Activation of Rac1 promotes hedgehog-mediated acquisition of the myofibroblastic phenotype in rat and human hepatic stellate cells

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Hepatic accumulation of myofibroblastic hepatic stellate cells (MF-HSCs) is pivotal in the pathogenesis of cirrhosis. Two events are necessary for MF-HSCs to accumulate in damaged livers: transition of resident, quiescent hepatic stellate cells (Q-HSCs) to MF-HSCs and expansion of MF-HSC numbers through increased proliferation and/or reduced apoptosis. In this study, we identified two novel mediators of MF-HSC accumulation: Ras-related C3 botulinum toxin substrate 1 (Rac1) and Hedgehog (Hh). It is unclear whether Rac1 and Hh interact to regulate the accumulation of MF-HSCs. We evaluated the hypothesis that Rac1 promotes activation of the Hh pathway, thereby stimulating signals that promote transition of Q-HSCs into MF-HSCs and enhance the viability of MF-HSCs. Using both in vitro and in vivo model systems, Rac1 activity was manipulated through adenoviral vector-mediated delivery of constitutively active or dominant-negative rac1. Rac1-transgenic mice with targeted myofibroblast expression of a mutated human rac1 transgene that produces constitutively active Rac1 were also examined. Results in all models demonstrated that activating Rac1 in HSC enhanced Hh signaling, promoted acquisition/maintenance of the MF-HSC phenotype, increased MF-HSC viability, and exacerbated fibrogenesis. Conversely, inhibiting Rac1 with dominant-negative rac1 reversed these effects in all systems examined. Pharmacologic manipulation of Hh signaling demonstrated that profibrogenic actions of Rac1 were mediated by its ability to activate Hh pathway-dependent mechanisms that stimulated myofibroblastic transition of HSCs and enhanced MF-HSC viability. Conclusion: These findings demonstrate that interactions between Rac1 and the Hh pathway control the size of MF-HSC populations and have important implications for the pathogenesis of cirrhosis. HEPATOLOGY 2010

Chronic liver injury induces regenerative responses in the liver, including the activation of quiescent hepatic stellate cells (Q-HSCs) to an activated, myofibroblastic phenotype that promotes scar formation and liver dysfunction.1 Hepatic accumulation of myofibroblastic hepatic stellate cells (MF-HSCs) is pivotal in the pathogenesis of cirrhosis. Consequently, considerable research has been directed toward delineating mechanisms that promote the transition of Q-HSCs to MF-HSCs and enhance the subsequent growth of MF-HSC populations.2 Recently, we showed that Rac1 is one factor that modulates hepatic accumulation of MF-HSCs.

Ras-related C3 botulinum toxin substrate 1 (Rac1) belongs to a subfamily of small guanosine triphosphate–binding proteins that includes Rac1, Rac2, Rac3, and several rac-related proteins.3 Rac1 regulates many cellular processes, including the cell cycle, cell–cell adhesion, and motility.3-6 It also maintains epidermal stem cells that generate epithelial tissues.7 We demonstrated that Rac1 is activated as Q-HSCs become MF-HSCs.8 Using a genetic approach to maintain high levels of Rac1 activity in MF-HSCs, we showed that this enhanced MF-HSC growth in culture. We also demonstrated that transgenic mice expressing constitutively active human Rac1 in MF-HSCs accumulated more myofibroblastic cells and developed worse liver fibrosis than littermates during liver injury.8 These data established activated Rac1 as an important mediator of MF-HSC accumulation and fibrosis.

The Hedgehog (Hh) pathway also modulates MF-HSC accumulation and liver fibrosis.9, 10 Hh ligands are lipid-modified morphogens that interact with Patched (Ptc), a membrane-spanning receptor on Hh-responsive cells. This ligand–receptor interaction prevents Ptc from inhibiting its coreceptor, Smoothened (Smo). Smo, in turn, initiates a series of intracellular events that culminate in activation and nuclear localization of Glioblastoma (Gli) family transcription factors.11 Hh signaling is antagonized by Hedgehog-interacting protein (Hhip), a factor that binds to Hh ligands and blocks Hh ligand–Ptc interactions.12 We have shown that Hh signaling is involved in adult liver repair.10, 13 Factors produced during liver injury stimulate MF-HSCs to produce Sonic hedgehog (Shh) ligand.9 Shh acts in an autocrine fashion to promote MF-HSC proliferation and viability,9 and in a paracrine fashion to stimulate the viability of ductular-type progenitor cells.13 Genetically altered mice with an overly active Hh pathway accumulate more MF-HSC and develop worse liver fibrosis after bile duct ligation (BDL) than littermate controls.13 Whether Hh signaling interacts with—or requires—Rac1 to increase MF-HSCs remains unclear. Therefore, the primary goal of the present study was to evaluate the hypothesis that Rac1 and the Hh pathway interact to promote the formation of MF-HSCs and/or to enhance accumulation of existent MF-HSCs. The secondary goal was to begin to characterize mechanisms involved in these processes.

