Hepatic Fibrosis Is Enhanced and Accompanied by Robust Oval Cell Activation after Chronic Carbon Tetrachloride Administration to Egr-1-Deficient Mice

Abstract

The transcription factor early growth response (Egr)−1 regulates the expression of genes required for execution of the wound healing response. Multiple cycles of injury, coupled to incomplete wound healing, lead to fibrosis. Therefore, we hypothesized that Egr-1 is required for the development of hepatic fibrosis. To test this hypothesis, we exposed wild-type and egr-1^−/− mice to acute or chronic carbon tetrachloride (CCl₄). Acute CCl₄ exposure established a profibrotic milieu in the liver, including activation of hepatic stellate cells as well as expression of type 1 collagen genes and tissue inhibitor of matrix metalloproteinase 1 in both wild-type and egr-1^−/− mice. This response was exacerbated in egr-1^−/− mice. After chronic CCl₄ exposure, hepatic fibrosis was established in both genotypes; however, the fibrotic response was profoundly worsened in Egr-1–deficient mice. Importantly, enhanced fibrosis in egr-1^−/− mice was accompanied by a robust activation of the oval cell response, suggesting more severe liver injury and/or reduced hepatocyte proliferation when compared with wild-type mice. Hepatic expression of genes indicative of oval cell activation, as well as the number of cells expressing A6, a mouse oval cell marker, was greater in egr-1^−/− mice. Taken together, these data reveal novel roles for Egr-1 as a negative regulator of both CCl₄-induced hepatic fibrosis and the oval cell response.

The wound healing response is a rapidly induced and finely orchestrated series of cellular and molecular events necessary for prompt tissue repair and reinstatement of tissue/organ integrity after injury.1 Perturbations in this system lead to aberrant tissue repair. In the case of prolonged exposure to injury, tissue fibrosis, defined by excessive extracellular matrix (ECM) deposition, is established. Liver fibrosis is a common end point of chronic hepatic diseases of all etiologies.2,3 Although there has been significant progress in defining the roles of the many cell types, soluble mediators and transcription factors that contribute to fibrosis and its resolution in animal models, these advances have not yet led to improvements in the therapeutic arsenal available for patients with fibrosis or cirrhosis. Indeed, liver transplant remains the only viable treatment option for the burgeoning population of patients with fibrosis and is hampered by a limited supply of available organs.4

Early growth response (Egr)−1 (also known as NGFI-A, zif268, TIS8, and Krox24) is an 80- to 82-kDa transcription factor and prototypic member of the early growth response gene family.5 Egr-1 is an immediate early gene induced in response to a number of growth and differentiation factors and in response to tissue injury.5,6,7,8,9,10 Its swift induction and rapid decay place Egr-1 at a position to induce the expression of genes necessary for rapid yet transient responses to many forms of acute cellular stress. Using its three zinc-finger DNA binding domain, Egr-1 binds to G/C-rich regions found in the promoters of genes involved in inflammation (tumor necrosis factor [TNF] α, intercellular adhesion molecule-1) and also indirectly affects the expression of other inflammatory mediators (macrophage inflammatory protein-2, monocyte chemoattractant protein [MCP]−1).11,12,13 Egr-1 is characterized as a ‘master regulator’ of inflammatory gene expression in mouse models of ischemia/reperfusion, endotoxemia, ethanol-induced fatty liver injury, and acute hepatitis.6,8,9,10

Egr-1 expression is also a component of the wound healing response.14,15,16 Successful repair after injury involves inflammation, coagulation, angiogenesis, new tissue formation, and ECM remodeling/scarring.1 Egr-1 contributes to this process by regulating genes involved in coagulation (tissue factor, plasminogen activator inhibitor-1), ECM deposition and metabolism (transforming growth factor β, fibroblast growth factor-2, type 1 collagen, matrix metalloproteinase 14, platelet-derived growth factor A and B, thrombospondin-1, and fibronectin), cell division (p53 and cell division cycle protein 20 homolog [Cdc20]) and angiogenesis (vascular endothelial growth factor).17,18,19,20,21,22,23,24,25 This broad array of gene expression suggests that Egr-1 has the ability to affect the wound healing response of many, if not all, cell types in a given tissue. For example, in the liver, Egr-1 could potentially promote hepatic stellate cell (HSC) activation via regulation of platelet-derived growth factor, endothelial cell proliferation by vascular endothelial growth factor, and hepatocyte proliferation by Cdc20. Indeed, Egr-1–regulated expression of Cdc20 is required for hepatocyte proliferation in response to partial hepatectomy.20 Although transient, regulated expression of these gene products is required for appropriate wound healing responses; chronic expression of repair-associated genes is associated with the development of fibrosis.

