RPMI Medium 1640 was from GIBCO/Invitrogen (Grand Island, NY, USA). FBS was from Hyclone/Thermo Scientific (VIC, Australia). PCR reagents were from Applied Biosystems (Foster City, CA, USA). ACEA (special CB1 agonists), AM281 (CB1 antagonist), AM630 (CB2 antagonist), SB203580 (p38 inhibitor), PD98059 [extracellular signal-regulated kinase (ERK) inhibitor], and compound C [adenosine monophosphate-activated protein kinase (AMPK) inhibitor] were from TOCRIS/R&D (Minneapolis, MN, USA). Y27632 (the inhibitor of Rho-associated protein kinase ROCK), pertussis toxin (PTX), and other common reagents were from sigma (St. Louis, MO, USA). Clodronate liposome (CL) was from YeaSen (Shanghai, China).

ICR mice aged 3 weeks were sacrificed by cervical dislocation at the time of BM harvest. BM cells were extracted from the tibias and femurs by flushing with culture medium using a 25 G needle. The cells were then passed through a 70-µm nylon mesh (BD Bioscience) and washed three times with PBS. Extracted BM cells were implanted with 2.4 × 10⁷/100 mm culture dish (BD Falcon) and cultured for 7 days in the presence of L929-conditioned medium (replacing culture medium at the third, fifth day). All animal work was performed under the ethical guidelines of the Ethics Committee of Capital Medical University. After 7 days, IL-4 (5 ng/mL) was used to homogenization of BMMs for 5 h.

ICR male mice aged 6 weeks received lethal irradiation (8 Grays) and immediately received transplantation by a tail vein injection of 1.5 × 10⁷ whole BM cells obtained from 3-week-old EGFP transgenic mice. Four weeks later, mice of BM-rebuild were subjected to CCl₄-induced liver injury. After another 4 weeks, mice were sacrificed and liver tissues were harvested.

A CCl₄ (1 μL/g BW)/olive oil (OO) mixture (1:9 v/v) was injected into abdominal cavity of mice twice per week. Mice were sacrificed at 1, 3 days and 1, 2, 4 weeks. The liver tissues were harvested. The intraperitoneal injection of AM281 (2.5 mg/kg BW) was performed at 4 h before CCl₄ administration (n = 6). The CLs (from Vrije University Amsterdam) were injected 24 h before CCl₄ administration via caudal vein. All animal work was conformed to the Ethics Committee of Capital Medical University and in accordance with the approved guidelines (specific institutional approval number of animal experiment: AEEI-2014-131).

Isolation of mouse liver macrophages was as described in this section. Subsequently, antibodies: PE-CD86 (BD Biosciences, Franklin Lakes, NJ, USA), and its isotype-matched negative control antibodies were added to the cell suspension. After 15 min of incubation in the dark, the cells were washed with PBS and subjected to FACS. FACS was performed on a FACSAria and analyzed with FACSDiva4.1 (BD Biosciences).

Mouse chemically modified and stable siRNAs of CB1 and Invivofectamine 3.0 reagent were from Invitrogen (Thermo Scientific, PA, USA). The method is briefly described as follows. The solution of siRNA duplex (1.2 mg/mL) 50 µL was mixed with 50 µL Buffer. The solution was mixed with 100 µL Invivofectamine 3.0 reagent, then incubated at 50°C for 30 min. The complex was diluted with 1 mL PBS. The solution (10 µL/g BW) was injected into mouse tail vein at 4 h before the first time of CCl₄ administration (n = 6).

Western blot analysis was performed with 50 µg of protein extract. Antibodies were as follows: rabbit anti-iNOS polyclonal antibody (1:200, St. Louis, MO, USA), rabbit anti-NF-κB p65 polyclonal antibody (1:1,000), rabbit anti-ERK1/2 polyclonal antibody (1:1,000), and rabbit anti-phorspho-ERK1/2 polyclonal antibody (1:1,000) (Cell Signaling, Beverly, MA, USA); rabbit anti-β-tubulin and anti-GAPDH monoclonal antibodies (1:1,000, Abcam, UK). ODYSSEY goat anti-rabbit IRDye^® 800 CW antibody (1:10,000, LI-COR, NE, USA) was used as secondary antibody. The bands were displayed using ODYSSEY and quantified by Odyssey v3.0 software. β-tubulin or GAPDH was as reference.

