Genetic and Pharmacologic Manipulation of TLR4 Has Minimal Impact on Ethanol Consumption in Rodents

Abstract

Toll-like receptor 4 (TLR4) is a critical component of innate immune signaling and has been implicated in alcohol responses in preclinical and clinical models. Members of the Integrative Neuroscience Initiative on Alcoholism (INIA-Neuroimmune) consortium tested the hypothesis that TLR4 mediates excessive ethanol drinking using the following models: (1) Tlr4 knock-out (KO) rats, (2) selective knockdown of Tlr4 mRNA in mouse nucleus accumbens (NAc), and (3) injection of the TLR4 antagonist (+)-naloxone in mice. Lipopolysaccharide (LPS) decreased food/water intake and body weight in ethanol-naive and ethanol-trained wild-type (WT), but not Tlr4 KO rats. There were no consistent genotypic differences in two-bottle choice chronic ethanol intake or operant self-administration in rats before or after dependence. In mice, (+)-naloxone did not decrease drinking-in-the-dark and only modestly inhibited dependence-driven consumption at the highest dose. Tlr4 knockdown in mouse NAc did not decrease drinking in the two-bottle choice continuous or intermittent access tests. However, the latency to ethanol-induced loss of righting reflex increased and the duration decreased in KO versus WT rats. In rat central amygdala neurons, deletion of Tlr4 altered GABA_A receptor function, but not GABA release. Although there were no genotype differences in acute ethanol effects before or after chronic intermittent ethanol exposure, genotype differences were observed after LPS exposure. Using different species and sexes, different methods to inhibit TLR4 signaling, and different ethanol consumption tests, our comprehensive studies indicate that TLR4 may play a role in ethanol-induced sedation and GABA_A receptor function, but does not regulate excessive drinking directly and would not be an effective therapeutic target.

Significance statement: Toll-like receptor 4 (TLR4) is a key mediator of innate immune signaling and has been implicated in alcohol responses in animal models and human alcoholics. Members of the Integrative Neuroscience Initiative on Alcoholism (INIA-Neuroimmune) consortium participated in the first comprehensive study across multiple laboratories to test the hypothesis that TLR4 regulates excessive alcohol consumption in different species and different models of chronic, dependence-driven, and binge-like drinking. Although TLR4 was not a critical determinant of excessive drinking, it was important in the acute sedative effects of alcohol. Current research efforts are directed at determining which neuroimmune pathways mediate excessive alcohol drinking and these findings will help to prioritize relevant pathways and potential therapeutic targets.

Keywords: (+)-naloxone; Toll-like receptor 4 knock-out; chronic intermittent ethanol vapor; drinking-in-the-dark; lipopolysaccharide; operant self-administration.

Figure 1.

Tlr4 KO rats are resistant to LPS-induced toxicity. LPS produced transient reduction of food (A) and water (B) intake in male WT but not Tlr4 KO rats. The data are presented as mean ± SEM food and water intake (in grams) in 24 h periods. Day 0 represents intake measured during the 24 h before LPS (1 mg/kg, i.p.) injection. C, LPS induced a robust weight loss in WT but not Tlr4 KO rats. The data are presented as percentage change relative to preinjection baseline weight (error bars are obscured by the symbols). Data were analyzed by two-way repeated-measures ANOVA followed by Bonferroni post hoc tests; *p < 0.05, **p < 0.01, ***p < 0.001 compared with baseline; n = 4 per genotype. Arrows indicate injection of LPS (1 mg/kg, i.p.).

Figure 2.

Tlr4 KO rats trained to self-administer ethanol (10%) are resistant to LPS-induced toxicity. The data are presented as mean ± SEM of food (A) and water (B) intake in 24 h in male WT (n = 5) and Tlr4 KO (n = 5) rats. Day 0 corresponds to intake for the 24 h before LPS (1 mg/kg, i.p.) injection. C, Time course of the change in body weight over 24 h in male rats. The data are presented as the percentage change relative to the preinjection baseline weight measured before the first LPS injection. D, Time course of the change in body weight of LPS-treated WT and KO rats after the first and second LPS injections. LPS-induced changes in body weight are expressed as the percentage change relative to the most recent preinjection weight. In WT rats, paired Student's t test (t₍₄₎ = 6.0, p < 0.01) showed a significant decrease in loss of body weight 1–4 d after the second LPS injection (−3.4 ± 0.5%) compared with the weight loss after the first LPS injection (−7.1 ± 0.7%). There was no significant difference in the LPS effects between the first and second LPS injections in the Tlr4 KO rats. Arrows indicate injection of LPS (1 mg/kg, i.p.).

