Altered Activity of Lateral Orbitofrontal Cortex Neurons in Mice following Chronic Intermittent Ethanol Exposure

Abstract The lateral orbitofrontal cortex (LOFC) is thought to encode information associated with consumption of rewarding substances and is essential for flexible decision-making. Indeed, firing patterns of LOFC neurons are modulated following changes in reward value associated with an action outcome relationship. Damage to the LOFC impairs behavioral flexibility in humans and is associated with suboptimal performance in reward devaluation protocols in rodents. As chronic intermittent ethanol (CIE) exposure also impairs OFC-dependent behaviors, we hypothesized that CIE exposure would alter LOFC neuronal activity during alcohol drinking, especially under conditions when the reward value of ethanol was modulated by aversive or appetitive tastants. To test this hypothesis, we monitored LOFC activity using GCaMP6f fiber photometry in mice receiving acute injections of ethanol and in those trained in operant ethanol self-administration. In naive mice, an acute injection of ethanol caused a dose-dependent decrease in the frequency but not amplitude of GCaMP6f transients. In operant studies, mice were trained on a fixed ratio one schedule of reinforcement and were then separated into CIE or Air groups. Following four cycles of CIE exposure, GCaMP6f activity was recorded during self-administration of alcohol, alcohol+quinine (aversive), or alcohol+sucrose (appetitive) solutions. LOFC neurons showed discrete patterns of activity surrounding lever presses and surrounding drinking bouts. Responding for and consumption of ethanol was greatly enhanced by CIE exposure, was aversion resistant, and was associated with signs of LOFC hyperexcitability. CIE-exposed mice also showed altered patterns of LOFC activity that varied with the ethanol solution consumed.


Introduction
Alcoholism is a chronically relapsing disorder that is characterized by excessive alcohol consumption and a lack of behavioral flexibility surrounding alcohol consumption. Indeed, addicted individuals will continue to consume excessive amounts of alcohol despite the development of a variety of negative social, legal, and health related consequences. Furthermore, this behavioral inflexibility presents a significant barrier to the implementation of efficacious treatment strategies. Accordingly, it is important to obtain a deeper understanding of the brain systems that underly this behavior and to learn how chronic exposure to alcohol impairs neurons in these regions.
Behavioral flexibility associated with alcohol drinking can be examined experimentally in rodent models using reward devaluation protocols where stimulus-reward relationships are unexpectedly manipulated (Hamilton and Brigman, 2015). Previous studies have shown that mice will self-administer high levels of alcohol and then reduce consumption when the bitter tastant quinine is added to the alcohol solution or when mice are pretreated with lithium chloride to produce illness which is then associated with alcohol drinking (Lopez et al., 2014;den Hartog et al., 2016). However, animals exposed to repeated cycles of chronic intermittent ethanol (CIE) vapor show impaired behavioral flexibility during these procedures and continue to drink high levels of alcohol even following reinforcer devaluation (Lopez et al., 2014;den Hartog et al., 2016).
The lateral orbitofrontal cortex (LOFC), a subregion of the prefrontal cortex, plays important roles in associative reinforcement learning and modifying behavioral responses when action outcome relationships are changed or devalued (Teitelbaum, 1964;Baltz et al., 2018). Accordingly, lesions or inactivation of the LOFC, impair the acquisition of associative learning, and the modification of learned behaviors following reward devaluation (Schoenbaum et al., 2002;West et al., 2011;Gourley et al., 2013;Gardner et al., 2017;Panayi and Killcross, 2018). Furthermore, studies using in vivo electrophysiology have shown that LOFC neurons fire in response to reward predictive cues, signaling the predictive value of the reward and also following reward consumption providing information about the actual value of the reward (Schoenbaum et al., 1998(Schoenbaum et al., , 2003Padoa-Schioppa and Assad, 2006). Additionally, LOFC neurons integrate reward value over time and modify their firing in accordance with new contingencies (Riceberg and Shapiro, 2017).