Bone marrow-derived monocyte/macrophage that had been cultivated for 7 days were fixed by 4% paraformaldehyde for 30 min and penetrated by 0.5% Triton X-100 (Amresco, OH, USA) for 15 min. After blocked with 3% BSA (Roche, Switzerland), they were incubated with anti-NF-κB p65 rabbit polyclonal antibodies (1:40; Cell Signaling, Beverly, MA, USA) or anti-phorspho-ERK1/2 (1:40; Cell Signaling, Beverly, MA, USA). FITC-conjugate affinipure goat-anti-rabbit IgG (1:100, Jackson Immunoresearch, PA, USA) was as secondary antibody. Nuclei were stained with DAPI. For high content analysis, the plates were imaged on Thermo Scientific CellInsight personal cell imaging platform (Cellomics, Inc., Thermo Fisher Scientific Inc., Waltham, MA, USA). 48 fields were automatically acquired by the software, corresponding to at least 4,000 cells. Cytoplasmic or nuclear fluorescence intensity of each well was analyzed by Cellomics Cell Health Profiling BioApplication software.

The liver specimen was fixed in 4% paraformaldehyde and frozen sections of 6 mm were used for immunofluorescence and paraffin sections for H&E. Frozen sections were incubated with F4/80 antibody as the first antibody and Cy3-conjugated goat anti-rat IgG (1:100, Jackson ImmunoResearch Laboratories) as the secondary antibody. Finally, the sections were stained with DAPI and observed under a confocal microscope (LSM510, Carl Zeiss MicroImaging, Jena, Germany).

Total RNA was extracted from liver frozen specimens or cultured BMMs with or without treatments using an RNeasy kit (Qiagen, Hilden, Germany). RT-qPCR was performed in an ABI Prism 7300 sequence detecting system (Applied Biosystems, Foster City, CA, USA). Primers sequences are listed in Table 1.

Applied BD CBA Mouse Flex Kit (Catalog No.558266), briefly as follows: liver tissues (~40 mg) or BMM (a density of 7,000 cells/cm²) supernatant homogenized into 40 µL and mixed in 40 µL buffer with IL-6 (Catalog No. 5580301), TNF (Catalog No. 558299) and MIP-1β (Catalog No. 558449). The sample was acquired on a FACSAria and analyzed with FACSDiva4.1 (BD Biosciences).

Primary liver macrophages were isolated from treated adult mice. First, anesthetized and heparinized mice were subjected to a midline laparotomy and cannulation of the portal vein followed by liver perfusion with an EGTA-chelating perfusion buffer (EGTA: 190 mg, glucose: 900 mg, HEPES: 10 mL of 1 M stock solution, KCl: 400 mg, Na₂HPO₄–12H₂O: 305 mg, NaCl: 8 g, NaH₂PO₄–2H₂O: 88 mg, and NaHCO₃: 350 mg, made up to 1 L with dH₂O). After perfusion with 0.4% collagenase buffer (CaCl₂–2H₂O: 560 mg, HEPES: 10 mL of 1 M stock solution, KCl: 400 mg, Na₂HPO₄–12H₂O: 305 mg, NaCl: 8 g, NaH₂PO₄–2H₂O: 88 mg, NaHCO₃: 350 mg, and collagenase IV: 400 mg, made up to 1 L with dH₂O), livers were teared and cells dispersed in saline; non-parenchymal cells were separated using low-speed centrifugation and 40% percoll density gradient centrifugation. F4/80 was used to sorted macrophages using FACS. Isolated macrophages were used to perform qPCR and FACS.

The ON-TARGETplus mouse CB1 siRNA smart pool (L-042461-00), CB2 siRNA smart pool (L-062503-00), and non-targeting control pool (D-001810-10-05) were from Dharmacon (Thermo Scientific, PA, USA). 40–50% confluent BMMs were prepared in 60-mm dishes. Transient transfection of siRNA (40 nmol/L) was performed by using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) as recommended by the manufacturer. Control cells were treated with 40 nmol/L RNA interference negative control duplexes (scramble siRNA). After 48 h, cells were used to perform the following assay.

