FGF21 counteracts alcohol intoxication by activating the noradrenergic nervous system

Abstract

Animals that consume fermenting fruit and nectar are at risk of exposure to ethanol and the detrimental effects of inebriation. In this report, we show that the hormone FGF21, which is strongly induced by ethanol in murine and human liver, stimulates arousal from intoxication without changing ethanol catabolism. Mice lacking FGF21 take longer than wild-type littermates to recover their righting reflex and balance following ethanol exposure. Conversely, pharmacologic FGF21 administration reduces the time needed for mice to recover from ethanol-induced unconsciousness and ataxia. FGF21 did not counteract sedation caused by ketamine, diazepam, or pentobarbital, indicating specificity for ethanol. FGF21 mediates its anti-intoxicant effects by directly activating noradrenergic neurons in the locus coeruleus region, which regulates arousal and alertness. These results suggest that this FGF21 liver-brain pathway evolved to protect against ethanol-induced intoxication and that it might be targeted pharmaceutically for treating acute alcohol poisoning.

Keywords: FGF21; alcohol; amethystic; anti-intoxicant; hormone; inebriation; liver; locus coeruleus; noradrenergic.

Figure 1.. Fgf21^−/− mice have a prolonged righting reflex recovery time after a binge ethanol dose.

(A) Plasma FGF21 concentrations were measured in wild-type (WT) mice either 1 hour before administering ethanol (5 g/kg, oral gavage) or at the indicated times after ethanol (n = 6 mice/group). (B-E) WT and Fgf21^−/− mice were administered ethanol (5 g/kg, oral gavage) and the following measurements made: time to loss of righting reflex (LORR) (B); LORR duration (C); plasma ethanol concentrations (D); brain ethanol concentrations (E) (n = 17 mice/group for (B) and (C) and 6 mice/group for (D) and (E)). All data represent the mean ± SEM. *, P < 0.05, **, P < 0.01, ***, P < 0.001 compared to −1 hour (A; one-way ANOVA) or WT (C; Student’s t-test).

Figure 2.. Hepatocyte-specific Fgf21^−/− and neuron-specific Klb^−/− mice have prolonged righting reflex recovery times after a binge ethanol dose.

Control (Fgf21^fl/fl) and hepatocyte-specific Fgf21^−/− (Fgf21^Alb) mice (A-C) or control (Klb^fl/fl) and neuron-specific Klb^−/− (Klb^Camk2a) mice (D-F) were administered ethanol (5 g/kg, oral gavage) and the following measurements made: time to loss of righting reflex (LORR) (A, D); LORR duration (B, E); and plasma ethanol concentrations (C, F) (n = 12–16 mice/group for A, B, D, E and 11–12 mice/group for C and 10–11 mice/group for F). All data represent the mean ± SEM. P values in (B, D, E) by Student’s t-test.

Figure 3.. Pharmacologic FGF21 accelerates recovery from alcohol-induced loss of righting reflex and ataxia and is selective for ethanol.

(A, B) Wild-type (WT) male (A) or female (B) mice were administered ethanol (5 g/kg, oral gavage) followed 1 hour later by i.p. injection of vehicle or FGF21 at the indicated doses. Loss of righting reflex (LORR) duration was measured after FGF21 or vehicle administration (n = 5–6 mice/group). Data represent the mean ± SEM. *, P < 0.05, **, P < 0.01, ***, P < 0.001 compared to vehicle by one-way ANOVA (A) or Student’s t-test (B). (C, D) WT and Fgf21^−/− mice (n = 13–15 mice/group) (C) or control (Klb^fl/fl) and neuron-specific Klb^−/− (Klb^Camk2a) mice (n = 12–16 mice/group) (D) were administered ethanol (5 g/kg, oral gavage) followed 1 hour later by i.p. injection of FGF21 (1 mg/kg) or vehicle. LORR duration was measured after FGF21 or vehicle administration. Data represent the mean ± SEM. Different lowercase letters indicate statistical significance (P < 0.05 by two-way ANOVA). (E) WT mice were administered ethanol (2 g/kg, i.p.) followed 30 minutes later by injection of FGF21 (1 mg/kg, i.p.; indicated by arrow) or vehicle. The time mice could remain on a spinning rotarod was measured, with 60 seconds the maximum (n = 8 mice/group). Data represent the mean ± SEM. P value by likelihood ratio test as described in Methods. (F)WT and Fgf21^−/− mice were administered ethanol (2 g/kg, i.p.) and the time they could remain on a spinning rotarod measured as in (E) (n = 8 mice/group). Data represent the mean ± SEM. P value by likelihood ratio test as described in Methods. (G-J) Wild-type mice were administered ethanol (4.3 g/kg) (G) (n = 9–12 mice/group), ketamine (200 mg/kg) (H) (n = 6 mice/group), diazepam (30 mg/kg) (I) (n = 5–6 mice/group) or pentobarbital (55 mg/kg) (J) (n = 7–8 mice/group) by i.p. injection. After 30 minutes, mice were i.p. injected with either FGF21 (1 mg/kg) or vehicle. Loss of righting reflex (LORR) duration was measured after FGF21 or vehicle administration. Data represent the mean ± SEM. *, P < 0.05 compared to vehicle-treated mice by Student’s t-test.

Figure 4.. FGF21 is a physiologic regulator of noradrenergic neurons.

