Inhibition of extracellular signal-regulated kinase 1 by adenovirus mediated small interfering RNA attenuates hepatic fibrosis in rats


  • Potential conflict of interest: Nothing to report.


Extracellular signal-regulated kinase 1 (ERK1) is a critical part of the mitogen-activated protein kinase signal transduction pathway, which is involved in hepatic fibrosis. However, the effect of down-regulation of ERK1 on hepatic fibrosis has not been reported. Here, we induced hepatic fibrosis in rats with dimethylnitrosamine administration or bile duct ligation. An adenovirus carrying small interfering RNA targeting ERK1 (AdshERK1) was constructed to determine its effect on hepatic fibrosis, as evaluated by histological and immunohistochemical examination. Our results demonstrated that AdshERK1 significantly reduced the expression of ERK1 and suppressed proliferation and levels of fibrosis-related genes in hepatic stellate cells in vitro. More importantly, selective inhibition of ERK1 remarkably attenuated the deposition of the extracellular matrix in fibrotic liver in both fibrosis models. In addition, both hepatocytes and biliary epithelial cells were proven to exert the ability to generate the myofibroblasts depending on the insults of the liver, which were remarkably reduced by AdshERK1. Furthermore, up-regulation of ERK1 paralleled the increased expression of transforming growth factor β1 (TGF-β1), vimentin, snail, platelet-derived growth factor-BB (PDGF-BB), bone morphogenetic protein 4 (BMP4), and small mothers against decapentaplegic-1 (p-Smad1), and was in reverse correlation with E-cadherin in the fibrotic liver. Nevertheless, inhibition of ERK1 resulted in the increased level of E-cadherin in parallel with suppression of TGF-β1, vimentin, snail, PDGF-BB, BMP4, and p-Smad1. Interestingly, AdshERK1 treatment promoted hepatocellular proliferation. Conclusion: Our study provides the first evidence for AdshERK1 suppression of hepatic fibrosis through the reversal of epithelial-mesenchymal transition of both hepatocytes and biliary epithelial cells without interference of hepatocellular proliferation. This suggests that ERK1 is implicated in hepatic fibrogenesis and selective inhibition of ERK1 by small interfering RNA may present a novel option for hepatic fibrosis treatment. (HEPATOLOGY 2009.)

Hepatic fibrosis is the response of the liver to different chronic insults, and is characterized by the excess production and deposition of extracellular matrix (ECM) components, leading to tissue scarring and the destruction of normal hepatic parenchyma.1 A very large number of studies have identified the hepatic stellate cells (HSCs) as the predominant source of myofibroblasts.2–4 Nevertheless, recent studies suggest that myofibroblasts can be generated from a variety of sources including resident mesenchymal, epithelial, and endothelial cells.5, 6 In particular, substantial advances have been made to highlight that myofibroblasts can be supplemented from cholangiocytes and hepatocytes by epithelial-mesenchymal transition (EMT) during hepatic fibrosis,7, 8 indicating that the parenchymal epithelial cells of the liver have a significant role in the perpetuation of hepatic fibrosis.

Extracellular signal-regulated kinase 1 (ERK1) is an important kinase in the mitogen-activated protein kinase (MAPK) signal transduction pathway. In unstimulated fibroblasts, the majority of ERK proteins are associated with the microtubule cytoskeleton.9 Mitogenic stimulation causes a significant proportion of ERK1/2 to accumulate in the nucleus, in contrast to their relative exclusion from the nucleus in resting cells.10 It has been proven that chemical inhibition of ERK abolished the transforming growth factor-β1 (TGF-β1) induced small mothers against decapentaplegic homolog 2 (Smad2) phosphorylation and EMT in the proximal tubular epithelial cells.11 Aldosterone induced EMT of renal tubular epithelial cells via reactive oxygen species–dependent ERK1/2 activation, suggesting that ERK is involved in the EMT of renal tubular epithelial cells.12 The ERK1/2 pathway also participates in the development of liver fibrosis. For example, platelet-derived growth factor-D (PDGF-D) exerted mitogenic and fibrogenic effects in both cultured HSCs and myofibroblasts comparable to PDGF-B with PDGF receptor β autophosphorylation and activation of the downstream signaling molecules ERK1/2, c-Jun N-terminal kinase, p38 MAPK, and protein kinase B/Akt.13 Cao et al. demonstrated the synergistic actions of JAK/STAT (Janus kinase/signal transducer and activator of transcription) and ERK1/2 pathway in leptin-induced HSC fibrogenesis.14 Smart et al. described a profibrogenic pathway in which ERK1/2 activation stimulated JunD-mediated elevation of tissue inhibitor of metalloproteinase-1 expression in activated HSCs, whereas inhibition of ERK1/2 by PD98059 inhibited tissue inhibitor of metalloproteinase-1 expression.15 Our previous complementary DNA array work reveals that ERK1 expression was remarkably up-regulated during the development of hepatic fibrosis. Transfection of ERK1 small interfering RNA (siRNA) into an active HSC cell line (HSC-T6) led to a significant inhibition of the proliferation of HSCs, accompanied by the induction of HSC apoptosis and reduction of collagen synthesis and deposition.16 Taken together, all of these studies suggest that ERK1 is involved in the development of hepatic fibrosis, and selective inhibition of ERK1 might be a novel strategy for the treatment of hepatic fibrosis.

