Inhibition of phosphatidylinositol 3-kinase signaling in hepatic stellate cells blocks the progression of hepatic fibrosis


  • Gakuhei Son,

    1. Division of Gastroenterology and Hepatology, Department of Medicine, University of North Carolina, Chapel Hill, North Carolina
    Search for more papers by this author
  • Ian N. Hines,

    1. Division of Gastroenterology and Hepatology, Department of Medicine, University of North Carolina, Chapel Hill, North Carolina
    Search for more papers by this author
  • Jeff Lindquist,

    1. Division of Gastroenterology and Hepatology, Department of Medicine, University of North Carolina, Chapel Hill, North Carolina
    Search for more papers by this author
  • Laura W. Schrum,

    1. Department of Biology, University of North Carolina, Charlotte, North Carolina
    Search for more papers by this author
  • Richard A. Rippe

    Corresponding author
    1. Division of Gastroenterology and Hepatology, Department of Medicine, University of North Carolina, Chapel Hill, North Carolina
    • CB #7032, Medical Biomolecular Research Building, Room 7340B, Division of Gastroenterology and Hepatology, Department of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7032
    Search for more papers by this author
    • fax: 919-966-7468

  • Potential conflict of interest: Nothing to report.


The hepatic stellate cell (HSC) is the primary cell type in the liver responsible for excess collagen deposition during fibrosis. Following a fibrogenic stimulus the cell changes from a quiescent vitamin A–storing cell to an activated cell type associated with increased extracellular matrix synthesis and increased cell proliferation. The phosphatidylinositol 3-kinase (PI3K) signaling pathway has been shown to regulate several aspects of HSC activation in vitro, including collagen synthesis and cell proliferation. Using a targeted approach to inhibit PI3K signaling specifically in HSCs, we investigated the role of PI3K in HSCs using a rodent model of hepatic fibrosis. An adenovirus expressing a dominant negative form of PI3K under control of the smooth muscle α-actin (αSMA) promoter was generated (Ad-SMAdnPI3K). Transducing HSCs with Ad-SMAdnPI3K resulted in decreased proliferation, migration, collagen expression, and several additional profibrogenic genes, while also promoting cell death. Inhibition of PI3K signaling was also associated with reduced activation of Akt, p70 S6 kinase, and extracellular regulated kinase signaling as well as reduced cyclin D1 expression. Administering Ad-SMAdnPI3K to mice following bile duct ligation resulted in reduced HSC activation and decreased extracellular matrix deposition, including collagen expression. A reduction in profibrogenic mediators, including transforming growth factor beta, tissue inhibitor of metalloproteinase 1, and connective tissue growth factor was also noted. However, liver damage, assessed by alanine aminotransferase levels, was not reduced. Conclusion: Inhibition of PI3K signaling in HSCs during active fibrogenesis inhibits extracellular matrix deposition, including synthesis of type I collagen, and reduces expression of profibrogenic factors. These data suggest that targeting PI3K signaling in HSCs may represent an effective therapeutic target for hepatic fibrosis. (Hepatology 2009.)

Liver fibrosis represents a wound-healing process in response to a variety of acute and chronic stimuli, including ethanol, viral infections, cholestasis, and metabolic diseases.1, 2 Characterized by excess synthesis and deposition of extracellular matrix (ECM), fibrosis disrupts the normal architecture of the liver, leading to organ dysfunction, which can progress to cirrhosis and ultimately organ failure if left untreated. Currently, few therapeutic strategies exist to treat liver fibrosis.

The hepatic stellate cell (HSC) is primarily responsible for excess deposition of ECM proteins during fibrosis. Following a fibrogenic stimulus, HSCs lose their retinoid stores, proliferate, express smooth muscle α-actin (αSMA), and produce large amounts of ECM proteins, including type I collagen. Phosphatidylinositol 3-kinase (PI3K) is a key signaling molecule composed of an 85-kDa regulatory subunit and a 110-kDa catalytic subunit that is recruited to and activated by the activated platelet derived growth factor (PDGF) receptor following HSC activation and growth factor stimulation.3 The importance of PI3K signaling during the fibrogenic response in HSCs is becoming clear. Blocking PI3K activity, using either pharmacological (LY294002) or by a genetic approach, inhibits HSC proliferation and collagen gene expression through interruption of key downstream signaling pathways, including Akt and p70 S6 kinase (p70S6K).4, 5 Activation of Akt is associated with HSC proliferation and α1(I) collagen transcription and translation.4, 6–8 Likewise, interruption of the p70S6K/mTOR (mammalian target of rapamycin) signaling suppresses HSC proliferation. Therefore, interruption of PI3K signaling is capable of suppressing key components of HSC activation and proliferation and may represent a therapeutic target for treating hepatic fibrosis.