Using slice preparations, we have previously shown that the intrinsic excitability of LOFC neurons is reduced by low concentrations of ethanol (Badanich et al., 2013;Nimitvilai et al., 2020) and that withdrawal from CIE results in enhanced current-evoked spiking (Nimitvilai et al., , 2018. Furthermore, lesions to or chemogenetic inhibition of the LOFC promotes escalation of alcohol drinking in CIE-exposed mice (den Hartog et al., 2016). Considering the important role of this brain region in behavioral flexibility and its sensitivity to the acute and chronic effects of alcohol, it is not surprising that CIE also impairs performance on LOFC-dependent reversal learning (Badanich et al., 2011) and reward devaluation tasks (den Hartog et al., 2016).
Despite these findings, little is known regarding how chronic alcohol exposure impairs the way that neurons in the LOFC modulate their activity when there is a change in the stimulus-outcome relationship during alcohol selfadministration. The studies in this manuscript address this issue and use fiber photometry and a mouse model of alcohol dependence to assess LOFC neuronal activity during operant self-administration of ethanol in the absence and presence of aversive or rewarding tastants.

Animals
Male C57BL/6J mice at four weeks of age were obtained from Jackson labs and allowed one week to acclimate to the Medical University of South Carolina (MUSC) housing facility before surgery and operant ethanol selfadministration. Lights in the housing facility were set on a reverse light dark schedule with lights off at 9 A.M. and on at 9 P.M. All drinking and fiber photometry studies were conducted during the dark cycle. All procedures were approved by the MUSC Institutional Animal Care and Use Committee and are consistent with NIH guidelines concerning the use of animals in research. We chose to focus these initial studies on male mice as they show robust and reproducible increases in alcohol drinking and selfadministration following exposure to repeated cycles of CIE. In contrast, there is significant variability in CIE-induced drinking in female mice with many studies including our own (Zamudio et al., 2021) showing little to no change in drinking following CIE exposure. This may reflect the higher baseline levels of drinking usually observed in female mice or other as yet identified factors.

Surgery
Mice were deeply anesthetized with isoflurane vapor (Penlon vaporizer; 1 l/min, 3% induction, 1.5-2% maintenance) and 300 nl of AAV1-CaMKII-GCaMP6f (Addgene) was injected into the LOFC (AP: 12.4; ML: À1.35; DV: À2.4 mm). A custom-made optical fiber and ferrule (400-mm diameter patch cord in a 1.25-mm ferrule; Thorlabs) were implanted at these coordinates and secured in place using Herculite. Mice recovered in their home cage for 7 d before beginning lever-press training. All mice were inspected postmortem to ensure proper viral expression and implant location. Any mice with inaccurate placements or lacking viral expression were removed from the study. Mice with damaged fiber optic implants but accurate expression were excluded from photometry analysis because of poor signal quality but were included in drinking microstructure analysis (n = 3 Air, n = 2 CIE).

Postprandial drinking
Access to food and water was regulated throughout this experiment to increase ethanol consumption using a postprandial drinking protocol similar to that described previously (King et al., 2017). Mice were food restricted until 1 h before the start of the daily drinking session. They were then allowed to eat ad libitum for 1 h before and 1 h after the drinking session. Water was continuously available until 1 h before drinking sessions at the same time that mice were given access to food. The weight of the mice did not significantly change as a result of this feeding paradigm.

Operant ethanol self-administration
Mice self-administered ethanol on a fixed ratio one schedule in daily 30-min sessions, Monday through Friday. During these sessions, mice were placed in sound-attenuated Med Associates boxes, with a fan and house light turned on at the beginning of each session. Following an active lever press, mice received auditory (tone) and visual (house light off) cues for 1.5 s while 20 ml of 10% ethanol (v/v) solution were delivered to a drinking well using calibrated Med Associates pumps. The active lever had a 1.5-s time out during the presentation of the cue and reward delivery. A second inactive lever was included in the box that when pressed produced no cues or delivery of the ethanol solution. Drinking at the alcohol well was monitored using Med Associates Lickomoters and licking microstructure was analyzed by grouping bursts of licks into drinking bouts. A drinking bout was determined to be at least three licks occurring with ,1 s in between, similar to Renteria et al. (2020).