Bone marrow-derived monocyte/macrophage with or without AM281, PTX, or PD98059 pretreatment were exposed to ACEA for different times, then BMM was lysed in lysis buffer, active RhoA was extracted using pull-down and detection kit (Catalog No. 16116, Thermo Scientific Pierce Biotechnology, IL, USA), finally, active and total Rho was showed by Western blot.

Results are expressed as mean ± SEM. Statistical significance was determined by Student’s t-test or ANOVA. Comparisons between two independent groups were performed using a two-sample t-test. Comparisons between multiple groups were performed by one-way analysis of variance with post hoc Tukey’s multiple comparison tests. Correlation coefficients were calculated by Pearson’s test. P < 0.05 was considered significant.

During liver fibrogenesis, the amount of macrophages was significantly increased in livers (28), yet their characters of polarization or the dynamic changes of M1 and M2 types are unclear. Thus we first detected the dynamic changes of M1 and M2 gene signatures in mouse livers at different times after CCl₄ treatment. CD86 (M1 type macrophage marker) mRNA was significantly upregulated from 2 days of CCl₄ administration and reached the peak at 2 weeks. After that, CD86 mRNA sustained around 10-fold of the initial level. Furthermore, we also investigated other M1 gene signatures, including MIP-1β, TNF, IL-6, and iNOS in injured livers. The mRNA level of MIP-1β rose rapidly and reached the peak at 4 weeks. iNOS and TNF mRNA also increased and got to the topmost level after 2 weeks, whereas IL-6 mRNA was slightly increased with a maximal increase after 4 weeks (Figure 1A). Considering M2 gene signatures, Arg1 mRNA had no changes at the early injured period and generated a slight decrease during chronic stage. CD206 mRNA had little changes at early injured stage and generated a violent decline at 4 weeks. CD163 expressions had no changes (Figure 1B). Next, we detected the profile of M1 type macrophage amount. Liver macrophages were isolated from livers with different treatments by F4/80⁺ gating (Figure 1C left). FACS analysis showed that the proportion of total M1 type macrophages in CCl₄-induced damaged livers was upregulated markedly from 2 days of CCl₄ administration compared with that in OO-treated liver and reached the peak at 4 weeks (Figures 1C,D). These results indicated M1-, but not M2, type macrophages might play important roles in liver injury.

Our preceding study demonstrated that CB1 was involved in CCl₄-induced liver injury and CB1 expression was increased (22). Here, we undertook correlation analysis between mRNA expression level of CB1 and CD86 or Arg1. In mouse CCl₄-injured livers, the mRNA levels of CD86 positively correlated with CB1 (r = 0.993, P < 0.001), whereas the mRNA level of Arg1 showed no correlations with it (Figure 2A). We also noted that CB1 mRNA exerted positive correlations with MIP-1β, TNF, IL-6, or iNOS (Table 2). However, all of M2 gene signature (including Arg1, CD163, and CD206) expressions had no correlations with CB1 (Table 2). Collectively, these results suggested that CB1 might play a crucial role in regulating M1 polarization of macrophages during liver fibrogenesis.

To figure out the role of CB1 on macrophage polarization in insulted livers, AM281 (CB1 antagonist) administration was performed in the model. Four weeks of AM281 injection markedly reduced, the mRNA levels of M1 markers in injured livers compared with that in the injured livers without AM281 pretreatment (Figure 2B). Furthermore, the blockade of CB1 restrained the protein levels of MIP-1β, TNF, IL-6 (by CBA) (Figure 2C), and iNOS (by western blot) (Figure 2D) in injured livers. In addition, FACS analysis showed that the amount of total M1-type macrophages (F4/80⁺/CD86⁺) in CCl₄-induced damaged liver was decreased after AM281 administration (Figure 2E).

It has been reported that there are two kinds of macrophages in damaged livers, the tissue-resident macrophages, Kupffer cells, and the infiltrated macrophages, BMMs. To further clarify the role of CB1 on Kupffer cell and BMM polarization, BM in the irradiated mice were reconstituted by transplantation of the genetic EGFP-labeled BM cells. Then, the BM rebuilt mice were induced liver injury by CCl₄ administration. After 4 week administration, we isolated liver macrophages (F4/80⁺ cells). EGFP⁺/CD86⁺ (M1 type of BMM) proportion in macrophages of damaged liver was increased after CCl₄ administration for 4 weeks and the amount of M1 type of Kupffer cell (EGFP⁻/CD86⁺) was also upregulated. But the increase of Kupffer cells was much less than that of BMMs. Specifically, almost 80% of M1-polarized macrophages are BMMs, which implied M1 type of BMMs might act as an important part of liver fibrogenesis. After AM281 pretreatment, the increased amount of EGFP⁺/CD86⁺ was significantly limited, while AM281 pretreatment had no effect on the amount of M1-type Kupffer cells (Figure 2F). Therefore, we thought that CB1 was involved in mediating the proportion of M1 type BMMs.