During development, adult wound healing, and cancer metastasis, the Hh pathway is known to promote formation of mesenchymal cell types by inducing epithelial-to-mesenchymal transitions (EMT).14, 15 EMT is a process that permits tissue remodeling by repressing expression of adherens junction-proteins (such as E-cadherin), reducing cell–cell adhesion, and promoting cell motility.16 EMT may contribute to liver fibrosis.17 It was recently reported that treating MF-HSCs with adenoviral vectors for bone morphogenetic protein-7 (BMP-7) dramatically down-regulated expression of fibroblast markers, and that injection of BMP-7 adenoviral vectors into rats with advanced thioacetamide-induced cirrhosis abolished liver fibrosis.18 BMP-7 is a potent inhibitor of EMT, and epithelial cells that are capable of transitioning to mesenchymal cells typically express BMP-7 to repress mesenchymal gene expression and maintain expression of E-cadherin and other epithelial genes.19 Overexpression of BMP-7 in hepatocytes was shown to have no effect on hepatocyte E-cadherin expression, suggesting that the beneficial effects of BMP-7 on liver fibrosis were not due to inhibition of hepatocyte EMT.18 However, the possibility that BMP-7 inhibits EMT in some other type of liver cell was not examined.

Previously, we showed that Shh induces ductular-type progenitor cells to undergo EMT.20 We also showed that freshly isolated Q-HSCs express biliary epithelial markers (including E-cadherin) and the Hh antagonist Hhip and demonstrated that Q-HSCs down-regulate Hhip, activate Hh signaling, lose epithelial markers, and express mesenchymal genes as they become MF-HSCs. Blocking Hh signaling through pharmacological inhibition repressed mesenchymal gene expression while restoring expression of epithelial genes and other Q-HSC markers.21 In the present study, we show that manipulating Rac1 in rat primary HSCs, clonally derived rat and human MF-HSC lines, and two different animal models of liver fibrosis consistently modulates Hh signaling, EMT, accumulation of MF-HSCs, and fibrogenesis. These findings suggest that Rac1 and Hh signaling are crucial mediators of HSC activation, support our hypothesis that these factors interact to induce and maintain the myofibroblastic HSC phenotype, identify key Hh pathway regulators as targets of activated Rac1, and suggest a novel mechanism for liver fibrosis. The latter involves activated Rac1-dependent induction of Shh and concomitant repression of Hhip, resultant activation of Hh signaling, and consequent transition of Q-HSCs to MF-HSCs, as well as enhanced MF-HSC viability. Together, these events cause accumulation of MF-HSCs and increase fibrogenesis.

Abbreviations:

α-SMA, α smooth muscle actin; BDL, bile duct ligation; BMP-7, bone morphogenetic protein-7; CCl4, carbon tetrachloride; cDNA, complementary DNA; EMT, epithelial-to-mesenchymal transition; Gli, Glioblastoma; Hh, Hedgehog; Hhip, Hedgehog-interacting protein; HSC, hepatic stellate cell; MF-HSC, myofibroblastic hepatic stellate cell; NADPH, nicotinamide adenine dinucleotide phosphate, reduced form; PDGF, platelet-derived growth factor; Ptc, Patched; Q-HSC, quiescent hepatic stellate cell; qRT-PCR, quantitative reverse-transcription polymerase chain reaction; Rac1, Ras-related C3 botulinum toxin substrate 1; ROS, reactive oxygen species; SEM, standard error of the mean; Shh, Sonic hedgehog; Smo, Smoothened; Sna, Snail; TGF-β, transforming growth factor-β.

Materials and Methods

Animals.

Adult male Sprague-Dawley rats were obtained from Charles River Laboratories (Wilmington, MA). Wild-type C57Bl6 mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Rac-transgenic mice that express a constitutively active mutant form of human rac1 (V12rac1) in myofibroblasts were described.22 Animal experiments fulfilled National Institutes of Health and Duke University-Institutional Animal Care and Use Committee requirements for humane animal care.

Plasmids.

The expression vectors pEXV-V12rac1 and pEXV-N17rac1, encoding the constitutively active and dominant-negative myc epitope-tagged rac1 complementary DNAs (cDNAs), respectively, have been described.6

Adenoviral Vectors.

