Cannabinoid 2 (CB2) receptors expressed on immune cells are considered to be antifibrogenic. Hepatic stellate cells (HSCs) directly interact with phagocytosis lymphocytes, but the nature of this interaction is obscure. We aimed to study the effects of CB2 receptors on hepatic fibrosis via their role in mediating immunity. Hepatic fibrosis was induced by carbon-tetrachloride (CCl4) administration in C57BL/6 wild-type (WT) and CB2 knockout (CB2−/−) mice. Irradiated animals were reconstituted with WT or CB2−/− lymphocytes. Lymphocytes from naïve/fibrotic WT animals and healthy/cirrhotic hepatitis C virus were preincubated in vitro with or without CB2 antagonist, evaluated for proliferation and apoptosis, and then cocultured with primary mouse HSCs or a human HSC line (LX2), respectively. Lymphocyte phagocytosis was then evaluated. Following CCl4-administration, CB2−/− mice developed significant hepatic fibrosis but less necroinflammation. WT mice harbored decreased liver CD4+ and NK+ cells but increased CD8+ subsets. Naïve CB2−/− mice had significantly decreased T cell subsets. Adoptive transfer of CB2−/− lymphocytes led to decreased fibrosis in the irradiated WT recipient compared with animals receiving WT lymphocytes. Moreover, necroinflammation also tended to decrease. In vitro, a CB2-antagonist directly increased human HSC activation and increased apoptosis and decreased proliferation of mice/human T cells (healthy/fibrotic) and their phagocytosis. We concluded that CB2−/− lymphocytes exert an antifibrotic activity, whereas lack of CB2 receptor in HSCs promotes fibrosis. These findings broaden our understanding of cannabinoid signaling in hepatic fibrosis beyond their activity solely in HSCs.
- hepatic stellate cells
- liver injury
the endocannabinoid system is comprised of at least two types of receptors (cannabinoid 1 and 2, CB1 and CB2). Except the nervous system, the CB1 receptors are present in adipocytes, gut, endothelial, and liver cells (3, 4, 10, 20, 22). CB1 signaling regulates intake of high-energy-containing food and alcohol, energy homeostasis, and hepatic lipogenesis (3, 29). CB2 receptors are largely expressed in several lines of peripheral blood immune cells, tonsils, spleen, and testes (2). Moreover, their presence at low levels has been confirmed in various other tissues and cells, such as hepatic myofibroblasts (9).
Hepatic fibrosis is a wound-healing response to chronic liver injury of various etiologies. Hepatic stellate cells (HSCs) are activated and transformed to myofibroblast-like cells in the course of chronic liver injury (7). The CB1 and CB2 receptors are not expressed in normal hepatocytes. However, increased expression of CB1 and CB2 receptors was demonstrated in hepatic myofibroblasts and vascular endothelial cells in chronic liver diseases (11, 18). Although the CB1 receptor is believed to have profibrogenic effects (9), studies on the CB2 receptor mediate antifibrogenic activity in liver because CB2 knockout (CB2−/−) mice developed augmented cirrhosis when exposed to CCl4 compared with wild-type (WT) animals (9). Thus the balance between the two receptors is very relevant to the development of hepatic fibrosis.
Both in human and animal models, profibrotic CD8 cells (24) and antifibrotic NK cells (13, 15) interact directly with HSCs via cellular adhesion (16) and phagocytosis (17). Because CB2 receptors are primarily present on immune system cells and at low levels in the HSCs, the purpose of this work was to explore the immune effects of CB2 receptors on hepatic fibrosis.
In this study, our results indicated that absence or antagonism of CB2 receptors increases fibrosis through direct HSC activation. At the same time, CB2 antagonism or loss reduces fibrosis indirectly by attenuating inflammatory responses through T cell apoptosis and decreasing their phagocytosis by HSCs. Therefore, our results suggest a dual pro- and antifibrogenic function of CB2 receptors.
MATERIALS AND METHODS
WT and CB2−/− male mice on the C57BL/6 background, 12 wk of age, received care according to ethical regulations of the Hebrew University and also adhered to NIH guidelines. All animal protocols were approved by the institutional animal care ethical committee and housed in a barrier facility. WT rodents were purchased commercially from Harlan Laboratories (Jerusalem, Israel), whereas the CB2−/− mice (B6.129P2-Cnr2/J-ST.005786) were purchased from the Jackson Laboratory (Bar Harbor, ME).
