Article Text

Download PDFPDF

Monocyte chemoattractant protein-1 derived from biliary innate immunity contributes to hepatic fibrogenesis
  1. Kenichi Harada1,
  2. Mayumi Chiba1,
  3. Atsushi Okamura1,
  4. Maylee Hsu1,
  5. Yasunori Sato1,
  6. Saya Igarashi1,
  7. Xiang Shan Ren1,
  8. Hiroko Ikeda2,
  9. Hajime Ohta3,
  10. Satomi Kasashima4,
  11. Atsuhiro Kawashima4,
  12. Yasuni Nakanuma1
  1. 1Department of Human Pathology, Kanazawa University Graduate School of Medicine, Kanazawa, Japan
  2. 2Division of Pathology, Kanazawa University Hospital, Kanazawa, Japan
  3. 3Department of Gastroenterology, National Hospital Organization, Kanazawa Medical Center, Kanazawa, Japan
  4. 4Department of Pathology, National Hospital Organization, Kanazawa Medical Center, Kanazawa, Japan
  1. Correspondence to Dr Kenichi Harada, Department of Human Pathology, Kanazawa University Graduate School of Medicine, Kanazawa 920-8640, Japan; kenichih{at}


Aims Monocyte chemoattractant protein-1 (MCP-1) is a major chemotactic factor for hepatic stellate cells (HSCs) associated with hepatic fibrosis. In this study, among several fibrogenetic factors derived from biliary epithelial cells (BECs), MCP-1 produced by the biliary innate immune system was found to be most critical in the histogenesis of hepatic fibrogenesis.

Methods Using cultured human BECs, the expression of five fibrogenetic factors including MCP-1 on stimulation with Toll-like receptor ligands, inflammatory cytokines or bile acids was examined. Moreover, in situ detection of MCP-1 and α-smooth muscle actin proteins was performed using sections from normal and diseased livers by immunohistochemistry.

Results All fibrogenetic factors were detected in BECs, but only MCP-1 expression was upregulated, by all the Toll-like receptor ligands, IL-1β, and tumour necrosis factor-alpha. Proliferating bile ductules in interface areas expressed MCP-1 in diseased livers accompanying α-smooth muscle actin-positive activated HSCs.

Conclusions Bile ductules proliferate in various hepatobiliary diseases, and its significance is still unknown. This study demonstrated that BECs in bile ductules could produce MCP-1, particularly, via biliary innate immunity, suggesting that MCP-1 derived from BECs plays an important role in the recruitment of HSCs to interface areas and the activation of HSCs resulting in the progression of periportal fibrosis.

  • Hepatic stellate cell
  • hepatic fibrosis
  • biliary epithelial cells
  • Toll-like receptor
  • MCP-1
  • biliary innate immunity
  • colorectal cancer
  • gall bladder
  • oncogenes
  • p53
  • pancreas
  • biliary
  • liver
  • liver disease
  • liver cancer
  • histopathology
View Full Text

Statistics from


Hepatic fibrosis is a major feature of advanced liver diseases and defines the prognosis. Hepatic fibrosis spreads within portal tracts and also periportal areas in patients with hepatitic and cholestatic liver diseases. Although periportal fibrosis is thought to be associated with the accumulation and activation of hepatic stellate cells (HSCs), its mechanisms are not fully understood. HSCs undergo differentiation towards an activated phenotype, and this process is enhanced by soluble mediators such as platelet-derived growth factor (PDGF) and transforming growth factor-β (TGF-β), which mediate the increase in cell proliferation and extracellular matrix production.1 2 Then, HSCs migrate into damaged areas in response to a chemokine, secreting monocyte chemotactic protein-1 (MCP-1/CCL2) that recruits monocytes and lymphocytes.3 Moreover, MCP-1 triggers the migration of activated, but not quiescent, HSCs in a dose-dependent manner. Experiments in vitro using HSCs isolated from normal human livers revealed that MCP-1-dependent signals were not transduced by the chemokine receptor of MCP-1, CCR2, and may be mediated by alternative chemokine receptors.3 Recently, Ramm et al4 reported that hepatocyte-derived MCP-1 induced by a hydrophobic bile acid, taurocholate, could result in the recruitment of HSCs in cholestatic liver injury as the major fibrogenesis in a paediatric cholestatic liver disease such as biliary atresia. In contrast, biliary epithelial cells (BECs) may promote fibrogenesis by a number of mechanisms including the synthesis of matrix constituents and the release of mediators such as MCP-1, PDGF-BB, TGF-β, connective tissue growth factor (CTGF) and endothelin-1.5 The role of BECs in the pathogenesis of hepatic fibrosis, particularly periportal fibrosis accompanying interface hepatitis in various hepatobiliary diseases including chronic viral hepatitis (CVH) and primary biliary cirrhosis (PBC), is speculated to be important to the disease's progression, but its precise mechanism is still unknown.