In order to confirm our conclusion, we adopted an established model in which chemically modified and stable siRNAs of CB1 was injected rapidly into the mouse tail vein, using a “hydrodynamic transfection method” (29). CB1-siRNA successfully downregulated each mRNA level by 61.3% in liver (Figure 3A), which confirmed the effectiveness of the siRNA. CB1 siRNA pretreatment decreased the mRNA and protein levels of M1 markers in CCl₄-damaged livers (Figures 3A–C). Moreover, FACS analysis for CD86 showed that the results of CB1 siRNA administration were consisted with AM281 pretreatment results (Figures 3D,E).

To confirm the role of CB1 on BMMs and Kupffer cells in vivo, we deleted Kupffer cells using CLs (30) and performed CCl₄ administration with or without AM281 pretreatment. Four weeks later, we examined M1 marker expressions in macrophages isolated from injured livers to exclude the role of Kupffer cells. The results of immunofluorescence assay and RT-qPCR showed the elimination efficiency (Figures 4A,B). We found that AM281 pretreatment on reducing mRNA levels of M1 marker were not affected after Kupffer cell deletion (Figure 4C). Furthermore, we used FACS to detect the amount of M1-type BMM in the isolated cell. Our result showed that the decreased amount of CD86⁺ macrophages reduced by AM281 pretreatment was unchanged after CLs injection (Figure 5A). We also measured the protein levels of M1 markers. Consistent with the mRNA results, AM281 pretreatment on reducing protein levels of M1 marker were not affected after Kupffer cell deletion (Figures 5B,C). These results indicated that CB1 was especially involved in BMM polarization but have no effect on that of Kupffer cells. To figure out the reason of CB1 different ability on the two kinds of macrophages, we measured CB1 expression on Kupffer cells and BMM by FACS. We noted that the expression of CB1 on BMM was higher than that of Kupffer cell (Figure 5D). These results confirmed that CB1 was involved in M1 polarization of BMMs in CCl₄-induced liver fibrogenesis.

We then investigated the effects of CB1 on BMM polarization in vitro. ACEA (1 µmol/L) promoted the mRNA levels of CD86 and TNF at 2, 6, and 12 h, implying that the activation of CB1 effectively promoted M1 polarization. Furthermore, CB1 activation also increased iNOS, MIP-1β, and IL-6 expressions at 2 and 6 h (Figure 6A). When BMMs were pretreated with AM281, ACEA-induced elevations of M1 gene signature mRNA levels were suppressed (Figure 6B). To confirm the critical role of CB1, we employed CB1-specifc siRNA. We first measured that CB1 siRNA effectively silenced the corresponding target gene at protein levels (Figure S1A in Supplementary Material). Silencing CB1 expression with siRNA abrogated the M1 polarization of BMMs induced by ACEA (Figure 6C). Actually, ACEA also showed low CB2 affinity, so we applied CB2 antagonist (AM630) and CB2 siRNA to BMMs. But, we found that AM630 (10 µmol/L) and CB2 knockdown hardly inhibited ACEA-mediated M1 polarization of BMMs (Figures S1B,C in Supplementary Material). Furthermore, CB1 activation also upregulated the protein levels of MIP-1β, IL-6, TNF, and iNOS, especially TNF, which was upregulated over 10-fold. Additionally, the increase protein levels of M1 markers induced by ACEA were abrogated by AM281 or siCB1 (Figures 6D–F). TNF expression was even reverted to the initial phase after AM281 or siCB1 pretreatment. Altogether, these results suggested that CB1 activation exerted a powerful action on BMM polarization toward M1 type.