Because αSMA expression is induced following HSC activation, we developed an αSMA-driven expression vector to direct expression of a dominant negative form of PI3K to prevent HSC proliferation and collagen expression in activated HSCs. We demonstrate that inhibition of PI3K signaling in vivo, specifically in HSCs, dramatically attenuates experimentally induced hepatic fibrosis in mice.


αSMA, smooth muscle α-actin; ALT, alanine aminotransferase; βgal, β-galactosidase; BDL, bile duct ligation; CCl4, carbon tetrachloride; CTGF, connective tissue growth factor; DMEM, Dulbecco's modified Eagles medium; ECM, extracellular matrix; ERK, extracellular regulated kinase; FBS, fetal bovine serum; GFP, green fluorescent protein; HA, hemagglutinin; HSC, hepatic stellate cells; MAKP, mitogen activated protein kinase; MOI, multiplicity of infection; mRNA, messenger RNA; mTOR, mammalian target of rapamycin; p70S6K, p70 S6 kinase; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; PDGF, platelet derived growth factor; PI, propidium iodide; PI3K, phosphatidylinositol 3-kinase; PPARγ, peroxisome proliferator activated receptor γ; SEM, standard error of the mean; RT-PCR, reverse transcription polymerase chain reaction; TBS-T, Tris-buffered saline + Tween-20; TGF-β, transforming growth factor beta; TIMP-1, tissue inhibitor of metalloproteinase.

Materials and Methods

Adenovirus Preparation.

The dnPI3K coding sequence was amplified from Ad-dnPI3K9 using polymerase chain reaction (PCR) with a forward primer 5′-gatcatgatcggatccccaccatgtacccatacgatgttccaga-3′, containing the hemagglutinin (HA)-tag (bold), and a reverse primer 5′-gatcatgatcggatcctcatcgcctctgctgcgcgt-3′. The amplified product was digested with Bam HI and cloned into the Bam HI site in the pDNR plasmid (BD Bioscience, San Jose, CA). The insert was sequenced to assess the orientation of the insert and the correct sequence. The αSMA promoter was amplified from the pSMP8 plasmid10 using forward primer 5′-gatcatgatcgaattcacaccataaaacaagtgcatgag-3′ and reverse primer 5′-gatcatgatcgaattcagctgcaccagcgtctcagg-3′ and the PCR product cloned into the pTOPO plasmid (Invitrogen, Carlsbad, CA). The αSMA promoter was excised from the pTOPO plasmid using Eco RI and cloned into the Eco RI site in dnPI3K-pDNR upstream of the dnPI3K insert. Sequencing confirmed orientation and integrity of the insert. The recombinant adenovirus was generated using the BD Adeno-x Expression System Promoterless Vector (BD Bioscience) according to the manufacturers' protocol. HSCs transduced with the recombinant adenovirus expressed the HA-tag, as assessed by western blot analysis. Hepatocytes, endothelial cells, and Kupffer cells did not express the HA-tag when transduced by Ad-SMAdnPI3K, thus confirming cell-specific expression (data not shown).

Hepatic Stellate Cell Isolation.

Mouse HSCs were isolated as described11, 12 from pCOL9GFP-HS4,5 transgenic mice (collagen promoter-driven green fluorescent protein (GFP) transgenic mice) on a BALB/c background that have been described.13 Isolated HSCs were cultured in Dulbecco's modified Eagles medium (DMEM) supplemented with 10% fetal bovine serum (FBS) in a 95% air / 5% CO2 atmosphere. Hepatocyte, endothelial cell, and Kupffer cell isolation procedures are provided in the supporting materials. All animal procedures were performed under the guidelines set by the University of North Carolina Institutional Animal Care and Use Committee and are in accordance with those set by the National Institutes of Health.

GFP Quantification.

HSCs from pCol9GFP-HS4,5 mice were plated in 6-well plates (2 × 105 cells/well). After 5 days, cells were transduced with Ad-SMAdnPI3K, Ad-βgal (β-galactosidase), or no virus. After 12 hours the medium was changed to media containing 10% FBS. After 48 hours plates were scanned for GFP expression and quantified by phosphorimager analysis according to the recommended protocol of the manufacturer (GE Healthcare, Piscataway, NJ).

Fibrosis Model.

pCol9GFP-HS4,5 transgenic mice on a BALB/c background were subjected to double ligation of the common bile duct (BDL).14 Sham-treated animals underwent the same procedure without ligation. Some mice were administered carbon tetrachloride (CCl4) to induce fibrosis, with specific procedures provided in the supporting materials. All procedures were in accordance and approved by the Institutional Animal Care and Use Committee guidelines at the University of North Carolina at Chapel Hill. Ad-SMAdnPI3K, 1 × 109 plaque-forming units (PFU), was administered 4 days following BDL by way of tail vein injection. Mice were sacrificed 9 or 18 days after BDL and blood and liver samples were harvested.

