Distinct Chrna5 mutations link excessive alcohol use to types I/II vulnerability profiles and IPN GABAergic neurons

Abstract

Genome wide association and animal studies have implicated genetic variations in CHRNΑ5, encoding the α5 subunit-containing nicotinic acetylcholine receptors (α5*nAChRs), as a risk factor for developing alcohol use disorders (AUDs). To understand how α5*nAChR mutations may influence alcohol (EtOH) drinking behavior, we used a two-bottle choice procedure with intermittent access to alcohol in male and female transgenic mice expressing either the highly frequent human single nucleotide polymorphism (α5SNP/rs16969968) or a deletion of the Chrna5 gene (α5KO). AUDs-related preconsommatory traits (anxiety, sensation-seeking and impulsivity) were assessed with a battery of relevant tasks (elevated-plus maze, novel place preference and step-down inhibitory avoidance). The implication of the α5-expressing IPN GABAergic neurons in AUDs and related behavioral traits was verified using neurospecific lentiviral (LV)-induced reexpression of the α5 subunit in α5KOxGAD-Cre mice. Both α5SNP and α5KO mice showed over-consumption of EtOH, but displayed opposite vulnerability profiles consistent with Cloninger's subtypes of human AUDs. α5SNP mice showed Type I-like characteristics, i.e., high anxiety, novelty avoidance, whereas α5KOs exhibited Type II-like features such as low anxiety and high impulsivity. LV re-expression of the α5 subunit in IPN GABAergic neurons restored the control of EtOH intake and improved the impulsive phenotype. We demonstrate that the SNP (rs16969968) or null mutation of Chrna5 result in increased volitional EtOH consumption but opposite effects on anxiety, novelty-seeking and impulsive-like behaviors that match Cloninger type I and II of AUDs, including sex-related variations. IPN GABAergic neurons expressing α5*nAChRs play a key role in limiting both EtOH drinking and motor impulsivity.

Fig. 1. Overview of experiments 1 and 2.

A Experiment 1: Effects of α5 mutations on AUDs-related behavioral traits and alcohol drinking. Twenty-eight WT (15 ♂ 13 ♀), 22 α5SNP (10 ♂ 12 ♀) and 22 α5KO (11 ♂ 11 ♀) mice were tested in three consecutive behavioral tasks: (1) the elevated plus maze (EPM) (2) novelty place preference (NPP) and (3) step-down inhibitory avoidance task (SDIA); followed by (4) voluntary alcohol drinking behavior using a two-bottle choice intermittent access paradigm (EtOH-IA). Mice performed one behavioral task per week. Males and females were tested on separated days for the EPM, NPP and SDIA tasks but concurrently for the EtOH-IA. Fifty-two additional males and females of each genotype, not previously tested in the behavioral assessment battery, were included in the EtOH-IA to complete measures of the oral consumption. B Experiment 2: Effects of α5 subunit re-expression in IPN GABAergic neurons on EtOH drinking and impulsivity. Thirty-seven α5KO x GAD-Cre (11 ♂ 26 ♀) mice received intra-IPN stereotaxic injections of lentivirus expressing either bicistronic α5-IRES2-eGFP or enhanced green fluorescent protein (eGFP) cDNAs to express, respectively, the α5 subunit (α5KO-α5^IPN-GABA mice) or GFP (α5KO-GFP^IPN-GABA mice) in IPN GABAergic neurons. Seven weeks after stereotaxic surgery, 19 α5KO-α5^IPN-GABA (6 ♂ 13 ♀) and 18 α5KO-GFP^IPN-GABA (5 ♂ 13 ♀) were tested in the SDIA and EtOH-IA tasks to sex-matched WT, α5SNP and α5KO.

Fig. 2. Anxiety-related behavior in α5SNP and α5KO mice.