Adenovirus N17rac1 containing the dominant-negative rac1 cDNA was constructed using adenovirus-based plasmid JM17 and the expression vector pEXV-N17rac1, which encodes the dominant-negative myc-epitope-tagged rac1 cDNA.22 A similar strategy was used to construct an adenovirus encoding a constitutively active mutant of rac1 (V12rac1; P. Goldschmidt-Clermont, University of Miami, FL).22 The E1-deleted adenovirus dl312,22 which lacks a cDNA insert, served as a control for all gene transfer studies.

Rac1 Manipulation in Intact Mice.

Adult C57Bl6 mice (n = 36) received vehicle (100 μL) that contained empty adenoviral vectors (dl312) or adenoviral vectors bearing a constitutively active rac1 (V12rac1) or dominant-negative rac1 (N17rac1). Mice then underwent BDL (n = 24)23, 24 or sham surgery (n = 12). Mice were monitored until sacrifice 10 days after BDL. Ten-day survival of the 16 mice that received dl312 or V12rac1 was 75% in each group, whereas none of the eight mice treated with N17rac1 survived more than 4 days after BDL. Rac1-transgenic mice (n = 11) and age-matched/sex-matched wild-type littermates (n = 6) received twice-weekly intraperitoneal injections with 0.5 mg/kg carbon tetrachloride (CCl4) (Sigma-Aldrich, St. Louis, MO) and were sacrificed after 8 weeks.8 Following euthanasia, livers were harvested, snap-frozen in liquid nitrogen, or fixed in formalin and paraffin-embedded.8

Cell Isolation and Culture.

Primary HSCs were isolated from Sprague-Dawley rats, assessed for purity and viability, and seeded at a density of 3 × 102 cells/mm2 in Dulbecco's modified Eagle's medium supplemented with fetal bovine serum and penicillin/streptomycin.9, 21 The clonally derived rat HSC line 8B (M. Rojkind, George Washington University, Washington, DC), human HSC line LX-2 (S. L. Friedman, Mount Sinai School of Medicine, New York, NY), normal rat cholangiocyte line NRC (N. F. LaRusso, Mayo Clinic, Rochester, MN), and primary rat hepatocytes were cultured as described.25-28

Adenoviral Transduction of HSCs.

HSCs (primary, 8B, and LX-2) were grown to 70%-90% confluency in six-well plates and serum-starved for 4 hours prior to infection. Pilot studies demonstrated that maximally efficient transduction occurred at a multiplicity of infection of 50. Subsequent experiments were performed with this multiplicity of infection of 50 for 24 hours, then virus-containing media was aspirated and replaced with fresh medium.

Pharmacological Manipulation of Hh Signaling.

MF-HSCs (8B) were treated with Smo agonist, a chlorobenzothiophene-containing agonist (0.3 μM; Axxora, San Diego, CA),29 cyclopamine (5 μM; Toronto Research Chemicals, Toronto, ON, Canada), or tomatidine, a catalytically inactive analog (5 μM), for 24 to 48 hours.

Cell Migration.

MF-HSCs (8B) were cultured for 7-days; a standard wound healing assay was performed after treatment with adenoviral vectors with/without Smo agonist/antagonist.20, 29

Cell Viability Assays.

Cell viability was measured with the Cell Counting Kit-8 (Dojindo Molecular Technologies, Gaithersburg, MD).30

Messenger RNA Quantification by Real-Time Reverse-Transcription Polymerase Chain Reaction.

Total RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA), reverse-transcribed using Superscript reverse transcriptase (Invitrogen). cDNA samples were used for quantitative reverse-transcription polymerase chain reaction (qRT-PCR) using iQ-SYBR Green Supermix (Bio-Rad Laboratories).21

Rac1 Activation Assay and Western Blotting.

Rac1 activation assay was performed.8, 31 Western blots were used to demonstrate changes in relevant proteins; results were normalized to β-actin expression.21

Quantification of Hepatic Collagen Content.

This was evaluated by morphometry of Sirius Red–stained liver sections and hydroxyproline assay.8

Statistical Analysis.

Results are expressed as the mean ± standard error of the mean (SEM). Comparisons between groups were performed using a nonparametric Wilcoxon rank-sum test (SAS version 9.1 software; SAS Institute, Cary, NC). P values were two-tailed, and significance was set at the 5% level.

Results

Rac1 Activity Increases During HSC Activation in Culture and Promotes Acquisition and Retention of the Myofibroblastic Phenotype.

