The Critical Role of Chemokine (C–C Motif) Receptor 2-Positive Monocytes in Autoimmune Cholangitis

The therapy of primary biliary cholangitis (PBC) has lagged behind other autoimmune diseases despite significant improvements in our understanding of both immunological and molecular events that lead to loss of tolerance to the E2 component of pyruvate dehydrogenase, the immunodominant autoepitope of PBC. It is well known that Ly6Chi monocytes are innate immune cells infiltrating inflammatory sites that are dependent on the expression of C–C motif chemokine receptor 2 (CCR2) for emigration from bone marrow. Importantly, humans with PBC have a circulating monocyte pro-inflammatory phenotype with macrophage accumulation in portal tracts. We have taken advantage of an inducible chemical xenobiotic model of PBC and recapitulated the massive infiltration of monocytes to portal areas. To determine the clinical significance, we immunized both CCR2-deficient mice and controls with 2OA-BSA and noted that CCR2 deficiency is protective for the development of autoimmune cholangitis. Importantly, because of the therapeutic potential, we focused on inhibiting monocyte infiltration through the use of cenicriviroc (CVC), a dual chemokine receptor CCR2/CCR5 antagonist shown to be safe in human trials. Importantly, treatment with CVC resulted in amelioration of all aspects of disease severity including serum total bile acids, histological severity score, and fibrosis stage. In conclusion, our results indicate a major role for Ly6Chi monocytes and for CCR2 in PBC pathogenesis and suggest that inhibition of this axis by CVC should be explored in humans through the use of clinical trials.

The therapy of primary biliary cholangitis (PBC) has lagged behind other autoimmune diseases despite significant improvements in our understanding of both immunological and molecular events that lead to loss of tolerance to the E2 component of pyruvate dehydrogenase, the immunodominant autoepitope of PBC. It is well known that Ly6C hi monocytes are innate immune cells infiltrating inflammatory sites that are dependent on the expression of C-C motif chemokine receptor 2 (CCR2) for emigration from bone marrow. Importantly, humans with PBC have a circulating monocyte pro-inflammatory phenotype with macrophage accumulation in portal tracts. We have taken advantage of an inducible chemical xenobiotic model of PBC and recapitulated the massive infiltration of monocytes to portal areas. To determine the clinical significance, we immunized both CCR2-deficient mice and controls with 2OA-BSA and noted that CCR2 deficiency is protective for the development of autoimmune cholangitis. Importantly, because of the therapeutic potential, we focused on inhibiting monocyte infiltration through the use of cenicriviroc (CVC), a dual chemokine receptor CCR2/CCR5 antagonist shown to be safe in human trials. Importantly, treatment with CVC resulted in amelioration of all aspects of disease severity including serum total bile acids, histological severity score, and fibrosis stage. In conclusion, our results indicate a major role for Ly6C hi monocytes and for CCR2 in PBC pathogenesis and suggest that inhibition of this axis by CVC should be explored in humans through the use of clinical trials.
Accumulative data suggest involvement of the innate immune system in PBC pathogenesis (9). The comprehension of liver macrophages has been revolutionized by the delineation of heterogeneous subsets of these cells exhibiting functional plasticity. Kupffer cells (KCs) are a self-sustaining, liver-resident population of macrophages and can be distinguished from the recruited monocyte-derived macrophages (MoMFs), which arise from circulating Ly6C hi monocytes that rapidly accumulate in the diseased liver (10). With respect to the latter, monocytes are members of the mononuclear phagocyte system and constitutes a critical effector component of the innate immune arm. Classical Ly6C hi monocytes originate from a common myeloid progenitor cell in the bone marrow, emigrate from the bone marrow in a C-C motif chemokine receptor 2 (CCR2)-dependent manner, and released into the bloodstream as non-differentiated cells that circulate in the blood for 1-3 days (11)(12)(13)(14). Following their recruitment to tissues, monocytes undergo differentiation into tissue macrophages or myeloid dendritic cells (15). Following arrival to their destination, monocyte or cells differentiated from monocytes conduct versatile functions including initiation and progression of inflammation, homeostasis maintenance, host defense, and tissue remodeling and repair, depending on specific environmental signals (16)(17)(18)(19). Several studies have suggested altered monocyte/macrophage function in PBC and accumulation of these cells in the livers of PBC patients (20,21). Mononuclear cells expressing the low-density lipoprotein binding glycoprotein CD68 were detected in the biliary epithelial layer and periductal tissue of PBC patients, whereas in patients with viral hepatitis, these cells were scattered, and in normal livers were rarely seen around bile ducts (21). These studies suggest involvement of monocyte-derived cells in PBC pathogenesis; however, solid data evaluating the importance and function of these cells in the in vivo context were missing so far.
Here, we demonstrate, by using a xenobiotic murine model of PBC, that infiltrating Ly6C hi monocytes are massively recruited into the portal zone of livers in a CCR2-dependent fashion. CCR2-deficient mice subjected to this model exhibit milder disease phenotype. Consequently, pharmacological inhibition of monocyte infiltration by blocking CCR2/CCR5 with Cenicriviroc attenuated disease severity as assessed by serum bile acids, histological severity scoring, and liver fibrosis quantification. Our results indicate a major role for monocyte-derived cells in PBC pathogenesis and proved that targeted inhibition of these cells significantly ameliorated both liver inflammation and fibrosis. These results reinforce further investigation of this therapeutic option in clinical trials.