Carbon tetrachloride (CCl4, C-5331; Sigma, St. Louis, MO) fibrosis model was induced by intraperitoneal injections of 0.5 μl pure CCl4/g body wt (1:9 dilution in mineral oil) biweekly, for 4 or 6 wk. Resolution WT and CB2−/− groups were included by CCl4 withdrawal along last 2 wk. Euthanasia was performed 2 days after the final CCl4 injections; mice were weighed and anesthetized intramuscularly with 0.1 ml of ketamine:xylazine:acepromazine (4:1:1) per 30 g of body weight.
Histological assessments of liver injury.
The posterior one-third of the liver was fixed in 10% formalin for 24 h and then paraffin embedded in an automated tissue processor. Seven-millimeter liver sections were cut from each animal. Sections (15 mm) were then stained in 0.1% Sirius red F3B in saturated picric acid (both from Sigma). Hematoxylin and eosin (H&E) staining was performed for each animal. Knodell score was assessed blindly by an expert hepatic pathologist based on H&E and Sirius red staining, using the modified Histological Activity Index criteria, incorporating semiquantitative assessment of periportal/periseptal interface hepatitis (0–4), confluent necrosis (0–6), focal lytic necrosis/apoptosis and focal inflammation (0–4), portal inflammation (0–4), and architectural changes/fibrosis and cirrhosis (0–6). F4/80 stain was applied in liver sections to allocate macrophage infiltrations.
Relative fibrosis area (expressed as a percent of total liver area) was assessed by analyzing 36 Sirius red-stained liver sections per specimen. Each field was acquired at ×10 magnification and then analyzed using a computerized Bioquant morphometry system. To evaluate relative fibrosis area, measured collagen area was divided by net field area and then multiplied by 100. The net field area was calculated by subtraction of lumen area from total field area.
Serum alanine aminotransferase.
Blood samples were collected from the inferior vena cava, and alanine aminotransferase (ALT) was measured using an automated enzymatic assay with the Vistros Chemistry Systems 950.
Mice lymphocyte isolations.
The removed spleen was grinded through mesh; pellet was then treated for 3 min with 1 ml lysis buffer to remove red blood cells. Then splenocytes, mostly lymphocytes, were washed and counted before analysis. Intrahepatic lymphocytes were isolated by perfusion of the liver with digestion buffer. After perfusion, the liver was homogenized and incubated at 37°C for 30 min. The digested liver cell suspension was centrifuged to remove hepatocytes and cell clumps. The supernatant was then centrifuged to obtain a pellet of cells depleted of hepatocytes to a final volume of 1 ml. Lymphocytes were then isolated from this cell suspension using 24% metrizamide gradient separation (17).
Adoptive transfer model.
Four groups of WT animals underwent sublethal irradiation (6 animals in each group) and served as recipients for lymphocyte (from spleen) reconstitution. Sublethally irradiation was achieved with a single total-body dose of 700 cGy as described previously (14). Donor lymphocytes were harvested from WT and CB2−/− naïve donors and were reconstituted into four groups of sublethally irradiated WT recipients. Recipients were therefore reconstituted with WT (groups A and C) or CB2−/− (groups B and D) lymphocytes. Million cells were repeatedly reconstituted by intraperitoneal injections once per week for 4 wk as previously described (13, 25). Recipients were then randomized either with (groups A and B) or without (groups C and D) fibrosis induction with the same CCl4 regimen for 4 wk. Groups C and D were used as naïve nonfibrotic controls that received vehicle. Following the previously described isolation (13, 15, 16, 24), lymphocytes were immediately transferred intraperitoneally to recipients or to the in vitro studies.
Tissue RNA extraction.
Total cellular RNA was extracted from target frozen tissues using Trizol reagent as previously described (24). Synthesized β-actin and α-smooth muscle actin (α-SMA) were detected by real-time PCR: β-actin (as a housekeeping-gene) Forward: 5′-GAT-GAG-ATT-GGC-ATG-GCT-TT-3′, β-actin Reverse: 5′-AGA-GAA-GTG-GGG-TGG-CTT-TT-3′; α-SMA Forward: 5′-TCC-TCC-CTG-GAG-AAG-AGC-TAC-3′, α-SMA Reverse: 5′-TAT-AGG-TGG-TTT-CGT-GGA-TGC-3′.
Primary HSCs isolation and culture for in vitro studies.
HSCs were isolated from naïve WT mice using sequential pronase/collagenase digestion followed by Nycodenz density gradient centrifugation as described previously in rats with minor modifications (26).
