If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
School of Biomedical Sciences, CHIRI Biosciences, Curtin University, Bentley, Western Australia, AustraliaSchool of Medicine and Pharmacology, University of Western Australia, Fremantle, Western Australia, Australia
School of Biomedical Sciences, CHIRI Biosciences, Curtin University, Bentley, Western Australia, AustraliaDepartment of Gastroenterology & Hepatology, Fiona Stanley and Fremantle Hospitals, Perth, Western Australia, AustraliaFaculty of Health Sciences, Edith Cowan University, Perth, Western Australia, Australia
Hepatic Fibrosis Group, QIMR Berghofer Medical Research Institute, Brisbane, Queensland, AustraliaFaculty of Medicine, The University of Queensland, Brisbane, Queensland, AustraliaDepartment of Gastroenterology, The University of Queensland, Brisbane, Queensland, Australia
Cystic fibrosis liver disease (CFLD) in children causes progressive fibrosis leading to biliary cirrhosis; however, its cause(s) and early pathogenesis are unclear. We hypothesized that a bile acid–induced ductular reaction (DR) drives fibrogenesis. The DR was evaluated by cytokeratin-7 immunohistochemistry in liver biopsies, staged for fibrosis, from 60 children with CFLD, and it demonstrated that the DR was significantly correlated with hepatic fibrosis stage and biliary taurocholate levels. To examine the mechanisms involved in DR induction, liver progenitor cells (LPCs) were treated with taurocholate, and key events in DR evolution were assessed: LPC proliferation, LPC biliary differentiation, and hepatic stellate cell (HSC) chemotaxis. Taurocholate induced a time-dependent increase in LPC proliferation and expression of genes associated with cholangiocyte differentiation (cytokeratin 19, connexin 43, integrin β4, and γ-glutamyltranspeptidase), whereas the hepatocyte specification marker HNF4α was suppressed. Functional cholangiocyte differentiation was demonstrated via increased acetylated α-tubulin and SOX9 proteins, the number of primary cilia+ LPCs, and increased active γ-glutamyltranspeptidase enzyme secretion. Taurocholate induced LPCs to release MCP-1, MIP1α, and RANTES into conditioned medium causing HSC chemotaxis, which was inhibited by anti-MIP1α. Immunofluorescence confirmed chemokine expression localized to CK7+ DR and LPCs in CFLD liver biopsies. This study suggests that taurocholate is involved in initiating functional LPC biliary differentiation and the development of the DR, with subsequent induction of chemokines that drive HSC recruitment in CFLD.
Cystic fibrosis–associated liver disease (CFLD) is characterized by indolent intrahepatic cholestasis that usually manifests by early adolescence.
Cystic fibrosis (CF) is caused by a defect in the gene coding for the cystic fibrosis transmembrane conductance regulator (CFTR) which, among other functions, maintains the flow of bile in contact with cholangiocytes by facilitating the movement of water and bicarbonate ions into bile. Abnormal ion/water transport leads to inspissated bile and cholestasis with focal fibrosing destruction of intrahepatic bile ducts.
The resulting pattern of injury is focal biliary cirrhosis for which there is no known treatment and which may progress to cause portal hypertension, multilobular biliary cirrhosis, and liver failure. However, only 10% of children with CF will develop portal hypertension and cirrhosis, a higher prevalence than the overall CF population, indicating a strong survival disadvantage. The cause and mechanism of this variable presentation and severity of pathology are not known.
The ductular reaction (DR) is characterized by the expansion of a population of bipotential liver progenitor cells (LPCs) which can differentiate toward the cholangiocyte lineage into reactive bile ductules, as well as intermediate hepatocytes.
The DR is observed along with activation of hepatic stellate cells (HSCs) and an increase in hepatic fibrosis, which is seen in adult liver diseases that show periportal damage in cholestatic liver diseases,
This has led to the hypothesis that the two processes of LPC expansion and HSC activation with the resultant DR and fibrosis, respectively, are closely related. Although histological evidence of a DR has previously been demonstrated in children with extrahepatic biliary atresia,
the potential role of LPCs driving HSC activation and fibrosis via development of the DR has not previously been assessed in CFLD.
Bile acids are synthesized in hepatocytes from cholesterol and are usually conjugated to taurine or glycine to form bile salts and secreted into the bile canaliculi. These bile secretions flow into bile ducts, where they are stabilized by bicarbonate ions and water secreted by cholangiocytes as they are excreted from the liver into the gall bladder. They aid lipid digestion in the duodenum by forming micelles. However, their partial hydrophobic nature, which is essential in forming micelles, also makes them toxic to the liver in higher concentrations.
