Supplemental: Astrocytes Guide Responding for Reward and Ethanol

Supplemental: Astrocytes guide responding for reward and ethanol

Supplemental Legends

S1. Further characterization of nucleus accumbens astrocytes following abstinence. A) Examples of GFAP+ staining following the various EtOH self-administration paradigms studied. Tissues were Nissl counter-stained. B) No NAcore GFAP expression change was detected with immunoblotting 3 wks after operant EtOH self-administration (p=0.99). C) In the same sections used for the NAcore analysis presented in Figure 1, we found increased NAshell astrocyte density (1-way ANOVA: F(4,25)=9.27, p=0.0001). A Dunnett’s post-hoc test revealed a significant increase (p<0.05) following either 24h or 3 wks abstinence from EtOH-IA. This is distinct from NAcore astrocytic adaptations. In further contrast, no change in NAshell astrocyte density was detected following either intermittent sucrose consumption or operant EtOH self-administration. D) No change was observed in NAshell non-GFAP positive cell packing density following any treatment (p=0.32) or E) volume (p=0.20). F) Representative staining of a second astrocytic marker aldehyde dehydrogenase 1L1 (ALDH1L1) (Cahoy et al., 2008). Bar = 10μm. G) Sections that were near-adjacent to those used to count GFAP+ cells were used for proportional counting of cells expressing ALDH1L1 (ALDH1L1+). No change was detected in the number of ALDH1L1+ cells (p=0.16). H,I) Immunoblotting was performed for two other markers using the same homogenate that was used to measure GFAP. No change was observed for H) ALDH1L1 expression (p= 0.52), but I) glutamine synthetase expression was increased (F(1,20)=16.88, p<0.001). β-tubulin III served as a procedural control. n: naive, e: 3 wks abstinence from operant EtOH. S-IA: intermittent sucrose (suc) access, EtOH-IA: intermittent EtOH access. Data represent mean±SEM, *p<0.05, **p<0.01.

S2. Supporting data for gap-junction hemichannel blocker microinjection study. A one-way ANOVA did not reveal any effect on A) inactive-lever responding (p=0.31) or B) latency to the first press (p=0.37) following microinjection of gap-junction hemichannel blockers into the NAcore. C-E) Microinjection sites. 18-α: 18-α-glycyrrhetinic acid. Data represent mean±SEM.

S3. Supporting data for Gαq-DREADD-modulated rewarding brain stimulation. A) Representative volume of transfected cells. B) A lack of chromatolysis in Nissl stained near-adjacent sections suggested that overt toxicity did not occur following experimentation. A two-way ANOVA did not reveal a significant treatment effect of C) 1.5mg/kg Clozapine-N-oxide (CNO: p=0.86) or D) 6mg/kg CNO (p=0.10) on lever responding for rewarding brain stimulation by Gαq-DREADD expressing rats suggesting that the modulation occurred within a narrow range of astrocytic activity. Due to solubility limits, the 6mg/kg dose was given at 2mL/kg. No effect of E) 3mg/kg CNO was observed in non-DREADD expressing rats (p=0.22). As a positive control, cocaine was evaluated. F) Cocaine significantly facilitated reward thresholds regardless of Gαq-DREADD expression. Responding by rats expressing Gαq-DREADD is illustrated (frequency: F(9,45)=43.59, p<0.0001, treatment F(1,5)=68.50, p=0.0004. G-H) Electrode and cannulae placements. % MCR: percent maximum control rate, which was calculated by (Response Rate During a Frequency Trial/MCR)x100. Data represent mean±SEM, *p<0.05.

S4. Supporting data for Gαq-DREADD-modulated EtOH breakpoint. A) A one-way ANOVA did not reveal any difference in latency to the first press when comparing vehicle to CNO in Gαq-DREADD expressing rats (p=0.68) or Gαq-DREADD to GFP following CNO administration (p=0.14). B-C) Cannulae placements.