Splenocytes from naïve and fibrotic animals (WT and CB2−/−) were incubated with or without CB2 receptor antagonist (1 μM, SR144528, no. 9000491-5; Cayman Chemical, Ann Arbor, MI). The diluents for the antagonist were ethanol:chremophor:saline, 1:1:18. Stock solutions were 10-2 M (in ethanol) and stored at −20°C. The concentration of solvent in assay never exceeded 0.1% (vol/vol).
One million splenocytes from each of the four groups were cocultured with 106 primary isolated mice HSCs from naïve WT donor. After 48 h of coculture, harvested cells were washed and an analyzed for apoptosis and phagocytosis.
Human peripheral-blood-lymphocyte isolation.
Heparinized blood samples of healthy volunteers and patients with chronic hepatitis C virus (HCV) were obtained. No patient had evidence of HBV/HCV dual infection or HIV coinfection. All specimens were collected with informed consent from donors. Only patients with advanced fibrosis (F3 and F4) were included. Mononuclear cells were isolated by centrifugation over ficoll-hypaque (Pharmacia, Uppsala, Sweden). After three washes in saline, cells were resuspended in medium of Roswell Park Memorial Institute 1640 with 10% FBS as described previously (16, 17).
Human lymphocyte-HSC interactions.
To investigate the role of CB2 receptors in the lymphocyte/HSC interactions in vitro, human LX2 (8, 32) (HSCs) was used. Quiescent activation was achieved when LX2 cells were cultured to confluence in six-well flasks (Nunc Brand Products, Roskilde, Denmark) with 1% FCS. To assess the direct cellular CB2 receptor role, CB2 receptor antagonist (SR144528) (23) at concentrations of 0, 1, 2 and 4 μM were incubated for 2 h with 106 healthy or HCV lymphocytes. Cells were then washed and cocultured for 48 h with LX2 cells. Following cocultures, cells were harvested, washed and detected for LX2 and lymphocytes apoptosis and LX2 α-SMA expressions by fluorescence-activated cell sorting (FACS).
Following cocultures, adhered LX2 cells were trypsinized, and the isolated mice lymphocytes or human lymphocytes were adjusted to 106/ml in staining buffer (in saline containing 1% bovine albumin; Biological Industries, Kibbutz Bet-Haemek, Israel).
To determine HSC activation, LX2 cells were fixed with 4% paraformaldehyde for 10 min and permeabilized with 0.1% saponine in PBS for 20 min and then stained with anti-human α-SMA-PE monoclonal-antibody (R&D Systems, Minneapolis, MN) for 30 min at room temperature. For human lymphocytes, anti-CD45 (Per-CP) were used (IQ Products, Groningen, Germany; diluted 1:40). For murine lymphocytes, anti-CD4 (FITC), anti-NK1.1 (PE), anti-CD45 (cy5), anti-CD8 (APC), and anti-CD3 (pacific blue) (IQ Products) were used. NK cells were determined as the NK1.1+CD3 population. Another set of stains was used to determine macrophage infiltrations. Macrophages were stained with the panmacrophage marker F4/80. For apoptosis measurements of mice splenocytes, human lymphocytes, and LX2 cells, propidium iodide staining of fragmented DNA and phosphatidylserine staining by annexin V-conjugated to FITC (R&D Systems) were used according to the manufacturer's instruction. Apoptosis was defined as annexin-V(+) but propidium-iodide(−). Phagocytosis was defined as CD45+ and α-SMA+ of the simultaneous gating of lymphocytes and LX2, which showed one homogenous population following coculture. All stained cells were analyzed with a flow cytometer (FACS-calibur Immunofluorometry systems; Becton-Dickinson, Mountain View, CA) as we previously described (13, 15).
T cell proliferation.
Following cocultures, lymphocytes were cultured in RPMI 1640 medium for 72 h, followed by an 18-h pulse with 1 μCi 3H-thymidine (ICN, Costa Mesa, CA). Incorporated radioactivity was measured on a flatbed μβ-counter (Wallac, Gaithersburg, MD). Readings are expressed as counts per minute, which correlates linear with proliferation activity.
Results are presented as means ± SD. Standard error was used for the Bioquant analysis. Student's t-test and ANOVA were used for statistically significant correlations.
Fibrotic CB2−/− mice develop severe fibrosis but with reduced inflammation and decreased intrahepatic lymphocytes.