We previously demonstrated that taurocholate induced the expression of monocyte chemotaxis protein-1 (MCP-1) in normal hepatocytes, and that MCP-1 produced by cholestatic hepatocytes from bile duct-ligated rats induced the chemotaxis of HSCs.
In this study, we investigated whether the DR is present in the liver of children with CFLD and examined mechanisms associated with its induction and role in fibrogenesis. We compared the severity of the DR in CFLD liver biopsies with those from other, more aggressively cholestatic pediatric liver disease and then specifically examined a potential mechanism of DR induction in CFLD through investigating the relationship between biliary taurocholate concentrations and the DR in liver biopsies from children with CFLD in vivo. We then used in vitro cell culture models and ex vivo liver slices to investigate the in vivo observed effects of elevated taurocholate on LPC differentiation, the secretion of chemokines by LPCs, and their chemotactic effects on HSCs in driving hepatic fibrogenesis.
Materials and Methods
This study is part of a long-term prospective cohort study of patients with CFLD, identified and referred from the CF Clinic, Royal Children's Hospital, Brisbane, that is previously well documented.
CFLD was defined as two of the following: i) hepatomegaly with or without splenomegaly; ii) persistent (>6 months) elevation of serum alanine aminotransferase (>1.5× upper limit normal); and iii) abnormal liver ultrasound (abnormal echogenicity or nodular edge). Those with liver synthetic dysfunction or a history of hepatobiliary surgery were excluded. The study conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the ethics committees of the Royal Children's Hospital, Brisbane, and the QIMR Berghofer Medical Research Institute. Informed consent was obtained from parents or patients.
Liver tissue was available from 60 children with CFLD (age, 11.8 ± 4.5 years). For comparison with other pediatric cholestatic liver diseases, liver tissue was also available from three infants with biliary atresia (age, 1.2 ± 1.4 years), five with idiopathic neonatal hepatitis (age, 3.4 ± 3.6 years), as well as four pediatric donor liver controls (age, 6.8 ± 4.5 years). Bile was available from 7 children with CFLD.
Formalin-fixed, paraffin-embedded liver sections were subjected to heat antigen retrieval, before exposure to 3% hydrogen peroxide in phosphate-buffered saline and Background Sniper (Biocare Medical, Pacheco, CA). Sections were then stained with primary antibody: mouse anti-human CK7 monoclonal antibody (dilution 1:100; Dako, Glostrup, Denmark), secondary antibody: rabbit anti-mouse biotinylated IgG (dilution 1:400, Dako), with detection using a streptavidin-biotin/horseradish peroxidase kit (Dako) and chromogenic substrate 3,3′-diaminobenzidine tetrahydrochloride (Sigma-Aldrich, St. Louis, MO), and counterstained with hematoxylin. Digital images were captured using Aperio ScanScope XT slide scanner (Leica Biosystems, Buffalo Grove, IL) under ×20 magnification at 0.5 μm/pixel. Image analysis was performed using Aperio Positive Pixel Count algorithm software version 9.1. Positively stained cells were counted, normalized to the total section area, and results correlated with Scheuer's hepatic fibrosis staging.
Biliary Taurocholate Measurement
Taurocholate concentration was measured in endoscopic bile samples available from seven patients with CFLD, as previously described.
Liver biopsies were processed for histochemical CK7 quantification as above. All patients had histological evidence of hepatic fibrosis. Biliary taurocholate levels were correlated with CK7 quantification.
was cultured in Williams' E medium containing 2 mmol/L glutamine, penicillin (100 U/mL), streptomycin (100 μg/mL), insulin (10 μg/mL), epidermal growth factor (10 ng/mL), insulin-like growth factor II (30 ng/mL), and 10% fetal bovine serum (FBS). The human HSC line LX-2 was cultured in Dulbecco's modified Eagle's medium with 2 mmol/L glutamine, penicillin (100 U/mL), streptomycin (100 μg/mL), and 2% FBS. PIL-2 cells were treated with taurocholate at 0 to 300 μmol/L concentration to establish a dose-response, with an optimal concentration of 150 μmol/L taurocholate chosen for all in vitro and ex vivo experiments, as previously described.