As a G protein-coupled receptor, the biological function of CB1 depends on multiple signaling pathways, such as mitogen-activated protein kinases (MAPKs) and AMPK signaling pathways (31). Our preceding study noted that CB1 activated RhoA to promote migration and phagocytosis in macrophages (22, 32). To identify the responsible signaling pathways for M1 polarization of BMMs regulated by CB1, we applied pharmacological inhibitors of signal transduction, PTX (G(α)_i/o inhibitor), Y27632 (ROCK inhibitor), SB203580 (p38 inhibitor), PD98059 (ERK inhibitor), and compound C (AMPK inhibitor). The CB1-mediated increase of CD86 mRNA expression was apparently impaired by PTX, Y27632, and PD98059 in BMMs, while SB203580 and compound C had no such effect (Figure 7A). Apart from this, PTX, Y27632, and PD98059 also markedly attenuated the mRNA levels of M1 other gene signatures increased by ACEA (Figures 7B–D). PTX, Y27632, and PD98059 pretreatment also decreased the protein levels of MIP-1β, IL-6, TNF (by CBA) (Figures 8A–C), and iNOS (by western blot) (Figure 8D) induced by ACEA. These findings implied that G(α)_i/o, RhoA, and ERK were involved in macrophage polarization mediated by CB1.

We investigated the effect of CB1 on signal pathway activation of its downstream. pull-down analysis showed that ACEA promoted active GTP-bound Rho protein levels of BMMs in 1, 2, and 4 h and this added GTP-bound RhoA conformation was impaired by AM281 pretreatment. We noted that PTX weakened GTP-bound RhoA protein with ACEA. However, ERK antagonist made no difference on the active GTP-bound Rho protein expression induced by ACEA, which implied that ERK was not involved the effect of CB1 on RhoA activation (Figure 9A).

Meanwhile, we measured the effect of CB1 on ERK activation. ACEA induced a significant increase in the protein level of phosphor-ERK1/2 and kept longer excitation of ERK1/2. Pretreatment with AM281 or PTX inhibited the increase of phosphor-ERK1/2, while Y27632 did not affect the activation of ERK in BMMs, which implied that ERK and RhoA had no influence on their activation of each other (Figure 9B). Taken together, these findings indicated that ACEA-activated G(α)_i/o coupled CB1, and then enlarged GTP-bound Rho and phosphor-ERK1/2 independently, finally promoting BMM polarization toward M1 phenotype.

Owing to the center role of NF-κB p65 nuclear translocation in M1 polarization, we wondered whether it was involved in CB1 regulation on M1 polarization. We detected NF-κB p65 localization in the presence of CB1 agonist by western blot and high content analysis. As shown in Figure 9C and Figure S1D in Supplementary Material, nuclear NF-κB p65 protein was increased in ACEA-treated cells, while cytoplasmic NF-kB p65 was decreased slightly. If the ratio of nuclear to cytoplasmic NF-κB p65 (Nuc/Cyto) in untreated cells was set as 1.00, the Nuc/Cyto was 3.32 in ACEA-treated cells. Furthermore, AM281 markedly attenuated CB1-mediated NF-κB p65 nuclear translocation, as the results of western blot analysis revealed that ACEA-induced NF-κB p65 nuclear translocation was reduced after AM281 pretreatment (Nuc/Cyto = 1.26). Moreover, PTX also weakened the NF-κB p65 nuclear translocation with ACEA (Nuc/Cyto = 0.92), which implied ACEA activated NF-κB p65 depending on G(α)_i/o. Next, we examined whether phosphor-ERK1/2 and GTP-bound Rho protein was involved in ACEA-induced NF-κB activation. The ACEA-induced nuclear translocation of NF-kB p65 was apparently impaired by Y27632 in BMMs (Nuc/Cyto = 1.78), while PD98059 has no such effect (Nuc/Cyto = 3.11). High content analysis showed similar results (Figures 10A,B). These data demonstrated that CB1-mediated NF-kB p65 nuclear translocation only depending on G(α)_i/o/RhoA signaling pathway.

To clarify the role of CB1 on ERK activation, we also detected phosphor-ERK1/2 nuclear translocation by high content analysis. Nuclear phosphor-ERK1/2 protein was increased in ACEA-treated macrophages, which indicated that phosphor-ERK1/2 could regulate M1 gene signature expressions directly (Figures 10C,D). In conclusion, CB1 mediated M1 polarization of BMMs, depending on two independent signaling pathways: G(α)_i/o/RhoA/NF-κB p65 and G(α)_i/o/ERK1/2 pathways (Figure 10E).