Quantitation of the Rank-Rankl Axis in Primary Biliary Cholangitis

There is substantial data that suggests an abnormality of innate immunity in patients with primary biliary cholangitis (PBC) which includes the transcription factor nuclear factor-kB (NF-kB) and well as downstream inflammatory signaling pathways. In addition, ImmunoChip analysis has identified a novel PBC-associated locus near the receptor activator of NF-kB ligand (RANKL) gene. Based on these observations, we investigated the role of the RANKL axis in the liver of patients with PBC compared to controls. We used immunohistochemistry to quantitate liver expression of RANKL, its receptor (RANK), and importantly the decoy receptor osteoprotegerin (OPG), including a total of 122 liver samples (PBC = 37, primary sclerosing cholangitis = 20, autoimmune hepatitis = 26, chronic hepatitis B = 32 and unaffected controls = 7). In addition, we studied RANKL-RANK-OPG co-localization in CD4 and CD8 T cells, B cells, dendritic cells, macrophages, NK, NKT cells, hepatocytes, and cholangiocytes. We report herein that RANK is constitutively expressed by cholangiocytes in both unaffected and diseased liver. However, cholangiocytes from PBC express significantly higher levers of RANK than either the unaffected controls or liver diseased controls. CD4, CD8 and CD19 cells with in the portal areas around bile ducts in PBC express significantly higher levels of RANKL compared to controls. Importantly, the overall hepatic RANKL level and the ratio of hepatic RANKL/OPG correlated with disease severity in PBC. In conclusion, our data indicate a role of RANK-RANKL axis in the innate immune activation in PBC and we hypothesize that the damaged cholangiocytes, which express high levels of RANK, lead to the recruitment of RANKL positive cells and ultimately the classic portal tract infiltrates.


Introduction
Primary biliary cholangitis (PBC) [1], is a chronic liver disease characterized by progressive destruction of intrahepatic bile ducts resulting in cholestasis, portal tract inflammation, and fibrosis that may progress to cirrhosis and ultimately end-stage liver disease [2,3]. Although the exact etiology of PBC remains unknown, it is widely recognized that the development of PBC requires one or more environmental factors that initiate an autoimmune response in genetically predisposed individuals [4][5][6]. Previous work from our group has identified a novel disease-associated locus near the TNFSF11 gene, upstream of the gene encoding the receptor activator of nuclear factor κB (NF-κB) ligand (RANKL) [7]. Even though it is inherently difficult for genome wide association studies to identify a specific gene, a large amount of data suggest that RANKL may play a relevant role in PBC.
Since the discovery of the receptor activator of NF-κB (RANK) in the late 1990s, the RANKL/RANK/osteoprotegerin (OPG) system has been implicated in regulating immune responses; RANK, also known as TRANCE Receptor, is a type I membrane protein that is expressed on the surface of osteoclasts, dendritic cells and mammary gland epithelial cells physiology [8][9][10][11]. The cognate ligand for RANK, RANKL, is a member of the tumor necrosis factor (TNF) cytokine family, which functions as a key factor for osteoclast differentiation and activation. In addition, the RANKL/RANK pathway serves a critical role in the immune system and augments the ability of dendritic cells (DC) to stimulate naive T cell proliferation and enhance DC survival [12]. It also has important regulatory functions in lymph node organogenesis and lymphocyte differentiation [13]. OPG is a decoy receptor for RANKL. Through the binding of RANKL, OPG inhibits NF-kB, a key regulator of inflammation, innate immunity and immune cell survival and differentiation [9,14].
The essential function of the RANKL-RANK-OPG axis in immune system is relevant to autoimmunity. The gene encoding RANKL, or TNFSF11, is positioned within a confirmed loci implicated in Crohn's disease [15,16]. Moreover, elevated serum levels of OPG and soluble RANKL have also been reported in rheumatoid arthritis (RA) [17,18]. Interestingly, RANKL mRNA is present in the synovial lining layer in RA but not in normal synovia [19]. The role of OPG and RANKL in bone remodeling and development of osteoporosis in PBC has been previously investigated [20]. We report herein data that strongly suggest abnormal RANK--RANKL signaling in PBC; further investigation is needed to better understand the recruitment of the inflammatory infiltrate that targets biliary cells.