For differentiation assays, cells were lysed 1 to 4 days after taurocholate treatment. LPC differentiation was assessed using real-time quantitative PCR (qPCR) and/or Western blot analysis evaluating markers for cholangiocyte (connexin 43, Cx43; cytokeratin 19, Ck19; integrin β 4, Itgb4; γ-glutamyl transpeptidase 1, Ggt1; epithelial cell adhesion molecule, EpCam; SYR-related HMG box transcription factor 9, Sox9) and hepatocyte (albumin; hepatocyte nuclear factor 4α, Hnf4α) phenotype.
qPCR for LPC Differentiation Marker and Chemokine mRNA Expression
RNA was extracted from both untreated and taurocholate-treated PIL-2 cells using the RNeasy kit (Qiagen, Hilden, Germany), and 1 μg was reverse transcribed into cDNA using SensiFast (Bioline, Taunton, MA). qPCR was performed on a Bio-Rad CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA) with HPRT used as the reference gene. Primers are listed in Table 1.
Absorbance was measured at 570 nm with a 620-nm reference filter.
Western Blot Analysis
Cell lysates were prepared from PIL-2 cells disrupted in RIPA buffer with Complete EDTA-free protease inhibitors (Roche, Basel, Switzerland). Lysates were resolved on 10% SDS-PAGE gels as described in Laemmli et al
and transferred to polyvinyl difluoride membranes (Immobilon-FL; Merck Millipore, Billerica, MA). Membranes were probed with anti-CK19 (ab15463; dilution 1:2000; Abcam, Cambridge, MA); anti-SOX9 (AB5535; dilution 1:1000; Merck Millipore); anti-HNF4α (Santa Cruz sc6556; dilution 1:1000; Santa Cruz Biotechnology, Dallas, TX); anti–acetylated α-tubulin (T7451; dilution 1:1000; Sigma-Aldrich); with anti–β-actin used to detect the β-actin loading control (dilution 1:2000; Cell Signaling Technology, Danvers, MA), in TBST LI-COR Odyssey blocking buffer (LI-COR, Lincoln, NE). Blots were probed with an anti-mouse 680RD or anti-rabbit 800CW secondary at dilution 1:20,000 in TBST LI-COR Odyssey blocking buffer. Secondary localization was detected with the LI-COR Odyssey CLx Infrared Imaging System. Band quantification was performed using Image Studio Lite version 5.2 (LI-COR).
To examine primary cilia, PIL-2 cells were grown on glass coverslips to 60% to 80% confluence and treated for 3 days with and without 150 μmol/L taurocholate, and were then serum-starved overnight before fixation. Cells were placed at 4°C for 30 minutes to destabilize microtubules and then fixed in 4% paraformaldehyde for 10 minutes. Cells were permeabilized with 0.5% Triton X-100, blocked with 5% bovine serum albumin, and probed with mouse anti–acetylated α-tubulin (dilution 1:600; Sigma-Aldrich) and rabbit anti–γ-tubulin (dilution 1:400; Sigma-Aldrich) antibodies overnight in phosphate-buffered saline with 0.05% Tween 20 with 1% bovine serum albumin. Primary binding was probed with AlexaFluor 594 donkey anti-rabbit (Invitrogen, Carlsbad, CA) and AlexaFluor 594 goat anti-mouse (Invitrogen) at dilution 1:500 for 1 hour in phosphate-buffered saline with 0.05% Tween 20 with 1% bovine serum albumin. Images were captured with a Zeiss AxioScop2 (Carl Zeiss, Oberkochen, Germany). Quantification of acetylated α-tubulin spots was performed using Particle Analysis in ImageJ software version 1.50i (NIH, Bethesda, MD; http://imagej.nih.gov/ij).
To identify the cellular source of HSC chemokines in liver biopsies from children with CFLD, we performed dual immunofluorescence for chemokines MCP-1; regulated upon activation, normal T cell expressed and secreted (RANTES); or macrophage inflammatory protein 1α (MIP1α), together with the LPC/DR marker, CK7. Primary antibodies used were: mouse anti-CK7 monoclonal antibody (dilution 1:150; Abcam), goat anti-human MCP-1 polyclonal antibody (dilution 1:8; R&D Systems, Minneapolis, MN), goat anti-human CCL5/RANTES polyclonal antibody (dilution 1:83; R&D Systems), goat anti-human MIP1α (dilution 1:50; R&D Systems), incubated overnight at 4°C followed by a 30-minute incubation step at room temperature with donkey anti-mouse IgG AlexaFluor 555 (dilution 1:300; Invitrogen) and donkey anti-goat IgG AlexaFluor 488 (dilution 1:300; Molecular Probes, Eugene, OR). Sections were mounted in Prolong Gold (Invitrogen) with DAPI and examined by confocal microscopy (Zeiss 780NLO).
