β-Hydroxybutyrate protects from alcohol-induced liver injury via a Hcar2-cAMP dependent pathway

Abstract

Background & aims: Sterile inflammation resulting in alcoholic hepatitis (AH) occurs unpredictably after many years of excess alcohol intake. The factors responsible for the development of AH are not known but mitochondrial damage with loss of mitochondrial function are common features. Hcar2 is a G-protein coupled receptor which is activated by β-hydroxybutyrate (BHB). We aimed to determine the relevance of the BHB-Hcar2 pathway in alcoholic liver disease.

Methods: We tested if loss of BHB production can result in increased liver inflammation. We further tested if BHB supplementation is protective in AH through interaction with Hcar2, and analyzed the immune and cellular basis for protection.

Results: Humans with AH have reduced hepatic BHB, and inhibition of BHB production in mice aggravated ethanol-induced AH, with higher plasma alanine aminotransferase levels, increased steatosis and greater neutrophil influx. Conversely supplementation of BHB had the opposite effects with reduced alanine aminotransferase levels, reduced steatosis and neutrophil influx. This therapeutic effect of BHB is dependent on the receptor Hcar2. BHB treatment increased liver Il10 transcripts, and promoted the M2 phenotype of intrahepatic macrophages. BHB also increased the transcriptional level of M2 related genes in vitro bone marrow derived macrophages. This skewing towards M2 related genes is dependent on lower mitochondrial membrane potential (Δψ) induced by BHB.

Conclusions: Collectively, our data shows that BHB production during excess alcohol consumption has an anti-inflammatory and hepatoprotective role through an Hcar2 dependent pathway. This introduces the concept of metabolite-based therapy for AH.

Lay summary: Alcoholic hepatitis is a life-threatening condition with no approved therapy that occurs unexpectedly in people who consume excess alcohol. The liver makes many metabolites, and we demonstrate that loss of one such metabolite β-hydroxybutyrate occurs in patients with alcoholic hepatitis. This loss can increase alcohol-induced liver injury, and β-hydroxybutyrate can protect from alcohol-induced liver injury via a receptor on liver macrophages. This opens the possibility of metabolite-based therapy for alcoholic hepatitis.

Keywords: Alcohol-induced liver injury; Hcar2; Therapy; β-Hydroxybutyrate.

Fig. 1. Low liver BHB in humans with alcoholic hepatitis and a combination of etomoxir and ethanol induces liver injury in mice

(A) BHB levels from normal and alcoholic hepatitis human livers. Data are shown as means +/− SEM. Control, n = 15; AH, n=10. (B–H) Mice were injected i.p. with etomoxir (20mg/kg) or saline. 30 minutes later they were gavaged with PBS or ethanol (6g/kg). Plasma and livers were harvested 16 hours after ethanol gavage. (B) Representative liver histology. (C) Statistical analysis of steatosis scores. U.D. Undetectable. (D) Statistical analysis of liver TG levels. (E) Plasma ALT. (F) Statistical results of neutrophil (CD45⁺CD11b⁺Ly6G⁺) numbers in the liver by flow cytometry. (G) Representative images of Ly6B⁺ neutrophils in the liver. (H) Statistical analysis of Ly6B⁺ neutrophils in the liver by immunohistochemistry. C–H, n=6–8 per group. (I) Mice were injected i.p. with etomoxir (20mg/kg) or saline. 30 minutes later they were gavaged with PBS or ethanol (6g/kg). Livers were harvested 6 hours after ethanol gavage and the liver BHB levels were detected (n=7~8 per group). Data were analyzed with nonparametric Wilcoxon rank-sum tests.* P<0.05; ** P<0.01; ***P<0.001.

Fig. 2. BHB protected mice from etomoxir and ethanol induced liver injury

(A–G) Mice were given an i.p. injection of BHB (3mmol/kg body weight) 6 hours after ethanol gavage and livers were collected 10 hours later. (A) Plasma ALT (n=8 per group for the EtOH gavage group with and without BHB; n=11~12 for the EtOH plus etomoxir group). (B) Representative liver histology. (C) Statistical analysis of steatosis scores (n=11~12 each group). (D) Statistical analysis of liver TG levels. (n=11~12 each group) (E) Statistical results of neutrophil (CD45⁺CD11b⁺Ly6G⁺) numbers in the liver (n=8~9 per group). (F) Representative images of Ly6B+ neutrophils in the liver. (G) Statistical analysis of Ly6B+ neutrophils in the liver by immunohistochemistry (n=10~11 per group). Data were analyzed with nonparametric Wilcoxon rank-sum tests.* P<0.05; ** P<0.01; ***P<0.001.