Our previous study demonstrated that human BECs possess several inflammatory cytokine receptors and Toll-like receptors (TLRs) and produced cytokines and chemokines in response to inflammatory cytokines and pathogen-associated molecular patterns (PAMPs), respectively.6–9 In particular, bile ductules located between interlobular bile ducts in portal tracts and bile canaliculi in hepatocytes are frequently increased in number under a variety of pathological conditions of the liver and take part in various immunological responses and the pathogenesis of biliary diseases.10–12 This ductular reaction is speculated to be closely associated with hepatic fibrosis, particularly periportal fibrosis accompanying interface hepatitis, but its precise mechanism is still unknown. In this study, we examined the possibility that BECs produce MCP-1 and play a role in fibrogenesis in periportal areas via MCP-1 production.

Materials and methods

Cultured human BECs

Two human intrahepatic BEC lines were established from explanted livers with PBC as described previously.7 These cell lines had been confirmed to be BECs by the expression of biliary-type cytokeratins.


Cultured BECs were stimulated with inflammatory cytokines, PAMPs and bile acids (table 1). As an NF-κB inhibitor, isohelenin (30 μM, Calbiochem, Darmstadt, Germany) was added to the culture medium before the stimulation. We previously confirmed that BECs possess the receptors for all these cytokines and PAMPs.6 Cell samples for the examination of mRNA as shown below were prepared 2 h after the stimulation.

Table 1


RT-PCR and real-time PCR

For the evaluation of mRNAs of MCP-1, PDGF-B, CTGF, TGF-β1, endothelin-1 and glyceraldehyde 3-phosphate dehydrogenase (internal control) in cultured BECs, total RNA was isolated from BECs, and RT-PCR and real-time quantitative PCR were performed according to a standard protocol using specific primers (table 2).

Table 2

Primers used for RT-PCR and quantitative PCR


Cultured BECs were stimulated with lipopolysaccharide, Pam3CSK4, poly(I:C) and IL-1β for 24 h, and supernatants were tested for human MCP-1 by ELISA (Biosource International, Camarillo, California).

Liver tissues

A total of 26 surgical or wedge liver biopsy specimens were obtained from patients with CVH (n=8, hepatitis C virus-related, male/female=5/3, average age 58-year-old., F1/F2=4/4), PBC (n=5, all female, average age 68-year-old, Nakanuma's classification13 Stage 2/3=4/1, Scheuer's classification Stage 1/2=2/3) and congenital hepatic fibrosis (CHF, n=5, used as a case of activated HSC-poor case14, male/female 4/1, average age 23-year-old), and six cases with no significant histopathological change (‘normal liver’).


Deparaffinised sections were pretreated in Target Retrieval Solution (Dako, Tokyo, Japan) and incubated with a primary antibody against MCP-1 (10 μg/ml; Abcam, Tokyo, Japan) or α-smooth muscle actin (αSMA, activated HSC marker) (1 μg/ml, Dako), and then Envision-HRP (Dako) was used. No positive staining was obtained when the primary antibody was replaced with an isotype-matched, non-immunised immunoglobulin.

The simultaneous detection of combined CCR2 and αSMA, and CCR2 and CD68, was evaluated by double fluorescence immunohistochemical staining. After incubation with the antibodies for these combinations, rabbit antibody (CCR2, 1:400, Epitomics, Burlingame, California) and mouse antibodies (αSMA and CD68, Dako) were visualised by Alexa Fluor 488 and 594 (Molecular Probe, Eugene, Oregon), respectively.


Detection of MCP-1 in cultured human BECs

RT-PCR revealed that all fibrogenic cytokines were detected in human BECs in the basal cultures (figure 1). Quantitative PCR revealed that only MCP-1 was upregulated following stimulation with all PAMPs and two cytokines (IL-1β and tumour necrosis factor-alpha (TNF-α); figure 2). Moreover, ELISA demonstrated that the concentration of MCP-1 protein in the supernatants was increased by these stimulants (figure 3). None of the bile acids upregulated the expression of MCP-1 (figure 2). The Pam3CSK4-induced upregulation of MCP-1 mRNA was significantly inhibited by pretreatment with the NF-κB inhibitor (figure 2).