Hepatic fibrosis was induced in WT and CB2−/− mice by biweekly intraperitoneal CCl4 injections for 6 wk and was compared with naive vehicle-treated mice. Fibrosis marker of the α-SMA expressions, ALT levels, lymphocyte inflammatory profile by FACS, and histological assessments of their infiltrations were performed. Figure 1A shows expressions of α-SMA-mRNA in livers that lack fibrosis in both naïve WT and CB2−/− mice. CCl4 induction provoked increased fibrosis in WT and was more prominent in CB2−/− animals. α-SMA expression was significantly increased 9.2 ± 3.5-fold in WT (P = 0.008) and 17 ± 3.9-fold in CB2−/− mice (P = 0.001) compared with naïve states. Fibrosis was significantly higher in CB2−/− mice compared with WT (P = 0.03). There were no differences in the cytochrome P-450 activity of liver extracts from two fibrotic strains (data not shown). Serum ALT levels (Fig. 1B) were significantly (P = 0.05) increased from 45.7 ± 17.5 in naïve mice to 168 ± 26 IU/l in fibrotic WT mice. On the other hand, serum ALT levels in the CB2−/− group were similarly low in naïve (51.7 ± 26) and fibrotic (60.8 ± 32.5 IU/l) groups, respectively. They were both lower than ALT serum levels seen in fibrotic WT mice (P = 0.03).
Polymorphonuclear cells including intrahepatic lymphocytes were isolated and identified for the anti-CD45 (panleukocyte marker) (Fig. 1C) using FACS. Naïve WT animals showed the well-recognized CD4, CD8, and NK fibrotic patterns (22). However, naïve CB2−/− animals had significantly lower levels of lymphocyte subsets that were not altered following fibrosis induction. CD4 T cells significantly dropped from 26.8 ± 11% in naïve WT mice to 14 ± 4.4% following fibrosis induction (P = 0.03). Naïve CB2−/− liver CD4 cells were 12.7 ± 7.8% of total but were significantly decreased to 7 ± 3% following fibrosis induction (P = 0.05). Naïve and fibrotic CB2−/− liver CD4 levels were significantly lower than WT mice (P values were 0.03 and 0.002, respectively).
WT liver CB8 subsets showed to be elevated (P = 0.02) in fibrosis to 23.9 ± 9.1% compared with naïve animals that had 11.4 ± 9.3%; naïve and fibrotic CB2−/− CD8 subsets were unchanged (8.1 ± 7.7 and 8.2 ± 3.6, respectively). On the other hand, intrahepatic NK cells were decreased from 11.5 ± 8.2% in naïve WT to 9.2 ± 6.5%, 5.3 ± 5%, and 4.8 ± 2% in fibrotic WT, naïve CB2−/−, and fibrotic CB2−/− groups, respectively. Although naïve CB2−/− NK cells tend to be lower than WT (P = 0.06), fibrotic CB2−/− NK cells were significantly lower (P = 0.03) compared with naïve WT. We also assessed liver CD4- and CD8-expressing NKT cells. CD8+/NK+ cells were 6.4 ± 6.8%, 5.3 ± 6.2%, 0.1 ± 0.2%, and 0.03 ± 0.07%, respectively. Naïve and fibrotic CB2−/− liver CD8+/NK+ cells were significantly lower than WT control readings (P values were 0.01 and 0.02, respectively). Liver CD4+/NK+ cells were 3.6 ± 2.9%, 1 ± 0.9%, 1.4 ± 0.7%, and 1.4 ± 0.3%, respectively. Compared with naives, all other three groups were similar but showed significant decrease (P values were 0.02, 0.03, and 0.05, respectively). Panmacrophage marker (F4/80) from the total CD45+ cells showed an increase in their percentage in the fibrotic CB2−/− (38 ± 2.6%) compared with the fibrotic naïve (25.3 ± 4.7%) counterpart (P = 0.01), emphasizing that other fractions of CD45+ infiltrate these livers.
To determine CB2−/− alterations in distribution of cellular infiltrates, isolated liver and spleen T cells were stained for CD3, and their percentages were determined by FACS. They were assessed in naive, fibrotic, and resolution phase of the CCl4 model in WT and CB2−/− mice. Figure 1D shows a low T cell percentage in the liver extracts of the naïve and fibrotic CB2−/− mice compared with the WT ones. No differences in the T cells infiltrated were seen in the fibrotic resolution groups in both animal strains. Spleen T cell (Fig. 1E) showed to be unchanged in the WT as well as the CB2−/− mice. The fibrotic resolution groups also did not show significant differences compared with the fibrotic groups in both strains. Subpopulations of T cells did not show any relevant pattern restricted to resolution (data not shown). These results support an originated hepatic decrease in T cell of CB2−/− mice.