Time spent on the EPM open arms (OA) per time segments (A) or on total test time (B) and in their extremities (Ext; C). There was a strong effect of Genotype on the three recorded parameters (OA time: F(2,66) = 10.25, p = 0.0001 and Time x Genotype interaction: F(18,594) = 10,952, p < 0.0001; %OA time: F(2,66) = 17.98, p < 0.0001; Ext time: F(2,66) = 6.39, p = 0.0029; A–C). α5KO spent more time, while conversely α5SNP mice tend to spend less time than WT in the OA (40 ± 4% vs 18 ± 2% vs 22 ± 2% respectively; α5KO vs WT: p < 0.0001, α5KO vs α5SNP: p < 0.0001; cumulated in sec, α5KO vs WT: p = 0.0020, α5KO vs α5SNP: p < 0.0001). Males and females α5KO showed an increase in OA time as compared to their sex-matched WTs (Genotype effect, in males: F(2,33) = 9.37, p = 0.0006; Time x Genotype interactions, in males: F(18,297) = 7.99, p < 0.0001; in females: F(18,297) = 3.50, p < 0.0001), with the females having a more moderate response than males (Time x Sex interaction: F(9,594) = 2.99, p = 0.0017; A). The total OA time % was much higher in α5KOs than WTs and α5SNPs in male (F(2,33) = 9.49, p = 0.0006, α5KO vs WT: p = 0.0020, α5KO vs α5SNP: p = 0.0002) and female mice as well (F(2,33) = 8.46, p = 0.0008, α5KO vs WT: p = 0.0009, α5KO vs α5SNP: p = 0.0008; B). A similar pattern was observed with “Ext” time parameter (C) in males but not in females showing instead a WT-like response (Ext % time: 14 ± 3% vs 7 ± 1% vs 5 ± 1%; α5KO vs WT: p = 0.0085, α5KO vs α5SNP: p = 0.0011; in males: F(2,33) = 4.41, p = 0.019, α5KO vs WT: p = 0.027, α5KO vs α5SNP: p = 0.0084). Females spent less time in OA-Ext regardless of the genotype (Effect of Sex: F(1,66) = 6.45, p = 0.013). Novelty seeking in α5SNP and α5KO mice. Relative percentage of entries in the center versus corners of the habituation compartment (FC: D); entry latency in novel compartment (NC: E); and novelty place preference score during test trial (Test). There was a Zone x Genotype interaction during habituation (F(2,66) = 5.28, p = 0.0074). WT and α5KOs (males and females) crossed the FC center more often than corners (respectively: 58 ± 2 vs 41 ± 2, paired t-test: p = 0.0005; and 62 ± 2 vs 38 ± 2, paired t-test: p < 0.0001; D), but this effect was absent in α5SNP mice (50 ± 3 vs 50 ± 1; p = 0.95 ns). Similarly to WT males, α5KO males crossed the center more often than corners of the FC (α5KO: t(14) = 4.22, p = 0.0040; WT: t(10) = 3.70, p = 0.0009), while α5SNP males tended to do the opposite. In females, α5SNPs and WTs entered the FC center as frequently as its corners (α5SNP: p = 0.60 ns; WT: p = 0.12 ns), while α5KOs crossed significantly more often the FC center than its corners (64 ± 4 vs 36 ± 4; t(10) = 3.50, p = 0.0056). During the test there was an effect of Time (F(9,594) = 20.14, p < 0.0001) and a Genotype x Time interaction (F(18,297) = 2.49, p = 0.0008, Test), due the avoidance of NC by α5SNP male mice.

Fig. 3. Impulsive-like behavior and alcohol drinking in α5SNP and α5KO mice.