Ex Vivo Liver Slice Studies
Precision-cut liver slices were made as previously described.
A minimum of three liver slices each from three separate C57BL6 WT mice for each time point and each treatment group, were incubated in Williams' E medium containing 2 mmol/L glutamine, penicillin (100 U/mL), streptomycin (100 μg/mL), and 10% FBS at 37°C for 2 and 4 days with and without 150 μmol/L taurocholate. Conditioned medium was examined for GGT enzyme activity. Slices were harvested for mRNA quantitation using qPCR.
γ-Glutamyl Transpeptidase Activity Assay
GGT activity was determined from in vitro cell cultures and ex vivo liver slice conditioned media using the GGT activity colorimetric assay kit (Sigma-Aldrich). Protein for normalization was measured using the Pierce BCA protein assay (Thermo Fisher Scientific, Waltham, MA).
Quantification of Cytokine Protein Secretion from LPCs with or without Taurocholate
Chemokines MCP-1, MIP1α, and RANTES, released from LPCs, were quantified using the Cytokine Bead Array (BD Biosciences, San Jose, CA), according per manufacturer's protocol. The concentrations of transforming growth factor-1β (TGF-β1) and platelet-derived growth factor-BB (PDGFB) secreted by LPCs were determined by enzyme-linked immunosorbent assay (R&D Systems).
Induction of HSC chemotaxis via taurocholate treatment of LPCs was examined by measuring cell migration through 8-μm 24-well cell culture inserts (Greiner Bio-One, Kremsmünster, Austria) in modified Boyden chambers. Assays were performed using 0.5 × 106 LX-2 cells (serum starved overnight) per mL within the insert, and the inserts were placed suspended within test medium. Test medium was either normal nonconditioned serum-free Williams' E medium or conditioned medium from LPCs with and without taurocholate. Chemokine neutralization was performed by pretreating conditioned medium with 2.5 μg/mL anti-mouse MCP-1 (AF-479-NA), 1 μg/mL anti-mouse MIP1α (AF-450-NA), or 1 μg/mL anti-mouse RANTES (AF478) polyclonal neutralization antibodies (R&D Systems), 1 hour before the assay. Testing of MIP1α as chemokine used murine (40-337; ProSci, Poway, CA) or human (40-364; ProSci) recombinant purified MIP1α proteins. Blocking of the CCR5 receptor used the neutralization/blocking antibody MAB182 (R&D Systems) applied to the LX-2 cells within the Boyden chamber insert for 1 hour before the assay at an insert concentration of 20 μg/mL. Chemotaxis assay were performed for 24 hours, and migrated cells on the underside of the insert membrane were then stained with Cell Stain (90144; Merck Millipore), treated with Extraction Buffer (90145; Merck Millipore) and absorbance measured at 560 nm. Results from separate experimental replicates were normalized to the mean absorbance of control samples (nonconditioned serum-free assay medium and FBS positive control).
Results are presented as means ± SEM. Statistical analysis was performed with GraphPad Prism software version 6 (GraphPad Software, La Jolla, CA). The following tests were performed: For immunohistochemical analysis, the U-test and one-way nonparametric analysis of variance with Kruskal-Wallis and Dunn's multiple comparison tests. For chemotaxis assay, one-way analysis of variance with Sidak's multiple comparison. For qPCR time course, analysis of variance with Sidak's post hoc test and two-way analysis of variance to examine the overall treatment effect. For qPCR chemokine expression, unpaired t-test was used. For cytokine bead array assay, unpaired t-test with Welch's correction was used. Differences were considered significant for P < 0.05.
Ductular Reaction Correlates with Fibrosis and Biliary Taurocholate in CFLD
The DR, as assessed by CK7 expression, was markedly increased in CFLD and contained expanded reactive bile ducts, and both single LPCs and strings of biliary cells/LPCs (Figure 1). The DR was mainly present in the periphery of portal tracts as irregular strands/chords of CK7+ cells and as isolated single CK7+ LPCs in adjacent areas. This DR was increased in CFLD (P < 0.05), as well as in biliary atresia (P = 0.114) and idiopathic neonatal hepatitis (P = 0.016), versus controls (Figure 2A). When stratified for fibrosis stage, there was a highly significant positive correlation between CK7 staining and increasing fibrosis stage in CFLD (Spearman r = 0.59, P < 0.0001) (Figure 2B). Bile samples were available from 7 of the 60 children with CFLD. There was a positive statistical correlation between biliary taurocholate concentrations and the extent of the DR as assessed via CK7 quantification (r = 0.528, P = 0.0321) (Figure 2C).