H&E immunohistochemical staining (Fig. 1F) for the four major animal groups showed lymphocyte infiltration in the fibrotic WT mice that were attenuated in the fibrotic CB2−/− mice. No inflammatory infiltrates were seen in H&E staining of naïve WT and naïve CB2−/−. The H&E staining results were in line with obtained serum ALT levels. Figure 1G shows the lack of fibrosis stained by Sirius red in both naïve WT and CB2−/− mice. CCl4 induction showed increased formation of red fibrosis septae in WT and more prominence in CB2−/− mice. These results confirmed the same pattern of measured liver α-SMA expression.
To define constituents for other fraction of CD45+ cells that may infiltrate these livers, F4/80 stain for the macrophage marker was assessed. Figure 1H shows cell infiltrate including macrophages (F4/80-positive cells) in the fibrotic WT and CB2−/− animals. CB2−/− mice showed a high macrophage infiltrate similar to the pattern obtained from the FACS analysis.
Taking this all together, although CB2−/− mice had lower lymphocyte subsets and lower inflammation, they displayed increased fibrosis compared with WT mice. The increased fibrosis is a direct effect of the absent CB2 receptor on HSCs (9).
CB2−/− splenocytes have lower fibrotic properties following adoptive transfer.
To examine the fibrogenic potential of CB2−/− lymphocytes, an adoptive transfer model was used. WT recipients were sublethally irradiated and then were reconstituted with WT (groups A and C) or CB2−/− (groups B and D) lymphocytes (Fig. 2). Recipients were then randomized either with (groups A and B) or without (groups C and D) fibrosis induction. Both fibrotic recipient groups that were reconstituted with WT or CB2−/− lymphocytes (groups A and B, respectively) exert a significant increase of serum ALT levels to 4,136.7 ± 1,095.2 (P = 0.001) and 3,366.7 ± 825 (P = 0.0002) IU/l (respectively, Fig. 2A), compared with 55.5 ± 0.5 (group C) and 101 ± 33.1 IU/l (group D) in the nonfibrotic recipients. Fibrotic groups showed a nonsignificant decreased in serum ALT levels in the CB2−/− recipients. Hepatic fibrosis was assessed by α-SMA expressions (Fig. 2, B–D). Hepatic α-SMA mRNA expression (Fig. 2B) was compared with nonfibrotic naives and showed a significant (P = 0.01) increase to 16.4 ± 7.2-fold in WT lymphocytes recipients (group A) and a nonsignificant trend (P = 0.1) of 4.9 ± 4.6-fold increase in CB2−/− lymphocytes recipients (group B). Fibrotic WT lymphocyte recipients (group A), however, showed a significant (P = 0.04) fibrosis increase compared with the fibrotic CB2−/− lymphocyte recipients (group B). Hepatic α-SMA quantitation was increased by fibrotic groups (groups A and B) compared with nonfibrotic ones (groups C and D, data not shown). Figure 2C shows α-SMA expressions in the upper bands and β-actin expressions in the lower bands. Although β-actin expressions were similar in all tested samples, α-SMA bands were stronger in fibrotic-WT lymphocyte recipients (group A) compared with fibrotic CB2−/− lymphocyte recipients (group B). Density of bands was assessed as described (12–15), and α-SMA/β-actin ratio was calculated to express fibrosis severity. The α-SMA/β-actin ratio (Fig. 2D) was 0.72 ± 0.07 in the fibrotic WT lymphocyte recipients (group A) and significantly (P = 0.0007) decreased to 0.43 ± 0.18 in fibrotic CB2−/− lymphocyte recipients (group B). Histological assessments of the H&E stains of the liver sections showed similar lymphocyte infiltrate patterns between fibrotic WT and fibrotic CB2−/− mice (data not shown).
It looks that, in the CB2−/− fibrotic animals, a fibrotic homeostasis between direct HSC activation vs. decreased activation via inflammatory and lymphocyte ameliorations was balanced in favor of increased fibrogenesis. The results of the adoptive transfer model, therefore, suggest isolated antifibrotic properties of the CB2−/− lymphocytes compared with WT.
Reduced phagocytic properties of CB2−/− splenocytes attributable to increased apoptosis.
