Hepatic stellate cells express functional CXCR4: Role in stromal cell–derived factor-1α–mediated stellate cell activation†
Article first published online: 9 FEB 2009
Copyright © 2009 American Association for the Study of Liver Diseases
Volume 49, Issue 6, pages 2055–2067, June 2009
How to Cite
Hong, F., Tuyama, A., Lee, T. F., Loke, J., Agarwal, R., Cheng, X., Garg, A., Fiel, M. I., Schwartz, M., Walewski, J., Branch, A., Schecter, A. D. and Bansal, M. B. (2009), Hepatic stellate cells express functional CXCR4: Role in stromal cell–derived factor-1α–mediated stellate cell activation. Hepatology, 49: 2055–2067. doi: 10.1002/hep.22890
Potential conflict of interest: Nothing to report.
- Issue published online: 28 MAY 2009
- Article first published online: 9 FEB 2009
- Accepted manuscript online: 9 FEB 2009 12:00AM EST
- Manuscript Accepted: 28 JAN 2009
- Manuscript Received: 8 JUL 2008
Chemokine interactions with their receptors have been implicated in hepatic stellate cell (HSC) activation. The hepatic expression of CXCR4 messenger RNA is increased in hepatitis C cirrhotic livers and plasma levels of its endogenous ligand, stromal cell–derived factor-1α (SDF-1α), correlate with increased fibrosis in these patients. The expression of CXCR4 by HSCs has not been reported. We therefore examined whether HSCs express CXCR4 in vivo and in vitro and explored whether SDF-1α/CXCR4 receptor engagement promotes HSC activation, fibrogenesis, and proliferation. The hepatic protein expression of both CXCR4 and SDF-1α is increased in hepatitis C cirrhotic livers and immunoflourescent and immunohistochemical staining confirms that HSCs express CXCR4 in vivo. Immortalized human stellate cells as well as primary human HSCs express CXCR4, and cell surface receptor expression increases with progressive culture-induced activation. Treatment of stellate cells with recombinant SDF-1α increases expression of α-smooth muscle actin and collagen I and stimulates a dose-dependent increase in HSC proliferation. Inhibitor studies suggest that SDF-1α/CXCR4-dependent extracellular signal-regulated kinase 1/2 and Akt phosphorylation mediate effects on collagen I expression and stellate cell proliferation. Conclusion: HSCs express CXCR4 receptor in vivo and in vitro. CXCR4 receptor activation by SDF-1α is profibrogenic through its effects on HSC activation, fibrogenesis, and proliferation. Extracellular signal-regulated kinase 1/2 and phosphoinositide 3-kinase pathways mediate SDF-1α–induced effects on HSC expression of collagen I and proliferation. The availability of small molecule inhibitors of CXCR4 make this receptor an appealing target for antifibrotic approaches. (HEPATOLOGY 2009.)
Chronic hepatitis C is a chronic inflammatory disease of the liver wherein necroinflammatory activity correlates with disease progression.1 Cytokines and chemokines are major regulators of tissue inflammation. Chemokines are small, proinflammatory, chemoattractant cytokines that bind to a large family of G-protein coupled receptors, characterized by seven transmembrane spanning domains and activate a variety of intracellular signaling pathways based on the specific receptor, target cell, and context.2, 3 They have been implicated in a range of physiologic and pathologic conditions, including inflammation, development, leukocyte trafficking, homing of stem/progenitor cells, angiogenesis, and cancer.4 Although chemokines play a prominent role in the inflammation associated with hepatitis C virus (HCV),5 increasing evidence suggests direct profibrogenic effects on hepatic stellate cells (HSCs), a key cell type in fibrosing liver injury. Thus far, three functional chemokine receptors have been identified on stellate cells: CXCR3,6 CCR7,7 and CCR5.8 These chemokine receptors each have at least three different endogenous ligands and in all cases, receptor engagement by one but not all of the endogenous ligands stimulates HSC proliferation and chemotaxis.