Mice
Female C57BL/6 WT mice were purchased from ENVIGO. Cx3cr1 gfp/+ , Ccr2 −/− Cx3cr1 gfp/+ mice were kindly provided by Prof. Steffen Jung, the Weizmann Institute of Science and maintained in the animal facility of the Tel-Aviv Sourasky Medical Center. Mice had unrestricted access to food and water, were housed in temperature and humidity-controlled rooms, and were kept on a 12-h light/dark cycle. Use of animals was in accordance with the National Institutes of Health policy on the care and use of laboratory animals and was approved by the Tel-Aviv Sourasky Medical Center Animal Use and Care Committee.

Preparation of immunogen
2-octynoic acid was purchased from Sigma-Aldrich (St. Louis, MO, USA) and was conjugated with BSA (EMD Chemicals, Gibbstown, NJ, USA), as described previously (8). Briefly, 2OA was dissolved in dry dimethyl ether. N-hydroxysuccinimide (NHS) was then added, and the solution was cooled to 0°C and stirred for 20 min. Dicyclohexylcarbodiimide was then added, and the mixture was allowed to warm to ambient temperature overnight. The solution was filtered, concentrated by roto-evaporation under reduced pressure, re-dissolved with ethyl ether, washed with water, NaHCO3 (1 M), brine, dried under magnesium sulfate, filtered, and concentrated. The product was then purified using flash chromatography (30% ethyl acetate/hexane). NHS-activated 2OA was dissolved in dimethyl sulfoxide and then coupled to the lysine residues of BSA. The solution was allowed to react for 3 h followed by dialysis [phosphate-buffered saline (PBS)].

immunization and cVc Treatment
Female mice at 8 weeks of age were immunized with 2OA-BSA at 1 mg/ml per animal intraperitoneally (i.p.) in the presence of complete Freund's adjuvant (CFA) (Sigma-Aldrich) containing 10 mg/ml of Mycobacterium tuberculosis strain H37Ra. In addition, pertusis toxin (Sigma-Aldrich, 100 ng/animal) was delivered i.p. at the day of immunization and 2 days after. All animals were given boost immunizations 2 weeks following the initial immunization with 2OA-BSA in incomplete Freund's adjuvant (Sigma-Aldrich). Animals were sacrificed 8-12 weeks after initial immunization.
For testing the efficacy of CVC, immunized mice received daily i.p. injections of CVC (20 mg/kg) (Allergen plc) or vehicle only as control, starting the day after the boosting until the end of the experiment (8 weeks from initial immunization).