To understand the anti-inflammatory and antifibrotic mechanism of CB2−/− lymphocytes, their apoptosis was assessed by FACS. Naïve and fibrotic WT lymphocytes from WT and CB2−/− mice were incubated in vitro with or without CB2 receptor antagonist (Fig. 3). The use of the CB2 antagonist on the CB2−/− lymphocytes was to explore the selectivity of the antagonist. The CB2 receptor antagonist was selected to mimic the in vivo CB2−/− results. Apoptosis of splenocytes (Fig. 3A) was significantly (P = 0.01) increased in naïve WT lymphocytes from 7.9 ± 2.4% to 20.8 ± 5.8% following splenocyte incubation with the CB2 antagonist. Apoptosis was also significantly increased (P = 0.002) in fibrotic WT lymphocytes from 1.4 ± 0.2% to 25.6 ± 7% following CB2 antagonist incubation. Fibrosis levels (without and with CB2 antagonism) were significantly different (P value of 0.005 and 0.007, respectively) compared with naïve splenocytes without CB2 antagonism. However, apoptosis following CB2 antagonism was similar (P = 0.2) in fibrotic and nonfibrotic lymphocytes. The incubations of the CB2−/− splenocytes with the antagonism (Fig. 3B) showed no changes in their apoptosis in the naïve and fibrotic groups, suggesting the specificity of the SR144528 for the CB2 receptor.
Primary mice HSCs isolated from naïve WT animals were used for the in vitro coculture with lymphocytes. WT and CB2−/− and fibrotic lymphocytes from naïve and fibrotic mice were incubated for 2 h with vehicle or with the CB2 antagonism before coculture with primary HSCs. After 48 h of cocultures, phagocytosis of lymphocytes was assessed by FACS. We previously reported that lymphocytes are either adhered or engulfed by HSCs (17). We defined phagocytosis as lymphocytes either adhered (CD45+ α-SMA-) or engulfed (CD45+ α-SMA+) by the primary HSCs. In Fig. 3C, CB2 antagonist-treated WT splenocytes showed less engulfment by the HSCs. This decrease in phagocytosis, although more prominent in the fibrotic splenocytes (P = 0.008), was also significant in the naive counterparts (P = 0.04). The preincubations of the CB2−/− splenocytes with the antagonism (Fig. 3D) before cocultures showed no changes in their phagocytosis in the naïve and fibrotic groups, confirming a selectivity of the antagonist with the CB2 receptor.
To assess a direct cellular CB2 receptor role on human HSCs, CB2 receptor antagonist was incubated in different concentrations for 2 h with healthy lymphocytes and LX2 cells (each with 106 cells/ml) (Table 1). Whereas CB2 receptor antagonist at 1 μM did not alter healthy lymphocyte apoptosis as untreated groups, the 2 μM concentration (P = 0.06, almost significant) and 4 μM concentration (P = 0.03, significant) increased their apoptosis (Table 1). Whereas 2 and 4 μM concentrations dramatically (P < 0.0001) increased the apoptosis of α-SMA-expressing LX2 cells, the 1 μM concentration achieved apoptosis levels similar to the LX2 cells without CB2 antagonism effect (Table 1). Therefore, we consider the 1 μM concentration of the used CB2 antagonist as optimal for our culture conditions (similar data were obtained with mice splenocytes, data not shown). We also assessed the α-SMA mean fluorescence intensities of LX2 cells as a marker to their activation severity. Although 2 and 4 μM concentrations significantly (P < 0.0001) increased the α-SMA densities as nonphysiological conditions, the 1 μM concentration significantly increased LX2 activation (P = 0.004) as α-SMA densities increased from 33.8 ± 2 without CB2 antagonism effect to 83.6 ± 6.3 arbitrary units following incubation of LX2 cells with 1 μM CB2 antagonist (Table 1).
Following the selection of 1 μM as the optimized concentration of CB2 receptor antagonist, healthy and HCV-derived PBL (106 cells each) were incubated 2 h with or without CB2 receptor antagonist. Following incubation, cells were harvested, washed, and stained. To assess outcome of CB2 antagonism, PBL proliferation (Fig. 4A) and apoptosis (Fig. 4B) were then analyzed. T cell proliferation of healthy PBL significantly (P = 0.05) increased from 393.0 ± 122.6 to 637.9 ± 338.7 CPM-B. T cell proliferation of HCV-PBL, however, was 678.2 ± 362.4 CPM-B. Although it was significantly (P = 0.03) higher than healthy lymphocytes, it was similar to healthy PBL proliferation levels following CB2 antagonism. HCV lymphocyte proliferation significantly (P = 0.003) decreased to 265.6 ± 94.6 CPM-B following CB2 receptor antagonism. Million washed PBL from each of the four conditions were cocultured with million LX2 cells. After 48 h of coculture, harvested cells were washed and analyzed for apoptosis (Fig. 4B) and phagocytosis (Fig. 4C) by FACS. Unlike results from Fig. 4A, the current results (Fig. 4B) are derived from adhered PBL to LX2 cells. It shows similar apoptosis rates of both healthy PBL conditions as well as the HCV lymphocytes (without CB2 antagonism). Incubation with CB2 receptor antagonism significantly increased HCV lymphocyte apoptosis (P < 0.0001). Phagocytosis of HCV-PBL by LX2 was defined as CD45+ adhered cells CD45+ cells. Phagocytosis was significantly (P = 0.009) decreased from 71.4 ± 2.1 to 55.1 ± 3.7% of total cultured LX cells (Fig. 4C).