Figure 3.

LPS induces transient reduction in ethanol (10%) self-administration in WT but not Tlr4 KO rats. Effect of LPS injections on ethanol (A) and water (B) self-administration in male WT and Tlr4 KO rats. The data are presented as mean ± SEM lever presses during 30 min sessions. Baseline (BL) was calculated as the average of the last 4 sessions before each LPS injection. Rats were given two LPS (1 mg/kg, i.p.) injections 14 d apart and the data from both injections were averaged. The average lever presses 1–4 d and 5–12 d after LPS are shown. Data were analyzed by mixed-factorial ANOVA followed by Newman–Keuls post hoc test; *p < 0.05 compared with KO; n = 5 per genotype.

Figure 4.

There was no effect of genotype on ethanol self-administration in WT or Tlr4 KO rats before or after dependence. A, Male WT and Tlr4 KO rats do not differ in ethanol (10%) self-administration before ethanol vapor exposure. The data are presented as mean ± SEM lever presses during 30 min sessions. Data were analyzed by two-way repeated-measures ANOVA with group as the between-subjects factor and session as the within-subjects factor. B, Male WT and Tlr4 KO rats do not differ in ethanol (10%) self-administration after CIE exposure. Ethanol-dependent rats underwent cycles of 14 h on (BEC: 146.8 ± 3.6 mg%) and 10 h off. The same rats were used to measure predependent versus postdependent ethanol self-administration. For baseline, the number of lever presses for ethanol (last six sessions before ethanol vapor exposure) was averaged and compared with the average number of lever presses over six self-administration sessions during CIE exposure. The data are presented as mean ± SEM lever presses during 30 min sessions. Data were analyzed using two-way ANOVA with group as the between-subjects factor and time (predependent vs dependent) as the within-subjects factor (n = 8 for WT and n = 7 for KO, *p < 0.05 compared with baseline).

Figure 5.

2BC ethanol intake in male and female WT and Tlr4 KO rats. A, Ethanol (10%) intake (g/kg/24 h), (B) water intake (ml/kg/24 h), (C) ethanol preference, and (D) body weight (g) in male and female WT and Tlr4 KO rats during 2BC drinking sessions measured over 5 weeks. The data are presented as mean ± SEM. Data were analyzed using 2 × 2 × 5 (genotype by sex × week) mixed ANOVAs (WT male, n = 23; WT female, n = 16; KO male, n = 15; KO female, n = 11). *Significant (p < 0.05) difference between male and female rats within genotype; #significant (p < 0.05) difference between week 1 and a subsequent week for the respective sex within genotype.

Figure 6.

Effect of (+)-naloxone on ethanol intake in the DID test in C57BL/6J mice. A, Male C57BL/6J mice were injected with (+)-naloxone (0, 30, or 60 mg/kg; n = 8 per group) 30 min before the start of the drinking session (20% ethanol) on day 4 of the DID protocol and intake was measured after 2 or 4 h. B, Male mice were injected with saline (vehicle) or 60 mg/kg (+)-naloxone 30 min before the start of the DID (20% ethanol) session for 4 d (n = 8 per group). The data are presented as mean ± SEM. Data were analyzed separately at the 2 and 4 h time points using one-way ANOVA with (+)-naloxone dose as the between-subjects factor.

Figure 7.

Effect of (+)-naloxone on ethanol intake in the 2BC and CIE-2BC tests in C57BL/6J mice. A, Effect of individual doses of (+)-naloxone on 2BC intake (g/kg/2 h). Increased ethanol (15%) consumption was observed in the CIE-2BC group compared with the control 2BC group (t₍₂₈₎ = 5.8, p < 0.0001). Male C57BL/6J mice (n = 15 per group) were administered (+)-naloxone (0, 3, 10, 30, 60 mg/kg) in a within-subject manner (every 3–4 d) 30 min before 2BC testing. The data are presented as mean ± SEM. Data were analyzed by two-way repeated-measures ANOVA (group as the between factor; dose as the within factor). B, Effect of 4 d of 60 mg/kg (+)-naloxone treatment on 2BC drinking (g/kg/2 h). The mice were then divided into two groups to receive (+)-naloxone or vehicle based on equal drinking averages. Male mice in a between-subject design (n = 4–6 per group: control-vehicle, control-60 mg/kg (+)-naloxone, CIE-vehicle, CIE-60 mg/kg (+)-naloxone) were injected 30 min before 2BC testing across 4 d. The data are presented as mean ± SEM. Data were analyzed by three-way repeated-measures ANOVA (group and dose as the between factors; day as the within factor). **p < 0.01 compared with baseline.