Assessment of Apoptosis.

In vitro, cell death was determined by propidium iodide (PI, Sigma, St. Louis, MO) and Hoechst 33342 (Sigma) double fluorescent staining. HSCs were isolated from pCol9GFP-HS4,5 transgenic mice and cultured in 6-well plates. After 5 days in culture, HSCs were transduced with Ad-SMAdnPI3K, Ad-βgal, no virus, as above. After 24 hours the cells were incubated with Hoechst (1 μg/mL) and PI (5 μg/mL) for 20 minutes. Afterwards, cells were visualized by fluorescent microscopy. Paraffin-embedded liver sections were stained for DNA fragmentation using the In Situ Cell Death Detection Kit, TMR red (Roche, Indianapolis, IN) according to the manufacturer's recommendations.

Histological Analysis and Immunohistochemistry.

Portions of the liver were collected at the time of sacrifice and fixed in 10% buffered formalin for 24 hours at 4°C. Expression of αSMA was determined in paraffin-embedded sections by immunohistochemistry as described.15 For fibrosis, liver sections were stained with Picrosirius Red (Sigma) as described.15 For immunofluorescent staining, liver specimens were fixed using 4% paraformaldehyde (PFA) for 16 hours at 4°C, incubated in PBS containing 30% sucrose 24 hours, then frozen at −80°C in OCT Compound (Sakura Finetek USA, Torrance, CA). Sections were incubated with monoclonal antibody to HA (1:100; Cell Signaling Technology, Danvers, MA) using the MOM kit (Vector Laboratories, Burlingame, CA) and Alexa Fluor 546 Streptavidin (Invitrogen) used for visualization. Hepatic GFP expression was assessed using frozen tissue sections and viewed by fluorescent microscopy. For triple staining, PFA-fixed frozen sections were stained with anti-GFP (chicken polyclonal; Aves Labs), anti-HA (mouse monoclonal), and anti-desmin (rabbit polyclonal; Thermo Scientific), all at a dilution of 1:100. For visualization, goat antichicken conjugated to FITC (to stain GFP), goat antirabbit conjugated to AlexaFluor 350 (to stain desmin), and goat antimouse conjugated to AlexaFluor 546 (to visualize HA) were incubated on the sections for 30 minutes at room temperature. For immunofluorescence cell staining, 24 hours after Ad-SMAdnPI3K transduction (multiplicity of infection [MOI] 300), HSCs were fixed with 4% PFA and −20°C methanol (10 minutes, room temperature). The cells were then incubated with anti-HA antibody 1:100 using the MOM Kit. Sirius red–positive area, αSMA-positive area, and GFP-positive area were quantified using Image J software (NIH Image; from five random nonoverlapping 100× fields for each animal with six animals assessed in each group.


Protein extracts were prepared from liver tissue (50 μg) or cultured HSCs (20 μg) using RIPA buffer and separated on 8%-16% Tris-glycine gels and transferred to polyvinylidene fluoride membranes as described.16 Antibodies against phospho-PI3K (Tyr458), PI3K, phospho-AktSer473, Akt, phospho-p70S6K (Thr421/Ser424), p70S6K, phospho-Erk (Thr202/Tyr204), Erk, Cyclin D1, CD31, HA-Tag (1:1,000; all from Cell Signaling), αSMA (1:5,000; Sigma), albumin (1:5,000; ICN), F4/80 (1:1,000; Serotec), and β-actin (1:1000; Sigma) were used for western blot procedures, incubated overnight at 4°C in Tris-buffered saline containing 0.05% Tween-20 (TTBS) and 5% nonfat dry milk. Appropriate horseradish peroxidase–conjugated secondary antibodies diluted 1:2,000-5,000 in TTBS + 5% nonfat dry milk were then incubated with blots for 1 hour at room temperature. Antibody staining was visualized by enhanced chemiluminescence (ECL plus, Amersham, Indianapolis, IN) exposed to film. Band density was quantified from digital images using Image J software.

Other Methods.

Procedures for cell isolation, proliferation, and migration, hydroxyproline and alanine aminotransferase (ALT) assay, western blotting, and reverse-transcription polymerase chain reaction (RT-PCR) are provided in the supporting materials.

Statistical Analysis.

All data are presented as means ± SEM (standard error of the mean). Differences between the Ad-SMAdnPI3K group and control groups were assessed by unpaired Student's t test. P-values <0.05 were considered statistically significant.


Inhibition of PI3K Signaling Reduces Cell Proliferation and Migration in Isolated HSCs.