Supplemental Materials and Methods

Astrocyte Cultures. Primary astrocyte cultures were prepared from P0-P1 CD-1 mice. Striatum was digested (2.5mg/ml, DNase 15μg/ml, 30min), triturated (10% FBS, 6g/L glucose, 0.1% NaHCO3, 50U/mL pen/strep), and plated onto petri dishes that were freshly coated with poly-L-lysine. After 8 days, microglia and progenitors were mechanically removed and astrocytes subcultured onto glass-bottom (35mm MatTek, Ashland, MA) dishes at 3K cells/ plate. After 3 additional days in culture, cells were transfected with AAV8 expressing either Gαq-DREADD or GFP under the GFAP promoter.

ICSS. Apparatus. Experiments were conducted in sound-attenuating chambers that enclosed standard operant chambers (29.2×30.5×24.1cm), which were equipped with a response lever (4.5cm wide that extended 2.0cm through the center of one wall located 3cm above the floor), three stimulus lights were positioned 7.6cm above the lever. Chambers were equipped with a 2W white house light and an ICSS stimulator (Med Associates, St. Albans, VT). Electrodes were connected to the stimulator via a bipolar cable and commutator (Model SL2C, Plastics One, Roanoke, VA).

Training. Sprague Dawley were used in order to facilitate comparison of ICSS data with earlier literature. Rats were trained under a fixed-ratio 1 (FR1) schedule for electrical brain stimulation. Each lever press resulted in a 0.5s train of square wave cathodal pulses (0.1ms pulse duration). Stimulation was accompanied by stimulus light illumination. Responding during the 0.5s stimulation period had no programmed consequence. During the initial training phase, session duration was 30-60min, stimulation frequency was held constant at 158Hz, and stimulation intensity was adjusted to the lowest value that would sustain reinforcement with at least thirty stimulations per minute. Once this criterion was met, frequency manipulations were introduced during training sessions. These sessions consisted of sequential 10min components. During each component, a descending series of ten current frequencies (158-56Hz in 0.05 log increments) was presented, with a 60s trial at each frequency. Each frequency trial began with a 5s timeout followed by a 5s “priming” phase, during which animals received five non-contingent stimulations with a 0.5s interval between each stimulation. This non-contingent stimulation was followed by a 50s “response” phase, during which responding produced electrical stimulation under a FR1. Training continued with three to twelve sequential components per day, during which current intensity was adjusted until reliable responding was achieved for at least the first three and no more than six frequency trials over at least three consecutive days. This intensity (range: 110-220μA) was held constant for the remainder of the study. Additionally, rats were habituated to saline injections until these injections had no significant effect on ICSS frequency-rate curves as determined by a two-way ANOVA. Next, for virus-treated rats, a pre-virus ICSS baseline was established over three consecutive days. After virus injection, training continued for 2 wks.

Testing. Testing proceeded over a period of 11 days in two phases. For all phases, test sessions consisted of six ICSS components. The first component of each daily session was considered acclimation, and data from this component were discarded. Data from the second and third components were averaged to generate baseline data for that session. Next, drug or vehicle was administered and testing resumed 30min later. Data from the following three components were averaged to generate treatment data for that session. Data were expressed as percent maximum control rate (% MCR), which was calculated by (Response Rate During a Frequency Trial/MCR)x100. Vehicle or CNO (0.05% DMSO, 1.5, 3 or 6mg/kg) was evaluated during phase 1, which occurred on days 14-23 after virus injection. All rats received all doses in a crossover, within-subjects design. Test sessions were separated by at least 48h, and three-component training sessions were conducted on weekdays between test days. Cocaine (10 mg/kg, ip) was evaluated as a positive control in Phase 2, which occurred on Day 24, using session parameters just described. A second group served as a non-virus control. Training and testing proceeded as described above except that only 0.05% DMSO, 3 mg/kg CNO and 10mg/kg cocaine were evaluated.