Rat primary HSCs were cultured for up to 7 days; affinity purification was performed to identify guanosine triphosphate–bound Rac1, the biologically active form of Rac1.6 Although active Rac1 was not identified in freshly isolated Q-HSCs, it increased steadily during culture (Fig. 1A). Rac1 activity was not detectable in normal rat cholangiocytes or hepatocytes, proving that Rac1 activity is a unique characteristic of myofibroblastic liver cells (Fig. 1B).

Figure 1.

Rac1 activation increases during HSC activation and promotes myofibroblastic gene expression. Primary rat HSCs were isolated and cultured in serum-containing media. (A) Affinity purification was performed to demonstrate endogenous, active (guanosine triphosphate–bound) Rac1. Immunoblot results are representative of data from triplicate experiments. (B) Affinity purification of active Rac1 in primary rat hepatocytes, normal rat cholangiocytes (NRC), and HSCs. (C-D) Primary rat MF-HSCs were treated with adenoviral vectors containing constitutively active rac1 (V12rac1), dominant negative rac1 (N17rac1), or empty adenoviral vector (dl312). Affinity purification was performed (C) and RNA was isolated after 72 hours of infection (D). Gene expression was assessed by way of qRT-PCR. Data are presented as the mean ± SEM of triplicate experiments. *P < 0.05, P < 0.005 versus the dl312-treated group. (E) Representative western blot analysis of parallel cultures analyzed in (D). β-Actin was assessed to control for variations in protein loading.

To determine whether Rac1 activation plays a causal role in myofibroblastic transition, 7-day culture-activated MF-HSCs were treated with adenoviral vectors encoding either constitutively active rac1 (V12rac1), dominant-negative rac1 (N17rac1), or empty cassette (dl312). Preliminary experiments confirmed that V12rac1 increased Rac1 activity whereas N17rac1 decreased it (Fig. 1C). V12rac1-treated HSCs demonstrated significantly higher expression of the myofibroblastic markers α-smooth muscle actin (α-SMA) and type 1 α1-collagen, whereas treatment with N17rac1 significantly lowered levels of these genes (Fig. 1D,E). Clonally derived rat (8B) and human (LX-2) HSC lines responded similarly to infection with these adenoviral vectors (Supporting Figs. 1 and 4), assuring that the findings in primary HSC cultures were not an idiosyncratic response of rodent HSCs or caused by selective outgrowth of rare, contaminating cell types. Thus, Rac1 activation not only occurs as Q-HSCs transition to MF-HSCs, it also actively contributes to acquisition and maintenance of the myofibroblastic phenotype in both rat and human HSCs.

Rac1 Induces the Myofibroblastic Phenotype by Promoting EMT.

Transition of Q-HSCs to MF-HSCs involves an EMT-like process.21 Therefore, we treated rat primary HSC cultures with adenoviral vectors for V12rac1, N17rac1, or dl312, and determined if modulating Rac1 activity influenced expression of various factors that are involved in EMT. Compared with dl312-treated HSCs, V12rac1-treated HSCs exhibited greater changes in EMT-related genes. For example, the EMT inhibitor BMP-7 was significantly down-regulated, as was expression of desmoplakin (epithelial gene). Conversely, expression of Snail (Sna), an EMT inducer, and S100A4, a marker of mesenchymal cells of epithelial origin, was enhanced (Fig. 2A). In contrast, treating HSCs with N17rac1 to repress Rac1 activity exerted the opposite effects, up-regulating BMP-7 and desmoplakin while down-regulating Sna and S100A4 (Fig. 2B). Similar results were noted when 8B and LX-2 were treated with these vectors (Supporting Figs. 2 and 4). Because cells that have undergone complete EMT acquire a migratory phenotype,16 we treated clonally derived rat MF-HSCs (8B) with V12rac1, N17rac1, or dl312 and used a standard wound healing assay to determine whether altering Rac1 activity influenced cell migration.20 Compared with treatment with dl312, V12rac1-treated cells exhibited enhanced migration into the wound, whereas N17rac1-treated cells showed inhibited cell migration (Fig. 33). Hence, Rac1 activation causes HSCs from adult livers to complete an EMT-like process that leads them to acquire a migratory, mesenchymal phenotype.

Figure 2.

Increasing Rac1 activity enhances gene expression changes associated with epithelial-to-mesenchymal transitions in primary rat HSCs. Primary rat HSCs were isolated and cultured in serum-containing media for 7 days to produce a myofibroblastic phenotype. Triplicate plates of day 7 culture-activated MF-HSCs were then treated with adenoviral vectors for either (A) V12rac1 or B) N17rac1. To control for nonspecific effects of adenoviral infection, triplicate plates were treated in parallel with dl312. RNA was isolated after 72 hours of infection. gene expression was assessed by way of qRT-PCR. Each experiment was performed a total of three times; data are presented as the mean ± SEM. *P < 0.05, P < 0.005 versus dl312-treated cultures.