A Step-down latency (SDL) during the acquisition, 24 h and 26h-delay retention trials. WT male and female mice learned the task efficiently (Trial effect in Males: F(2,66) = 70.22, p < 0.0001; Females: F(2,66) = 37.88, p < 0.0001). α5KOs exhibited shorter SDL than WT and α5SNP mice at either 24 h or 26h-delay trials (Genotype effect: F(2,66) = 7.35, p = 0.0013, α5KO vs WT: p = 0.0050, α5KO vs α5SNP: p = 0.0007; Trial x Genotype interaction: F(4,132) = 5.01, p = 0.0009). When analyzed by sex, this effect was significant only in females (Genotype effect, total: F(2,33) = 5.91, p = 0.0064, α5KO vs WT: p = 0.0060, α5KO vs α5SNP: p = 0.0043; 24 h: F(2,33) = 5.98, p = 0.0061, α5KO vs α5SNP: p = 0.0015; 26 h: F(2,33) = 5.00, p = 0.012, α5KO vs WT: p = 0.0052) and was more pronounced over trials (Trial x Genotype interaction: F(4,66) = 4.51, p = 0.028). All females were faster than males to step down (Sex effect: F(1,66) = 11.64, p = 0.0011 and Trial x Sex interaction: F(2,132) = 9.24, p = 0.0002). Female α5KO and α5SNP exhibited shorter SDL over trials than their male counterparts (α5KO: F(1,20) = 7.01, p = 0.015; α5SNP: F(1,20) = 5.78, p = 0.026). B Time to return on the platform (=escape latency: EL) for retention trials performed 24 h (top) and 26 h (bottom) after. This parameter decreased over trials in WT male and female mice (effect of Trial on EL, WT males: F(2,66) = 54.63, p < 0.0001; WT females: F(2,66) = 64.36, p < 0.0001). There was no effect of Genotype on EL at any delay, however female mice took generally more time to get back on the platform (effect of Sex on EL, 24 h: F(1,66) = 6.46, p = 0.0134; 26 h: F(1,66) = 16.27, p = 0.0001). C Fear-related behavior in the SDIA task as assessed by the freezing time expressed on the platform during trials at 24 h (top) and 26 h (bottom). The total freezing time was higher in α5SNPs than WTs (and α5KO) in either males (F(2,33) = 7.04, p = 0.0028, α5SNP vs WT: p = 0.012, α5SNP vs α5KO: p = 0.00080) or females (F(2,33) = 6.54, p = 0.0040, α5SNP vs WT: p = 0.0049, α5SNP vs α5KO: p = 0.0028). During the 24 h delay trial, α5SNPs displayed more freezing than WTs (F(2,66) = 3.61, p = 0.032, α5SNP vs WT: p = 0.012; top). During the 26 h delay trial, the SNP group displayed more freezing than WTs, whereas α5KOs expressed less freezing than their WT littermates (F(2,66) = 13.35, p < 0.0001, α5SNP vs WT: p = 0.0004; α5SNP vs α5KO: p < 0.0001; bottom). D–E Dose-response curves for EtOH intake (EI in g/kg/24 h) and EtOH preference ratio over water (EP) in α5SNP, α5SKO and WT mice. WT mice (males and females) increased their EI from 1.04 ± 0.1 g/Kg/24 h 12.3 ± 1.2 g/Kg/24 h over sessions 3% to 60%-EtOH; and EtOH preference (EP) from 0.20 ± 0.02 to 0.39 ± 0.05. Maximum EI was reached during 40%-EtOH in females and 60%-EtOH sessions in males (Sex x Dose interaction: F(9,1044) = 21.61, p < 0.0001). EP was higher in females during sessions with 3-20%-EtOH (Sex x Dose interaction: p = 0.051). Total fluid intake (EtOH + H₂O) over 24h-period mice (3 ml) did not differ between males and females, but females being lighter than males (25 ± 0.1 g vs 33 ± 0.2 g, F(1,116) = 146.67, p < 0.0001) total fluid intake was higher in females when reported to the weight (2.6 ± 0.05 ml/20 g vs 2.0 ± 0.3 ml/20 g, F(1,46) = 46.74, p < 0.001). EtOH consumption was higher in both α5KO and α5SNP mice with a right-shift in EtOH doses from 40% in females to 60% in males (Genotype effect on EI: F(2,116) = 14.32, p < 0.0001; and on EP: F(2,116) = 8.53, p = 0.0003). α5KO and α5SNP males differed the most from WT males at concentration as high as 60%, with respective EI of 23.3 ± 1 and 21.9 ± 1 versus 13.8 ± 1 g/kg/24 h in WTs (3 sessions at 60%: F(2,58) = 9.82 p = 0.0002, α5KO vs WT: p = 0.0003, α5SNP vs WT: p = 0.0006; 3D left) and respective 60%-EP ratio of 0.42 ± 0.02 and 0.39 ± 0.02 versus 0.25 ± 0.02 in WTs (3 sessions at 60%, F(2,58) = 9.33, p = 0.0003, α5KO vs WT: p = 0.00040, α5SNP vs WT: p = 0.00070; 3E left). Only α5SNP females differed from WTs at 60% (EP = 0.25 ± 0.02 vs 0.18 ± 0.02; Genotype effect: F(2,58) = 3.18, p = 0.048; α5SNP vs WT: p = 0.015, α5KO vs WT: p = 0.19 ns; 3E right). Both α5KO and α5SNP females differed the most from WT females at 40%, with respective EI of 20.9 ± 1 and 18.1 ± 1 versus 13.6 ± 1 g/kg/24 h in WTs (3 sessions at 40%: F(2,58) = 8.83, p = 0.0004, α5KO vs WT: p = 0.0002, α5SNP vs WT: p = 0.0043; 3D right) and respective 40%-EP ratio of 0.56 ± 0.04 and 0.49 ± 0.03 versus 0.34 ± 0.02 in WTs (3 sessions at 40%, F(2,58) = 6.23, p = 0.0035, α5KO vs WT: p = 0.0020, α5SNP vs WT: p = 0.0123; 3E right). No difference with WTs was found in term of total fluid intake (α5KO: 2.4 ± 0.05, α5SNP: 2.4 ± 0.04, WT: 2.3 ± 0.03 ml/20 g).