Operant EtOH. Apparatus. Standard test chambers (29x25x29cm, Coulbourn) were enclosed by sound-attenuating cubicles. Inside the test chamber, a 100μL liquid dipper was located between two retractable levers (7.5cm above floor), and cue lights were located 8.5cm above each lever. Activation of a house light, fan, and lever extension indicated session start. After completion of the fixed ratio schedule of reinforcement, the dipper cup containing the reinforcer was elevated, a stimulus light above the active lever was illuminated, a tone was activated, and the dipper cup was illuminated; all for 4 sec. If the cup was not actively licked during the first 2 sec, the cup fell and this event was recorded as a null response and not included in g/kg calculations. For the cup to rise again, the rat had to complete another fixed ratio schedule.

Breakpoint. After 3 wks abstinence, rats were presented with an EtOH odor cue generated by 15ml of 87% EtOH (v/v) sprinkled beneath the previously EtOH-paired lever. After 2min, the levers were extended and a compound cue was presented (tone, stimulus light above the active lever, and illumination of the EtOH-filled dipper cup). Sucrose-trained rats were similarly treated, but bedding was sprinkled with water instead of EtOH. The reinforcement schedule was rounded from [5e(reinf # x 0.1)]-5. Breakpoint was defined as active lever depressions in the last completed ratio. After 15min of inactivity sessions timed out. Sessions generally ended within 1hr.

Intermittent EtOH or sucrose access. Rats in the intermittent EtOH cohort (EtOH-IA) had EtOH (20% v/v) and water access (24h starting at 6pm every Monday, Wednesday, and Friday). No EtOH was available Tuesday, Thursday, Saturday, and Sunday. Bottle position was alternated with each presentation. For the sucrose intermittent access cohort (S-IA), sucrose (2% w/v) was substituted for EtOH. After 3mo of intermittent access, all rats underwent 5-7, 15h operant training sessions for either EtOH or sucrose. This allowed determination of a breakpoint in the intermittent access cohorts, as we have previously reported (Hopf et al., 2010).

Stereology. Tissue preparation. The experimenter was blinded to treatment. Rats were deeply anesthetized with pentobarbital (100mg/kg, ip) and transcardially perfused in treatment pairs (300mL cold PBS and 300mL 4% paraformaldehyde), brains were postfixed (2% paraformaldehyde, 15h, 4˚C), cryoprotected (20% sucrose followed by 30% sucrose, 4˚C), and rapidly frozen (-80˚C). Next, 40μm thick serial sections were collected on a cryostat and stored in cryoprotection solution (25% glycerin, 25% ethylene glycol in 46mM NaH2PO4, 154mM Na2HPO4, −20˚C). Every sixth section throughout the region of interest was selected for analysis. Sections were permeabilized and blocked (0.02% TritonX-100, 1% normal goat serum, PBS, 1hr, RT), incubated with primary GFAP antiserum in blocking buffer (1:15,000, polyclonal, Dako) at 4˚C for 15h. Next, sections were washed with PBS and incubated in biotinylated secondary followed by avidin-biotin solution (Vectastain Elite, Vector Labs) prior to developing with DAB (Vector Labs) according to the manufacturer’s instructions. For the Nissl counterstain, sections were mounted onto gelatin coated slides, dried, and then rehydrated through steps of 95%, 70%, and 50% EtOH in tap water. Following rehydration, sections were incubated in cresyl violet (1.5min),rinsed in tap water, and differentiated in acetic acid/EtOH. Tissues were then dehydrated through 50%, 70%, 95%, and 100% EtOH prior to clearing in xylene and coverslipping with Cytoseal (Thomas Scientific).