Fiber photometry
Data were acquired using custom-built imaging equipment based on that described by the Deisseroth laboratory (Lerner et al., 2015), with modifications. Illumination was provided by 405-and 490-nm fiber collimated LEDs (Thorlabs; 30 mW per channel) connected to a four-port fluorescence mini-cube (Doric Lenses). The combined LED output passed through a 400-mm optical fiber (0.48 NA) pigtailed to a rotary optical swivel (Doric Lenses) and connected to the implanted fiber using a ceramic sleeve or pinch connector (Thorlabs). Emission light was focused onto a photodetector (Newport, model 2151; DC low setting) and sampled at 6.1 kHz by a RZ5P lock-in digital processor (Tucker-Davis Technology) controlled by Synapse software. Excitation light was sinusoidally modulated at 531 Hz (405 nm) and 211 Hz (490 nm) via software control of an LED light driver (Thorlabs). Real-time demodulated emission signals from the two channels were acquired at a frequency of 0.93084 kHz and stored off-line for analysis (Fig. 1B). Lever-presses, head entries and licking at the drinking port were time-locked to fiber photometry data using inputs from the Med Associates hardware to the digital processor. Data were processed using custom-written functions in MATLAB (MathWorks) software. The signals for each channel were first fitted to a polynomial versus time curve and then subtracted from one another to calculate the DF/F time series. Video of the test sessions were recorded using a C930e webcam (Logitech) affixed to the top of the operant chamber. Figure 1A shows the timing of experimental manipulations, operant self-administration training and fiber photometry sessions. Mice (five weeks old) received surgery and were given one week to recover before beginning ethanol drinking. Mice were then exposed to 30-min noncontingent ethanol drinking sessions for two weeks to establish stable drinking of 10% ethanol before the establishment of operant responding for 10% ethanol. Mice then self-administered 10% ethanol for eight weeks; fiber optic recordings were taken weekly for each mouse. Mice were then separated into Air or CIE groups. These groups were counterbalanced by ethanol consumption during baseline drinking sessions. The CIE protocol was similar to those used in previous studies (Becker and Lopez, 2004;Zamudio et al., 2021). Briefly, all mice were injected with the alcohol dehydrogenase inhibitor pyrazole (1 mmol/ kg) before being exposed to either air or ethanol vapors for 16 h/d, Monday through Friday. Mice underwent forced abstinence over the weekend. This pattern of exposure continued for four consecutive weeks. Beginning 72 h after the final ethanol or air-exposure mice, underwent five daily self-administration sessions during which they consumed different solutions. On the first 2 d mice drank 10% ethanol to re-establish drinking behavior, on the next 2 d, they drank 10% ethanol 1 60 mM quinine, and on the last day they drank 10% ethanol 1 5% sucrose (Fig. 1A). During drinking sessions, licking at the ethanol port was monitored using lickometer circuitry and this data were used to evaluate the drinking microstructure including lick rate and bout size.

Statistical analysis
Data were analyzed using Prism 8 software (GraphPad Inc.). Average values were obtained from all drinking bouts or lever presses in each session and used for comparison across group and drink type. Outliers were identified using Grubb's test. When appropriate, repeated measures (RM) one-way ANOVAs with the Geisser-Greenhouse correction were used to compare effects of different drinking solutions on calcium signals within groups. In other instances, RM two-way ANOVAs with the Geisser-Greenhouse correction were used to compare data between groups with the repeated measure being within subject values. Post hoc tests corrected for multiple comparisons were conducted following all ANOVA analyses with significant main effects or interactions. In all cases, values were considered statistically significant when p , 0.05.