Human Subjects
A total of 115 patients with chronic liver diseases and 7 unaffected controls were enrolled in this study. These patients included 37 subjects with well characterized PBC, 26 subjects with autoimmune hepatitis (AIH), 32 subjects with chronic hepatitis B (CHB), and 20 subjects with primary sclerosing cholangitis (PSC). Based on Scheuer's classification, 37 PBC patients were staged, including stage I and II (n = 14) and stage III and IV (n = 23) [21]. The clinical data are summarized in Table 1 and the diagnosis was based on established criteria for PBC [22], PSC [23], AIH [24], and CHB [25], respectively. Unaffected subjects were included as controls (HC); these samples were derived from patients undergoing liver surgery for benign lesions (haemangioma resection, n = 3) and liver donors for Orthotropic Liver Transplantation (n = 4). None of these individuals had evidence of liver pathology. Samples from PBC subjects as well as AIH, CHB, and HC were collected in China, whereas PSC samples were obtained from subjects followed at the Center for Autoimmune Liver Diseases, Humanitas Clinical and Research Center.
Informed consent in writing was obtained from all subjects and the study protocol was approved by the Ethics Committee of Renji Hospital, Shanghai Jiao Tong University and Humanitas Clinical and Research Center, Rozzano, Milan, Italy.

Immunohistochemistry
All liver samples were fixed in 10% formalin and embedded in paraffin before cutting into 4μm sections. Immuohistochemistry staining of RANKL, RANK and OPG was performed with a Leica BondTM system (Leica, Germany). The liver sections were first treated for heat mediated antigen retrieval with sodium citrate buffer (pH6, epitope retrieval solution 1) for 20 minutes, followed by incubation with anti-RANKL antibody(1:100), anti-RANK antibody (1:150) and anti-OPG antibody (1:200) (Abcam, Cambridge, MA) for 20 minutes at room temperature, then developed with a horse radish peroxidase-conjugated compact polymer system. 3, 3 0 -diaminobenzidine was used as the chromogen. The liver sections were then counterstained with haematoxylin. All sections were visualized using a light microscope (Olympus, Japan) with five fields randomly selected. The numbers of RANKL or OPG positive cells were quantified at high-power field (HPF) (40×10 magnification). The level of RANK expression in each section was scored using a scale of 1 to 4 based on the positive area in each portal tract area: 1, 0-25%; 2, 26-50%; 3, 51-75%; and 4, 76-100%. According to the intensity, the signal level of RANK in the biliary epithelial cells was scored. Weak signal was scored as 1; Mild signal was scored as 2; Strongly positive was scored as 3. Each slide was given an average score based on all of the evaluated portal tracts. A single unbiased "blinded" pathologist analyzed all samples.

Statistical Analysis
All data are reported as mean ± standard error of the mean (SEM). Group means were compared by ANOVA. All the statistical analysis were performed with Prism 5.0 (GraphPad Software La Jolla, CA). A p value of less than 0.05 was considered statistically significant.

Hepatic RANKL Expression Correlates with Stage of Disease in PBC
Importantly, a significant correlation was found between hepatic expression of RANKL and stage in PBC. Indeed, patients with advanced PBC demonstrated a significantly higher expression of RANKL (Fig 4A), whereas expression of RANK and OPG were not associated with stage in PBC (Fig 4B and 4C). Since OPG is the decoy receptor of RANKL, we calculated the hepatic RANKL/OPG ratio, which was found to be significantly higher in late stage PBC compared with early stage (p<0.001) (Fig 4D).

Identification of Cellular Sources of RANKL/RANK/OPG in PBC
We also performed immunofluorescence double staining for CD4, CD8, CD19, CD11c, CD56 and CD68 to identify the cell type(s) expressing RANK, RANKL, and OPG. In PBC, RANKL predominantly co-localized with CD4+, CD8+ and CD19, suggesting that T and B lymphocytes were the primary sources (Fig 5). In contrast, RANK co-localized with CD19 but was not detected on CD4+ or CD8+ T lymphocytes (Fig 6). CD56 and CD68 positive cells were RANK positive (S1 Fig). OPG was expressed by CD4+ T lymphocytes but not the CD8+ T lymphocytes or B lymphocytes (Fig 7). Lymphocyte subsets were also studied in samples from AIH, PSC, and CHB patients, as well as HC, and were found to be not specific, and not related to the inflammation and fibrosis grade (data not shown).

RANK Is Expressed by Damaged Cholangiocytes in PBC
The immunohistology staining demonstrated that RANK was also expressed in biliary epithelial cells, especially the degenerated cholangiocytes of interlobular bile ducts in PBC (Fig 8A). To determine whether RANK(+)ve cholangiocytes were specific for PBC, we determined the level of RANK expression in cholangiocytes from PBC, unaffected controls and chronic liver diseases. Importantly, RANK expression in cholangiocytes was significantly higher in PBC than AIH, CHB or PSC (Fig 8B, p<0.001).