Figure 3.

Rac1 induces myofibroblastic phenotype by promoting EMT. Triplicate cultures of the clonally derived rat HSC line 8B were treated with adenoviral vectors containing constitutively active rac1 (V12rac1), dominant-negative rac1 (N17rac1), or empty adenoviral vector (dl312). A standard wound healing assay was performed and 24 hours later, cell migration into the wound was examined. Representative results are shown.

Rac1 Promotes EMT and Survival in HSCs by Enhancing Hh Pathway Activity.

In many cell types, including adult liver HSCs, EMT is regulated by the Hh pathway.20, 21 Our data show that Rac1 activation induced EMT in HSCs, suggesting potential interactions between Rac1 and the Hh pathway. To address this more directly, HSCs were treated with V12rac1, N17rac1, or dl312, and effects on Hh signaling were examined by way of qRT-PCR and western blot analysis. Compared with dl312-treated HSCs, HSCs treated with V12rac1 demonstrated enhanced Hh signaling, with significant down-regulation of Hhip, and increased expression of Shh (Hh ligand), Gli2, and sFRP1 (both Hh target genes) (Fig. 4A,C). Conversely, blocking Rac1 activity with N17rac1 increased Hhip expression and decreased expression of Shh and Gli2, reducing Hh signaling and expression of the Gli-regulated gene sFRP1 (Fig. 4B,C). Similar effects on the Hh pathway were noted when Rac1 activity was manipulated in the clonally derived 8B and LX-2 lines (Supporting Figs. 3 and 5). Thus, Rac1 activity represses HSC expression of Hhip and increases HSC expression of Shh ligand, leading to activation of Hh signaling that results in increased expression of Hh target genes (Gli2). Gli family transcription factors promote the transcription of other factors, such as Sna, that orchestrate complex, global changes in cellular gene expression to affect EMT.32 To verify that Hh signaling-regulated events were causally involved in HSC EMT, a wound healing assay was performed with clonally derived rat HSCs (8B) treated with cyclopamine, a specific antagonist of Hh signaling, or tomatidine, an inactive cyclopamine analog.30 Cyclopamine (but not tomatidine) decreased HSC migration (Fig. 5A), confirming that Hh signaling is necessary for HSCs to undergo complete EMT in culture. Furthermore, adding a Smo agonist restored the migratory capabilities and changes in gene/protein expression of MF-HSCs that had been treated with N17rac1, whereas blockage of Hh signaling inhibited migration and normalized gene/protein expression in V12rac1-treated MF-HSCs (Fig. 5B, Supporting Fig. 6). More protracted treatment of MF-HSCs with either N17rac1 or cyclopamine reduced cell survival compared with respective controls (Fig. 5C). Therefore, Rac1 activation initiates a cascade of events, which result in induction of Gli transcription factors, Sna, and other factors that cause the cells to acquire a mesenchymal, migratory phenotype and remain viable in vitro.

Figure 4.

Increasing Rac1 activity promotes Hh pathway signaling in primary rat HSCs. Primary rat HSCs were cultured for 7 days to generate MF-HSCs. Triplicate day 7 cultures were treated with adenoviral vectors for either (A) V12rac1 or (B) N17rac1 as described in Fig. 2. RNA was isolated for qRT-PCR. Results are compared with triplicate cultures of MF-HSCs that were treated with dl312. Each experiment was performed a total of three times. Data are presented as the mean ± SEM. *P < 0.05, P < 0.005 versus dl312-treated cultures. (C) Representative western blot analysis of cultures described in (A, B). β-Actin was used as a protein loading control.

Figure 5.

Hh pathway manipulation influences cell migration and inhibition reduces cell survival. (A) Triplicate cultures of the clonally derived rat HSC line 8B were treated with either the Hh pathway inhibitor, cyclopamine (5 μM) or its biologically inert analog, tomatidine (5 μM), for 24 hours. Cell migration was assessed using a standard wound healing assay. Representative results are shown. (B) Triplicate cultures of 8B were treated with adenoviral vectors containing constitutively active rac1 (V12rac1) or dominant-negative rac1 (N17rac1) and then treated with vehicle or Smo agonist (SAG) or tomatidine or cyclopamine, respectively. A standard wound healing assay was performed, and cell migration into the wound was examined 24 hours later. Representative results are shown. (C) Triplicate cultures of MF-HSC 8B were treated for 48 hours with either N17rac1 (to inhibit Rac1 activity) or cyclopamine (to block Hh signaling). Cell viability was assessed by way of CCK8 assay. Experiments were performed a minimum of three times. Data are presented as the mean ± SEM. P < 0.005 versus respective dl312-treated cultures or tomatidine-treated control cultures.