To determine the role of PI3K signaling in HSCs in vivo, we constructed an adenovirus that allows for HSC-specific expression of a dominant negative form of PI3K9 using the αSMA promoter.10 This recombinant adenovirus expressed the HA-tag in HSCs, but not in hepatocytes, Kupffer cells, or endothelial cells (Fig. 1A). In isolated HSCs, αSMA expression was first detected following 2 days in culture (data not shown). Therefore, we transduced isolated HSCs with Ad-SMAdnPI3K at Day 2 and noted weak expression of the HA-tag 1 day after viral transduction, which progressively increased by 5 days in culture (data not shown). Therefore, for our in vitro studies, we transduced HSCs after 4 days in culture to assure high levels of expression of HA-dnPI3K from the αSMA promoter.

Figure 1.

Inhibition of PI3K signaling with Ad-SMAdnPI3K reduces HSC proliferation, migration, and collagen gene expression. (A) Each cell was cultured in proper medium containing 10% FBS for 5 days and was transduced at an MOI of 300 with Ad-SMAdnPI3K. After 48 hours of incubation, whole cell extracts were prepared for western blots of HA, albumin, F4/80, CD31, αSMA, and β-actin. HC: hepatocyte, KC: Kupffer cell, LSEC: liver sinusoidal endothelial cell, HSC: hepatic stellate cell. Results are representative of three independent experiments. (B) Culture-activated HSCs were treated with PBS or transduced with Ad-βgal or Ad-dnSMAPI3K and cell proliferation assessed after 24 hours, 48 hours, and 72 hours following stimulation with 10% FBS. P = 0.0007, P = 0.0006, P = 0.0006, respectively, compared to Ad-βgal. (C) Cell migration, using a wound assay, was assessed in culture-activated HSCs treated with PBS in cells transduced with Ad-βgal or Ad-SMAdnPI3K after 24 hours. P = 0.0008 compared to Ad-βgal. (D) Colocalization of GFP (green) and HA (red) expression in HSCs isolated from pCol9GFP-HS4,5 transgenic mice transduced with Ad-SMAdnPI3K. (E) GFP expression, assessed by fluorescence microscopy, in HSCs isolated from pCol9GFP-HS4,5 mice 2 days following Ad-SMAdnPI3K transduction. Original magnification, ×100 with quantification of GFP fluorescence using phosphorimager analysis, in culture-activated HSCs 24 hours, 48 hours, and 72 hours following PBS treatment or transduction of pCol9GFP-HS4,5 HSCs with either Ad-βgal or Ad-SMAdnPI3K. Original magnification ×200. (F) Steady-state levels of α1(I) collagen mRNA was determined by quantitative real-time RT-PCR. Data are expressed as the mean ± SEM from three independent experiments. *P < 0.01 versus Ad-βgal transduced cells.

Previous studies have shown that PI3K positively regulates rat HSC proliferation.4 To confirm that similar effects occur in murine HSCs, mouse HSCs were isolated and cultured for 4 days. Cells were left untreated or transduced with Ad-βgal or Ad-SMAdnPI3K in serum-free media for 48 hours. Cell proliferation was stimulated using media containing 10% FBS and cells monitored for 72 hours. Control cells, or cells transduced with Ad-βgal, showed similar increases in growth rate during the experimental period; however, cells transduced with Ad-SMAdnPI3K showed a significant reduction in cell numbers compared to Ad-βgal–transduced cells (Fig. 1B).

To determine whether PI3K signaling influences HSC migration, a wound assay was performed and the distance cells migrated from the wound was assessed following a 24-hour culture period. Control HSCs or cells transduced with Ad-βgal showed similar migratory responses; however, cells transduced with Ad-SMAdnPI3K showed a significant reduction in cell migration compared to control Ad-βgal–transduced cells (Fig. 1C).

Inhibition of PI3K Decreases Expression of α1(I) Collagen in HSCs Mediated in Part by Inhibiting Gene Transcription.

We first examined the expression pattern of the HA-tag present on dnPI3K and GFP, a marker for collagen expression in HSCs from pCol9GFP-HS4,5 transgenic mice transduced with Ad-SMA-dnPI3K. Some cells expressed only GFP, other cells only expressed the HA-tag; however, only a few cells expressed both the HA-tag and GFP (Fig. 1D). This result is similar to that previously reported where HSCs express either αSMA or collagen, with only a few cells expressing both genes.10

To confirm that blocking PI3K also inhibits α1(I) collagen gene expression in murine HSCs, similar to that shown in rat HSCs,4 we examined the effect of Ad-SMAdnPI3K on collagen expression in HSCs from pCol9GFP-HS4,5 transgenic mice. HSCs were either left untreated or transduced with Ad-βgal or Ad-SMAdnPI3K after 5 days in culture. A significant reduction in GFP fluorescence, a marker for collagen expression in these cells, was observed following viral transduction with Ad-SMAdnPI3K (Fig. 1E); a similar reduction in α1(I) collagen messenger RNA (mRNA) expression was also observed (Fig. 1F). Therefore, blocking PI3K activity in HSCs inhibits α1(I) collagen gene expression, at least partially at the transcriptional level.