The various ligands of a single chemokine receptor may induce similar, alternate, or even antagonistic effects on the receptor.9 Chemokine–chemokine receptor interactions, such as CCL21/CCR7,7 are important in promoting both inflammation and fibrosis in HCV. Therefore, understanding the mechanisms by which chemokines may directly affect stellate cells has important therapeutic implications.
Gene array studies indicate that CXCR4 messenger RNA (mRNA) expression is increased in patients with HCV10 compared with normal liver, and increased expression differentiates patients with moderate from mild fibrosis.11 Furthermore, both plasma levels of stromal cell–derived factor-1α (SDF-1α; CXCL12), an endogenous ligand for CXCR4, as well as hepatic staining for SDF-1α positively correlate with increased fibrosis in patients with chronic HCV.12 Interestingly, in the setting of chronic injury, SDF-1α staining is anatomically redistributed and becomes greatest along fibrotic septae, and is thought to be expressed by endothelium of neo-blood vessels.12 In addition, more recently murine liver sinusoidal endothelial cells have been shown to express significant SDF-1α and contribute to the transmigration of CD4+/CXCR4+ T cells.13 Because stellate cells—the resident stromal cells of the liver and major mediator of the fibrotic process—are located between the liver sinusoidal endothelial cells and hepatocytes and align themselves along fibrotic septae in chronic liver injury, endothelial-derived SDF-1α may have paracrine effects on stellate cells. HSCs themselves may also be a cellular source of SDF-1α in chronic liver injury. The expression of CXCR4 on stellate cells has not been reported.
The aims of this study were to (1) compare the protein expression of CXCR4 and SDF-1α in HCV cirrhotic livers compared with normal livers; (2) determine whether stellate cells express CXCR4 in HCV-infected livers and in culture; and (3) determine whether SDF-1α/CXCR4 signaling promotes stellate cell activation, proliferation, and production of collagen I and understand the underlying mechanisms.
Materials and Methods
Normal control liver tissue (five samples) was obtained during surgical resection of isolated colon cancer metastases. The tissue was obtained at a minimum of 5 cm from the tumor. HCV cirrhotic livers (seven samples) were obtained at the time of orthotopic liver transplantation. Tissues were obtained with permission of the Mount Sinai Institutional Review Board (GCO #99-247  PBC Pathology, Course and Treatment Project 2 of the Patient-Based Studies: A Tissue and Serum Bank for Studies of Liver Diseases) and used for western blot and immunofluorescent staining. For CXCR4 immunohistochemistry, formalin-fixed, paraffin-embedded liver sections obtained from patients with chronic HCV as part of their standard of care were provided by the Mount Sinai Department of Pathology with the approval of Mount Sinai IRB (GCO #05-1343).
Expression of CXCR4 and SDF-1α was examined in whole liver homogenate from patients with either HCV cirrhosis (n = 7) or control (n = 5). Briefly, protein was extracted from whole liver by homogenization using a dounce homogenizer followed by centrifugation (14,000 rpm) at 4°C for 10 minutes. Supernatant was collected, and 50 μg of protein was subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and membrane probed for CXCR4 and SDF-1α. β-Actin used as a loading control. For western blots examining expression of CXCR4 in LX-2 cells or primary HSCs and effects of SDF-1α on expression of alpha-smooth muscle actin (α-SMA), collagen I, phospho–extracellular signal-regulated kinase (ERK) 1/2, and phospho-Akt, 40 μg of whole cell lysate was subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis as described.14 The following antibodies with corresponding dilutions were used: rabbit anti-human CXCR4 polyclonal antibody (1:1,000; Affinity Bioreagents, catalog #OPA1-01101); anti–SDF-1α (1:1,000; R&D systems; catalog #MAB350); anti–phospho-ERK1/2 (1:1,000; Thr202/Tyr204, Cell Signaling, catalog #20G11) and total ERK 1/2 (1:1,000; Cell Signaling, catalog #9102); anti–phospho-Akt (1:1,000; ser473, Cell Signaling, catalog #4058S) and total Akt (1:1,000; Cell Signaling, catalog #39292); anti–α-SMA(1:1,000; Sigma, catalog #2547); anti-collagen I (1:1,000; Rockland, catalog #600-401). Species appropriate horseradish peroxidase–conjugated secondary antibodies were used at 1:2,000 to 1:5,000 dilutions and blots developed using an enhanced chemiluminescence western blotting detection system (Amersham Pharmacia Biotech, UK). Given the high-level expression of CXCR4 on Jurkat T cells,15 20 μg Jurkat T cell protein extract was used as a positive control.