Detection of serum aMas
Serum AMAs were detected using enzyme-linked immunosorbent assay (ELISA) utilizing recombinant human pyruvate dehydrogenase complex-E2 (PDC-E2) to coat individual wells of a 96-well microtiter plate (8). Samples were tested at a dilution of 1:1,000. Immunoreactivity was determined by measuring the optical density at 450 nm after incubation with 100 µl of 3,3′,5,5′-tetramethylbenzidine (Merck, Kenilworth, NJ, USA) for 30 min followed by the addition of 50 µl of 18N H2SO4 stopping solution.
histopathology Livers from euthanized mice were fixed in formaldehyde 4% buffered (pH 7.2), embedded in paraffin, cut into 4-µm sections, deparaffinized, and stained with hematoxylin and eosin. Disease severity was evaluated by experience pathologist (Eli Brazowski) who was blinded to treatment allocation. Scoring was done considering four parameters (portal infiltrate, bile duct damage and loss, granulomas formation, and lobular inflammation) ranging between 0 and 2 (normal, moderate, or severe, respectively).

Fibrosis evaluation
Picro-Sirius red staining was performed using a commercially available kit (ab150681, Abcam, Cambridge, UK). The staining procedure was carried out according to the manufacturer's protocol. Briefly, deparaffinized sections were stained with Picro-Sirius red solution for 1 h. Slides were then rinsed twice in acetic acid solution, followed by dehydration with absolute alcohol and mounting in synthetic resin.
Tissue collagen accumulation was quantitatively assessed by a hydroxyproline accumulation assay (Sigma-Aldrich; MAK008) according to the manufacturer's instructions.

isolation of hepatic non-Parenchymal cells for Flow cytometry
Isolation of hepatic non-parenchymal cells was performed as previously described (22). In brief, mice were anesthetized; the livers were perfused with cold PBS and excised. Livers were cut into small fragments, incubated (37°C, 250 rpm for 45 min) with 5 ml digestion buffer [5% FBS, 0.5 mg/ml collagenase IV (Sigma-Aldrich, Rehovot, Israel, C5138-500MG), 0.1 mg/ml deoxyribonuclease I from bovine pancreas (Sigma-Aldrich, USA) in PBS], and filtered through 200-µm wire mesh. This was followed by three cycles of washings with PBS at 400 rpm, 4°C, 5 min, harvesting the supernatant, and discarding the parenchymal cell pellet. The supernatant was centrifuged at 1,500 rpm, 4°C, 5 min. The erythrocytes in the cell pellet were lysed by 2 min incubation with ACK Lysing buffer (0.15 M NH4Cl, 0.01 M KHCO3) and washed with PBS.

immunofluorescence of Frozen Tissue sections
Cx3cr1-GFP immunofluorescent was performed on frozen liver sections of 13 µm. Slides were incubated in cold acetone for 7 min and dried at room temperature. Samples were closed with Fluorescent Mounting Medium with or without 4′,6′diamidino-2-phenylindole (DAPI, GBI Labs). Images were taken with ZEISS Confocal Microscope (MicroImaging GmbH, ZEISS, Germany). Processing was performed with ZEN 2010 Software.

TnF-α levels analysis by elisa
THP-1 cells suspended in RPMI medium, as described above, were placed in 24-well culture plates at a concentration of 1 × 10 6 cells/ml and incubated for 24 h with various concentrations of CVC (0.1-10 µM) prior to LPS stimulation (10-100 ng/ml) for 3 h. Cell-free supernatants were recovered, and the concen trations of TNF-α were measured using a commercially avai lable sandwich ELISA kit according to the manufacturer's instructions (R&D Systems Inc.).