Taurocholate Induces Biliary Differentiation of LPCs
Because in CFLD, the extent of the DR was demonstrated to correlate with both fibrosis severity and biliary taurocholate levels, it was hypothesized that taurocholate may induce LPC biliary differentiation and the DR. To interrogate this hypothesis, the murine LPC line PIL-2 was treated with 0 to 300 μmol/L taurocholate for 4 days to establish the optimal conditions for all in vitro and ex vivo studies. Taurocholate induced a dose-dependent increase in mRNA expression of the cholangiocyte markers Cx43 (analysis of variance, P < 0.0001) (Figure 3A) and Ck19 (analysis of variance, P < 0.0001) (Figure 3B). A taurocholate concentration of 150 μmol/L was chosen for all subsequent experiments as has previously been described in in vitro studies,
PIL-2 cells were treated with 150 μmol/L taurocholate for 0 to 4 days, and markers of cholangiocyte differentiation were further assessed. Taurocholate induced a time-dependent increase in mRNA expression of cholangiocyte markers Cx43 (analysis of variance, P < 0.0001) (Figure 3C), Ck19 (analysis of variance, P < 0.0001) (Figure 3D), Itgb4 (analysis of variance, P < 0.029) (Figure 3E), and Ggt1 (analysis of variance, P < 0.001) (Figure 3F). Taurocholate treatment did not significantly alter expression of two other cholangiocyte markers, EpCam and Sox9 (data not shown).
Of interest, a concomitant decrease in mRNA expression of the master regulator of hepatic differentiation, Hnf4α was observed (analysis of variance, P = 0.0255) (Figure 4A). A second hepatocyte marker, albumin (Alb) showed no significant difference with taurocholate treatment (Figure 4B). Taken together, these results suggest that taurocholate augments the cholangiocytic differentiation of the LPC line PIL-2.
LPC Proliferation Is Induced by Taurocholate
Integral to the formation of the DR is the expansion of LPCs in chronic liver disease. To assess the potential effect of taurocholate on LPC proliferation, PIL-2 cells with and without 150 μmol/L taurocholate were subjected to an MTT cell proliferation assay every 24 hours for 1 to 4 days. Taurocholate induced a time-dependent increase in LPC proliferation (analysis of variance, P = 0.0003) (Figure 5), which was most notable early, that is, after 1 to 2 days.
Taurocholate Induces Differentiation of LPCs into Functional Cholangiocytes
Although Figure 3 demonstrates that taurocholate induced the expression of cholangiocyte-associated genes in LPCs, an assessment of actual cholangiocyte function remained to be confirmed. Thus, we examined expression of proteins involved in various aspects of biliary cell function. We observed that CK19, a protein responsible for the structural integrity of cholangiocytes is significantly induced by taurocholate (P = 0.0006) (Figure 6, A and B). Taurocholate also induced a significant >2.5-fold increase in SOX9 protein, a key transcriptional regulator of cholangiocyte differentiation and bile duct formation (P = 0.001) (Figure 6, A and C). Conversely, HNF4α protein expression, a marker and transcriptional regulator of hepatocyte specification, was significantly decreased by taurocholate (P = 0.019) (Figure 6, A and D).
One key morphological characteristic of cholangiocytes is the presence of primary cilia with acetylated α-tubulin being a key structural component. Western blot analysis revealed a significant induction of acetylated α-tubulin protein expression in taurocholate-treated LPCs (P = 0.026) (Figure 6, A and E). Because acetylated α-tubulin can also be found within other cellular structures such as microtubule networks, the cellular localization of acetylated α-tubulin was further investigated using immunofluorescence of LPCs with and without taurocholate. Expression of acetylated α-tubulin causes primary cilia to appear as discrete dots or lines, whereas microtubule networks form thin filament-meshed patterns.
As can be seen in Figure 7A, taurocholate induced a marked increase in the number of acetylated α-tubulin+ primary cilia when compared to controls, with quantitation of the relative fluorescence showing a 2.4-fold increase (P = 0.022) (Figure 7B).