Knockdown of CXCR4 and CXCR7 in Cultured Cells.
To knock down CXCR4 and CXCR7, short hairpin RNA (shRNA) was purchased from Sigma (St. Louis, MO). LX-2 and primary human HSCs were grown in six-well culture plates to 60%-70% confluence and transfected with either CXCR4 shRNA (Sigma, Clone ID: NM_003467.x-346s) or CXCR7 shRNA (Sigma, Clone ID: NM_020311.1-901s1c1) and shRNA Negative Control Med GC (Sigma, Plko.1-puro) according to the manufacturer's instructions using TransIT-LT1 reagent (Mirus Bio, Catalog: MIR2300). After 6 hours, the medium was replaced with normal growth medium, and efficiency of knockdown was determined via western blot analysis.
Enhanced chemiluminescent images of immunoblots were analyzed via scanning densitometry and quantified with a BIOQUANT NOVA imaging system (BIOQUANT NOVA PRIME). All values were normalized to housekeeping protein and expressed as fold changes relative to control.
Immunofluorescent Cell Staining.
For the detection of CXCR4 on stellate cells, primary HSCs grown in chamber slides or liver frozen sections (4 μM) were fixed with cold acetone (−20°C) for 5 minutes, rapidly rehydrated with 0.04% sodium azide/phosphate-buffered saline (PBS), permeabilized with 0.1% Triton/PBS, and blocked for 20 minutes at room temperature with blocking solution (PBS with 10% bovine serum albumin, 5% normal goat serum, 0.1% Triton). Subsequently, cells or liver sections were incubated with mouse monoclonal anti-human CXCR4 at 10 μg/mL (MAB172, R&D Systems, MN) or isotype control at 10 μg/mL (MAB003, R&D Systems). Liver sections were also incubated with monoclonal anti-human α-SMA at a dilution of 1:200 (D2547, Sigma). Cells and tissue were extensively washed with 0.1% Triton/PBS and incubated for 30 minutes at room temperature with Alexa Fluor 594 goat anti-mouse (for cells), or Alexa Fluor 488 goat anti-mouse and Alexa Fluor 594 goat anti-rabbit for frozen sections (Molecular Probes, Invitrogen, CA) at a dilution of 1:1,000. The cells and sections were mounted with Vectashield mounting media for fluorescence (Vector Laboratories, UK). 4',6-Diamidino-2-phenylindole (catalog # P36931, Molecular Probes) was used to stain for host cell DNA. Images were acquired with a Nikon Eclipse E600 fluorescence microscope (Nikon, Tokyo, Japan) for cells, and at the Mount Sinai Microscopy Shared Research Facility with Zeiss Axiophot 2 fluorescence microscope (Zeiss, Oberkochen, Germany) for tissue sections.
Formalin-fixed, paraffin-embedded liver tissue obtained from patients with chronic HCV (n = 5) were deparaffinized, rehydrated, blocked with 3% hydrogen peroxide for 30 minutes at room temperature, and thoroughly washed with PBS. Mouse anti-human CXCR4 polyclonal antibody (eBioscience, catalog #14-6009-81) at a dilution of 1:500 was used for overnight incubation at 4°C. Both primary isotype control antibody and secondary antibody with no primary antibody were used as negative controls. Slides were washed three times and a super-sensitive link-label ICH detection system (BioGenex) was used according to the manufacturer's instructions. Hematoxylin stained the nuclei blue.