statistical analysis
The results are presented as mean ± SEM. Statistical significance was assessed using a two-tailed Student's t-test; p values <0.05 were considered as significant.
resUlTs characterization of hepatic Macrophages in steady state and experimental chronic autoimmune cholangitis The liver macrophage compartment harbors a main selfmaintaining population of resident macrophages named KCs and an infiltrating population of MoMFs that accumulate to be the dominant liver macrophage population under various pathological settings. Others and we have demonstrated that infiltrating MoMFs can be distinguished from resident KCs by their expression of the fractalkine chemokine receptor (CX3CR1) (22,23). To identify and characterize macrophages populations in the liver under the setting of autoimmune cholangitis, we took advantage of Cx3cr1 gfp/+ reporter mice (24). Cx3cr1 gfp/+ and Cx3cr1 gfp/+ Ccr2 −/− mice were subjected to the 2OA-BSA PBC model and 8 weeks following the initial immunization liver macrophages subsets were evaluated. As previously reported, flows cytometry analysis of non-parenchymal liver cells in steady-state livers revealed two main types of CD64 + MHC II + macrophages: CD11b int F4/80 hi CX3CR1 neg KCs and the CD11b hi F4/80 int CX3CR1 hi MoMFs (Figure 1A). In vast contrast, livers from the immunized mice exhibited profound accumulation of MoMFs and reduction in the occurrence of KCs. As predicted, in CCR2-deficient mice, monocytederived cells did not infiltrate the chronically inflamed liver (Figures 1B,C). Visualization of liver section derived from immunized mice by fluorescent microscopy, revealed accumulation of CX3CR1-GFP-positive cells around the portal triads of 2OA-BSA immunized Cx3cr1 gfp/+ mice, whereas in Cx3cr1 gfp/+ Ccr2 −/− immunized mice and Cx3cr1 gfp/+ naive mice very few CX3CR1-GFP-positive cells were detected in the portal area ( Figure 1D).

ccr2-Deficient Mice Display attenuated Disease severity in the 2Oa-Bsa autoimmune cholangitis Model
To evaluate the impact of MoMF deficiency on autoimmune cholangitis, liver sections of Cx3cr1 gfp/+ reporter mice and Cx3cr1 gfp/+ Ccr2 −/− mice immunized with 2OA-BSA were examined. Cx3cr1 gfp/+ mice demonstrated periportal infiltration of lymphocytes and mononuclear cells and bile ducts destruction, whereas, in Cx3cr1 gfp/+ Ccr2 −/− mice, no abnormalities were observed (Figure 2A). Histological disease severity scoring revealed a significant attenuation in disease severity in CCR2-deficient mice ( Figure 2B). Serum AMA titers were comparable in Cx3cr1 gfp/+ reporter mice and Cx3cr1 gfp/+ Ccr2 −/− mice immunized subjected to the 2OA-BSA model and found to be significantly higher then in the serum of naïve mice ( Figure S1A in Supplementary Material).
To further investigate the effect of MoMF deficiency on liver inflammation, expression of pro-inflammatory cytokines and chemokines was examined in liver tissue using quantitative real time PCR. The levels of Tnfα, Ifnγ, Il1β, Ccl2, and Cx3cl1 were significantly increased in Cx3cr1 gfp/+ mice but not in Cx3cr1 gfp/+ Ccr2 −/− mice subjected to the 2OA-BSA model, indicating that MoMFs profoundly contribute to the development of pro-inflammatory response in the immunized livers ( Figure 2C).

ccr2-Deficient Mice are Protected From the Development of liver Fibrosis in the 2Oa-Bsa autoimmune cholangitis Model
To assess the extent of liver fibrosis, Sirius red staining for detection of collagen distribution was performed at 8 weeks following immunization with 2OA-BSA. Cx3cr1 gfp/+ mice exhibited periportal collagen deposition while Cx3cr1 gfp/+ Ccr2 −/− or Cx3cr1 gfp/+ control mice displayed normal collagen distribution similar to the non-immunized Cx3cr1 gfp/+ control mice ( Figure 3A). This finding was supported by a lower pro-fibrotic hepatic gene expression signature in Cx3cr1 gfp/+ Ccr2 −/− mice including significant reduced expression levels of genes encoding for αSMA, Coll1, and TIMP-1 in CCR2-deficient mice ( Figure 3B). There was also clear reduction trend in the expression levels of the gene encoding for TGFβ, previously pointed as a key pro-fibrotic factor provided by infiltrating Ly6C hi monocytes (25).