Despite the key regulatory role that Egr-1 plays in the timely expression of genes induced during the tissue response to injury, the role of Egr-1 in hepatic fibrosis is not known. Because Egr-1 regulates the expression of many profibrotic genes, we hypothesized that Egr-1 is required for hepatic fibrosis in response to chronic injury and that Egr-1–deficient mice would be protected from hepatotoxin-induced fibrosis. To test this hypothesis, we used the well-characterized carbon tetrachloride (CCl₄)-induced hepatic fibrosis model26 in both wild-type and egr-1^−/− mice. We predicted that, in the absence of Egr-1, fibrosis would be attenuated by the contribution of Egr-1 to expression of many genes associated with the wound healing response. To our surprise, hepatic fibrosis in egr-1^−/− mice after chronic CCl₄ was enhanced and recovery delayed compared with wild-type mice. In addition, chronic CCl₄ exposure in egr-1^−/− mice induced a striking activation of the hepatic progenitor cell or ‘oval cell’ compartment. Oval cell activation is not typically associated with hepatocyte replenishment after CCl₄ exposure. Undamaged hepatocytes are widely appreciated to enter the cell cycle and recover lost liver mass in this model; only when liver injury is severe and/or hepatocyte proliferation is attenuated will oval cells participate in liver regeneration.27 To our knowledge, this is the first study to report a robust progenitor cell response during hepatic recovery after chronic CCl₄ exposure. Our recent work (unpublished data) support the hypothesis that very severe liver injury in egr-1^−/− mice after an initial exposure to CCl₄ contributes to oval cell activation after chronic CCl₄ exposure.

Materials and Methods

Materials

CCl₄ and olive oil were purchased from Sigma-Aldrich (St. Louis, MO). All primers for real-time reverse-transcription PCR were synthesized by Integrated DNA Technologies (Coralville, IA). Primary antibodies were purchased from the following companies: Cytochrome P450 2E1 (CYP2E1) Research Diagnostics, Inc (Flanders, NJ); Hsc 70 Santa Cruz (Santa Cruz, CA); α smooth muscle actin (αSMA, clone 1A6 Sigma-Aldrich, St. Louis, MO); A6 antibody was a generous gift from Dr. Valentina Factor (NIH/NCI).

Animals

Female wild-type (C57BL/6NTac) or egr-1^−/− (B6.129-Egr1^tm1Jmi N12) mice (8–12 weeks) were purchased from Taconic Farms (Germantown, NY). The egr-1^−/− mouse was originally developed in the laboratory of Jeffrey Milbrandt.28,29 Animals were housed in standard microisolator cages and fed standard laboratory chow (rodent diet #2918, Harlan-Teklad, Madison, WI). All animal procedures were approved by the Cleveland Clinic Institutional Animal Care and Use Committee. The following numbers of animals were used in the acute studies: olive oil, 4 wild-type and 4 egr-1^−/− mice; 18 hours, 5 wild-type and 5 egr-1^−/− mice; 72 hours, 5 wild-type and 5 egr-1^−/− mice. The following numbers of animals were used in chronic CCl₄ experiments: olive oil, 7 wild-type and 6 egr-1^−/− mice; 3 days, 9 wild-type and 6 egr-1^−/− mice; 7 days, 5 wild-type and 5 egr-1^−/− mice.