Inhibiting PI3K Induces Cell Death in Murine HSCs.

Because HSC proliferation was reduced when PI3K activity was blocked (Fig. 1B), we examined whether this inhibition is associated with increased cell death. HSCs were isolated and cultured for 5 days, then transduced with Ad-βgal or Ad-SMAdnPI3K. Transduction with Ad-SMAdnPI3K markedly increased PI staining compared to cells transduced with Ad-βgal (Fig. 2A-F). Therefore, inhibition of PI3K activity in activated HSCs promoted membrane permeability and nuclear envelope disruption, processes associated with cell death.

Figure 2.

Transduction of Ad-SMAdnPI3K induces cell death in activated HSCs and reduces profibrotic gene expression. Hoechst 33342 (A-C) and PI (D-F) staining in culture-activated (Day 5) pCol9GFP,HS4,5 HSCs. (A,D) control cells; (B,E) cells transduced with Ad-βgal; (C,F) cells transduced with Ad-SMAdnPI3K. After 24 hours cells were incubated with Hoechst (1 μg/mL) and PI (5 μg/mL) for 20 minutes. Figures are representative data from three independent experiments. Original magnification ×100. (G) Culture-activated (Day 5) HSCs from pCol9GFP-HS4,5 mice were transduced with Ad-SMAdnPI3K, Ad-βgal, or no virus and incubated for 48 hours. Steady-state mRNA levels for TGF-β, αSMA, TIMP-1, CTGF, and PPARγ were quantified by RT-PCR. Data are expressed as the mean ± SEM from three independent experiments. *P<0.05 versus Ad-βgal-transduced cells.

Expression of Genes Associated with a Fibrogenic Response Is Reduced in HSCs Following Inhibition of PI3K Signaling.

Because inhibiting PI3K activity reduced collagen gene expression in HSCs, we examined the effect of PI3K inhibition on the expression of several other profibrogenic genes. HSCs were cultured for 5 days, then either left untreated or transduced with Ad-SMAdnPI3K or Ad-βgal. After 48 hours, mRNA expression of transforming growth factor beta (TGF-β), αSMA, tissue inhibitor of metalloproteinase (TIMP-1), and connective tissue growth factor (CTGF) were significantly reduced in cells transduced with Ad-SMAdnPI3K, compared to Ad-βgal transduced cells (Fig. 2G). In contrast, expression of peroxisome proliferator activated receptor γ (PPARγ) was increased 2.3-fold (Fig. 2G). These findings correlate with a reduced fibrogenic response of the activated HSC when PI3K signaling is interrupted.

Inhibition of PI3K Signaling in Activated HSCs Blocks Activation of Akt and p70S6K Signaling and Cyclin D1 Expression.

To examine a potential mechanism by which inhibition of PI3K signaling may mediate an antifibrogenic response, we assessed the effects of blocking PI3K on extracellular regulated kinase (ERK), PI3K, Akt, and p70S6K signaling pathways, known to influence the fibrogenic response in HSCs. Day 5 HSCs were transduced with Ad-βgal or Ad-SMAdnPI3K, serum-starved for 24 hours, then stimulated with 10% serum and activation of PI3K and mitogen-activated protein kinase (MAPK) signaling analyzed. Increased phosphorylation of the PI3K p85 subunit was observed after 30 minutes in both PBS and Ad-βgal–treated cells (Fig. 3A). In contrast, phosphorylation of PI3K, Akt, and p70S6K was completely blocked in cells transduced with Ad-SMAdnPI3K, but not in PBS-treated or Ad-βgal–transduced cells. Phosphorylation of ERK was only modestly attenuated by Ad-SMAdnPI3K (Fig. 3A).

Figure 3.

Inhibition of PI3K activity blocks the activation of PI3K/Akt/p70S6K signaling and reduces cyclin D1 expression. (A,B) HSCs from pCol9GFP-HS4,5 were cultured for 5 days and then transduced with indicated adenoviruses. Cells were serum-starved for 48 hours, then stimulated with 10% FBS. Whole cell extracts were prepared at the indicated timepoints and western blots for P-PI3K, PI3K, P-Akt, Akt, P-p70S6K, P-ERK, ERK, cyclin D1, and β-actin performed. Results are representative of three independent experiments.

To investigate a potential mechanism for reduced HSC proliferation when PI3K signaling is blocked, we examined the influence of PI3K signaling on cyclin D1 expression, previously shown to regulate HSC proliferation.17, 18 Isolated HSCs were cultured for 5 days, transduced with Ad-βgal or Ad-SMAdnPI3K, then serum-starved for 48 hours. Cells were then stimulated to proliferate using 10% serum for 24 hours and cyclin D1 gene expression assessed. Control cells and cells transduced with Ad-βgal showed increased cyclin D1 expression over the 24-hour time period (Fig. 3B). In contrast, cells transduced with Ad-SMAdnPI3K failed to induce cyclin D1 expression (Fig. 3B), thus demonstrating that PI3K induces cyclin D1 expression in activated HSCs, which is partially responsible for inducing HSC proliferation in activated HSCs.