(A) Representative confocal images of immunostaining for c-Fos (white) and norepinephrine transporter (NET; green) in locus coeruleus sections prepared from wild-type (WT) and Fgf21^−/− mice 2.5 hours after oral gavage with either water or ethanol (5 g/kg). Scale bar represents 50 μM. (B) Quantification of c-Fos/NET co-expression (upper panel) and total NET-positive cell number (lower panel) (n = 3 sections/mouse, 4 mice/group). Data represent the mean ± SEM. Different lowercase letters indicate statistical significance (P < 0.05) by two-way ANOVA.

Figure 5.. Pharmacologic FGF21 activates noradrenergic neurons in the locus coeruleus.

(A) Confocal images of immunostaining performed in locus coeruleus (LC) sections prepared from transgenic mice expressing tdTomato fused to the C-terminus of KLB. Antibodies against tdTomato (magenta) and norepinephrine transporter (NET; green) were used. Scale bars represent 50 μM. (B and C) Immunostaining for c-Fos (white) and NET (green) in LC sections prepared from wild-type mice treated for 2 hours with vehicle or FGF21 (1 mg/kg, i.p.). Representative confocal images are shown in (B). Scale bars represent 50 μM. Quantification of c-Fos/NET co-expression (top panel) and total NET-positive cell number (lower panel) is shown in (C) (n = 4 sections/mouse, 3 mice/group). Data represent the mean ± SEM. ****, P < 0.0001 by Student’s t-test. (D) Immunostaining for c-Fos and NET in LC sections from groups of control (Klb^fl/fl) and neuron-specific Klb^−/− (Klb^Camk2a) mice treated for 2 hours with vehicle or FGF21 as in (B). Quantification of c-Fos/NET co-expression (left panel) and total NET-positive cell number (right panel) is shown (n = 3 sections/mouse, 3 mice/group). Data represent the mean ± SEM. Different lowercase letters indicate statistical significance (P < 0.05) by two-way ANOVA. See also Figure S1.

Figure 6.. FGF21 exerts its anti-intoxicant activity through noradrenergic neurons.

(A) Immunostaining for DBH in locus coeruleus and adrenal from control (Dbh^fl/fl) and neuron-specific Dbh^−/− (Dbh^Camk2a) mice. DAPI staining (blue) and merge of the two is shown below. Scale bars represents 50 μM. (B) Dbh^fl/fl and Dbh^Camk2a mice were administered ethanol (5 g/kg, oral gavage) followed 1 hour later by injection of vehicle or FGF21 (1 mg/kg, i.p.). Loss of righting reflex (LORR) duration was measured after FGF21 or vehicle administration (n = 7–9 mice/group). (C) Wild-type (WT) mice were injected with DSP-4 (50 mg/kg, i.p.) or vehicle. Two days later, the mice were administered ethanol (5 g/kg, oral gavage) followed 1 hour later by injection of FGF21 (1 mg/kg, i.p.) or vehicle (n = 16 mice/group). LORR duration was measured after FGF21 or vehicle administration. (D) WT mice were administered ethanol (5 g/kg, oral gavage) followed 1 hour later by injection of vehicle, FGF21 (1 mg/kg, i.p.), prazosin (0.8 mg/kg) or FGF21+prazosin (n = 8–10 mice/group). LORR duration was measured after vehicle or FGF21 administration. (E) The experiment was performed as in (B) except mice were injected with vehicle, FGF21 (1 mg/kg, i.p.), propranolol (10 mg/kg, i.p.) or FGF21+propranolol (n = 15–17 mice/group). LORR duration was measured after vehicle or FGF21 administration. All data represent the mean ± SEM. Different lowercase letters indicate statistical significance (P < 0.05) by two-way ANOVA. See also Figure S2.

Figure 7.. FGF21 acts directly on noradrenergic neurons in the locus coeruleus region.

(A) Control (Klb^fl/fl) and noradrenergic neuron-specific Klb^−/− (Klb^Dbh) mice were administered ethanol (5 g/kg, oral gavage) followed 1 hour later by injection of vehicle or FGF21 (1 mg/kg, i.p.). LORR duration was measured after FGF21 or vehicle administration (n = 12–17 mice/group). (B) Klb^fl/fl mice bilaterally injected in the locus coeruleus (LC) region with adeno-associated viruses (AAV) expressing either control (GFP) or GFP-Cre were administered ethanol (5 g/kg, oral gavage) followed 1 hour later by injection of vehicle or FGF21 (1 mg/kg, i.p.). LORR duration was measured after FGF21 or vehicle administration (n = 8–11 mice/group). (C, D) Immunostaining for c-Fos and norepinephrine transporter (NET) was performed in LC sections from groups of control (Klb^fl/fl) and Klb^Dbh mice (C) or Klb^fl/fl mice bilaterally injected in the LC region with AAV-GFP or AAV-GFP-Cre (D) treated for 2 hours with vehicle or FGF21 (1 mg/kg, i.p.). Quantification of c-Fos/NET co-expression (left panel) and total NET-positive cell number (right panel) is shown. For (C), n = 3 sections/mouse, 6–7 mice/group. For (D), n = 7 sections/mouse, 3 mice/group. All data represent the mean ± SEM. Different lowercase letters indicate statistical significance (p < 0.05) by two-way ANOVA. See also Figure S3.