Fig. 4. Basal plasmatic corticosterone and blood ethanol concentrations in α5SNP and α5KO mice.

A WT mice exhibited similar pCORT in males and females (respectively: 21.4 ± 4 ng/ml and 25.9 ± 10 ng/ml), α5SNP showed higher pCORT levels (40.2 ± 5 ng/ml) than WT while conversely α5KOs exhibited lower pCORT levels (12.7 ± 1 ng/ml; Genotype effect: F(2,36) = 9.83, p = 0.0004, α5SNP vs WT: p = 0.0074, α5KO vs WT: p = 0.10 ns, α5SNP vs α5KO: p = 0.0002). There was no effect of sex, but when analyzed separately the genotype effect reached statistical significance only in males (F(2,21) = 9.47, p = 0.0012, α5SNP vs WT: p = 0.0078, α5KO vs WT: p = 0.13 ns, α5SNP vs α5KO: p = 0.0004; in females: p = 0.065 ns). Blood EtOH concentration (BEC) induced by the EtOH-IA protocol as a function of EI data with 40% (B) and 60% EtOH solutions (C) in male and female mice for each genotype. After a 30 min-session, no difference was found between α5 mutants and WT in terms of either 40%-EI/EP or 60%-EI/EP. BEC was positively correlated to their relative 40%-EI and 60%-EI in both male (D) (F(1,37) = 8.14, p < 0.0001) and female mice (E) (F(1,33) = 8,45, p = 0.0001). To make direct comparisons of BEC between mice, we divided the BEC (in mg/mL) of each mouse by their corresponding EI (g/Kg/30 min) to assess the BEC for 1 g/Kg/30 min EI (F). No Genotype or Sex difference was found with BEC for 1 g/Kg/30 min EI of: 0.29 ± 0.04 (males) and 0.34 ± 0.05 (females) in WT, 0.26 ± 0.04 (males) and 0.25 ± 0.03 (females) in α5KO and 0.31 ± 0.03 (males) and 0.22 ± 0.04 (females) in α5SNP (F).

Fig. 5. Re-expression of the α5 subunit in GABAergic neurons of the IPN improves impulsive-like behavior and reduces alcohol drinking.