Expression of αSMA Driven dnPI3K in Experimentally Induced Hepatic Fibrosis.

To examine cell-specific expression of Ad-SMAdnPI3K during experimentally induced liver fibrosis, pCol9GFP-HS4,5 transgenic mice underwent BDL and 4 days later were administered PBS or 1 × 109 PFU of either Ad-βgal or Ad-SMAdnPI3K. Two days following adenovirus administration expression of HA was detected in liver extracts from mice administered Ad-SMAdnPI3K, whereas no HA expression was detected in the PBS- or Ad-βgal-treated animals (Fig. 4A). BDL induced significant GFP expression in both PBS- and Ad-βgal-treated mice; however, mice administered Ad-SMAdnPI3K showed a marked attenuation of GFP fluorescence, indicative of reduced collagen gene expression (Fig. 4B). Interestingly, GFP expression did not appreciably colocalize to expression of the HA tag expressed by Ad-SMAdnPI3K. A discordant expression of collagen expression and αSMA expression in activated HSCs has been reported.10 Because expression of dnPI3K in cultured HSCs induced cell death (Fig. 2A-F), we examined if HSC death was induced in the Ad-SMAdnPI3K-treated animals. Indeed, HSC death was increased, assessed by transferase dUTP nick end labeling (TUNEL) staining, in the animals in which Ad-SMAdnPI3K was administered, but not in PBS- or Ad-βgal-treated animals (Fig. 4C). To further assess the colocalization in vivo of the adenovirus within HSCs, triple staining for GFP, the marker of collagen expression, desmin, a marker of hepatic stellate cells, and HA, a marker of virus expression was performed on Day 6, 2 days following adenoviral administration. As shown in Fig. 4D, there is appreciable colocalization of desmin, collagen-GFP, and HA at this timepoint following adenoviral administration. Given the effects of PI3K inhibition of HSC survival both in vitro and in vivo, we chose to also look at colocalization of HA and GFP at an earlier timepoint hypothesizing that HSC death had already occurred 4 days following adenovirus administration. Shown in Supporting Fig. 1, colocalization of HA and GFP could be observed 1 day following adenovirus administration (Day 5 post-BDL). Together, these data demonstrate the ability of dnPI3K driven by the αSMA promoter to be expressed in stellate cells in vivo and induce their apoptosis within the cholestatic liver.

Figure 4.

Ad-SMAdnPI3K expresses the HA-tag in the liver and induces apoptosis in fibrotic liver. (A) BDLs were performed on Day 0 and adenovirus administered on Day 4. Proteins were collected from whole liver 2 days after adenoviral administration. HA-tag expression was detected in the liver extracts by western blot analysis. (B) HA expression was assessed 2 days after adenoviral administration by immunofluorescent staining using a monoclonal HA antibody (Red). Original magnification, ×200. (C) Liver cell death was assessed on Day 9 after BDL TUNEL (Day 5 after adenoviral transduction) in liver tissue. Original magnification, ×200. (D) Immunofluorescence staining shows localization of collagen (green), desmin (blue), and HA (red) 2 days after adenoviral administration. Original magnification, ×200. n = six animals per group.

Inhibiting PI3K Signaling Attenuates Hepatic αSMA Expression in BDL-Induced Liver Fibrosis.

To determine the extent of activated HSCs in the liver following BDL-induced liver fibrosis, immunohistochemical analysis for αSMA expression, a classical marker of activated HSCs, was performed. Induction of liver fibrosis resulted in an increase in αSMA expression in the livers of animals treated with either PBS or Ad-βgal 9 and 18 days following BDL, respectively (Fig. 5A,B). Both significant interstitial and vascular αSMA expression was observed. αSMA expression was attenuated in the Ad-SMAdnPI3K-treated animals where the interstitial expression pattern was mostly resolved, but vascular expression remained relatively high (Fig. 5A,B). Western blot analysis confirmed the immunohistochemical staining (Fig. 5C).

Figure 5.

Transduction of Ad-SMAdnPI3K inhibits αSMA expression in BDL-treated mice. (A) αSMA expression was examined in paraffin-embedded liver tissue of pCol9GFP-HS4,5 mice 9 days after BDL (5 days after adenoviral transduction) and 18 days after BDL (14 days after adenoviral transduction) stained using an αSMA antibody. Original magnification, ×100. (B) Quantification of αSMA-stained area. (C) Western blot analysis for αSMA expression 18 days after BDL (14 days after adenoviral transduction). n = six animals per group.

Inhibition of PI3K Signaling Attenuates Collagen Expression and Deposition Following BDL.