LOFC activity during intraperitoneal injection of alcohol
Results from our previous studies using ex vivo slice preparations showed that low concentrations of ethanol reduce current-evoked spiking of LOFC neurons (Badanich et al., 2013). In the present study, we used in vivo fiber photometry to examine whether ethanol also reduces LOFC activity in the awake, freely moving animal. Activity of the calcium indicator GCaMP6f was recorded in the homecage over a 10 min baseline period and then mice were given an intraperitoneal injection of ethanol followed by an additional 20 min of recording. Mice showed spontaneous calcium transients during the baseline period that were detected and quantified using the findpeaks function in MATLAB. There was a sharp increase in the amplitude of the GCaMP 6F signal at the time of ethanol injection followed by a decrease in activity within 10 min ( Fig. 2A, 2D), consistent with the time course of distribution of ethanol to the brain following intraperitoneal injection (Nurmi et al., 1994). A dose of ethanol (2 g/kg, 20% v/v) that is associated with locomotor stimulation (Smoothy and Berry, 1985) reduced the frequency of GCaMP6f transients by ;25% (RM one-way ANOVA, F (1.156,4.624) = 13.73, p = 0.015, Dunnett's multiple comparisons test was significant at 10 min: q = 4.91, p = 0.014 and 20 min: q = 3.69, p = 0.036, n = 5; Fig. 2C). A higher dose of ethanol (3.5 g/kg, 20% v/v) that produces sedation (Aguayo et al., 2014) reduced the frequency of GCaMP6f peaks by ;80% (RM one-way ANOVA, F (1.991,7.965) = 19.74, p , 0.001, Dunnett's multiple comparisons test was significant at 10 min: q = 4.68, p = 0.016, and 20 min: q = 5.82, p = 0.008, n = 5; Fig. 2F). Ethanol had no effect on the amplitude of GCaMP6f events at either the 2 g/kg (RM one-way ANOVA, F (1.181,4.725) = 2.52, p = 0.18, n = 5; Fig. 2B) or 3.5 g/ kg dose (RM one-way ANOVA, F (1.095,4.380) = 2.50, p = 0.18, n = 5; Fig. 2E). These data are consistent with those from our slice electrophysiology studies showing an acute inhibitory effect of ethanol on LOFC neuron firing (Badanich et al., 2013).

Operant ethanol self-administration
Mice were trained to lever press for 10% ethanol under an FR1 schedule of reinforcement. After eight weeks of baseline drinking in daily 30-min sessions, they were separated into CIE or Air groups counterbalanced by baseline ethanol consumption. Mice were then exposed to four consecutive weeks of CIE or air exposure, as described in previous studies (Becker and Lopez, 2004; den Hartog et  Nimitvilai et al., 2016;Zamudio et al., 2021). CIE-exposed animals reached an average blood ethanol concentration of 217.08 mg/dl (SEM = 8.75). Three days after the final CIE exposure, mice returned to operant cages and underwent a week of self-administration trials. During the first 2 d, they lever pressed for 10% ethanol to re-establish self-administration behaviors. On days 3 and 4, the ethanol solution was adulterated with the bitter tastant quinine (60 mM) while on day 5, mice self-administered 10% ethanol sweetened with 5% sucrose.
The microstructure of alcohol drinking in Air-exposed and CIE-exposed mice was assessed by analyzing the number of licks per bout, the number of bouts per session and the number of licks per second. Bout size in air-exposed mice varied significantly across the different drinking solutions (RM one-way ANOVA, F (1.666,13.33) = 12.35, p = 0.0014, Dunnett's multiple comparisons test was significant for sucrose1ethanol: q = 3.12, p = 0.035; Fig. 3D). Likewise, bout sizes were significantly changed following CIE exposure, primarily because of increased drinking of the appetitive solution (RM one-way ANOVA, F (1.648,18.13) = 24.4, p , 0.0001, Dunnett's multiple comparisons test was significant during sucrose1ethanol drinking, q = 5.86, p = 0.0003, and trended toward significance with quinine1ethanol drinking (Dunnett's q = 2.54, p = 0.068; Fig. 3E). When comparing between groups, bout sizes were not significantly altered by CIE exposure (RM two- Figure 2. LOFC activity during intraperitoneal ethanol injection. A, Example trace of 30-min LOFC GCaMP6f activity before and after intraperitoneal injection of 2 g/kg ethanol. The ethanol injection causes a large spike in activity followed by recovery. B, 2 g/kg ethanol