In this study, we investigated the effect of ERK1 siRNA on hepatic fibrosis and clarify its mechanism. Our results demonstrated that inhibition of ERK1 by adenovirus-delivered siRNA significantly attenuated hepatic fibrosis in rats induced by both bile duct ligation (BDL) and dimethylnitrosamine (DMN) through inhibition of HSC proliferation and activation, meanwhile, reversing the EMT of biliary epithelial cells (BECs) in BDL-induced biliary fibrosis or EMT of hepatocytes in DMN-induced chemical fibrosis. These results suggest that gene-silencing therapy with ERK1 siRNA might lead to new therapeutic options for hepatic fibrosis.


ECM, extracellular matrix; HSCs, hepatic stellate cells; EMT, epithelial-mesenchymal transition; ERK1, extracellular signal-regulated kinase 1; MAPK, mitogen-activated protein kinase; TGF-β1, transforming growth factor-β1; Smad, small mothers against decapentaplegic homolog; PDGF, platelet-derived growth factor; siRNA, small interfering RNA; BDL, bile duct ligation; DMN, dimethylnitrosamine; BECs, biliary epithelial cells; NC, negative control; AdshERK1, the adenovirus expressing ERK1 shRNA; α-SMA, α-smooth muscle actin; GFAP, glial fibrillary acidic protein; CK19, cytokeratin-19; BMP4, bone morphogenetic protein 4; PCNA, proliferating cell nuclear antigen.

Materials and Methods

See Supporting Methods for detailed experimental methods.

Construction of siRNA-Expressing Adenovirus Vector and Cell Infection.

Three siRNAs targeting rat ERK1 messenger RNA (mRNA) and a scrambled siRNA used as a negative control (NC) were designed with software found on the Ambion website and synthesized by GenePharma (Shanghai GenePharma Co., Ltd., Shanghai, China). The siRNA sequences are listed in Supporting Table 1. The adenovirus vectors, containing ERK1 siRNA (AdshERK1) and the NC (AdshNC), were constructed using pShuttle as described.17 HSC-T6 cells, kindly provided by Dr. Friedman,18 were infected with recombinant virus for 2∼3 days.

Real-Time Reverse Transcription Polymerase Chain Reaction.

Total RNA was extracted from the cells or tissues with Trizol reagent (Invitrogen, Carlsbad, CA), and complementary DNA was synthesized with an oligo(deoxythymidine) primer and Moloney murine leukemia virus reverse transcriptase according to the manufacturer's instructions. Transcript levels were detected via real-time reverse transcription polymerase chain reaction (RT-PCR) with a SYBR Green PCR Kit (Applied Biosystems, Foster City, CA). Primers for these transcripts are listed in Supporting Table 2.

Western Blotting Analysis.

Western blot analysis of ERK1/2 and p-ERK1/2 was performed according to the manufacturer's recommended method (Bio-Rad Laboratories, Hercules, CA).

Determination of HSC Proliferation.