Given the profound impact of PI3K inhibition on HSC αSMA expression in vivo following BDL, expression of collagen was examined. As shown in Fig. 6A,B, Sirius red staining was significantly reduced in Ad-SMAdnPI3K-treated mice subjected to BDL for 9 or 18 days when compared to control-treated or Ad-βgal-treated mice subjected to BDL. Moreover, tissue hydroxyproline levels were significantly reduced following 18 days of BDL in AD-SMAdnPI3K when compared to control or βgal-treated mice subjected to BDL (Fig. 6C). GFP expression was significantly reduced as was steady-state α1(I) collagen mRNA expression following either 9 or 18 days of BDL, compared to control treated or Ad-βgal-treated mice subjected to BDL (Fig. 6D-F). Consistent with its effects in the BDL model of liver fibrosis, administration of Ad-SMAdnPI3K reduced fibrogenesis in the CCl4 model of liver injury and fibrosis to a similar degree (Supporting Fig. 2). Interestingly, a dnPI3K driven by the CMV promoter, and thus expressed in all cells, led to significant reduction in collagen gene expression (Supporting Fig. 3) in conjunction with significant morbidity in animals in both models of liver fibrosis, a process that could be correlated with a significant reduction in functional liver mass. These data demonstrate the ability of PI3K to support fibrogenesis and highlight important cell-specific functions of this signaling protein. Additional studies are under way to understand the importance of PI3K in other cell types, including hepatocytes within the chronically damaged liver.

Figure 6.

Administration of Ad-SMAdnPI3K attenuates liver fibrosis in BDL-treated mice. (A) Paraffin-embedded tissue of pCol9GFP-HS4,5 mice 9 days after BDL (5 days after adenoviral transduction) and 18 days after BDL (14 days after adenoviral transduction) were stained with Sirius red. Original magnification, ×100. (B) Quantification of Sirius red-stained area. (C) Quantitation of hepatic hydroxyproline content. (D) GFP expression in pCol9GFP-HS4,5 mice after 9 days and 18 days following BDL (5 days and 14 days, respectively, after adenoviral transduction) visualized by fluorescent microscopy. Sham tissue was examined at Day 9. Original magnifications, Day 18 ×40, Sham and BDL Day 9 ×100. (E) Percentage of GFP-positive area, in ×100 fields, was quantified using the NIH Image J software. (F) Steady-state α1(I) collagen mRNA levels at Day 18 after BDL (n = six animals per group). *P<0.05 versus Ad-βgal-treated mice.

Ad-SMAdnPI3K Inhibits Hepatic Expression of Fibrogenic Genes During BDL-Induced Liver Fibrosis, but Does Not Reduce Liver Injury.

To assess the effect of Ad-SMAdnPI3K on expression of other genes associated with liver fibrosis, we examined the expression of TGF-β, αSMA, TIMP-1, and CTGF in the livers of animals from each experimental group. Following 9 days of BDL, expression of TGF-β, αSMA, TIMP-1, and CTGF were all increased and all were significantly reduced in animals administered Ad-SMAdnPI3K, but not in control animals (Fig. 7A). Liver injury was also increased in the livers of the experimental animals; however, despite a reduction in the progression of fibrosis, ALT levels were not reduced following administration of Ad-SMAdnPI3K (Fig. 7B).

Figure 7.

Administration of Ad-SMAdnPI3K inhibits profibrotic gene expression in the livers of BDL-treated mice, but does not affect ALT levels. (A) Steady-state mRNA levels of TGF-β, αSMA, TIMP-1, and CTGF 9 days after BDL (Day 5 after adenoviral transduction) as determined by quantitative real-time RT-PCR. Data are presented as mean ± SEM from six animals per group. *P<0.05 versus animals administered Ad-βgal. (B) Serum ALT levels 9 days after BDL (Day 5 after adenoviral transduction; n = six animals per group).


Liver fibrosis is a key risk factor for the development of cirrhosis and chronic liver failure. Activation of HSCs is a crucial component of this process.1, 2 Here we show the importance of PI3K signaling in the development and progression of hepatic fibrogenesis in vivo and during HSC activation, proliferation, and profibrogenic collagen production in vitro. In doing so, an HSC-specific adenoviral vector was created that directed strong selectivity for expression in activated, smooth muscle actin–positive HSCs.