does not alter the amplitude of GCaMP6f events. C, 2 g/kg ethanol significantly decreases the frequency of GCaMP6f-mediated events. D, Example trace of 30-min GCaMP6f activity before and after intraperitoneal injection of 3.5 g/kg ethanol. The ethanol injection causes a large spike in activity followed by a sustained reduction in activity. E, 3.5 g/kg ethanol does not alter the amplitude of GCaMP6f events. F, 3.5 g/kg ethanol injection inhibits the frequency of GCaMP6f events; * indicates significant main effects, *p , 0.05, ***p , 0.001, # indicates significant post hoc tests, #p , 0.05, ##p , 0.01. Figure 3. Drinking microstructure. A, Air-exposed mice are sensitive to the devaluing effects of 60 mM quinine and consumed significantly less of that solution compared with baseline. B, Following CIE exposure, mice consumed significantly more ethanol, etha-nol1quinine, and ethanol1sucrose solutions. C, Compared with air-exposed mice, CIE mice consumed significantly greater amounts of the ethanol, ethanol1quinine, and ethanol1sucrose solutions. D, In air-exposed mice, drinking bout size changed depending on the drinking solution, posttests revealed that this was driven by larger bouts during ethanol1sucrose drinking. E, In CIEexposed mice, drinking bout size changed depending on the drinking solution, posttests revealed that this was driven by larger bouts during ethanol1sucrose drinking. F, Bout sizes were significantly larger in CIE animals, posttests revealed that this was driven by increased bout sizes in ethanol1quinine and ethanol1sucrose drinking. G, The number of bouts during a drinking session was not significantly different in any of the drinking solutions for air-exposed mice. H, CIE mice initiated significantly more drinking bouts following CIE exposure when drinking any of the ethanol solutions. I, Compared with air-exposed mice, CIE mice initiated significantly more drinking bouts following CIE exposure when drinking ethanol1quinine or ethanol1sucrose. J, Lick rates during drinking sessions were not significantly different in any of the drinking solutions for air-exposed mice. K, Lick rates were significantly different in CIE-exposed mice with slower rates during ethanol1quinine. L, When compared with air-exposed mice, lick rates were not way ANOVA, significant main effect of drink: F (1.755,33.35) = 32.41, p , 0.0001; but not CIE: F (1,19) = 1.768, p = 0.20; or interaction: F (3,57) = 2.45, p = 0.073, Sidak's multiple comparisons test was insignificant across all solutions, p . 0.05; Fig. 3F).
Finally, there were no differences in lick rate across drinking solutions for air-exposed animals (RM one-way ANOVA, F (2.203,17.62) = 0.34, p = 0.74). CIE treatment produced a small but significant effect on lick rate (RM one-way ANOVA F (2.207,24.27) = 4.89, p = 0.014; Fig. 3J) that was driven by a decrease in lick rate during the ethanol1quinine drinking sessions (Dunnett's multiple comparisons test, q = 3.48, p = 0.013). However, this effect did not reach statistical significance when lick rates between Air-treated and CIE-treated mice were compared with one another (RM two-way ANOVA non-significant main effects of CIE: F (1,19) = 0.63, p = 0.4369, and drink: F (1.851,35.17) = 0.28, p = 0.74, interaction F (3,57) = 2.41, p = 0.076; Fig. 3L). Collectively, these findings demonstrate that CIE exposure alters both the consumption and drinking structure of ethanol containing solutions.

LOFC activity during ethanol drinking
The genetically encoded calcium indicator GCaMP6f was combined with fiber photometry to measure LOFC neural activity during the same self-administration sessions described above. Analysis of these recordings under baseline conditions revealed three main characteristics of LOFC activity associated with drinking bouts. As illustrated in Figure 4A, LOFC activity increased immediately preceding the initiation of a drinking bout (Ramp), decreased during active drinking (Dip), and rebounded immediately after the bout ended (Spike). These characteristic patterns of activity are generally similar to those reported by other investigators during sucrose drinking in rats (Moorman and Aston-Jones, 2014) and non-human primates (Tremblay and Schultz, 1999) and ethanol drinking in non-dependent rats (Hernandez and Moorman, 2020).