Isolation of Primary Human and Murine Hepatic Stellate Cells.
Primary stellate cells were isolated from wedge sections of normal human liver in patients undergoing hepatic resection for primary benign tumors or single metastasis from colon cancer. Briefly, immediately after hepatectomy, a pathologist (I. Fiel) provided investigators with wedge sections of normal liver placed in Dulbecco's modified Eagle's medium. The liver was washed and portal venules cannulated for in situ digestion with pronase and collagenase. Stellate cells were isolated via density centrifugation and plated on plastic.16 For isolation of murine HSCs, stellate cells from six C57BL6 mice (20 to 30 weeks old) were pooled for each experiment. Mice were anesthetized with ketamine and xylazine (protocol approved by Mount Sinai IACUC), portal veins were cannulated in situ, livers were perfused with pronase and collagnease, density centifugation was performed, and pooled cells were plated on plastic as described.17
Fluorescence-Activated Cell Sorting Analysis.
To examine cell surface and total CXCR4 receptor expression, LX-2 cells or primary HSCs were grown in six-well dishes, scraped, and fixed with 4% paraformaldehyde in PBS for 20 minutes at 20°C. Cells were washed extensively with wash/resuspension buffer, permeabilized using 0.1% saponin in PBS containing 1 mM 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid, and incubated with phycoerythrin (PE)–immunoglobulin G2a isotype control (BD Biosciences, catalog #555574) or PE-conjugated anti-CXCR4 (BD Biosciences, catalog #555974) for 1 hour at room temperature. Cells were washed with wash resuspension buffer and analyzed via a fluorescence-activated cell sorting (FACS) instrument (FACScan, Becton Dickinson). For cell surface CXCR7 expression, nonpermeabilized LX-2 cells were prepared as above and incubated with PE-conjugated anti-CXCR7 (R&D Systems, catalog #FAB42271P) or isotype control.
[3H]-Thymidine Incorporation Assay.
DNA synthesis was estimated using methyl-[3H]-thymidine incorporation. LX-2 or culture-activated human primary HSCs were plated at a density of 20,000 cells per well in 24-well plates. After 12 hours, the medium was changed to 0.2% bovine serum albumin Dulbecco's modified Eagle's medium for 48 hours, and the cells were treated with carrier-free human recombinant SDF-1α at 50 ng to 500 ng/mL (R&D Systems, catalog #350-NS) for 24-72 hours with 1 μCi/mL [3H] thymidine present during the final 4 hours. Cells were then washed three times with ice-cold PBS and fixed in methanol for 30 minutes at 4°C. Cells were solubilized in 0.25% sodium hydroxide/0.25% sodium dodecyl sulfate. After neutralization with hydrochloric acid (1N), radioactivity was measured using a scintillation counter. ERK1/2 (UO126) and phosphoinositide 3-kinase (PI3K) (LY294002) inhibitors were added to the serum-starved cells at concentrations of 10 nM and 25 μM, respectively, 30 minutes prior to treatment with SDF-1α. For CXCR4 and CXCR7 knockdown experiments, cells were transfected with shControl versus shCXCR4 or shCXCR7 for 72 hours prior to treatment with SDF-1α. Each experiment was performed a minimum of three times.
Human SDF-1α Immunoassay (Quantitative Sandwich Enzyme Immunoassay).
To evaluate whether stellate cells secrete SDF-1α, conditioned media from LX-2 cells and primary stellate cells (passage 1) was collected, centrifuged to remove particulate debris, and frozen at −20°C until time of assay. Samples were prepared and assay was performed according to the manufacturer's protocol (DSA00, R&D Systems). Serum-free media and PBS were used as negative controls. Standard curves generated using purified recombinant human SDF-1α.
Quantitative Real-Time Reverse Transcription Polymerase Chain Reaction.