Cannabinoid receptors (CB1, CB2) are upregulated on HSCs (9). It is assumed that upregulation of CB1 is involved in the development of hepatic fibrosis and steatosis (30). In a recent study, the expression of CB1 receptor was increased in chronic hepatitis C and associated with steatosis in humans (31). By contrast, CB2 receptor activation mainly protected from liver injury (19).
In the current study, the hepatic fibrosis model was generated through CCl4 inductions in WT and CB2−/− mice for 6 wk and were compared with naïve states. Similar to the reported antifibrogenic effects of CB2 receptor (9), we showed that CB2−/− mice exert severe fibrosis (Fig. 1A). Inflammatory activity, however, was significantly lower in the CB2−/− mice as detected by serum ALT levels (Fig. 1B) and H&E staining (Fig. 1D). Although the acute CCl4 model (0.5 ml/kg body wt, 1:5 dilution in mineral oil) showed increased ALT serum levels in CB2−/− compared with WT mice (9), it appeared that a chronic model might behave differently. Data from CB2−/− mice fed a high-fat diet support this proposition. Here, CB2 receptors contribute to obesity-associated inflammation, insulin resistance, and nonalcoholic fatty liver disease (NAFLD) and suggest that CB2 receptor antagonists may open a new therapeutic approach for the management of obesity-associated NAFLD (6, 20). Indeed, preclinical studies have demonstrated the critical role of CB2 receptors in different inflammatory process associated with rheumatoid arthritis, inflammatory bowel diseases, atherosclerosis, as well as liver ischemia-reperfusion injury (2, 4, 27).
It has been demonstrated in human cirrhotic liver samples that the expression of CB2 receptor is limited primarily to cells positive for α-SMA located within fibrotic septa; however, it is also detected in nonparenchymal and inflammatory cells adjacent to them (28). Thus we looked at the composition of intrahepatic lymphocytes. Whereas WT animals showed the well-recognized CD4, CD8, and NK fibrotic patterns (13, 15, 16), the naïve CB2−/− animals had significantly lower levels of hepatic lymphocyte subsets that were not altered following induction of fibrosis (Fig. 1C). Our results validate the reported (5) unchanged numbers of spleen T cells in the WT and CB2−/− animals (including CD4+ and CD8+ T cells, Fig. 1E). However, we show a reduced number of CB2−/− hepatic T cells compared with the WT animals. Although CB2−/− mice had lower liver lymphocyte subsets and less inflammation, they exhibited increased fibrosis compared with WT mice. Such increased fibrosis was shown previously to occur as a direct effect of the missing receptor on the HSCs (7, 9). CB2−/− animals, on the other hand, had fewer lymphocytes and less inflammation that may contribute to decreased fibrogenesis. Therefore, our current results suggest pivotal effects of CB2 receptors in hepatic fibrosis of CB2−/− mice.
A large number of in vitro and in vivo studies demonstrated that CB2 is capable of suppressing immune responses, suggesting that CB2 would make a good therapeutic target for the treatment of immune disorder (1). However, very little is known about the cellular and molecular mechanisms whereby CB2 modulates immune responses. Because CB2 receptors are found predominantly in the immune system cells (2), we utilized the fact that CB2 receptor is also expressed in splenocytes (11). Thus, to isolate the fibrogenic outcome of CB2−/− lymphocytes, the model of adoptive transfer of splenocytes was used (15, 25). Using the adoptive transfer model, we previously showed that intraperitoneal injections of lymphocytes to SCID mice were homed to the liver of the recipients, confirmed by histological sections and FACS analysis of the intrahepatic lymphocytes alterations (24). Both fibrotic recipient groups that were reconstituted with WT or CB2−/− splenocytes exhibited a significant increase of serum ALT levels, with a nonsignificant trend of decreased levels in the CB2−/− recipients (Fig. 2A). A significant increase of fibrosis was seen in recipients reconstituted with WT splenocytes compared with CB2−/− splenocytes (Fig. 2, B–D). The results of adoptive transfer model, therefore, imply isolated antifibrotic properties of the CB2−/− splenocytes compared with WT. Increased apoptosis of WT lymphocytes following incubation with CB2 antagonist (Fig. 3A) suggests a CB2−/−-mediated mechanism for their decrease and subsequent anti-inflammatory response. After splenocyte coculture with primary isolated HSCs, phagocytosis (Fig. 3B) tends to decrease in naïve samples but was significantly increased in fibrotic WT lymphocytes following incubation with CB2 antagonist. All together, it appears that CB2−/− fibrotic animals had a pivotal pro- and antifibrotic equilibrium that finally was decided in favor of increased fibrogenesis (Fig. 1A). There was a balance between direct HSC activation as a profibrotic pathway (7) vs. indirect decreased activation of inflammatory (Fig. 1B) and lymphocyte pathways (Fig. 1C). Apoptosis of splenocytes and decrease of their phagocytosis by HSCs provide, therefore, antifibrotic actions of lymphocytes following CB2 blockade (CB2−/− or CB2 antagonism). Similar imbalances in endocannabinoid system signaling have been observed in various pathological conditions, including nervous system disorders, metabolic disturbances, impaired immunological responses, cardiovascular and gastrointestinal diseases, and carcinogenesis (12, 22).