In both Air-exposed and CIE-exposed mice, there were no within group differences in the size of the postbout spike in GCaMP6f signal [air: RM one-way ANOVA, F (2.118,10.59) = 3.54, p = 0.07 (Fig. 4G); CIE: RM one-way ANOVA, F (2.199,19.79) = 1.98, p = 0.16 (Fig. 4H)]. However, when spike size was compared between groups, CIEexposed mice showed significantly larger spikes during ethanol1quinine and ethanol1sucrose drinking (two-way ANOVA, main effects of drink: F (2.382,33.35) = 1.39, p = 0.26, CIE: F (1,14) = 0.6, p = 0.45, interaction: F (3,42) = 4.8, p = 0.0058, Sidak's multiple comparisons test was not significant for any drinking solution; Fig.  4I). These data suggest that CIE exposure alters the activity of LOFC neurons after consuming alcohol with different gustatory properties.  Orange ticks are TTL pulses generated during licking at the ethanol port. A characteristic pattern of activity surrounded drinking bouts, with an increase in activity before drinking (Ramp) a decrease in activity during drinking (Dip) and an increase in activity after a drinking bout (Spike). B, No significant differences in the ramp up of activity preceding a drinking bout in air-exposed animals. C, The prebout ramp is significantly different in CIE animals with an earlier and larger increase in activity preceding drinking bouts with ethanol1quinine and ethanol1sucrose. D, There are no significant differences in the size of the dip during drinking bouts in air-exposed animals. E, The size of the dip is significantly different in CIE animals with a significantly larger dip during ethanol1sucrose drinking. F, There was a significant main effect of drink and a significant interaction showing larger dip size in CIE animals. G, The size of the after drinking bout spike was not significantly different across drinking sessions in air-exposed mice. H, The size of the after-drinking bout spike was not significantly different in CIE-exposed mice. I, There was a significant interaction indicating increased spike size in CIE animals. J, Exemplar traces of LOFC GCaMP6f activity in from the same air-exposed mouse during sessions of ethanol (left), ethanol1quinine (middle), and ethanol1sucrose (right) solutions. Exemplar traces of LOFC GCaMP6f activity in the same CIE-exposed mouse during sessions of ethanol (left), ethanol1quinine (middle), and ethanol1sucrose

Discussion
In this study, we examined the signaling characteristics of LOFC neurons in awake behaving mice using GCaMP6f fiber photometry. We report that LOFC neurons are acutely inhibited by non-contingent alcohol administration and become hyperactive following CIE exposure. We also found that LOFC neurons are responsive during both the seeking and consummatory phases of alcohol self-administration. The findings suggest that exposure to CIE may prime the LOFC to engage in the evaluation of gustatory stimuli associated with alcohol while air-exposed animals may rely on regions outside of the LOFC to guide their changes in consummatory behavior.
There is some disagreement in the literature about the effects of acute and chronic alcohol on orbitofrontal cortex activity. Some studies measuring immediate early genes or c-Fos expression have concluded that acute ethanol exposure increases OFC activity (Knapp et al., 2001;Liu and Crews, 2015), while functional studies show that OFC neurons are inhibited by acute ethanol exposure (Badanich et al., 2013). Some of these differences may be attributable to technical differences between studies, such as route of administration (e.g., bath application, intragastric infusion, intraperitoneal injection), time after last ethanol exposure (hours vs days), specificity of effect (individual glutamatergic neurons vs whole-brain region) and continued (right) solutions; * indicates significant main effects or interactions *p , 0.05, **p , 0.01, ***p , 0.001, # used for significant post hoc tests #p , 0.05, ##p , 0.01. Figure 5. LOFC activity during lever pressing. A, Compared with air-exposed mice, CIE-exposed mice had significantly more lever presses during ethanol1quinine and ethanol1sucrose drinking sessions. B, Inactive lever pressing was not significantly different between air-exposed and CIE-exposed mice during any of the drinking sessions, but there was a main effect of drinking session. C, Latency between a lever press and the initiation of a drinking bout was not significantly different across drinking sessions in air-exposed mice. D, In CIE-exposed mice, the latency between a lever press and initiation of a drinking bout was significantly shorter during ethanol, ethanol1quinine, and ethanol1sucrose drinking sessions than baseline drinking sessions. E, In air-exposed mice, there were no differences in LOFC GCaMP6f activity preceding the lever press during any drinking sessions. F, In CIE-exposed mice, there was an increase in LOFC GCaMP6f activity preceding a lever press following CIE exposure during ethanol or etha-nol1quinine drinking. G, In air-exposed mice, there was no difference in LOFC GCaMP6f activity immediately after a lever press during any drinking session. H, In CIE-exposed mice, there was no difference in LOFC activity immediately after a lever press during any drinking session; * indicates significant main effects or interactions **p , 0.01, ****p , 0.0001, # used for significant post hoc tests #p , 0.05, ##p , 0.01. preparation (in vivo vs ex vivo). Previous studies from our lab, using electrophysiological recordings from ex vivo slice preparations, have shown that LOFC neurons are acutely inhibited by bath application of ethanol at concentrations associated with voluntary consumption (Badanich et al., 2013;Nimitvilai et al., 2020). In order to determine whether these effects persist in glutamatergic LOFC neurons in vivo, we examined whether intraperitoneal injections of ethanol would inhibit LOFC activity. Using fiber photometry, we found that both locomotor stimulating (2 g/kg) and sedating (3.5 g/kg) doses of ethanol significantly reduced the frequency of GCaMP6f peaks with the higher dose inhibiting ;80% of activity. In further agreement with previous in vitro studies Radke et al., 2017), we found that CIE exposure and withdrawal increased LOFC activity in vivo as reflected by the increase in the frequency of detected GCaMP6f events.