LX-2 cells were treated with 500 ng/mL SDF-1α for 1, 2, 4, 6, 12, and 24 hours, and RNA was extracted using Qiagen mini-columns with an on-column DNAase treatment. Quantitative real-time Reverse Transcription Polymerase Chain Reaction (qRT-PCR) real-time polymerase chain reaction (PCR) was performed using the following primers: collagen I (α1) forward, 5′-GTCCCTGAAGTCAGCTGCATA-3′ and reverse, 5′- TGGGACAGTCCAGTTCTTCAT-3′; glyceraldehyde 3-phosphate dehydrogenase (GAPDH) forward, 5′-CAATGACCCCTTCATTGACC-3′ and reverse, 5′-GATCTCGCTCCTGGAAGATG-3′. Samples were analyzed in triplicate using an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems) and normalized to GAPDH expression. Optimal dose and duration were repeated on at least 4 independent days.
All results are expressed as the mean ± standard deviation. Statistical significance was tested using an unpaired Student's t test (SPSS software), and P < 0.05 indicated a significant difference.
Increased Protein Expression of CXCR4 and SDF-1α in HCV Cirrhotic Livers and Expression of CXCR4 by Stellate Cells In Vivo.
To determine whether protein expression of CXCR4 and SDF-1α was increased in HCV cirrhosis, western blotting was performed on whole liver homogenates from explanted HCV cirrhotic and normal control livers (Fig. 1A). Protein expression of CXCR4 was an average 1.8-fold higher (P < 0.005) and SDF-1α protein expression 2.4-fold higher (P < 0.005) in patients with HCV cirrhosis. In order to specifically examine whether stellate cells, the cell type primarily responsible for liver fibrogenesis, express CXCR4 in vivo, frozen liver sections from three patients with HCV cirrhosis were coimmunostained with α-SMA (red), a marker of activated stellate cells, and CXCR4 (green). Immunofluorescence microscopy revealed a population of dual-stained cells (Fig. 1C). Immunohistochemistry for CXCR4 on paraffin-embedded liver biopsies from patients with chronic HCV demonstrated staining on bile duct epithelial cells, lymphocytes within the portal tracts, and stellate cells (Fig. 1D).
Immortalized and Primary HSCs Express CXCR4, and Expression Increases with Progressive Stellate Cell Activation.
To confirm in vivo findings, we first examined the expression of CXCR4 by LX-2 cells, an immortalized human HSC, with features resembling an activated stellate cell.18 Expression of CXCR4 by LX-2 cells was confirmed via PCR with sequencing (data not shown), western blot, immunofluoresence, and FACS (Fig. 2A -C). Protein expression of CXCR4 by passage 3 primary HSCs was also examined (Fig. 2A). Immunofluorescence microscopy demonstrated both cell surface and cytoplasmic expression of CXCR4 in stellate cells (Fig. 2B). FACS analysis revealed that over 50% of LX-2 cells express CXCR4 on the cell surface, and virtually all cells contain a cytoplasmic pool of CXCR4 (Fig. 2C). To characterize whether CXCR4 expression was affected by HSC activation, we used a model of in vitro activation of isolated primary HSCs.16 When primary stellate cells are isolated, they are rounded, lipid-rich, and therefore autofluoresce under standard phase contrast microscopy. With progressive activation, they become more myofibroblast-like, achieving peak activation by passage 3 (Fig. 3A). FACS was performed on both fixed and fixed and permeabilized primary HSCs during these progressive stages of activation to examine cell surface and total expression of CXCR4, respectively. Both cell surface and total CXCR4 expression increased as cells become more activated (Fig. 3A,B). By the time cells reached peak activation, their CXCR4 expression profile closely resembled that observed in LX-2 cells, an immortalized activated stellate cell (Fig. 2C). As proof of concept, primary murine stellate cells were isolated, and FACS for cell surface expression of CXCR4 was performed. As was seen with human primary stellate cells, CXCR4 expression increased with progressive culture-induced activation (Fig. 3C).
Stellate Cells Secrete SDF-1α and Exogenous Recombinant SDF-1α Promotes Further Activation.