PI3K represents a key signaling molecule that controls many cellular functions such as proliferation, survival, adhesion, and migration.3, 4 In the liver, macrophage-associated PI3K activation promotes cytokine production and subsequent hepatocyte proliferation early following partial hepatectomy.19 Hepatocyte-associated PI3K regulates growth following a reduction in liver volume, a process involving Akt activation. In the setting of fibrosis, PI3K is activated by way of PDGF receptor ligation following CCl4 treatment and correlates with collagen production.20 In the present study, inhibition of PI3K activity, by way of a dominant negative form of PI3K within HSCs, reduced hepatic fibrogenesis following BDL. Inhibition of fibrogenesis was associated with reduced TGF-β expression, αSMA expression, and collagen production. Several mechanism(s) may exist by which PI3K might promote this activation. First, PI3K-induced Akt activation leads to cell survival and proliferation. Sustained activation of Akt has been shown to induce significant cellular proliferation, likely involving p70S6K activation and increased cyclin D1 expression.4 This is consistent with the present study showing that inhibition of PI3K in HSCs in vitro decreased cyclin D1 expression and reduced cellular proliferation. Lending to the connectivity of this pathway, inhibition of PI3K in HSCs inhibited p70S6K activation. Previous studies have shown that inhibition of the mTOR/p70S6K pathway by rapamycin in HSCs reduces G1 to S phase progression.17 Therefore, inhibiting PI3K, and thus its downstream signaling mediators, significantly impacts cell growth and proliferation.21–23

In addition to reduced proliferation, impairment of PI3K activity in HSCs was associated with the induction of cell death. Interruption of PI3K activity in HSCs led to a substantial increase in TUNEL staining in vivo and membrane permeability in vitro, suggesting the induction of cell death rather than a transition back to quiescence. The mechanism(s) promoting this response are not clear, although inhibition of MAPK pathways, specifically ERK, could be partially responsible. ERK activation is associated with suppression of caspase induction and mitochondrial permeability changes.24 Moreover, Akt activation suppresses BAD function, a proapoptotic factor.25 Indeed, it has been postulated that clearance of activated HSCs is a critical step in fibrosis resolution within the liver and, from data presented here, PI3K inhibition may represent a novel mechanism for their elimination.

A third mechanism by which PI3K may function to promote the fibrogenic response is through the regulation of profibrogenic mediator production by HSCs. Interruption of PI3K activity suppressed TGF-β and PDGF expression by HSCs in vitro and hepatic expression of these factors during cholestasis and their production of these factors likely promotes autocrine activation. Moreover, PI3K activation may suppress PPARγ expression. Previous studies have demonstrated the importance of PPARγ in suppressing hepatic fibrogenesis in vivo and HSC activation in vitro.26In vitro, PPARγ expression is reduced as HSCs become activated and its overexpression leads to suppression of collagen production and αSMA expression, a process involving inhibition of p300 transactivation.27–29 In our study, restoration of PPARγ function could be one mechanism by which PI3K inhibition leads to suppression of collagen production and HSC activation. In sum, interruption of PI3K activity is capable of influencing multiple cellular levels, from cell survival and proliferation to the promotion of growth factor production, to drive hepatic fibrogenesis.

Another interesting finding in the present study was the observed heterogeneity of HSCs both in vitro and in vivo. Previous studies by Magness et al.10 identified cells within the damaged liver that expressed either αSMA, collagen, or both, suggesting the potential for either multiple cell populations with characteristics similar to HSCs. In the setting of BDL, it is clear that extrahepatic myofibroblasts are recruited periportally and contribute significantly to hepatic fibrogenesis, in the absence of significant αSMA expression.30 For in vitro studies, however, HSCs are enriched from undamaged livers. Similar to the in vivo studies in the BDL model, HSCs express either αSMA or collagen or both. Thus, the use of αSMA-driven dnPI3K may only function in a subset of the collagen-producing cells within the liver. Alternatively, when PI3K activity is blocked in HSCs, cell death is induced (Fig. 2A-F). Nevertheless, our approach has led to a large and significant reduction in collagen expression and deposition, and inhibition of profibrogenic growth factor production, both in vitro and in vivo. The overall significance of multiple HSC populations, or intermediates within this lineage, remains to be determined.

In summary, we have demonstrated that selective inhibition of PI3K in activated HSCs, by way of a αSMA-driven dominant negative form of PI3K, reduces hepatic fibrogenesis, likely through suppression of fibrotic gene transcription, as well as through promotion of HSC death. In vitro studies further demonstrate the ability of PI3K inhibition to suppress HSC proliferation, migration, and profibrogenic gene expression, processes that likely involves interruption of downstream signaling molecules including Akt and p70S6K. Furthermore, this inhibition appears to be critical for the induction of HSC death, thereby eliminating important collagen-producing cells from the liver. In addition, the current investigation has also revealed, as previous studies would indicate, a potential heterogeneity of collagen-producing cells within the liver, specifically the presence of αSMA-negative collagen-producing cells. Finally, the current study has led to the creation and validation of an activated HSC-specific adenoviral vector that will provide a platform for future investigation into the functions of other cellular proteins within HSCs, both in vitro and in vivo. In conclusion, PI3K is critical for HSC activation and collagen production and, as such, therapies directed at inhibiting PI3K, specifically in HSCs, may be useful to treat hepatic fibrogenesis.