A Histological control of LV-induced re-expression of the α5 subunit in the IPN. Analysis of all LV-injected α5KO-GFP^IPN-GABA and α5KO-α5^IPN-GABA mice included in this study (n = 37) revealed that LV infection was restricted to IPN neurons. Strong eGFP signals were visualized in the rostral subnucleus of the IPN (IPR) known for expressing the highest density of GABAergic neurons which are highly enriched with native α5*nAChRs. IPC: central subnucleus; IPL: lateral subnucleus; IPDL: dorsolateral subnucleus. B Step-down inhibitory avoidance task (SDIA). As previously observed (Fig. 3A), the step-down latency (SDL) increased over trials in α5KO-GFP^IPN-GABA, α5KO-GFP^IPN-GABA and WT mice, yielding a Trial effect (F(2,158) = 82.28, p < 0.0001) and a Trial x Genotype interaction (F(6,158) = 2.31, p = 0.036). In females, there was a strong Genotype effect on SDL at 26 h (F(3,46) = 3.16, p = 0.033) and α5KO and α5KO-GFP^IPN-GABA exhibited shorter SDL than their WT littermates (α5KO vs WT: F(1,22) = 9.42, p = 0.0056; α5KO-GFP^IPN-GABA vs WT: F(1,24) = 4.98, p = 0.035). In contrast, female α5KO-α5^IPN-GABA mice displayed a WT-like phenotype (in sec: 110.1 ± 30 and 122.4 ± 25 in α5KO-α5^IPN-GABA and WT respectively; versus 32.8 ± 11 and 53.2 ± 18 in α5KO and α5KO-GFP^IPN-GABA respectively). C Evolution of escape latencies (EL) during SDIA retention trials at 24 h (top) and 26 h (bottom) after acquisition. EL decreased over trials: F(2,158) = 75.84, p < 0.0001, with no difference between WT, α5KO, α5KO-GFP^IPN-GABA and α5KO-α5^IPN-GABA mice during trials at 24 h or 26 h. D Dose-response curves for EtOH drinking assessed with EtOH intake (EI) and (E) preference (EP). Overall sessions and males/females combined, α5KO and α5KO-GFP^IPN-GABA (13.5 ± 0.7 and 12.5 ± 0.7 g/kg/24 h) on one side and α5KO-α5^IPN-GABA and WT (9.2 ± 0.4 and 9.9 ± 0.5 g/kg/24 h) on the other side showed similar levels of EI (Genotype effect: F(3,79) = 9.36, p < 0.0001, WT vs α5KO: p < 0.0001, WT vs α5KO-GFP^IPN-GABA: p = 0.0017, α5KO-α5^IPN-GABA vs α5KO: p = 0.0010). In males (left), α5KO-α5^IPN-GABA mice showed lower levels of EI than α5KO-GFP^IPN-GABA and α5KO males overall sessions (F(3,33) = 14,50, p < 0.0001, α5KO-α5^IPN-GABA vs α5KO-GFP^IPN-GABA: p = 0.0022, α5KO-α5^IPN-GABA vs α5KO: p < 0.0001, α5KO vs WT: p < 0.0001). There was also a Genotype effect for 40%-EI (F(3,33) = 4,03, p = 0.015), 60%-EI (F(3,33) = 14.10, p < 0.0001, α5KO-α5^IPN-GABA vs α5KO: p < 0.0001, α5KO-GFP^IPN-GABA vs α5KO: p = 0.0053, α5KO vs WT: p < 0.0001), and 60%-EP as well (F(3,33) = 6,38, p = 0.0016, α5KO-α5^IPN-GABA vs α5KO: p = 0.0008, α5KO vs WT: p = 0.0009). In females (right), there was a Genotype effect for both 40%-EI (F(3,46) = 2.86, p = 0.047) and 40%-EP (F(3,46) = 2.94, p = 0.042, α5KO-α5^IPN-GABA vs α5KO: p = 0.0052). Females α5KO-α5^IPN-GABA consumed less than α5KOs during session 2 and 3 of 40%-EI (α5KO-α5^IPN-GABA vs α5KO: F(1,22) = 4.87, p = 0.038; WT vs α5KO: F(1,21) = 11.57, p = 0.0027).