Rac1 Promotes Fibrosis in a BDL Model of Fibrosis.

To determine if a similar Rac1-regulated process occurs when injury provokes Q-HSCs to become MF-HSCs in intact animals, C57Bl6 mice were pretreated with dl312, V12rac1, or N17rac1 by portal venous injection, and then subjected to BDL or sham surgery. Surviving mice were sacrificed after 10 days. Compared with sham surgery, BDL caused liver injury and fibrosis in dl312-treated mice (data not shown). No mice treated with N17rac1 survived beyond 4 days after BDL, so the effects of inhibiting Rac1 activity could not be analyzed 10 days after BDL. However, compared with BDL mice pretreated with dl312, V12rac1-treated BDL mice demonstrated significantly greater mesenchymal gene expression (α-SMA, type 1 α1-collagen, and S100A4), as well as decreased expression of the EMT inhibitor BMP-7 (Fig. 6A). V12rac1-treated BDL mice also demonstrated greater Hh pathway activity than dl312-treated BDL mice (Fig. 6B,C). Consistent with these findings, BDL mice pretreated with V12rac1 demonstrated greater liver fibrosis, as evidenced both by Sirius Red staining of liver sections (Fig. 7A-C) and quantification of hepatic hydroxyproline content (Fig. 7D).

Figure 6.

Rac1 enhances Hh signaling and myofibroblastic markers in a BDL model of fibrosis. C57Bl6 mice were pretreated with adenoviral vectors for dl312 (n = 8 mice) or V12rac1 (n = 8 mice) by portal venous injection, subjected to BDL or sham surgery, and sacrificed after 10 days for liver harvest. Liver RNA and protein were analyzed by way of qRT-PCR and western blot analysis. (A, B) Messenger RNA expression of fibrosis and EMT markers (A), and Hh pathway (B). Data are presented as the ± SEM. *P < 0.05, P < 0.005 versus dl312-treated group. (C) Representative western blot analysis. β-actin confirms equal protein loading.

Figure 7.

Rac1 promotes fibrosis during BDL. (A,B) Sirius Red staining of liver sections from representative dl312-treated control mice (A) and V12rac1-treated mice (B) at 10 days after BDL. (C) Quantitative morphometry of Sirius Red–stained sections from all 16 mice. (D) Hepatic hydroxyproline content in both groups. Data are presented as the mean ± SEM. *P < 0.05 versus dl312-treated control group.

Overexpression of Rac1 Enhances Expression of Markers of EMT and Hh Pathway Activity in CCl4-Induced Liver Injury.

Although rac1 transgene expression was verified in primary HSCs that were isolated from rats that received portal vein injections of rac1 transgene-bearing adenoviral vectors (Fig. 8A) and results of the in vivo study recapitulated our findings in cultured primary HSCs and clonal HSC lines, the outcomes observed after BDL might have been unique to this type of liver injury and/or mediated by other liver cell types that were also transduced during the process. Therefore, responses to another type of chronic liver injury were examined in another animal model in which overactivation of Rac1 was targeted to α-SMA–expressing cells (myofibroblasts). In such rac1-transgenic mice, regulatory elements of the α-sma gene control expression of a human rac1 transgene that produces constitutively-active Rac1 protein, resulting in accumulation of activated Rac1 in α-SMA–expressing cells.8 We reported that rac1-transgenic mice developed greater liver fibrosis after 8 weeks of CCl4-induced liver injury.8 Analysis of livers from CCl4-treated rac1-transgenic mice demonstrated that selectively increasing Rac1 activity in α-SMA–expressing cells exacerbated injury-related increases in both Hh signaling and expression of various mesenchymal genes (such as α-SMA, type 1 α1-collagen, and S100A4), whereas it decreased expression of BMP-7 (Fig. 8B,D). Thus, results from studies that used two different approaches to activate Rac1 in HSCs in intact animals complement data that were generated by gain-of-function or loss-of-function studies in cultured primary HSCs and clonal HSC lines, providing compelling evidence that Rac1 activation is critically involved in both generation and maintenance of MF-HSCs.

Figure 8.