To test the effect of ERK1 siRNA on HSC proliferation, HSC-T6 cells were plated in triplicate wells on a 96-well plate at 1.5 × 103 cells/well and cultured for 24 hours. The cells were then infected with adenovirus at a multiplicity of infection of 1200. The number of metabolically active mitochondria and viable cells was determined colorimetrically at 490 nm using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay as described.19

Treatment of Hepatic Fibrosis in Rats.

Two distinct models of hepatic fibrosis were induced either by BDL or injection of DMN (Supporting Fig. 1) as described previously.20 Group 1 served as the control with sham surgery. Two days before or 3 days after the operation, rats in groups 2, 3, and 4 (eight rats in each group) were infused with a single dose of phosphate-buffered saline, 2 × 109 plaque-forming units (pfu) AdshNC, or AdshERK1 via the tail vein, respectively. For the DMN model, group 1 served as a normal control that received intraperitoneal saline injection. The remaining rats were injected intraperitoneally with 1% DMN (10 μg/kg; Sigma, St. Louis, MO) for 3 consecutive days per week up to 4 weeks and infused with phosphate-buffered saline, 2 × 109 pfu AdshNC, or the same amount of AdshERK1 after six DMN injections via the tail vein, respectively. The animals were sacrificed 3 weeks after BDL or 4 weeks after DMN administration.

Histological Examination and Immunohistological Staining.

All paraffin-embedded liver tissues were stained with hematoxylin and eosin, Masson's trichrome, and Sirius red staining. For the semiquantitative analysis, connective tissues stained blue with Masson's trichrome staining were measured on an image analyzer (Image J, National Institutes of Health) by a technician blinded to the samples. Three fields were selected randomly from each of two sections, and six rats from each group were examined. The percentage was calculated as previously described.21

Immunohistochemistry and double immunostaining were performed on paraffin-embedded liver sections according to the manufacturer's recommendation. The primary antibodies used for staining are shown in Supporting Table 3.

Measurement of Hepatic Hydroxyproline Content.

Total hepatic hydroxyproline levels were determined in the hydrolysates of liver samples as described.22

Statistical Analysis.

The arithmetic mean and standard deviation were calculated for the data, and were statistically evaluated using two-tailed unpaired t test. P < 0.05 was considered statistically significant.


Adenovirus-Mediated siRNA Targeting ERK1 Reduces the Expression of ERK1 in HSC-T6.

According to the real-time RT-PCR and western blot results, the siRNA 189 showed the best inhibition of ERK1 expression in HSC-T6 cells (data not shown), and was selected for subsequent experiments. AdshERK1 decreased ERK1 mRNA in HSC-T6 by 74% and 77% at 72 and 96 hours after the infection, respectively, compared with AdshNC (Fig. 1A). In addition, AdshERK1 reduced both ERK1 protein and phosphorylated ERK1 by approximately 65%, but ERK2 expression was not affected (Fig. 1B,C). Suppression of ERK protein expression by AdshERK1 was also detected by immunofluorescence (Fig. 1D).

Figure 1.

ERK1 suppression by AdshERK1 in HSC-T6. HSC-T6 cells were treated with medium (Control), AdshNC, or AdshERK1. (A) Real-time PCR was used to detect the effect of ERK1 siRNA on ERK1 expression in HSC-T6 at 72 and 96 hours after infection. All of the mRNA expression was normalized against β-actin. *P < 0.05, **P < 0.01. (B,C) ERK1/2 and phosphorylated ERK1/2 (p-ERK1/2) levels were checked by western blot. The values at the bottom of the lanes are the relative density of the ERK1 band and the p-ERK1 band normalized by that of GAPDH (glyceraldehyde 3-phosphate dehydrogenase). (D) Immunofluorescence staining was performed to examine the level of ERK1/2 (×400).

ERK1 siRNA Inhibits Cell Proliferation and Fibrosis-Associated Gene Expression.

MTT analysis revealed that AdshERK1 significantly suppressed HSC-T6 proliferation from the third day after infection. On the fifth day, the suppression rate was up to 50% (P < 0.001, data not shown). Furthermore, compared with AdshNC, ERK1 silencing suppressed collagen types I and III, tissue inhibitor of metalloproteinase-1, and TGF-β1 mRNA levels 96 hours after the infection by 44% (P < 0.01), 57% (P < 0.01), 35% (P < 0.05), and 47% (P < 0.05), respectively (Supporting Fig. 2).