In this study, mice showed an escalation in drinking following repeated exposures to ethanol vapor but not air. Additionally, we found that while air-exposed mice were sensitive to the devaluing effects of quinine adulteration, CIE-exposed mice were not and actually showed increased drinking of this solution relative to ethanol alone. These data are consistent with previous reports (Becker and Lopez, 2004; den Hartog et al., 2016; Rose et al., Figure 6. Background LOFC activity during drinking sessions. A, In air-exposed mice, there was no difference in the interevent interval of GCaMP6f calcium transients during any of the drinking sessions. B, In CIE-exposed mice, interevent intervals of GCaMP6f calcium transients were significantly shorter following CIE exposure during all drinking sessions relative to pre-CIE activity. C, In airexposed mice, there were no differences in the amplitude of GCaMP6f calcium transients during any drinking session. D, In CIE-exposed mice, there were no differences in the amplitude of GCaMP6f calcium transients during any drinking session. E, In air-exposed mice, there were no differences in the widths of GCaMP6f calcium transients during any drinking session. F, In CIE-exposed mice, there were no differences in the widths of GCaMP6f calcium transients during any drinking session; * indicates significant main effect *p , 0.05, # used for significant post hoc test, #p , 0.05. 2016). In the current study, we used a relatively low concentration of quinine (60 mM) compared with other studies (Hopf et al., 2010;Timme et al., 2020), as we have previously shown (den Hartog et al., 2016) that it clearly separates the groups without completely eliminating drinking in air-exposed mice. Interestingly, CIE-exposed but not air-exposed mice also increased consumption when sucrose was added to the alcohol solution, potentially because of the order that the solutions were presented, with sucrose following quinine. Air-exposed mice may have developed an aversion during the quinine drinking that blunted any effect of sucrose. Although additional sessions might have revealed an effect of sucrose on drinking in the air-exposed mice, CIE induced increases in drinking begin to return to control levels within 7 d following the last CIE exposure thus limiting the period of comparison between groups.
We also examined the microstructure of drinking episodes and found that CIE exposure increased the number of drinking bouts per session, consistent with some previous findings (Robinson and McCool, 2015). Interestingly, a human laboratory drinking study found that an increase in the number of drinking bouts was highly correlated with reported levels of liking a drink (Gero et al., 2019). However, Barkley-Levenson and Crabbe (2015) found that two different strains of mice, selected for high alcohol consumption, consumed high levels of alcohol through either larger bouts or a higher frequency of bouts. Indeed, others have shown that CIE treatment of C57BL/6J mice increases the size of their drinking bouts without any effect on the number of bouts (Renteria et al., 2020). However, these differences may be because of the amount of access that the mice were given; 30 min in our study and 16 h in the study by Renteria et al. (2020) or the concentration of the ethanol solution 10% in our study compared with 15% (Renteria et al., 2020). Further, the addition of sucrose to the ethanol solution in the present study increased the size of drinking bouts in both Air-exposed and CIE-exposed mice, consistent with an increased preference for the sweetened solution (Barkley-Levenson and Crabbe, 2012).