Measurement. The experimenter was blinded to treatment during stereological quantification. GFAP+ cell density was measured on a bright field microscope (BH-2, Olympus Corporation) and CCD camera (TXD13C, Baumer Optronic GmbH) that were interfaced to a computerized stereology system (Stereologer, Stereology Resource Center) and integrated with a motorized stage(Prior Scientific). The systematic-random sampling of the region of interest (described above) generated7-10 sections through the NAcore and7-9 sections through the NAshell. The reference space was outlined at low power (4x). Both the atlas of Paxinos and Watson and a 1:6 series of calbindin D28k-stained adjacent sections (1:5,000, monoclonal, Sigma) were consulted to aid tracing (Hedou et al., 2002).The subventricular zone and rostral migratory stream werecarefullyexcluded from analysis. Cells were counted at high magnification (63x, 1.4NA, oil-immersion, PlanApo, Zeiss). Quantification was performed using the optical fractionator. Quantification adhered to both the dissector principle and unbiased counting rules (Mouton, 2002). Only GFAP+ cell bodies falling inside the counting frame or touching the green line of the counting frame were counted. In addition, Nissl positive, but GFAP negative, cell bodies were counted.On average, about 200 GFAP+ cells in 100-200 disectors per subject and structure were countedwith an optical disector height of at least 10μm anda guard zone of 1μm.This resulted in 10% coefficient of error. Packing density was calculatedas the total number of cells counted in the reference space divided bythe product of total disector number and the disector volume (Courchesne et al., 2011; Miguel-Hidalgo, 2005).

Proportional Counting. The experimenter was blinded to treatment during quantification. Sections that were near-adjacent to those used for stereological analysis of GFAP+ cells were used to count ALDH1L1+ cells. Sections underwent antigen retrieval: cryoprotectant was rinsed off in PBS and buffer exchanged to citrate buffer (10mM citric acid, 0.05% Tween20, pH 6.0) prior to uncloaking in pre-warmed citrate buffer (98˚C, 20min). Sections were allowed cool to room temperature and buffer exchanged to PBS. Sections were permeabilized and blocked (0.02% TritonX-100, 1% normal goat serum, PBS, 1hr, RT), incubated with primary ALDH1L1 antiserum in blocking buffer (1:100, AbCam) at 4˚C for 15h. Tissues were probed with secondary (1:10,000, AF488, Invitrogen, 90min, RT), and counterstained with Hoechst (1:10,000, Invitrogen, 8min, RT). For analysis, a 5x2 tile scan z-stack was acquired on a confocal (Zeiss), and labeled cells were manually counted. The region analyzed was centered above the most dorsal point of the anterior commissure. Across treatments, 131.19±9.22 ALDH1L1+ cells and 720.78±17.93 nuclei were counted.

Immunoblotting. Protein concentrations were performed as described (Bowers and Kalivas, 2003), Protein expression was determined in crude homogenate following a Bradford assay using the following conditions: GFAP (10 μg, 1:20,000, DAKO), ALDH1L1 (30μg, 1:7,000, AbCam), glutamine synthetase (7μg, 1:7,000, Millipore) and near-infrared secondaries (1:7,000, AlexaFlor680, Invitrogen or 1:7,000 IRdye800, Rockland). Immunolabeling was measured (Odyssey, Licor). Integrated intensity was normalized to intralane β-tubulin III (1:10,000, AbCam for GFAP, 1:20,000, AbCam for ALDH1L1, and 1:10,000, Sigma for glutamine synthetase).

Blood EtOH concentrations. Immediately after the test session, lateral rostral tail vein blood was collected under light isoflurane anesthesia and blood ethanol concentrations (BEC) determined, as previously described (Bowers et al., 2008). Briefly, serum was shaken under heat (70˚C, 10min) and 1mL headspace subjected to isothermal chromatography (50˚C, 30m BAC-1 capillary, 40ml/min He carrier). EtOH was detected by flame ionization (200˚C, H2/Air) on an Agilent 6980 with a 0.01mM detection limit. The integrated area under the curve for each sample was normalized to a 1-propanol (0.1mg/mL) external standard and compared to known EtOH standards in order to calculate mg% BEC. Samples were run in triplicate.