Objective:To investigate the function, mechanism and therapeutic potential of macrophages in non-alcoholic steatohepatitis (NASH).Methods:Eight-week-old male foz/ foz (Alms mutant) mice were fed with a high fat diet (HFD) for 6, 8 and 10 weeks and 8-week-old male C57BL/6 mice were fed with a methionine and choline-deficient (MCD) diet for 7 d, 3 weeks and 4 weeks to establish NASH models. The mice of control group were fed with normal diet or MCD control diet. The expression of F4/80 mRNA level in the livers of mice of NASH model group and control group was detected by fluorescence quantitative polymerase chain reaction. Macrophages in the livers of mice of NASH group and control group were determined by immunofluorescence staining. After transgenic lysM-Cre/DTR mice were fed with MCD diet for 5 weeks, they were divided into transgenic experimental group (ablation of macrophages induced by diphtheria-toxin (DTox) injection) and transgenic control group (phosphate buffer saline injection). The levels of triglyceride and lipid peroxide in the livers of transgenic experimental group and transgenic control group were detected, and the inflammation of the livers of the mice was scored. The mechanism of macrophages regulating inflammation in NASH was investigated by cytokine profiliny analysis and Western blotting. The interaction between hepatocytes and macrophages were determined by co-culturing the conditional medium of hepatocytes AML-12 and macrophages RAW264.7. Macrophages of mice of control group and NASH model group were depleted by liposomal clodronate to confirm its value in NASH prevention. Independent sample t-test was used for statistical analysis. Results:F4/80 mRNA level in the livers of NASH model foz/ foz mice fed with HFD for 6 weeks, 8 weeks and 10 weeks was higher than that of control group (1.49±0.19, 1.70±0.15 and 1.93±0.04 vs.1.05±0.22), and the differences were statistically significant ( t=3.06, 4.92 and 7.92, all P<0.05). The expression of F4/80 mRNA level of the livers of NASH model mice fed with MCD for 7 d and 3 weeks was higher than that of control group (2.70±0.99 and 3.08±1.71 vs.1.00±0.83), and the differences were statistically significant ( t=3.43 and 3.54, both P<0.01). The results of immunofluorescence demonstrated that compared with that of control group, the number of F4/80 + inducible nitric oxide synthase (iNOS) + M1 macrophages were significantly increased, while F4/80 + CD206 + M2 macrophages were significantly decreased in the livers of NASH model mice fed with MCD for 4 weeks. After macrophages depletion, the inflammation score, the levels of triglyceride and lipid peroxide in the liver of transgenic experimental mice were all lower than those of transgenic control mice (0.69±0.32 vs. 1.95±0.74, (43.97±13.24) g/mg vs. (63.09±14.85) g/mg, (24.84±6.21) nmol/mg vs.(37.91±8.91) nmol/mg), and the differences were statistically significant ( t =3.14, 2.72 and 2.41, all P<0.05). The results of cytokine profiling analysis showed that macrophage depletion could lower the levels of interleukin (IL)-12 and macrophages inflammatory protein-1α (the difference between multiples: -3.98, -2.74, both P<0.05). CCAAT/enhancer binding protein β was defected in the nuclear of transgenetic experimental mice. In vitro study showed that RAW264.7 macrophages conditional medium could promote lipid accumulation in AML-12 hepatocytes, while conditional medium from MCD medium-treated AML-12 hepatocytes could promote RAW264.7 macrophages to M1 polarization. After treated with liposomal clodronate, the levels of triglyceride and lipid peroxidation in the liver of control mice were both lower than those of MCD-induced NASH model mice((45.33±14.59) g/mg vs. (63.10±16.02) g/mg, (2.11±0.48) nmol/mg vs. (2.73±0.17) nmol/mg), and the differences were statistically significant ( t=2.84 and 2.73, both P<0.05). The results of Western blotting indicated that after treating with liposomal clodronate, the relative content of phosphorylated protein kinase R-like endoplasmic reticulum kinase, inositol requiring enzyme-1α, protein disulfide isomerase, glucose regulatory protein 78, phosphorylated eukaryotic initiation factor 2α in the liver of NASH model mice were all lower than those of NASH model mice without liposomal clodronate treatment (1.84±0.36 vs. 3.05±0.83, 1.50±0.84 vs. 6.65±1.47, 0.87±0.12 vs. 2.28±0.52, 1.68±0.43 vs. 4.76±1.13, 1.42±0.19 vs. 2.75±0.79), and the differences were statistically significant( t=2.32, 5.28, 4.56, 4.41 and 2.85, all P<0.05). Conclusions:Macrophages are polarized into M1 phenotype in NASH. M1 macrophages contributed to NASH progression by interacting with hepatocyets to promote the secretion of inflammatory cytokines, up-regulation of lipogenic factors, oxidative stress and endoplasmic reticulum stress, resulting in the progression of NASH. Macrophages depletion by liposomal clodronate is a potential noval approach for NASH prevention.