Figure 1

Detection of fibrogenic cytokines in cultured human BECs. The two human biliary epithelial cell lines (BEC1 and BEC2) express all mRNAs of fibrogenic cytokines (monocyte chemoattractant protein-1 (MCP-1), platelet-derived growth factor (PDGF)-B, connective tissue growth factor (CTGF), transforming growth factor-β (TGF-β)1, and endothelin-1 (GAPDH)) and the internal control (glyceraldehyde 3-phosphate dehydrogenase) under basal conditions. Each PCR product showed the predicted size as a single band. Negative controls (NC) were obtained by replacing reverse transcriptase with RNase- and DNase-free water.

Figure 2

Quantitative PCR analysis for mRNA of fibrogenic cytokines (monocyte chemoattractant protein-1 (MCP-1), platelet-derived growth factor (PDGF)-B, connective tissue growth factor (CTGF), transforming growth factor-β (TGF-β)1, and endothelin-1). MCP-1 expression alone was upregulated by stimulation with Toll-like receptor (TLR) ligands (lipopolysaccharide (LPS), Pam3CSK4, and poly(I:C)) and cytokines (interleukin-1β (IL-1β) and tumour necrosis factor-alpha (TNF-α)). No bile acids (taurochenodeoxycholic acid (TCDC), taurocholic acid (TC), taurodeoxycholic acid (TDC) or tauroursodeoxycholic acid (TUDC)) significantly upregulated MCP-1 mRNA expression. Other fibrogenic cytokines were not affected by any stimulants. A lower figure denotes that Pam3CSK4-induced upregulation of MCP-1 mRNA expression was inhibited by pretreatment with an NF-κB inhibitor, isohelenin. Bars indicate the mean±SEM. *<0.05. INF-γ, interferon-γ.

Figure 3

Measurement of monocyte chemoattractant protein-1 (MCP-1) protein by ELISA. After stimulation with lipopolysaccharide (LPS), Pam3CSK4 (Pam3), poly(I:C), and IL-1β for 24 h, the concentration of MCP-1 protein in cultured supernatants of human biliary epithelial cells was significantly increased. Bars indicate the mean±SEM. *<0.05.

In situ detection of MCP-1 and activated HSCs

MCP-1 was mainly expressed in bile ducts, and the majority of proliferating bile ductules predominantly expressed MCP-1 in periportal and interface areas in cases of CVH and PBC (figure 4). Moreover, αSMA-positive cells (activated HSCs) were scattered around these bile ductules (figure 4). These characteristic findings were made in CVH and PBC accompanying interface hepatitis and ductular reaction, but rare or absent in normal livers. Periportal hepatocytes were also frequently positive for MCP-1, but the intensity was similar to or less than that in bile ductules. In contrast, MCP-1-positive bile ductules were not found in CHF. Moreover, the expression of αSMA was limited in portal tracts, and αSMA-positive HSCs were not found in CHF.

Figure 4

Immunohistochemistry for monocyte chemoattractant protein-1 (MCP-1) (A, C, E) and α-smooth muscle actin (αSMA) (B, D, F). (A, B) Hepatitis C virus-related chronic viral hepatitis. Proliferating bile ductules at the interface of periportal areas express MCP-1 (A, arrows) and αSMA-positive cells showing hepatic stellate cell (HSC) morphology are found in the same area (B, arrows). In contrast, αSMA-positive HSCs are not found in periportal parenchyma lacking bile ductules (asterisks). (C, D) Primary biliary cirrhosis. MCP-1-positive proliferating bile ductules and αSMA-positive cells are intermingled (arrows). (E, F) Congenital hepatic fibrosis. No biliary bile ducts or ductules express MCP-1, and no αSMA-positive HSCs are found in periportal areas (asterisks). The expression of αSMA is limited in portal tracts (lower right).

Characterisation of CCR2-expressing cells

Fluorescence double immunostaining revealed that cells double positive for CD68 and CCR2 were scattered mostly in interface areas of CVH and PBC, but cells double positive for αSMA and CCR2 were not found in any area (figure 5).