Pharmacological inhibition of ccr2/ccr5 Blocks Monocytes infiltration and attenuates experimental autoimmune cholangitis
In light of our findings regarding the role of CCR2 in the development of autoimmune cholangitis, we aimed at assessing the efficacy of the dual CCR2/CCR5 inhibitor Cenicriviroc (CVC) as a potential treatment for this pathology (26). WT-C57BL/6 and Cx3cr1 gfp/+ reporter mice were subjected to the 2OA-BSA model. CVC treatment was initiated 1 day after 2OA-BSA boosting (day 15) by daily intra-peritoneal injections (2 mg/kg), for the remaining 6 weeks. Control mice received vehicle only. Flow cytometry analysis of non-parenchymal liver cells demonstrated a significant reduction in the occurrence of infiltrating Ly6C hi monocytes in the CVC-treated mice (Figures 4A,B). This was further manifested by the reduced fraction of the MoMFs descendants of infiltrating Ly6C hi monocytes. CVC treatment was specific to the Ly6C hi monocyte-derived liver macrophages as there was increased representation of resident KCs in livers from CVC-treated mice (Figures 4C,D).
Next, we evaluated the effect of CVC treatment on parameters of disease severity. Histologic evaluation of liver sections from CVC-treated mice revealed normal liver tissue with almost no infiltration of leukocytes, whereas mice administered with vehicle only exhibited periportal infiltration and accumulation of immune cells around the bile ducts ( Figure 5A). Serum AMA titers were comparable in CVC and vehicle-treated groups subjected to the 2OA-BSA model and found to be significantly higher then in the serum of naïve mice, indicating that this intervention did not interfere with AMA production ( Figure  S1B in Supplementary Material). Blinded histological severity score assessment revealed significant improvement in CVCtreated mice (Figure 5B). In addition, serum level of total bile  acid, a biochemical marker of PBC, was markedly decreased in mice treated with CVC ( Figure 5C). Assessment of hepatic pro-inflammatory cytokines and chemokines milieu revealed lower gene expression levels of Tnfα, Ifnγ, Ccl2, and Il1β in CVCtreated mice (Figure 5D).

cenicriviroc ameloriates liver Fibrosis in the 2Oa-Bsa autoimmune cholangitis Model
Evaluation of fibrosis severity was performed by Sirius red staining, hepatic hydroxyproline quantification, and expression levels analysis of pro-fibrotic genes in the liver (Figures 6A-C, respectively). While portal fibrosis was clearly evident in the livers of vehicle-treated group, livers from CVC-treated mice displayed normal collagen deposition ( Figure 6A). In addition, hydroxyproline levels were significantly higher in the vehicletreated group versus CVC-treated group, indicating higher collagen content in vehicle-treated mice ( Figure 6B). Furthermore, real time-PCR analyses for pro-fibrotic genes in liver samples revealed significantly higher expression levels of genes encoding for TIMP-1, Coll1, and TGFβ in vehicle-treated versus CVCtreated mice ( Figure 6C). Collectively, these results highlight a therapeutic potency for CVC in the amelioration of hepatic fibrosis in experimental PBC.