Another measure of functional LPC-cholangiocyte differentiation is the production of enzymatically active GGT, which is expressed and released by cholangiocytes within bile ducts. PIL-2 cells were treated with and without 150 μmol/L taurocholate for 1 to 4 days with conditioned medium removed for analysis every 24 hours. Taurocholate induced a significant time-dependent increase in GGT activity (analysis of variance, P < 0.0001), with maximal production between days 3 to 4 (Figure 7C). At the conclusion of the experiment (day 4), this result remained highly significant when corrected for cellular protein concentration (Supplemental Figure S1).
The Effect of Taurocholate in ex Vivo Precision-Cut Liver Slices
To demonstrate potential physiological relevance of our in vitro observations, precision-cut liver slices from normal control mice, treated with and without 150 μmol/L taurocholate for up to 4 days, were examined. The hepatic mRNA expression of Cx43, as a representative cholangiocyte marker, was significantly increased in a time-dependent manner (analysis of variance, P = 0.0354) (Figure 8A). Conditioned medium collected was assayed for GGT activity and normalized to liver slice wet weight. Taurocholate again induced a significant increase in GGT activity, peaking at day 2 (analysis of variance, P = 0.0021) (Figure 8B).
Chemokine Expression Induced by Taurocholate in LPCs
PIL-2 cells treated with 150 μmol/L taurocholate showed significantly increased expression of MCP-1, MIP1α, and RANTES mRNA at 6 hours (Figure 9A). We also examined the gene expression of the profibrotic cytokine TGF-β1 and HSC mitogen PDGFB, which both have HSC chemotactic potential. Although taurocholate induced increased expression of both TGF-β1 and PDGFB, this did not reach statistical significance (Figure 9A).
Conditioned medium from PIL-2 cells with and without 150 μmol/L taurocholate for 6 hours or 24 hours was assessed for protein secretion of key hepatic chemokines. Rantes, Mcp-1, Mip1α, Tgf-β1, and Pdgfb were all expressed constitutively by PIL-2 cells. After 6 hours, taurocholate treatment resulted in a twofold increase in MIP1α protein release (P < 0.0001), whereas RANTES and PDGFB increased by approximately 1.5-fold (Figure 9B). Taurocholate induced an increase in MCP-1 secretion; however, this did not reach statistical significance. By 24 hours, chemokine levels were indistinguishable from controls, except for PDGFB, which remained elevated in response to taurocholate (P < 0.05). No effect was observed on TGF-β1 release.
HSCs Migrate in Response to Chemokines Secreted by Taurocholate-Treated LPCs
Chemotaxis of the LX-2 HSC line was significantly induced by conditioned medium from PIL-2 cells treated with taurocholate for either 6 hours (P < 0.001) or 24 hours (P < 0.01) (Figure 10A). Similar levels of chemotaxis were demonstrated using primary rat HSCs (data not shown). To determine the identity of chemokines responsible for this chemotaxis, conditioned medium was pretreated with neutralizing antibodies to various different chemokines. Chemotaxis was completely inhibited by anti-MIP1α neutralizing antibody (nAb); however, anti-RANTES nAb had no effect (Figure 10B). Although chemotaxis of LX-2 cells was partially suppressed using anti-MCP-1 nAb, this did not reach statistical significance.
HSCs Migrate in Response to MIP1α Mediated by CCR5
As shown in Figure 10B, chemotaxis of HSCs in response to conditioned medium from PIL-2 cells treated with taurocholate was inhibited by anti-MIP1α nAb. To strengthen the case for the involvement of MIP1α as a HSC chemotactic agent, a chemotaxis assay was performed on LX2 cells using either murine or human recombinant purified MIP1α. Both murine (analysis of variance, P = 0.022) (Figure 11A) and human (analysis of variance, P = 0.0009) (Figure 11B) MIP1α produced a significant dose-dependent increase in HSC migration, approaching the effect observed with the positive control (FBS).
LX-2 cells were examined for the expression of both known major hepatic MIP1α receptors, C-C chemokine receptor type 1 and 5 (CCR1 and CCR5). Using qPCR, expression of both Ccr1 and Ccr5 was demonstrated, with Ccr5 showing higher mRNA levels versus Ccr1 (Supplemental Figure S2). To investigate whether CCR5 was the major contributing receptor in HSC chemotaxis, LX-2 cells were pretreated with an anti-CCR5 nAb. This antibody effectively attenuated the migration of LX-2 HSCs in response to both recombinant murine MIP1α (Figure 11C) and conditioned medium from PIL-2 cells treated with 150 μmol/L taurocholate (Figure 11D).