Because stellate cells express CXCR4 and human smooth muscle cells also produce SDF-1α,19 the only endogenous ligand identified for CXCR4, we examined whether activated stellate cells, which have myogenic features, also secrete SDF-1α. Enzyme-linked immunosorbent assay was performed on 72-hour conditioned media from both LX-2 cells and primary human HSCs (passage 1) revealing production of SDF-1α by both cell lines (Fig. 4A). To determine effects of exogenous recombinant SDF-1α on the expression of α-SMA expression, a marker of activated stellate cells, LX-2 cells were serum-starved for 24 hours, then treated with increasing concentrations of SDF-1α (100-750 ng/mL) for 6 to 24 hours, and western blotting was performed to examine expression of α-SMA. We observed a dose-dependent increase in α-SMA expression, which peaked at 750 ng/mL (Fig. 4B,C). To determine whether reduction in CXCR4 expression could decrease SDF-1α–induced stellate cell activation, LX-2 cells were transfected with either control shRNA (shControl) or CXCR4 targeted shRNA (shCXCR4) and western blotting performed to determine the efficiency of knockdown (Fig. 4D). Because peak reduction in CXCR4 protein expression was noted 72 hours after transfection, cells were then treated with SDF-1α (500 ng/mL) and western blotting was performed for α-SMA, demonstrating an approximately 50% reduction in α-SMA expression in the presence of shCXCR4 (Fig. 4E).
SDF-1α Induces Stellate Cell Proliferation, Which Is Mediated by Both ERK 1/2 and PI3K-Akt Pathways.
LX-2 cells were treated with increasing concentrations of SDF-1α (50-500 ng/mL), and proliferation was examined via [3H]-thymidine incorporation after 24 to 72 hours. At all doses and time points, SDF-1α significantly induced stellate cell proliferation (Fig. 5A). Knockdown of CXCR4 using shRNA prior to SDF-1α significantly attenuated the proliferative response (Fig. 5B). Because both ERK 1/2 and PI3K-Akt pathways regulate stellate cell proliferation20 and SDF-1α is known to be a potent activator of ERK 1/2 and PI3K signaling in other cells,21 we first examined whether SDF-1α induces phosphorylation of ERK 1/2 and Akt in LX-2 cells. Treatment with SDF-1α induced phosphorylation of ERK 1/2 and Akt as demonstrated via western blotting (Fig. 5C,D). Moreover, treatment with either the ERK inhibitor (UO126; 10 nM)) or PI3K inhibitor (LY294002; 25 μM) significantly decreased SDF-1α–induced stellate cell proliferation by >50% (Fig. 5E; P < 0.001). Treatment with both inhibitors resulted in a synergistic reduction in SDF-1α–induced stellate cell proliferation, suggesting that both pathways are activated in concert. SDF-1α induced a similar proliferative response in primary HSCs (Fig. 6A) that was decreased in the setting of CXCR4 knockdown (Fig. 6C) and blocked in the presence of UO 126 and LY294002 (Fig. 6D). These data strongly implicate both the ERK 1/2 and PI3K-Akt pathways in promoting SDF-1α induced stellate cell proliferation.
SDF-1α Promotes Collagen I Expression Predominantly Through the PI3K-Akt Pathway.
Because plasma and hepatic levels of SDF-1α correlate with fibrosis in patients with HCV, we examined whether SDF-1α induces collagen I expression, the fibril-forming collagen characteristic of the cirrhotic liver. LX-2 cells were treated with SDF-1α (500 ng/mL) for 2, 6, and 12 hours, and RNA was extracted for qRT-PCR and protein extracted for western blotting to examine levels of collagen I α1 mRNA and total collagen, respectively. The maximal effect on collagen I mRNA expression was observed after 2 hours of SDF-1α treatment (Fig. 7A), and increases in collagen I protein were seen at all time points, peaking at 12 hours after SDF-1α treatment (Fig. 7B). Pretreatment of LX-2 cells with CXCR4-targeted shRNA significantly decreased SDF-1α–induced collagen I expression, suggesting that observed effects were largely mediated by CXCR4 (Fig. 7C). Because both the ERK 1/2 and PI3K-Akt pathway are known to be important for collagen I expression in stellate cells,22 we examined whether SDF-1α induction was ERK 1/2– or PI3K-Akt–dependent by preincubating LX-2 cells with the respective inhibitors, followed by SDF-1α treatment (500 ng/mL) for 6 hours and protein extraction for collagen I western blotting. Both inhibitors resulted in a reduction in collagen I expression (Fig. 7D), though the effect was more pronounced in the presence of the PI3K inhibitor (LY294002).
