molecules are also capable of modulating GLT-1 expression and ethanol drinking (Althobaiti et al., 2019), including some that have efficacy when given orally (Hakami et al., 2017). Taken together, these observations leave open the possibility of future studies with our model under different experimental conditions or using different beta-lactam compounds.

Despite not observing effects of ceftriaxone, we did find differences in GLT-1 and mGlu2 expression between the ethanol dependent and non-dependent mice in the vehicle-treated groups. The reductions in GLT-1 and mGlu2 in ethanol dependent mice are consistent with other reports also showing reductions in the context of ethanol exposure (Barker et al., 2016; Das et al., 2016; Das et al., 2015; Meinhardt et al., 2013). Specifically, in the data presented here, there was a decrease in GLT-1 expression in the NAc Shell of ethanol dependent mice and a decrease in the pre-synaptic mGlu2 receptor in the NAc Core of dependent mice. These findings are interesting because our earlier work, using microdialysis, targeted the border region between Shell and Core (Griffin et al., 2015; Griffin et al., 2014). Thus, reductions in two proteins with significant influence on glutamate homeostasis could explain, at least in part, the increased extracellular glutamate reported earlier (Griffin et al., 2015; Griffin et al., 2014). This idea is also consistent with a report from Ding and colleagues where it was shown that reduced mGlu2 expression occurred in concert with higher extracellular glutamate in the NAc Core in a model of rat ethanol drinking (Ding et al., 2013). It is important to note that, in the present study, reduced expression of GLT-1 and mGlu2 were not explicitly related to altered glutamate concentrations and future experiments will be needed to confirm this hypothesis. Further, these findings suggest that the neuroadaptive mechanisms leading to glutamatergic transmission in the context of ethanol dependence may differ between brain areas.

In conclusion, in this model of ethanol dependence and relapse, treatment with ceftriaxone did not affect ethanol drinking in either dependent or non-dependent mice nor did it strongly affect GLT-1 expression. However, ethanol dependence was associated with decreases in key proteins known to significantly influence glutamate reuptake (GLT-1) and glutamate release (mGlu2) that point to future studies examining the regulation of expression of these proteins within the context of ethanol dependence.

Highlights.

Ethanol dependence increased ethanol drinking and was associated with reduced GLT-1 expression in the nucleus accumbens shell and mGluR2 expression in the nucleus accumbens core.

Treatment with the beta-lactam antibiotic ceftriaxone did not influence ethanol drinking in dependent and non-dependent mice.

Treatment with the highest dose of ceftriaxone increased GLT-1 expression in the nucleus accumbens shell of dependent mice.