DR Cells and LPCs Express Chemokines in Liver from Children with CFLD
To assess the potential clinical relevance of taurocholate-induced chemokine expression demonstrated in in vitro and ex vivo studies, liver sections from patients with CFLD (n = 22) were examined for the cellular source of chemokines. Dual immunofluorescence confirmed that CK7+ DR cells, and single CK7+ LPCs, express MCP-1 (Figure 12A), RANTES (Figure 12B), and MIP1α (Figure 13).
The DR is proposed to be associated with the initiation of hepatic fibrosis in a number of chronic liver diseases.
However, the precise mechanisms involved in its induction and indeed in the interactions between LPCs and HSCs, which may drive both the DR and fibrogenesis, remain to be elucidated. This histological process has been reported in many adult liver diseases, including chronic hepatitis C,
However, the presence or role of the DR has not previously been investigated in children with CFLD. Here we show that the DR is indeed present in CFLD (and comparable to the DR present in more aggressive pediatric fibrosing cholestatic diseases such as biliary atresia
have elucidated a role for many different soluble mediators in hepatic fibrosis. Understanding the interaction between LPCs and HSCs is of great importance for clarifying the contribution of these cells in mediating fibrogenesis and for designing therapeutic strategies aimed at reduction of hepatic fibrosis.
Bile salt retention is part of the proposed mechanism of injury and biliary fibrosis in CFLD, and we previously demonstrated that biliary/serum levels of the conjugated bile acid taurocholate are elevated in children with CFLD.
In the present study, it was demonstrated that the extent of the DR correlates with biliary taurocholate concentration in children with CFLD and thus speculate a potential mechanistic link. To determine whether taurocholate plays a role in inducing the DR, both an LPC in vitro cell culture model and ex vivo precision-cut liver slices were used, and a reductionist strategy was employed to examine the impact of taurocholate on three key events in the evolution of the DR: LPC proliferation, biliary differentiation of LPCs, and HSC chemotaxis.
In this study, it was demonstrated that taurocholate increased LPC proliferation in vitro, suggesting that taurocholate may play a role in enhancing LPC expansion as part of the DR in CFLD. This observation is supported by studies from Clouzeau-Girard et al
showing that taurocholate leads to uncharacterized cell proliferation around the portal tracts in ex vivo liver slices. Our data are also consistent with the observation that feeding taurocholate to rats increases cholangiocyte numbers,
although this study did not distinguish whether these were mature adult cholangiocytes or derived from LPCs.
The biliary differentiation of LPCs and subsequent formation of reactive bile ductular structures appear to be an important component of DR development. A significant increase in expression of the cholangiocyte genes Ck19, Cx43, Itgb4, and Ggt1 was demonstrated in response to taurocholate treatment of LPCs. Conversely, the hepatocyte lineage marker Hnf4α was decreased, with albumin expression unchanged. This in vitro result may be consistent with LPC differentiation that occurs during the DR in vivo and provides one potential mechanism for induction of the DR in pediatric cholestatic liver diseases such as CFLD. Further supporting this concept, it was demonstrated that taurocholate induced functional cholangiocyte differentiation, evidenced through the increased expression of acetylated α-tubulin, CK19, and SOX9 protein, the increased numbers of LPCs elaborating primary cilia, as well as the secretion of enzymatically active GGT in vitro. The effect of taurocholate on cholangiocyte function was further confirmed in ex vivo precision-cut liver slices, showing induced expression of Cx43 mRNA and secretion of active GGT.
In addition, this study has demonstrated that taurocholate-exposed LPCs secrete a number of chemokines including MCP-1, MIP1α, RANTES, and PDGFB; however, HSC chemotaxis in response to taurocholate-treated LPC conditioned medium appears to be principally due to MIP1α. Neutralizing antibodies to either MIP1α in LPC conditioned medium or the MIP1α receptor CCR5 on HSCs attenuated taurocholate-treated LPC-induced HSC chemotaxis. Thus, we propose that MIP1α derived from bile acid–exposed LPCs may recruit HSCs (via CCR5) during the DR and aid in the initiation of hepatic fibrosis in CFLD. We did not demonstrate that RANTES aided in the chemoattraction of HSCs. Neutralization of MCP-1 resulted in a limited inhibition of HSC chemotaxis, however this did not reach statistical significance due to biological variability of the data. RANTES and MCP-1 may induce recruitment of other cells to the LPC niche as part of the fibrogenic process such as additional LPCs or inflammatory cells, as we previously described.
Migration of HSCs toward LPCs, as well as HSC activation, could explain why fibrosis begins in periportal regions of the liver, where there is decreased bile flow through bile ducts, and thus increased accumulation of hydrophobic bile acids.