Figure 8.

Effect of Tlr4 knockdown in NAc on ethanol intake in the continuous and intermittent access 2BC tests in Tlr4^F/F male mice. A–C, Ethanol intake (g/kg/24 h) (A), preference for ethanol (B), and total fluid intake (g/kg/24 h) (C) in the 2BC continuous access test in untreated control Tlr4^F/F mice (n = 10) and Tlr4^F/F mice injected with LV-Cre-EGFP (n = 20) or LV-Cre-Empty (n = 10). D–F, Ethanol (15%) intake (g/kg/24 h) (D), preference for ethanol (E), and total fluid intake (g/kg/24 h) (F) in the 2BC intermittent access test in untreated control Tlr4^F/F mice (n = 10) and Tlr4^F/F mice injected with LV-Cre-EGFP (n = 20) or LV-Cre-Empty (n = 11). Each point is the average of 2 d of drinking. The data are presented as mean ± SEM. Data were analyzed by two-way repeated-measures ANOVA.

Figure 9.

Verification of injection site after Tlr4 mRNA knockdown in NAc. A, Composite microscope image of a coronal section of the NAc after lentivirus injection using fluorescent microscopy (left) to show EGFP marker signal (green) and bright-field (right) to demonstrate neuroanatomy. B, Coronal brain atlas diagram of the NAc injection site with blue circles showing the NAc and green ovals illustrating the typical lentivirus location and spread.

Figure 10.

Verification of Tlr4 mRNA knockdown after lentivirus injections in NAc. Tlr4 mRNA levels in the NAc target site were assessed by qPCR and normalized relative to Gadph mRNA levels after the 2BC continuous access (n = 12, LV-EGFP-Cre; n = 7, LV-EGFP-Empty) and 2BC intermittent access tests (n = 11, LV-EGFP-Cre; n = 7, LV-EGFP-Empty). Values (mean ± SEM) are shown relative to LV-EGFP-Empty treated mice; ***p < 0.001 determined by Student's t test.

Figure 11.

The sedative effects of ethanol are reduced in Tlr4 KO male rats. Left panel, Time to LORR in male WT and Tlr4 KO rats (n = 10 per group, df = 18, t = 2.3, p < 0.05). Right panel, Time of LORR in WT and Tlr4 KO rats (n = 10 per group, df = 18, t = 3.9, p < 0.01). The data are presented as mean ± SEM in minutes. *p < 0.05 and **p < 0.01 determined by unpaired Student's t test.

Figure 12.

Acute ethanol potentiates spontaneous and miniature GABA release in CeA neurons from ethanol-naive WT and Tlr4 KO rats. A, Traces of representative sIPSC recordings of CeA neurons from naive WT (left) and Tlr4 KO (right) male rats before and after acute bath application of ethanol (44 mm). B, Acute ethanol increased the mean sIPSC frequency in CeA neurons by 37.6 ± 11.4% in WT (from 1.4 ± 0.5 to 1.6 ± 0.6 Hz, n = 9) and by 45.7 ± 14.2% in KO (from 0.9 ± 0.2 to 1.2 ± 0.3 Hz, n = 10) rats. There was a significant main effect of ethanol (F_(1,17)) = 11.0, p < 0.01), but no main effect of genotype or interaction between ethanol and genotype. Acute ethanol did not have significant effects on sIPSC amplitudes (WT: 111.3 ± 9.4% of control; KO: 115.8 ± 5.1% of control) or kinetics (rise time: WT, 97.0 ± 3.3% of control; KO, 106.6 ± 3.7% of control; decay time: WT, 112.5 ± 7.5% of control; KO, 117.6 ± 9.7 of control). C, Representative mIPSCs recorded from CeA neurons of naive WT and Tlr4 KO rats showing an increase in mIPSC frequencies after acute ethanol application. D, Acute ethanol increased mIPSC frequencies in CeA WT and Tlr4 KO neurons. There was a significant main effect of ethanol (F_(1,12) = 11.8, p < 0.05) on mIPSC frequencies in WT (26.4 ± 7.1%; from 0.5 ± 0.1 to 0.6 ± 0.1 Hz, n = 5) and KO neurons (43.6 ± 13.5%; from 0.3 ± 0.1 to 0.4 ± 0.1 Hz, n = 9), but no significant main effects of genotype or the interaction between ethanol and genotype. Ethanol had no effects on mIPSC amplitudes (WT: 100.3 ± 2.0% of control; KO: 93.6 ± 3.6% of control) or kinetics (rise time: WT, 90.9 ± 2.8% of control; KO, 99.9 ± 4.9% of control; decay time: WT, 102.9 ± 2.3% of control; KO, 96.7 ± 11.6% of control) in neurons from WT and Tlr4 KO rats. Statistical significance was calculated by two-way repeated-measures ANOVA (#p < 0.05). The data are presented as mean ± SEM.