Fig. 3. Expression of Hcar2 is required for the protective effect of BHB

Male Hcar2 deficient mice were treated with etomoxir and ethanol. Control saline (CT) or BHB were injected 6 hours after ethanol gavage. Plasma and livers were collected 10 hours later. (A) Representative liver histology (n=10~11 per group). (B) Statistical analysis of steatosis scores (n=10~11 per group). (C) Statistical analysis of liver TG levels (n=8 per group). (D) Plasma ALT (n=10~11 per group). (E) Statistical results of neutrophil (CD45⁺CD11b⁺Ly6G⁺) numbers in the liver (n=8 per group). Data were analyzed with nonparametric Wilcoxon rank-sum tests.

Fig. 4. BHB increased il10 mRNA levels in the liver

Mice were injected i.p. with Etomoxir (20mg/kg) or saline. 30 minutes later they were gavaged with ethanol (6g/kg). Mice were given an i.p. injection of BHB (3mmol/kg body weight) 6 hours after ethanol gavage and livers were collected 10 hours later. mRNA levels of cytokines in the livers were detected. n=8~11 per group. Data were analyzed with nonparametric Wilcoxon rank-sum tests. * P<0.05. U.D. Undetectable.

Fig. 5. Hcar2 is expressed on liver macrophages and BHB promoted an M2 macrophage phenotype in the liver

(A) Immunofluorescence of Hcar2 expression in the liver. Blue, DAPI; red(mRFP), Hcar2; Green, F4/80 or CD32. (B&C) Wild-type mice were injected i.p. with Etomoxir (20mg/kg) or saline. 30 minutes later they were gavaged with ethanol (6g/kg). Mice were given an i.p. injection of BHB (3mmol/kg body weight) 6 hours after ethanol gavage and livers were collected 10 hours later. (B) Expression of M1 markers on liver macrophages (CD45⁺CD11b⁺F4/80⁺) by flowcytometry. (C) Expression of M2 markers on liver macrophages (CD45⁺CD11b⁺F4/80⁺) by flow cytometry. (Quantified expression values of M1 and M2 markers on macrophage in the liver (n=7 per group). ΔMFI=(Mean Fluoresce Intensity)_{positive staining} – (Mean Fluoresce Intensity)_{isotype staining.} Data were analyzed with nonparametric Wilcoxon ranksum tests.* P<0.05; ** P<0.01.

Fig. 6. BHB promoted M2 gene expression through decreasing mitochondrial membrane potential

Bone marrow derived macrophages (BMDMs) from wild-type mice were cultured under M2 condition (IL-4 and IL-13) for 48 hours (A) and 24 hours (B) with and without BHB (15mM). BMDMs from IL-10 deficient mice were cultured under M2 condition (IL-4 and IL-13) for 24 hours (C) with and without BHB (15mM). (A and B) mRNA levels of M2 associated gene were detected in WT BMDMs. (C) mRNA levels of M2 associated gene were detected in IL-10 deficient BMDMs. (D) cAMP response element-luciferase activity in HEK cells treated under M2 condition for 24 hours with addition of control saline (CT), BHB or dbCAMP separately. (E) mRNA levels of M2 associated genes were detected WT BMDM cultured under M2 condition (IL-4 and IL-13) and BHB for 24 hours with and without Rp-cAMPS (100µM). (F) Mitochondrial membrane potentials were measured in WT BMDMs under M2 condition (IL-4 and IL-13) for 24 hours with and without BHB (15mM). (G) mRNA levels of M2 associated gene were detected in WT BMDMs treated with BHB (15mM) and oligomycin A (5µM) to maintain the mitochondrial membrane potential. Data are mean and SEM from three or four independent experiments. Data were analyzed with Student’s t-tests. * P<0.05; ** P<0.01; ***P<0.001.

Figure 7. BHB protected liver injury induced by chronic alcohol feeding and binge

Mice were fed with control (Control) liquid diet or 5% ethanol (EtOH) liquid diet for 10 days. On the 11^th day, mice on ethanol liquid diet were injected with etomoxir (i.p., 20mg/kg). Ethanol (5g/kg) was gavaged to mice on alcoholic liquid diet 30 minutes after etomoxir injection. BHB (i.p., 3mmol/kg) or Control saline (CT) was given 6 hours after ethanol gavage. Plasma and livers were harvested 16 hours after ethanol gavage. (A) Representative liver histology. (B) Statistical analysis of steatosis scores. U.D. Undetectable. (C) Statistical analysis of liver TG levels. (D) Plasma ALT. (A–D, n=8~10 per group). Data were analyzed with nonparametric Wilcoxon rank-sum tests. * P<0.05.