Overexpression of Rac1 in myofibroblastic cells increases hepatic expression of mesenchymal genes. Primary HSCs were isolated from animals 2 days after intraportal injection of adenoviral vectors carrying V12rac1, N17rac1, or dl312. Proteins were isolated and western blots were used to demonstrate the myc-tagged products of the two Rac transgenes. (A) Representative membrane that was probed with myc-tag primary antibody. (B) Rac1-transgenic mice (n = 11) were treated with CCl4 or vehicle (n = 6 mice) for 8 weeks. Liver RNA and protein were obtained for qRT-PCR and western blot analysis. (B) Messenger RNA levels of fibrosis and EMT markers and (C) Hh pathway. Data are presented as the mean ± SEM data. *P < 0.05, P < 0.005 versus vehicle-treated controls. (D) Western blot analysis. β-Actin was used as a loading control. Results shown are representative of triplicate immunoblot analyses.

Discussion

Hepatic accumulation of MF-HSCs is pivotal in the pathogenesis of cirrhosis. Two events are necessary for MF-HSCs to accumulate in damaged livers: transition of resident Q-HSCs to MF-HSCs, and expansion of MF-HSC numbers through increased proliferation and/or reduced apoptosis. Recently, we demonstrated that the first event (transition to MF-HSCs) involves a process that has classical features of EMT, and showed that HSC EMT is regulated by Hh signaling.21 Previously, we demonstrated that Hh ligands enhance the viability of MF-HSCs9 and reported that both Rac18 and Hh9 promote proliferation of MF-HSCs. However, it was unclear if Rac1 and Hh interacted to regulate the accumulation of MF-HSCs after liver damage. Therefore, in the present study, we evaluated the hypothesis that Rac1 promotes the activation of the Hh pathway, thereby stimulating signals that promote EMT in Q-HSCs and enhance the viability of MF-HSCs. Using both in vitro and in vivo systems, Rac1 activity was manipulated through adenoviral vector-mediated delivery of constitutively active or dominant-negative Rac1. Parallel studies were done with adenoviral vectors bearing an empty cassette to control for nonspecific effects of the adenovirus. Selected studies were repeated in transgenic mice in which overactivation of Rac1 was restricted to myofibroblastic cells. Results in all models demonstrated that increasing Rac1 activity enhanced Hh signaling, EMT, and fibrogenesis. Conversely, Rac1 inhibition reversed all of these effects in each model system examined. These findings extend existing knowledge about mechanisms that expand populations of MF-HSCs in injured livers, and prove that activated Rac1 and Hh are critical mediators of this process.

Many other factors are also known to play important roles in controlling the hepatic content of MF-HSCs. These include soluble growth factors, cytokines, chemokines, and their respective receptors, as well as pattern recognition receptors and their ligands.33 It is still unclear, however, whether there is a hierarchy of importance among these various mediators, or to what extent different factors might interact to modulate conserved signaling that results in the formation and/or growth of MF-HSCs. Previously published data demonstrate interactions between several of these factors and the Hh pathway. For example, transforming growth factor-β (TGF-β),10 platelet-derived growth factor (PDGF),9 and leptin (unpublished data) have been shown to induce different types of liver cells, including HSCs, to produce Hh ligands. TGF-β treatment activates the Hh pathway in cultured human A549 cells (which are used to interrogate EMT regulation)34 (Supporting Fig. 7); whether TGF-β, Rac1, and Hh interact to promote EMT has not yet been examined in HSCs. On the other hand, it is clear that the mitogenic actions of PDGF on MF-HSCs require Hh pathway activity, because blocking Hh signaling prevents PDGF from increasing proliferation of MF-HSCs.9 Similarly, leptin-mediated induction of myofibroblastic genes is blocked by inhibiting Hh signaling in HSCs, demonstrating that Hh pathway activity is also necessary for leptin to exert its fibrogenic effects (unpublished data). Interestingly, although primary HSCs from fa/fa rats (which have an inherited defect in the long form of the leptin receptor) are unresponsive to leptin-mediated fibrogenesis, they retain Hh pathway activity and remain capable of transitioning into MF-HSCs when cultured. This observation suggests that leptin, like PDGF and TGF-β, operates upstream of the Hh pathway and raises the possibility that HSCs may require a functional Hh pathway in order to respond optimally to other profibrogenic factors. This concept is supported by evidence that Hh pathway inhibition generally abrogates culture-induced transition of Q-HSCs into MF-HSCs and causes culture-activated MF-HSCs to reacquire a more quiescent (less myofibroblastic) phenotype.21 Similarly, treating cultured MF-HSCs with antibodies that neutralize endogenously produced Hh ligands dramatically reduces their viability, demonstrating that Hh ligands are autocrine viability factors for MF-HSCs.9 Hh ligands from MF-HSCs also act in a paracrine fashion to stimulate resident liver cells to produce factors that recruit HSCs (monocyte chemoattractant protein 1) and/or exert profibrogenic actions on HSCs (interleukin-4 and interleukin-13).35, 36 Thus, the Hh pathway operates at multiple levels to promote the formation of MF-HSCs from Q-HSCs and to enhance MF-HSC accumulation.