Figure 5

Double immunohistochemistry using chronic viral hepatitis livers. (A, B, C) Monocyte chemoattractant protein-1 receptor (CCR2, green), α-smooth muscle actin (red), and merged image, respectively. Double positive cells indicating CCR2-positive hepatic stellate cells (yellow) are not found. (D, E, F) CCR2 (green), CD68 (red) and merged image, respectively. Double positive cells indicating CCR2-positive Kupffer cells (yellow) are scattered (arrows).


Chronic liver injury is generally characterised by inflammation and subsequent fibrosis, and, regardless of the aetiology, a cytokine-rich environment caused by inflammation and infection is closely associated with hepatic fibrosis. HSC is considered the most important effecter cell associated with fibrogenesis in hepatic parenchyma including the interface between portal tracts and periportal hepatocytes, and the fibrous enlargement of portal tracts and fibrous extension from portal areas are closely associated with activated HSCs and their transformed version, myofibroblasts. Within portal tracts, in contrast, fibroblasts and fibrocytes are also important to the histogenesis of portal fibrosis. Several cytokines such as PDGF-B, CTGF, TGF-β, endothelin-1 and MCP-1, are reported as fibrogenetic factors in liver.5 These cytokines could chemoattract HSCs or directly activate the production of collagen in myofibroblasts.

We have previously reported that human BECs possess receptors for inflammatory cytokines (IL-1β, IL-6, Interferon-γ and TNF-α) and an innate immune system consisting of TLRs.6 8 9 Therefore, in this study, we examined whether BECs contribute to hepatic fibrosis using cultured human BECs. Consequently, although mRNAs of all fibrogenic cytokines were detected in BECs, only the expression of MCP-1 was upregulated, by two proinflammatory cytokines (by IL-1β and, to a lesser degree, by TNF-α) and all TLR ligands in an NF-κB-dependent manner. Because human BECs produced IL-1β when stimulated with TLR ligands,15 the biliary innate immune response is suggested to be a critical trigger of MCP-1 production.

Bile ductules and their proliferation occur non-specifically in various hepatobiliary diseases. In normal livers, a few bile ductules are recognisable in the portal tract, while in various hepatobiliary diseases, these ductular structures are often increased in number, to be termed ‘ductular proliferation’ or ‘proliferating bile ductules.’ Because ductular proliferation was found accompanying interface hepatitis and periductal fibrosis in chronic liver diseases, these biliary elements are thought to take part in disease progression. However, the exact association between ductular proliferation and hepatic fibrogenesis is still unknown. Immunohistochemistry revealed that MCP-1 was expressed in bile ductules in areas of interface hepatitis, whereas normal livers lacked these findings. The ductular reaction predominantly corresponded to interface hepatitis in various chronic hepatobiliary diseases, and these areas were rich in several cytokines caused by immune-mediated (necro)inflammatory reactions against virus-infected hepatocytes and bile-derived PAMPs. Therefore, these microenvironments are suitable for the production of MCP-1 in proliferating bile ductules at the interface.

In damaged liver, HSCs are activated, proliferate and migrate into the injured area in response to the chemoattractive effects of chemokines. MCP-1 is a chemokine attracting monocyte/macrophages and plays a role in persistent inflammation in chronic liver diseases. Marra et al3 reported that MCP-1 attracts HSCs, particularly activated HSCs, to the liver. Innate immunity is known to promote liver fibrosis, and as its mechanism, HSCs are reported to produce MCP-1 via TLR9 signalling.16 In this study, immunohistochemistry using liver sections from patients with CVH and PBC revealed that αSMA-positive activated HSCs (myofibroblasts) were scattered and accumulated around MCP-1-expressing bile ductules in areas showing interface hepatitis. This finding suggests that MCP-1 derived from BECs plays a role in the chemoattraction of HSCs.

A recent report demonstrated that hepatocytes are also a source of MCP-1 and that hepatocyte-derived MCP-1 induced by a hydrophobic bile acid, taurocholate, chemoattracts HSCs and is associated with the liver fibrosis under cholestatic conditions in cases of paediatric cholestatic liver disease such as biliary atresia. Our immunohistochemical study also confirmed the expression of MCP-1 in periportal hepatocytes, but, in the early stage of CVH and PBC, the MCP-1 derived from bile ductules is speculated to the major effecter, compared with that from hepatocytes, based on the intensity of MCP-1 immunostaining. We also examined the effect of bile acids on the production of MCP-1 in cultured BECs. No bile acids affected the MCP-1 expression in BECs, suggesting that cholestasis could not directly induce the production of MCP-1 in BECs, differing from hepatocytes.