DiscUssiOn
Here, we explored the role of CCR2-positive monocytes in the development of experimental PBC. We showed that these cells massively infiltrate the liver to become the dominant hepatic macrophage population in the setting of chronic cholangitis. Inhibition of the recruitment of these cells either by genetic deficiency in CCR2-deficient mice or by pharmacological inhibition of this pathway resulted in amelioration of all aspects of the pathology including portal inflammation, biochemical markers, and fibrosis.  In the setting of acute and chronic inflammation, monocyte infiltrate has the potential to aggravate the pathology by various mechanism including antigen presentation and secretion of various mediators including pro-inflammatory cytokine and factors supporting angiogenesis and fibrosis (19,27,28). The mechanism Ly6C hi monocytes and cells derived from these cells contribute to autoimmune cholangitis is probably multifactorial and has not been elucidated in this study. Future studies utilizing in vivo tools that target-specific functions of these cells are warranted and might discover new targets to treat this pathology. Of note, other murine models of hepatic fibrosis uncovered opposing functions of hepatic macrophages during progression and regression of fibrosis. Therefore, depletion of macrophages could not only prevent fibrosis development but also delayed repair processes after cessation of the injury (29). It has been suggested that fibrogenic macrophages can switch their phenotype toward restorative macrophages, which are characterized by high expression of anti-inflammatory mediators and matrix-degrading metalloproteinases (30). Thus, the overall effect of long-term inhibition of monocyte recruitment to the liver in PBC will have to be evaluated in appropriate clinical trial.  Cenicriviroc (CVC) is an orally available, dual inhibitor of the chemokine receptors CCR2 and CCR5 (26). CVC treatment has been recently shown to ameliorate disease severity in experimental models of liver disorders including nonalcoholic steatohepatitis and paracetamol induced liver injury (31,32). A phase 2b trial of CVC in patients with non-alcoholic steatohepatitis (NASH) showed improved fibrosis with safety and tolerability comparable to placebo (33), and a phase 3 study for the treatment of liver fibrosis in adults with NASH is currently recruiting (NCT03028740). We have shown here that CVC efficiently blocked the infiltration of Ly6C hi monocytes to the liver in a model of autoimmune cholangitis; however, CVC inhibits also CCR5 and potentially can affect other CCR5-expressing leukocytes. We have found no effect on the hepatic recruitment of T cells, and actually the proportion of hepatic T cells out of all hepatic leukocytes increased in CVC-treated mice without affecting the ratio between hepatic CD4 + and CD8 + T cells ( Figure S2 in Supplementary Material). This increase in T cell  proportion is most probably due to the significant effect of CVC on monocyte and monocyte-derived cells infiltration. Of note, two recently published studies showed no effect of CVC on the hepatic recruitment of other CCR5-expressing lymphocyte populations in experimental model of acute liver injury and NASH (31,34).
To exclude that CVC has additional effects on monocyte function, we have exposed LPS-activated THP-1 monocytic cells to various concentration of CVC and found no effect on TNFα secretion  ( Figure S3 in Supplementary Material), corroborating previous results that demonstrated no effect of CVC on bone marrowderived macrophage polarization (34). Primary biliary cholangitis is a classic autoimmune disease characterized by the existence of antigen-specific antibodies as well as antigen-specific T cells with an identified self-antigen expressed specifically by the target tissue-the biliary epithelium (1). Genome-wide association studies suggest a role for innate immunity and in particularly for mononuclear phagocytes in this pathology as manifested by polymorphism in key genes known to be expressed specifically by myeloid cells and having the potential to generate or augment immune response, e.g., Il12a, Irf5, cd80, Tnfrsf1a, and Tnfaip2 (35). We have previously observed an increase in the frequency and absolute numbers of circulating monocytes in patients with PBC compared with sex-matched controls of similar age (20). More important, we have demonstrated that CD14 pos monocytes, the human parallel to the mice Ly6C hi monocytes, isolated from patients with PBC are more sensitive to signaling via selected toll-like receptors stimulation, resulting in secretion of pro-inflammatory cytokines (20). All of the above suggested an important role for monocyte or monocyte-derived cells in PBC development; however, evidence so far was correlative. Here, we have demonstrated for the first time in the in vivo setting the crucial role of Ly6C hi monocytes in the development of this pathology, and by inhibiting the hepatic recruitment of these cells ameliorated all aspects of the disease.
Ursodeoxycholic acid (UDCA) has been the mainstay of therapy for PBC for more than two decades. However, about one-third of patients has persistent elevation of alkaline phosphatase or bilirubin and considered as inadequate responders (36). Obeticholic acid, a farnesoid X receptor agonist, was recently conditionally approved as an add-on therapy in patients with PBC who are inadequate responders to UDCA or who cannot tolerate this drug. However, about 50% of patients with PBC still lack an adequate response to a combination of UDCA and obeticholic acid (37), and therefore, there is still a substantial medical need to develop new therapies for PBC.
In conclusion, Ly6C hi monocyte and their progeny critically contribute to the pathogenesis of PBC and inhibition of their recruitment constitutes an attractive therapeutic avenue in patients non-responding to current available medications. These results reinforce investigation of this therapeutic option in clinical trials.

eThics sTaTeMenT
This study was carried out in accordance with the recommendations of National Institutes of Health policy on the care and use of laboratory animals and the protocol was approved by the Tel-Aviv Sourasky Medical Center Animal Use and Care Committee.