The current study proves that Rac1 controls Hh signaling in HSC and demonstrates that this occurs, in part, because Rac1 activation differentially regulates HSC production of Shh ligand and its inhibitor, Hhip. Rac1 activation induces Shh production while inhibiting expression of Hhip, skewing the ligand/inhibitor balance to favor activation of Hh signaling and consequent induction of Hh-mediated processes, including HSC survival and EMT. Fibrogenic stimuli, such as culture in serum-supplemented medium, are known to activate Rac1 in HSCs.8 The present study proves that blocking Rac1 activity inhibits activation of the Hh pathway, survival, and EMT in HSCs that are exposed to fibrogenic stimuli. Thus, it is reasonable to conclude that Rac1 functions as a common intermediate that permits profibrogenic stimuli to promote activation of the Hh pathway in HSCs.

Additional research is needed to delineate the mechanisms by which Rac1 activation alters HSC expression of Shh and Hhip. Rac1 regulates diverse cellular processes, including the cell cycle, cell-cell adhesion, and motility.3 However, to our knowledge, it has not yet been determined if these various outcomes are mediated by conserved Rac1-initiated signals. One function of Rac1 is to direct assembly of the nicotinamide adenine dinucleotide phosphate (reduced form) (NADPH) oxidase enzyme complex, which is an important source of reactive oxygen species (ROS) in many types of cells.4 NADPH oxidase activity increases as HSCs become myofibroblastic and inhibiting this enzyme also inhibits HSC activation, both in culture and in intact animal models of liver fibrosis.37 The concept that the profibrogenic effects of NADPH oxidase are mediated by ROS is further supported by evidence that antioxidants inhibit accumulation of MF-HSCs.38 Bone marrow chimeric mice were generated to prove that MF-HSCs were the most pathophysiologically relevant cellular source of NADPH oxidase-produced ROS in animal models of liver fibrosis.39 Given that Rac1 mediates NADPH oxidase assembly, their findings support our conclusion that it is the activation of Rac1 in HSCs (rather than some other type of liver cell) that drives MF-HSC accumulation and liver fibrogenesis after BDL or CCl4-induced liver injury.

Our literature search failed to identify any publications that link NADPH oxidase and the Hh pathway. However, transcription of Shh is known to be regulated by the redox-sensitive transcription factor nuclear factor-κB.40 Hence, it is conceivable that Rac1 might increase Shh expression in HSCs by activating NADPH oxidase, with resultant increase in ROS, activation of redox-sensitive transcription factors, and increased Shh transcription. Presumably, Hhip expression would be concomitantly repressed in a redox-sensitive manner, because Shh and Hhip are reciprocally regulated by Rac1 activation. Preliminary data do not support this concept, however, because antioxidants failed to attenuate the ability of V12rac1 to induce Shh in cultured HSCs (data not shown). Therefore, alternative mechanisms merit consideration, because PDGF induces HSC expression of Shh through mechanisms that are dependent upon activation of phosphoinositide 3-kinase/AKT.9 Considerable additional work will be required in order to delineate the precise molecular events that link Rac1 activation to altered expression of Hh pathway regulators, and to evaluate whether these mechanisms are important for stellate cell activation in extrahepatic tissues. Although these efforts extend beyond the scope of this publication, the present studies are important because they identified a previously unsuspected relationship between Rac1 and the Hh pathway in adult mammalian cells, thereby opening entirely novel lines of research that are likely to have general relevance to cellular biology and to liver fibrosis.

In conclusion, targeted manipulation of Rac1 activity in HSCs proves that activation of Rac1 is required for Q-HSCs to become—and remain—myofibroblastic. Rac1 may promote the myofibroblastic phenotype in HSCs through several mechanisms. Our results, however, demonstrate that key profibrogenic actions of Rac1 are mediated by its ability to activate the Hh pathway, with resultant Hh-dependent enhancement in cell viability and transition to a myofibroblastic phenotype through an EMT-like process. The findings also identify two of the early events in this process, namely Rac1 activation-dependent induction of Hh ligand production and coincident repression of the Hh ligand antagonist, Hhip. These discoveries have important implications for the pathogenesis of cirrhosis, because the latter results in large part from hepatic MF-HSC accumulation.

Ancillary