Portal fibroblasts located in portal tracts are fibrogenic cells distinct from HSCs and may be important mediators of biliary fibrosis and cirrhosis. Recently, Kruglov et al17 reported that portal fibroblasts express functional receptors for MCP-1 that are distinct from CCR2 and that the secretion of MCP-1 by BECs induces myofibroblastic transdifferentiation of portal fibroblasts. In fact, the expression of MCP-1 was found in some interlobular bile ducts as well as bile ductules within portal tracts in CVH and PBC (data not shown). This finding suggests that MCP-1-dervied from BECs of these interlobular bile ducts is associated with the migration and activation of portal fibroblasts. However, we speculate that the fibrogenesis associated with portal fibroblasts is mainly associated with the histogenesis of portal sclerosis and expansive fibrous enlargement of portal tracts, not the fibrous extension accompanying interface hepatitis from portal tracts. The close correlation between MCP-1-positive bile ductules and αSMA-positive activated HSCs in the interface area shown in this study supports our contention. Moreover, we examined CHF as a control diseased liver in this study. CHF is different from cirrhosis in which no abnormal biliary channels are seen. In portal tracts of CHF, irregular and newly proliferating bile ducts and ductules rather than congenitally abnormal ductal plates, and cholestasis in these bile ductules are found. However, in parenchyma, the features of chronic cholestasis and interface changes are not prominent. Our previous study reported that the fibrogenesis of CHF is associated with intraportal heparan sulfate proteoglycan and CTGF, but not periportal HSCs.14 It is true that although proliferating bile ductules were scattered within portal tracts, MCP-1 expression in bile ductules and αSMA-positive HSCs were not found in CHF, suggesting that the MCP-1-mediated migration of HSCs limited in proliferating bile ductules accompanying interface hepatitis of chronic inflammatory hepatobiliary diseases, but not associated with the fibrogenesis in sole cholestatic liver diseases lacking interface changes such as CHF.

Because the receptor of MCP-1, CCR2, is not expressed in human HSCs, HSC migration by MCP-1 occurs independent of CCR2 via an unknown receptor for MCP-1, instead of by CCR2.3 In fact, double immunohistochemistry in this study revealed that the expression of CCR2 is found not in αSMA-positive HSCs, but in CD68-positive Kupffer cells, suggesting that MCP-1-dervied from BECs could chemoattract Kupffer cells. As mentioned, portal fibroblasts within portal tracts also lack CCR2, but could be attracted by MCP-1 derived from BECs in a CCR2-independent manner.17 In contrast, Seki et al18 reported that both Kupffer cells and HSCs express CCR2 in mice but that these differences might be explained by differences between humans and mice.

Various types of inflammation including infection-triggered inflammation are causative factors to induce hepatic fibrosis in chronic liver diseases. Particularly, HSCs are critical for hepatic fibrogenesis, and MCP-1 is an important cytokine associated with HSC migration in fibrogenic areas. Therefore, the identification of MCP-1-producing cells and the clarification of the mechanism of MCP-1 production are mandatory to help regulate hepatic fibrosis and treat liver fibrosis. This study revealed that BECs are a source of MCP-1 in some hepatobiliary diseases, and the production of MCP-1 is induced by inflammatory cytokines and biliary innate immune responses. Proliferating bile ductules are thought to be part of a non-specific reaction in various hepatobiliary diseases, but this study suggests that they are closely associated with the progression of periportal fibrosis via MCP-1 derived from biliary innate immunity. Because MCP-1 is thought to be a key mediator of hepatic fibrosis, it is a potential therapeutic target in inflammatory hepatobiliary diseases with hepatic fibrosis.

Take-home messages

  • Biliary epithelial cells express several fibrogenic cytokines (MCP-1, PDGF-B, CTGF, TGF-β1 and endothelin-1), but only MCP-1 expression is upregulated by biliary innate immune reaction and proinflammatory cytokines (IL-1β and TNF-α).

  • Proliferating bile ductules in interface areas expressed MCP-1 in diseased livers accompanying αSMA-positive activated hepatic stellate cells.

  • MCP-1 derived from biliary epithelial cells plays an important role in the recruitment of hepatic stellate cells to interface areas and the activation of HSCs resulting in the progression of periportal fibrosis.


View Abstract


  • Funding This work was supported by Grants-in-aid for Scientific Research from the Ministry of Health, Labour and Welfare of Japan and Grants-in-Aid for Scientific Research (C) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was approved by the Kanazawa University Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.