ERK1 siRNA Attenuates Hepatic Fibrosis Induced by BDL or DMN in Rats.

We then tested the effect of AdshERK1 in two distinct rat models of hepatic fibrosis: BDL rats or rats administered DMN. As shown in Fig. 2A, only faint ERK1 staining and normal collagen distribution was detected in animals that underwent sham-operation. Overexpression of ERK1 and aniline blue–stained fibrils were associated with proliferating bile ducts that formed a continuous meshwork of connective tissue infiltrating the hepatic parenchyma with loss of the lobular architecture 3 weeks after BDL, suggesting that hepatic fibrosis was successfully established. Interestingly, a single dose of AdshERK1 suppressed ERK1 expression and prevented the development of hepatic fibrosis, as confirmed by hematoxylin and eosin, Masson's trichrome, and Sirius red staining. AdshERK1 reduced the ECM area (Masson's staining) by 62% after injection (P < 0.01). Moreover, hydroxyproline content decreased in the AdshERK1-treated group (199.84 ± 4.20 μg/g), compared with AdshNC group (272.59 ± 7.8 μg/g, P < 0.05) and model group (288.28 ± 11.46 μg/g, P < 0.05), whereas the hydroxyproline content in the sham-operated group was 176.91 ± 13.04 μg/g liver tissue. In addition, AdshERK1 significantly suppressed collagen types I and III in liver (Supporting Fig. 3).

Figure 2.

AdshERK1 ameliorates hepatic fibrosis induced by BDL and DMN in rats. ERK1 in liver sections was assessed by immunohistochemistry; hematoxylin and eosin, Masson's trichrome, and Sirius red staining was used to examine pathological alterations and collagen deposition. Data shown are the representative of eight animals. (A) BDL model, (B) DMN model.

To further evaluate the effect of ERK1 siRNA on fibrosis induced by BDL, we injected the virus 3 days after BDL, in which hepatic fibrosis has already developed, as demonstrated in the study by Beaussier et al.6 The results showed that AdshERK1 could ameliorate liver fibrosis as well (Supporting Fig. 4). The hydroxyproline content in the AdshERK1 group (230.12 ± 24.35 μg/g) decreased by 41% as compared with AdshNC group (392.05 ± 19.69 μg/g, P < 0.01) and by 40% as compared with model group (383.52 ± 12.04 μg/g, P < 0.05). The ECM area (Masson's staining) was reduced by 64% (P < 0.05) as compared with AdshNC group.

As expected, DMN injection induced prominent hepatic fibrosis in rats as shown by Masson's trichrome and Sirius red staining and increased ERK1 expression (Fig. 2B). AdshERK1 treatment blocked this increase in ERK1 and reduced ECM (Masson's staining) by 74% (P < 0.01). The hydroxyproline content in the AdshERK1-treated group (198.94 ± 26.33 μg/g liver tissue) also decreased compared with that in the AdshNC group (355.11 ± 54.82 μg/g, P < 0.05) and model group (359.41 ± 59.10 μg/g, P < 0.05).

ERK1 Is Associated with the Activation of Myofibroblasts.

To address the possible role of ERK1 in the activation of myofibroblasts, double-immunostaining was performed to detect the expression profile of ERK1 and α-smooth muscle actin (α-SMA) in both the BDL and DMN fibrosis models. There was no overlap of ERK1 and α-SMA in normal liver. ERK1 staining dramatically increased in the cytoplasm of proliferative BECs after BDL. Interestingly, ERK1 expression was also detected in the nuclei of the cells around the peribiliary region and colocalized with α-SMA (Figs. 2A and 3A). Similarly, increased staining of ERK1 was found around portal tracts and in fibrotic septa 4 weeks after DMN injection, especially in the nuclei of interstitial cells (Fig. 2B and 3B), which was accompanied by increased expression of α-SMA, implying that ERK1 is implicated in the activation of myofibroblasts. AdshERK1, however, prevented the coexpression of ERK1 and α-SMA in the two kinds of liver fibrosis (Fig. 3) and decreased mRNA levels of ERK1 and α-SMA in both models (Supporting Fig. 5).