In the current study, human LX2 cells mimic rodent HSC activation in the case of CB2-blockade (9). CB2 cannabinoid receptor antagonist, SR144528, has a 700-fold higher affinity for the CB2 receptor than for the CB1 receptor (23). Based on the ex vivo [3H]-CP 55,940 binding studies, SR 144528 appears to be effective in blocking the CB2 but not the CB1 receptors (ED50 value of 0.36–0.06 mg/kg in the spleen vs. no interaction in the brain up to 10 mg/kg po or 10 mg/mouse iv) with a long duration of action after oral administration in mice. In this study, we have confirmed this selectivity of the antagonism through its incubations with CB2−/− splenocytes (Fig. 3).
Although concentration of CB2 antagonist at 1 μM did not alter lymphocytes and LX2 apoptosis following incubation, it significantly increased LX2 activation (Table 1). Healthy or HCV-derived lymphocytes were incubated with or without CB2 receptor antagonist (Fig. 4). T cell proliferation of healthy lymphocytes significantly (Fig. 4A) increased to reach HCV levels. Conversely, HCV lymphocyte proliferation significantly decreased following CB2 receptor antagonism as observed in fibrotic rodent splenocytes. LX2 cells were cocultured with healthy or HCV-derived lymphocytes that were preincubated with or without CB2 receptor antagonist. After coculture, harvested lymphocytes showed similar apoptosis rates to both healthy lymphocyte conditions (with and without CB2 antagonism), as well as to the HCV lymphocytes without CB2 antagonism. Incubation with CB2 receptor antagonism only increased HCV lymphocyte apoptosis (Fig. 4B). Phagocytosis of HCV lymphocytes by the LX2 cells (HSCs), on the other hand, significantly decreased (Fig. 4C) following CB2 blocking.
Hence, our in vitro human culture analyses are in line with both in vivo and in vitro rodent results. CB2 blockade or knockout directly activates HSCs, whereas, post-CB2 blocking or knockout HCV lymphocytes as well as fibrotic mice lymphocytes show decreased HSC activation. Antifibrotic immune effects of CB2 blockade were mediated via increasing lymphocyte apoptosis, decreasing T cell proliferation, as well as decreasing lymphocyte phagocytosis by HSCs. The in vitro rodent models confirmed alleviation of inflammation as a result of lymphocyte alterations. Although a recent study established the curative properties of a CB2 agonist in cirrhotic rats (19), the understanding of multiple CB2 fibrotic pathways may lead to a better recognition of differential interventional outcomes for acute or chronic liver disease. Modulation of endocannabinoid receptor activity and the balance between CB1 and CB2 receptors, with appropriately timed agonists and antagonists, could have therapeutic implications in the management (and prevention) of hepatic fibrosis.
This work was supported by grants from Israel Scientific Foundation (ISF), Madaan Rashi, and the Israel-American Bi-national Scientific Foundation (BSF) Awards.
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: Y.A., E.M.B., and R.S. conception and design of research; Y.A., J.A., and R.S. edited and revised manuscript; J.A. analyzed data; J.A. interpreted results of experiments; J.A. prepared figures; J.A. and R.S. drafted manuscript; J.A. and R.S. approved final version of manuscript; S.D., L.A.-T., M.M., and A.A.K. performed experiments.
We thank Scott Friedman for input on the manuscript.
- Copyright © 2012 the American Physiological Society