Discussion
There has been an enormous effort in both human and murine systems to identify factors that lead to the development of PBC, both on the cellular and molecular basis [3,[27][28][29][30][31][32][33][34]. Such efforts have not translated into novel therapy or the use of biologics in PBC [35][36][37][38][39][40]. Our data reflects the hepatic expression and cellular localization of RANKL/RANK/OPG in PBC. We report that the expression of RANKL/RANK/OPG were all markedly elevated in the liver of PBC compared to extensive use of controls. Most importantly, the hepatic level of RANKL was associated with disease severity. There is limited information previously published on the role of RANK/RANKL in PBC; most previous data focused only on bone metabolism. Previous studies have suggested that a dysregulated innate immune response is mediated by members of the TNF superfamily and downstream inflammatory signaling pathways in PBC [41,42]. Recently novel members of this superfamily, including the RANKL/ RANK/ OPG molecules, have been found to be a link between inflammatory cells and autoimmunity. Both in vitro and in vivo studies have indicated that RANKL/RANK interaction activates the immune response by promoting the survival of DCs [43,44]. Moreover, the intricate RANKL/RANK/OPG signal pathways potentially mediates self-tolerance and oral-tolerance through regulatory T cell functions [45]. Of particular interest, accumulating evidence suggests the importance of RANKL signaling in the prevention and suppression of autoimmunity. Indeed, studies in murine models reflect that deficiencies in RANKL, RANK or their downstream molecules in the signal pathway is sufficient to induce autoimmunity, resulting in severe disease phenotypes [46][47][48][49]. Activation of RANKL/RANK/OPG system has been implicated in rheumatoid arthritis (RA), Crohn's disease (CD) and ulcerative colitis (UC) [16,50]; moreover RANKL blockade improves hepatic insulin resistance and prevents development of diabetes [51].
SNP analysis has revealed that the RANKL gene is near a novel risk locus for PBC, suggesting a role of RANKL/RANK/OPG system in the pathogenesis of PBC. PBC patients had higher serum OPG levels but lower serum RANKL levels compared to unaffected controls [52]. Our data however demonstrate a high expression of RANKL in liver, suggesting that the level of RANKL in circulation does not reflect the actual level of this protein in the disease site of PBC. As earlier reported, higher serum OPG levels were associated with advanced PBC disease, whereas lower serum RANKL level was associated with low osteocalcin level. However, no significant differences in hepatic OPG mRNA level were found between PBC and controls [53]. This latter study included only a relatively small sample size. We should emphasize that it is known that post transcriptional events play an important role in the overall expression of RANKL/RANK/OPG proteins [54].
Our data also indicate that RANKL is expressed predominantly by CD4+ and CD8+ T lymphocytes and CD19+ B lymphocytes, whereas B lymphocytes expresses RANK, the cognate receptor for RANKL. One of the most important findings in our study is that RANK is also expressed by cholangiocytes. As cholangiocytes are the major victims in PBC, RANKL-RANK interaction may directly affect the pathogenesis of cholangiopathy in PBC. OPG, the decoy receptor for RANKL, was primarily expressed by CD4+ T lymphocytes.
Little is known regarding the role of RANK and RANKL in lymphoid subpopulations. Tumor-infiltrating regulatory T cells have been reported to stimulate mammary cancer metastasis through RANKL-RANK signalling [55]; and RANKL-positive expression colocalized with CD4+ and CD8+ T lymphocytes in acutely rejected kidney tissue in an animal model of kidney transplantation [56]. To our knowledge, deficiencies in RANK/RANKL have not been investigated in PBC murine models. Based on our results, we postulate that the RANK/RANKL axis in PBC potentiates the inflammatory response against cholangiocytes. Indeed, damaged cholangiocytes in PBC express high levels of RANK which leads to the recruitment of RANKL positive T cells and exacerbates the immune attack against biliary cells. On the other hand, the presence of RANKL would recruit OPG, which could sequester the immunoregulatory activity of RANKL and promote the infiltration of inflammatory cells in the portal tract (Fig 9).
Our study has some weaknesses that should be discussed. First, the number of unaffected subjects included is limited. These extremely difficult to obtain samples derive from subjects undergoing both liver surgery for benign lesions and liver donors. In order to address the the lack of unaffected subjects in our study we included a sufficient number of subject with noncholestatic liver disease (CHB). Second, functional data are missing; indeed, further studies should investigate RANKL/RANK/OPG signaling in PBC murine model.
In conclusion, our findings provide the first evidence for an active RANKL/RANK/OPG signaling in PBC, a potential target of immunotherapy. If so, translation to the clinic may be feasible given current pharmacological strategies to lower RANKL activity to treat osteoporosis.