Figure 3.

ERK1 is up-regulated in myofibroblasts in vivo. Double immunostaining was used to analyze the expression of ERK1 and α-SMA in liver from control, model, AdshNC, or AdshERK1 groups. ERK1 expression was detected using a peroxidase-conjugated secondary antibody and colored with 3,3′-diaminobenzidine (brown). α-SMA was stained with Alexa Fluor-488 (green). Hematoxylin and 4′,6-diamidino-2-phenylindole were used for counterstaining of DNA (blue). The staining was visualized under a fluorescence microscope and the arrows point to the ERK1/α-SMA double positive cells.

BECs and Hepatocytes Contribute to the Accumulation of Myofibroblasts in Fibrogenesis In Vivo.

To further investigate the activation of myofibroblasts in fibrogenesis, immunostaining was performed in normal and fibrotic liver using serial section. As shown in Fig. 4A, the signal for α-SMA, desmin, and glial fibrillary acidic protein (GFAP) was restricted to the wall of blood vessels in sham-operated control rats. Nevertheless, the expression of α-SMA, desmin, and GFAP was increased around the bile ducts in BDL rats (Fig. 4A) and in fibrotic septa in DMN-induced hepatic fibrosis (Fig. 4B), which was remarkably attenuated by AdshERK1. Interestingly, we noticed that the desmin-labeled or GFAP-labeled HSCs only accounted for approximately 50% of the α-SMA–positive myofibroblasts, implying that not all of the myofibroblasts were derived from HSCs.

Figure 4.

AdshERK1 blocks activation of myofibroblasts in vivo. α-SMA, desmin, and GFAP immunoreactivities were analyzed using serial sections. All three pictures in each group were collected from the same field of the three serial sections (×400).

To detect the other source of the myofibroblasts in fibrotic liver, double immunostaining for the biliary epithelial marker cytokeratin-19 (CK19) and α-SMA was carried out in the BDL model. As shown in Fig 5A, CK19 staining was clearly detected in BECs. Additionally, a large amount of CK19-positive cells were also detected around proliferative bile ducts and in fibrotic septa, although with a weaker signal than those within the basement membrane. Interestingly, the cells with weaker CK19 signals were colocalized with α-SMA, whereas CK19-positive BECs within the basement membrane were negative for α-SMA. AdshERK1, however, inhibited the coexpression of CK19 and α-SMA in biliary epithelium.

Figure 5.

Colocalization of CK19/α-SMA and albumin/α-SMA in BDL-induced and DMN-induced liver fibrosis. (A) Double immunostaining for CK19 (brown) and α-SMA (green) in liver sections. The colocalization of CK19 and α-SMA is clearly evident around the bile ducts after ligation (arrows). The staining was visualized under a fluorescence microscope and the yellow bars represent 50 μm. (B) Double immunofluorescence staining for albumin (red) and α-SMA (green) 4 weeks after DMN administration. The arrows point to fusiform-shaped albumin-positive cells that express α-SMA. The staining was visualized under a laser scanning confocal fluorescent microscope and the white bars represent 20 μm.

To seek evidence that hepatocyts can convert into matrix-producing myofibrolasts, double immunofluorescence staining was used to show the colocalization of albumin and α-SMA in the DMN model. Besides definite albumin staining in hepatocytes, a large amount of fusi-shaped cells with weaker albumin expression were also detected in fibrotic septa that were also α-SMA–positive, whereas hepatocytes in the liver parenchyma were α-SMA–negative (Fig. 5B). Similarly, AdshERK1 attenuated the coexpression of albumin and α-SMA in hepatocytes. These observations suggested that both BECs and hepatocytes contribute to the accumulation of matrix-producing myofibrolasts in fibrogenesis, and AdshERK1 protects the epithelial cells from this transformation.

ERK1 siRNA Attenuates Experimental Hepatic Fibrosis Through Inhibiting EMT.