SDF-1α Effects on Stellate Cell Proliferation and Collagen I Expression Are Not Mediated by CXCR7.
Because CXCR7 has recently been identified as an additional receptor for SDF-1α,23, 24 we examined whether stellate cells express CXCR7 and whether some of the observed effects of SDF-1α on stellate cell proliferation and collagen I expression could be mediated in part by CXCR7. FACS was first performed to determine cell surface expression of CXCR7 on LX-2 cells (Fig. 8A). Having established expression of CXCR7 by stellate cells, LX-2 cells were transfected with nontargeting shRNA (shControl) or CXCR7-targeted shRNA (shCXCR7) and western blotting was performed to confirm decreased CXCR7 protein expression (Fig. 8B). To determine the role of CXCR7 on SDF-1α–induced proliferation, LX-2 cells were serum-starved and transfected with shControl versus shCXCR7. Seventy-two hours later, transfected cells were treated with SDF-1α for 24 to 72 hours, and proliferation assessed by thymidine incorporation (Fig. 8C), and collagen I expression via western blotting (Fig. 8D). No reduction in SDF-1α–induced proliferation or collagen I expression was noted in the presence of shCXCR7.
In this study, we demonstrate that (1) expression of both CXCR4 and SDF-1α protein is increased in patients with HCV cirrhosis; (2) stellate cells express CXCR4 in vivo and in vitro; (3) CXCR4 expression increases with stellate cell activation; and (4) SDF-1α binding to CXCR4 on stellate cells induces proliferation and collagen I production via both ERK 1/2 and PI3K-Akt pathways. During chronic HCV infection, liver injury is associated with persistent recruitment of inflammatory cells and activation of HSCs resulting in progressive fibrosis. The chemokine/chemokine receptor pair SDF-1α/CXCR4 is important for the homing of lymphocytes to sites of injury.21 Interestingly, SDF-1α is a more potent chemotactic stimulus for peripheral blood lymphocytes derived from patients with HCV than either CXCL10 or CCL21, highlighting the relative importance of SDF-1α in HCV-induced liver inflammation.5 In addition, SDF-1α promotes migration and adhesion of liver-infiltrating lymphocytes to fibronectin,5 and murine liver sinusoidal endothelial cells preferentially present SDF-1α on their cell surface and enhance T cell (CD4+/CXCR4+) transmigration.13 Taken together, these findings suggest that SDF-1α is involved in both recruitment of CXCR4+ cells from the circulation and the retention of CXCR4+ cells in the liver.5 Although biliary ductal epithelial cells are the predominant cellular source of SDF-1α in the normal liver, in the setting of chronic HCV-induced injury and fibrosis there is an anatomical redistribution of SDF-1α staining to proliferating bile ductules and endothelium of neo-vessels in fibrotic septae.12 Whereas our data suggest that stellate cells may also be a cellular source of SDF-1α during chronic liver injury and theoretically play a role in the recruitment and retention of lymphocytes during injury, the concentrations needed to elicit effects on stellate cell activation, proliferation, and collagen expression favor predominantly paracrine stimuli. Because stellate cells reside in the space of Disse between liver sinusoidal endothelial cells and hepatocytes and align along fibrotic septae during injury, it is more likely that endothelial-derived SDF-1α may provide the paracrine stimulus for stellate cell activation in vivo.