In this study, we identified a characteristic response of LOFC neurons during alcohol seeking and consumption. We demonstrated that activity increases preceding the initiation of a drinking bout, decreases during active alcohol consumption and spikes immediately after drinking. These patterns of activity in the LOFC are similar to those shown by other investigators, using in vivo electrophysiology, during sucrose drinking (Tremblay and Schultz, 1999;Moorman and Aston-Jones, 2014;Hernandez and Moorman, 2020). Interestingly, in the present study, exposure to CIE primed LOFC neurons to alter the magnitude of their activation in response to changes in the gustatory properties of the alcohol solution. Conversely, mice exposed to air did not have significant changes in neuronal activity surrounding alcohol drinking. Considering the important role of the LOFC in processing gustatory stimuli (Rolls, 2015), we were surprised to see that LOFC signaling was unchanged in air-exposed animals during the introduction of appetitive or aversive tastants. One possible explanation is that CIE exposure increases the salience of alcohol and its associated gustatory properties, and that the LOFC is more heavily involved in situations with higher salience. CIE-exposed mice consumed significantly more alcohol following exposure while Air-exposed mice showed a trend toward decreased drinking. Considering that the OFC appears to be especially involved in situations integrating novelty and reward (Elliott et al., 2000), LOFC neurons may be more engaged when the reward is perceived as being higher, such as when alcohol is available following the development of dependence.
A growing number of studies have reported that high concentrations of alcohol (achieved through binge models or vapor exposure) alter the structural, molecular and physiological framework of the LOFC (Coleman et al., 2011(Coleman et al., , 2014McGuier et al., 2015) as well as behaviors that are dependent on LOFC function (Obernier et al., 2002;Badanich et al., 2011;Kroener et al., 2012;Fernandez et al., 2017). While the LOFC is acutely sensitive to ethanol even at low concentrations (Knapp et al., 2001;Badanich et al., 2013), disruption of LOFC-dependent behaviors appears to require higher levels of drinking or chronic ethanol exposure (McMurray et al., 2014(McMurray et al., , 2016den Hartog et al., 2016). Furthermore, lesions or chemogenetic inhibition of the LOFC had no effect on drinking in air-exposed non-dependent mice but increased consumption in CIE-treated animals (den Hartog et al., 2016). These findings support the theory that high levels of alcohol exposure are necessary to disrupt LOFC function, or fully engage it in the evaluation of alcohol-related cues.
The LOFC is thought to maintain a pliable representation of the anticipated value of a reward (Yan et al., 2016). In this study, we found that CIE treatment resulted in enhanced LOFC activity in the seconds preceding active lever pressing during ethanol and ethanol1quinine drinking sessions, an effect that was not seen during baseline drinking or in air-exposed mice. Accordingly, CIE-exposed mice may have increased anticipation or motivation of obtaining the ethanol reward leading to elevated LOFC activity just before the lever press. Following CIE exposure, mice had a decreased latency between lever pressing and alcohol consumption likely reflecting a difference in reward anticipation. Interestingly, this anticipatory prelever activity returned close to baseline by the last day of testing with the ethanol1sucrose solution, although there remained a small but non-significant increase in activity during the second before lever pressing during these sessions. There are several studies that have shown that four cycles of CIE produce a number of effects which return to baseline within a week or two of the final exposure. For example, AMPA/NMDA ratios in LOFC neurons return to baseline levels by 7 d after CIE  and c-Fos activity in the LOFC is similar to air controls by 7 d of withdrawal (Smith et al., 2019). Also, CIE-induced escalations in drinking are attenuated the second week after the final CIE exposure (Lopez and Becker, 2005). Therefore, it is possible that the reduction in prelever activity during ethanol1sucrose drinking is related to the timing of these sessions 7 d after the last vapor exposure.
In conclusion, the results of this study show that repeated cycles of CIE exposure increase alcohol consumption, decrease the devaluing effect of quinine and produce time-locked alterations in the activity of LOFC neurons during operant ethanol self-administration. While these studies did not distinguish between projection specific subpopulations of LOFC neurons, future studies should examine how CIE alters the signaling characteristics of LOFC neurons in neural circuits involved in anticipation of reward delivery, evaluation of reward value, and implementation of flexible behavior.