We then analyzed the levels of EMT markers by immunohistochemical analysis. E-cadherin was decreased on the membranes of the parenchymal and nonparenchymal cells in both fibrosis models, whereas vimentin and snail were up-regulated (Fig. 6). These changes were reversed by AdshERK1 treatment. The localization of TGF-β in fusi-shaped myofibroblasts was also blocked by AdshERK1 (Fig. 6A,B). Consistent with the immunohistochemical results, the enhanced expression of TGF-β1 mRNA level was substantially decreased by AdshERK1 in both models (Supporting Fig. 5). In addition, the expression of PDGF-BB, bone morphogenetic protein 4 (BMP4), and phosphorylated Smad1 (p-Smad1) increased along with the up-regulation of ERK1 in DMN-induced hepatic fibrosis, which was reduced after AdshERK1 treatment (Supporting Fig. 6).

Figure 6.

AdshERK1 attenuates experimental hepatic fibrosis through inhibition of EMT. Immunohistochemical staining was performed to detect the expression of E-cadherin, vimentin, snail, and TGF-β in fibrotic livers from control, model, AdshNC, or AdshERK1 groups. (A) BDL model, (B) DMN model.

AdshERK1 Does Not Inhibit Hepatocyte Proliferation.

To address the effect of ERK1 suppression on hepatocytes, proliferating cell nuclear antigen (PCNA) staining was performed after gene delivery. We found that PCNA staining mostly emerged in the proliferative BECs or in fibroblast-like cells in the fibrotic septa, rarely in hepatocytes in Model and AdshNC groups. Intriguingly, AdshERK1 treatment increased the levels of PCNA-positive hepatocytes in both models (Fig. 7).

Figure 7.

AdshERK1 treatment promotes hepatocellular proliferation. (A,B) Immunohistochemistry was used to determine the expression of PCNA (×400) in fibrosis induced by BDL or DMN in liver tissues from sham, model, AdshNC, and AdshERK1 groups. Arrows point to the PCNA-positive hepatocytes. (C) PCNA-positive hepatocytes (%) were quantified on image analysis software (Image J, NIH) by a technician blinded to the samples. At a magnification of ×400, three fields were selected randomly from each section, and a total of 100 hepatocytes were evaluated in every field. *P < 0.05 (AdshERK1 versus AdshNC).


The MAPK signal pathway regulates diverse cellular events such as proliferation, growth, differentiation, and apoptosis.23, 24 In recent years, many studies have revealed that the MAPK pathway participates in the pathogenesis of fibrosis.25–29 As one of the key kinases in MAPK pathway, ERK1 is also proven to be implicated in the development of hepatic fibrosis.15, 16, 30 Moreover, a recent study by Schmitz et al. has demonstrated that activation of the ERK signaling pathway predicts poor prognosis in human hepatocellular carcinoma, indicating that ERK1 is potentially a more general molecular player that is involved in the outcome of liver damage.31 We have previously shown that inhibition of ERK1 by siRNA in HSCs inhibited cell proliferation, induced apoptosis, and reduced collagen production.16 In this study, we reveal for the first time that selective inhibition of ERK1 by siRNA attenuated the deposition of ECM in fibrotic liver induced by two mechanistically different models in rats. These data further support the role of ERK1 in hepatic fibrogenesis and suggest that siRNA targeting of ERK1 might serve as a novel effective agent for hepatic fibrosis treatment.

Recent studies have revealed that the parenchymal cells in fibrotic liver can contribute to the accumulation of myofibroblasts in fibrotic liver. α-SMA is a marker of activated myofibroblasts, whereas desmin and GFAP are markers of HSCs.32 We found that only half of the α-SMA–positive cells presented with desmin or GFAP staining in fibrotic liver in both fibrosis models, suggesting that the activated myofibroblasts in fibrotic liver are derived from an additional source. Using double immunostaining for α-SMA and CK19, a BEC marker, we observed that BECs around proliferative bile ducts and in fibrotic septa were stained for α-SMA but with weaker signal of CK19 compared with BECs in the basement membrane, implying that these BECs may gradually lose their epithelial phenotype and acquire the mesenchymal phenotype of myofibroblasts after BDL. On the other hand, in accordance with a study by Zeisberg et al.,33 we also demonstrated that about 40%–50% of the α-SMA–positive fibroblasts were costained with albumin in the DMN-treated fibrotic liver, indicating their hepatocyte origin. These data clearly suggest that myofibroblasts could be generated from both hepatocytes and BECs, depending on the insults to the liver.