In addition to its potential role in HCV-induced inflammation, our study suggests a novel role for SDF-1α/CXCR4 signaling in promoting HCV fibrosis via direct stimulation of HSC proliferation and production of collagen I. Although several factors are mitogenic for HSCs, the most potent is platelet-derived growth factor.20 Both mitogen-activated protein kinase and PI3K-Akt pathways are critical for HSC proliferation,25, 26 and our findings confirm the importance of these pathways in SDF-1α–mediated HSC proliferation. Knockdown of CXCR4 significantly decreased the proliferative response of stellate cells to SDF-1α, suggesting that SDF-1α partially exerts its effects via CXCR4. However, as inhibition was not complete, we examimed whether SDF-1α may be mediating some of its effects through CXCR7, previously considered an orphan receptor but now recognized as a receptor for SDF-1α.23, 24 Although we have shown that stellate cells express CXCR7, reduction in CXCR7 expression did not diminish SDF-1α effects on either proliferation or collagen I expression. In addition, preincubating stellate cells with platelet-derived growth factor-BB neutralizing antibody prior to SDF-1α treatment failed to block SDF-1α–induced effects on proliferation (data not shown), ruling out the possibility of off-target effects of SDF-1α on platelet-derived growth factor signaling. Blockade of downstream ERK 1/2 and PI3K-Akt pathways, however, significantly diminished SDF-1α–induced proliferation, highlighting the importance of these downstream pathways in mediating effects of SDF-1α on stellate cells.
In addition to its effects on proliferation, SDF-1α up-regulates collagen I expression both at the mRNA and protein levels. This up-regulation appears to be predominantly via CXCR4, because knockdown of CXCR4 effectively blocked SDF-1α–induced collagen I expression, whereas no effect was seen with reduced CXCR7 expression. Both ERK 1/2 and PI3K-Akt pathways are important for collagen I expression in HSCs, and cross-talk between the two pathways may be important.28 In this study, inhibitors for both the ERK 1/2 and PI3K-Akt pathways blocked SDF-1α induction of collagen I, although the PI3K inhibitor LY294002 was more effective. These findings are consistent with recent data highlighting the prominent role for PI3K-Akt in collagen gene expression and hepatic fibrosis in vivo and in vitro.27
Our data as well as those of others12 suggest that targeted blocking of this receptor may be both anti-inflammatory as well as antifibrotic and will require in vivo confirmation. The importance of this pathway may not be limited to patients with HCV, because SDF-1α and CXCR4 mRNAs are also increased in patients with HBV and primary biliary cirrhosis.10 In addition, CXCR4 is a coreceptor for human immunodeficiency virus (HIV). Because patients coinfected with HIV/HCV develop more rapid fibrosis than HCV-monoinfected patients, and fibrosis progression rate correlates with HIV viremia,28 our findings also raise the possibility of direct effects of HIV (X4-tropic) and/or its viral envelope protein (gp120) on stellate cells. It has recently been shown that HIV gp120 binds CXCR4 on hepatocytes, promoting HCV replication and initiating downstream signaling pathways independent of direct viral infection.29 The effects of gp120 (X4) have not been examined on stellate cells, but given our identification of functional CXCR4 on stellate cells, further investigation into this possibility is warranted.
In conclusion, our data provide the first evidence of functional CXCR4 on HSCs and implicates SDF-1α signaling in promoting fibrogenesis through direct effects on stellate cells. The commercial availability of orally bioavailable small molecule inhibitors specific for CXCR4 make this receptor a particularly attractive target.
- 16Isolation and culture of hepatic nonparenchymal cells. In: Tyson C, Frazier J, eds. In Vitro Biological Systems. Volume 1. San Diego, CA: Academic Press, 1993: 292–310..
- 26Extracellular signal-regulated kinase activation differentially regulates platelet-derived growth factor's actions in hepatic stellate cells, and is induced by in vivo liver injury in the rat. HEPATOLOGY 1999; 30: 951–958., , , , , , et al.