This pattern of chemokine expression upon in vitro taurocholate treatment may be indicative of the specific role of LPCs in generating a microenvironment that aids in the activation of HSCs. In the current study, immunofluorescence revealed that MCP-1, RANTES, and MIP1α are all expressed by DR cells and CK7+ LPCs in liver from children with CFLD, providing in vivo evidence validating our in vitro observations.
HSC activation is also influenced by other cell types present during chronic liver injury and inflammation, because the LPC niche harbors immune cells including liver macrophages, which also secrete MCP-1, MMPs, TIMPS,
and chemokines. The downstream chemoattractive effect caused by exposure of LPCs to increased levels of taurocholate appears likely to recruit HSCs into the LPC niche and places these two cell types within close proximity and potentially in direct contact. Supporting this, activated HSCs have been observed adjacent to LPCs in various mouse model of fibrosis,
demonstrating recruitment of each cell type toward the other during fibrosis. Various studies have also suggested that membrane contact (and cross talk) between LPCs and HSCs is a key mediator of fibrogenesis.
In this context, our previous work showed that HSCs treated with the TNF superfamily member, lymphotoxin-β, induces NF-κB signaling, up-regulating chemotaxis-associated factors such as RANTES and ICAM-1, and aiding in LPC and inflammatory cell recruitment.
Because HSCs express the lymphotoxin-β receptor on their cell surface, and lymphotoxin-β is a membrane-bound ligand on LPCs, this suggests that direct cell–cell contact between LPCs and HSCs may be required to initiate chemokine expression.
Adding to the complexity of this system, high concentrations of taurocholate have also been shown to induce inflammatory gene expression in hepatocytes, as well as cytokine/chemokine secretion, that may also lead to HSC recruitment and activation.
It is also possible that the two processes may be injury context-specific thus further investigations are warranted to fully elucidate the temporal nature of LPC niche development. In addition to HSCs, portal fibroblasts may also contribute to CFLD. However, whereas previous studies have shown that both activated HSCs and myofibroblasts produce collagen in this disease,
the role of myofibroblasts derived from portal fibroblasts and indeed the potential recruitment by, and interaction between, these cells and LPCs has not been previously investigated.
In conclusion, this study has shown that the DR is present in CFLD and is correlated with biliary taurocholate levels, as well as increasing hepatic fibrosis severity. Thus, we propose that the following model may in part help to explain the pathogenesis of CFLD (Figure 14). In this model, the CFTR defect causes thickened mucous secretions and obstruction of the intrahepatic bile ducts. This obstruction elevates biliary levels of bile salts such as taurocholate in the vicinity of LPCs, induced as a result of hepatic injury and potentially compromised hepatocyte replication associated with the chronic nature of liver injury in CFLD. In response to supraphysiological levels of bile salts, LPCs first undergo proliferation followed by differentiation toward cholangiocytes as part of the DR. Also, in response to altered bile salt levels, LPCs secrete chemokines, including MCP-1, MIP1α, and RANTES, and this likely attracts HSCs and other immune/inflammatory cells to the LPC niche. Cell–cell contact between LPCs and recruited HSCs may result in activation of HSCs and subsequent fibrogenesis, at least in part via interaction between the HSC lymphotoxin-β receptor via cognate ligands on LPCs, with further inflammatory niche chemotaxis, as we previously described.
Increased understanding of the role of hydrophobic bile acids in CFLD and their impact on the LPC niche, cellular cross talk with HSCs, and induction of the DR will provide important new knowledge of clinical relevance, which may aid in the development of targeted therapeutic strategies to reduce fibrosis in CFLD and potentially in other pediatric cholestatic disorders.
We thank Clay Winterford, Nigel Waterhouse, Tam Hong Nguyen, and Fernando Guimaraes (QIMR Berghofer Medical Research Institute) for technical support, and Prof. George Yeoh (University of Western Australia) for his gift of the PIL-2 murine liver progenitor cell line.
Liver disease in cystic fibrosis: a prospective study on incidence, risk factors, and outcome.
Supported by National Health and Medical Research Council (NHMRC) of Australia grant APP1048740 (G.A.R., P.J.L., R.W.S.), Practitioner Fellowship from the NHMRC of Australia grant APP1042370 (J.K.O.), and Senior Research Fellowship from the NHMRC of Australia grant APP1061332 (G.A.R.).
K.N.P., M.A.P., and T.N.P. contributed equally to this work.