It has been well proven that EMT regulates tissue construction during embryogenesis and is involved in carcinogenesis and metastasis.34, 35 Recent studies suggest that this process is also implicated in tissue remodeling after injury in the kidney, lung, and liver.36, 37 Intriguingly, our current study reveals that AdshERK1 treatment reduced the fusiform (myofibroblast-like) cells and simultaneously enhanced E-cadherin expression along with the inhibition of vimentin and snail, suggesting that reversion of EMT contributes to the resolution of hepatic fibrosis. TGF-β1 is considered as the most powerful inducer of EMT.38 BMP4, a member of the TGF-β superfamily, is demonstrated to mediate BDL-induced hepatic fibrosis through activation of the ERK and Smad1 pathways in HSCs.39 In this study, we showed that ERK1 up-regulation was associated with the increased expression of TGF-β1, BMP4 and p-Smad1 during hepatic fibrogenesis, which was inhibited by AdshERK1 treatment. Thereafter, we may infer that AdshERK1 reversed EMT at least partially through the reduction of TGF-β expression. It is known that TGF-β1 positively regulates its own expression in normal and fibrotic liver. Thus, autoinduction of TGF-β1 at sites of injury may result in a positive feedback loop that perpetuates the fibrotic process. The ERK1 pathway is known to be involved in the autoinduction of TGF-β1.40 In this case, we speculate that the inhibition of TGF-β1 by ERK1 siRNA may be achieved by the interruption of its autoinduction. On the other hand, we also demonstrated that the expression of PDGF-BB in fibrotic liver was suppressed after the delivery of AdshERK1, indicating that both pathways of PDGF and TGF-β are involved in the inhibitory effect of ERK1 siRNA on hepatic fibrosis.

It is generally accepted that the Ras-ERK–dependent signaling cascade participates in control of cell fate, proliferation, and survival in various mammalian organs, including the liver. Therefore, our major concern became whether silencing of ERK1 gene by siRNA inhibits the proliferation of hepatocytes, because it could be the crucial limiting factor for clinical application of this therapy. ERK2 and ERK1 are proteins of 42 and 44 kDa size that are nearly 85% identical overall. These two kinases are coexpressed in most tissues but have distinct functions.41 It has been demonstrated that ERK2 but not ERK1 is the key form involved in regulation of hepatocyte replication in vivo and in vitro.42 There was no evidence in our results supporting the proliferative inhibition of liver cells by AdshERK1, as shown by the PCNA immunohistochemistry. In contrast, it was likely to promote hepatocyte replication. It is well known that TGF-β1 can induce hepatocyte apoptosis and stimulation of ECM deposition in hepatic fibrosis.43 Therefore, the potential mechanism for the pro-proliferative effect of AdshERK1 may be attributed to the fact that TGF-β1 expression was suppressed by AdshERK1, which facilitated hepatocyte proliferation and protected hepatocytes from apoptosis. In addition, degradation of ECM may also contribute to liver regeneration.44

SiRNA has become a powerful tool for functional genetic studies and gene therapy in mammals.45, 46 Although gene knockdown by siRNA is highly effective, the off-target effect of siRNA may represent a major obstacle for therapeutic applications. However, the potential off-target effects could be minimized by choosing an siRNA with maximal sequence divergence from the list of genes with partial sequence identity to the intended mRNA target.47 There were no genes except ERK1 fully matched with the sense or antisense sequence of 189 siRNA used in this study, analyzed by the Basic Local Alignment Search Tool ( Moreover, even the few partially matched rat genes are unrelated with hepatic fibrosis. Therefore, the phenotype observed with ERK1 siRNA was attributed to ERK1 knockdown and not a bystander effect of the siRNA sequence.

In summary, this investigation indicates that liver fibrogenesis is a complicated process which involves not only the activation of HSCs, but also EMT of hepatocytes or BECs. More intriguingly, our study provides the first strong evidence for the striking suppression effect of AdshERK1 on hepatic fibrosis, which is associated with the suppression of HSCs, the reversal of EMT of hepatocytes or BECs, as well as promotion of hepatocyte proliferation. These results further implicate the involvement of ERK1 during the development of hepatic fibrosis and suggest that gene silencing of ERK1 might emerge as a novel option for hepatic fibrosis therapy.