Figure 13.

Acute ethanol enhances GABA release in CeA neurons from WT and Tlr4 KO rats differentially after CIE exposure. A, Traces of sIPSC recordings in CeA neurons from WT and Tlr4 KO male rats exposed to CIE vapor. B, Bath application of 44 mm ethanol facilitated sIPSC frequencies in WT neurons significantly by 57.2 ± 4.3% (from 1.0 ± 0.4 to 1.5 ± 0.6 Hz, n = 8), but enhanced frequencies in only 6 of 10 KO neurons (46.5 ± 23.1% of control; from 1.2 ± 0.7 to 1.7 ± 0.8 Hz, n = 6) and decreased (< 85% of control, n = 3) or had no effect (85–115% of control, n = 1) on sIPSC frequencies in the remainder (85.5 ± 9.3% of control; from 0.7 ± 0.1 to 0.5 ± 0.04 Hz) of KO neurons. Two-way repeated-measures ANOVA of the peak ethanol effects from all of the tested cells showed a significant main effect of ethanol (F_(1,16) = 9.3, p < 0.05), but no effect of genotype or interaction between ethanol and genotype. The acute ethanol effects on sIPSC frequencies were transient in the neurons from KO compared with WT rats. C, Representative mIPSCs in CeA neurons from WT and Tlr4 KO rats chronically exposed to ethanol. D, Acute ethanol (44 mm) application increased mIPSC frequencies in WT and Tlr4 KO neurons by 55.1 ± 19.9% (from 0.3 ± 0.1 to 0.4 ± 0.1 Hz, n = 5) and 58.6 ± 22.6% (from 0.4 ± 0.1 to 0.5 ± 0.8 Hz, n = 6), respectively. Two-way repeated-measures ANOVA showed a significant main effect of ethanol (F_(1,9) = 19.8, p < 0.05), but not genotype or the interaction between ethanol and genotype. Acute ethanol had no significant effects on the mean mIPSC amplitudes (WT: 104.8 ± 6.9% of control; KO: 109.9 ± 11.3% of control) or kinetics (rise time: WT, 105.4 ± 5.8% of control; KO, 112.0 ± 6.2% of control; decay time: WT, 115.1 ± 13.3% of control; KO, 101.4 ± 4.3% of control) in WT and Tlr4 KO rats. The data are presented as mean ± SEM (#p < 0.05).

Figure 14.

LPS injection alters GABA responses to acute ethanol in CeA neurons from WT but not Tlr4 KO rats. A, Representative recordings of sIPSCs from LPS-treated WT and Tlr4 KO male rats. B, Seven to 10 d after a single LPS injection (1 mg/kg, i.p.), acute 44 mm ethanol application reduced sIPSC frequencies by 41.4 ± 4.0% (from 1.3 ± 0.3 to 0.7 ± 0.2 Hz, n = 8) in CeA neurons from LPS-treated WT rats. Ethanol's effects on sIPSC frequencies in LPS-treated KO rats (an overall increase of 29.9 ± 21.6%, n = 7) were more varied, with ethanol-induced potentiation (> 115% of control, n = 3) or no effect (85–115% of control, n = 4) occurring in a cell-specific manner. There was a significant main effect of ethanol treatment (F_(1,13) = 5.9, p < 0.05), but not genotype, and no interaction between genotype and ethanol. There were no significant effects of acute ethanol on the amplitudes (WT: 93.6 ± 12.9%; KO: 112.3 ± 11.1% of control) or kinetics (rise time: WT, 109.9 ± 4.4%; KO, 97.8 ± 2.2%; decay time: WT, 90.2 ± 14.7%; KO, 115.2 ± 10.1% of control) in LPS-treated WT and Tlr4 KO rats. The data are presented as mean ± SEM. Statistical significance was calculated by two-way repeated-measures ANOVA (#p < 0.05).