CCl₄ Administration and Sample Collection

In both acute and chronic CCl₄ studies, CCl₄ was prediluted 1:3 in olive oil before administration; controls received injections of olive oil alone. For acute CCl₄ studies, mice received a single dose at 1 μl/g body weight of prediluted (1:3 in olive oil) CCl₄ administered by intraperitoneal injection using 100 μl Hamilton syringes and 26 G 5/8 inch needles. The dose was well tolerated by wild-type and egr-1^−/− mice. Both genotypes had lost 1 g of body weight relative to olive oil controls 18 hours after the CCl₄ injection. This weight loss was completely recovered in wild-type mice by 72 hours; egr-1^−/− mice recovered some but not all of the weight loss by 72 hours. For chronic CCl₄ studies, mice received a total of 10 injections (2 injections each week on Tuesdays and Fridays) over 5 weeks. Mice were ramped up to the full dose of CCl₄ over two doses (first injection at 0.25 μl/g body weight, the second at 0.5 μl/g body weight). The third through tenth injections were administered at the full 1 μl/g body weight dose. As in the acute studies, each dose was prediluted (1:3) in olive oil before administration. Injections were given on the opposite side of the peritoneal cavity from the previous injection. Wild-type mice tolerated the chronic CCl₄ administration paradigm well; egr-1^−/− mice were less resilient to subsequent doses of CCl₄ by the end of the third week. The difference in tolerance to the chronic treatment paradigm was paralleled by differences in body weight at 3 and 7 days after completion of CCl₄ exposure (see supplemental Figure S1 at http://ajp.amjpathol.org). In the acute studies, animals were sacrificed 18 and 72 hours after CCl₄ administration. In the chronic studies, animals were sacrificed at 3 and 7 days after the final CCl₄ exposure. At the time of sacrifice, mice were weighed, and once anesthetized, blood was collected from the posterior vena cava. Plasma was separated from whole blood and stored at −80°C. Mice were then euthanized, while under anesthesia, by exsanguination. Livers were removed, weighed, and portions fixed in formalin, frozen in Optimal Cutting Temperature medium (Sakura Finetek USA, Torrance, CA), snap frozen in liquid nitrogen, or stored in RNAlater (Ambion, Austin, TX) for further analysis.

Liver Histology and Assessment of Fibrosis

For histological analysis, formalin-fixed tissues were paraffin-embedded, sectioned (5 μm), and stained with hematoxylin and eosin (H&E), Sirius red, or Masson trichrome. Slides were coded before initial examination and viewed by two separate individuals. Quantification of Sirius red staining was performed using ImagePro plus software (Media Cybernatics, Silver Spring, MD).

Real-Time RT-PCR

Total RNA was isolated from liver and 4 μg of total hepatic RNA was reverse transcribed.9 Real-time PCR amplification was performed using Brilliant SYBR Green QPCR Master Mix (Stratagene, La Jolla, CA) and gene-specific primers (Table 1) in an Mx3000p PCR machine (Stratagene). Reaction volumes and thermal cycle parameters were as previously described.9 The relative amount mRNA was determined using the comparative threshold (Ct) method by normalizing target cDNA Ct values to that of 18S; 18S Ct values did not change from baseline values at time points after CCl₄ used in these studies. Fold induction ratios were calculated relative to basal conditions for each genotype using the formula: 2^−ΔΔCt.

Table 1.

Real-Time PCR Primer Sequences for Genes of Interest (GOI)

GOI	GenBank accession	Primer bank* accession	Primer sequences
18S	X00686	NA	F:5′–ACGGAAGGGCACCACCAGGA–3′
			R:5′–CACCACCACCCACGGAATCG–3′
Egr-1	NM_007913	6681285a2	F:5′–AGCGAACAACCCTATGAGCAC–3″
			R:5′–TCGTTTGGCTGGGATAACTCG–3″
αSMA	NM_007392	6671507a1	F:5′–GTCCCAGACATCAGGGAGTAA–3′
			R:5′–TCGGATACTTCAGCGTCAGGA–3′
desmin	NM_010043	33563250a1	F:5′–GTGGATGCAGCCACTCTAGC–3′
			R:5′–TTAGCCGCGATGGTCTCATAC–3′
COL1A1	NM_007742	NA	F:5′–CAAGAACAGCAACGAGTACCG–3′
			R:5′–GTCACTGGTCAACTCCAGCAC–3′
COL1A2	NM_007743	6680980a2	F:5′–GGTGAGCCTGGTCAAACGG–3′
			R:5′–ACTGTGTCCTTTCACGCCTTT–3′
CTGF	NM_010217	6753878a1	F:5′–GGGCCTCTTCTGCGATTTC–3′
			R:5′–ATCCAGGCAAGTGCATTGGTA–3′
TIMP1	NM_011593	6755795a1	F:5′–GCAACTCGGACCTGGTCATAA–3′
			R:5′–CGGCCCGTGATGAGAAACT–3′
AFP	NM_007423	31982513a2	F:5′–AACTCTGGCGATGGGTGTTTA–3′
			R:5′–ACACTGATGTCTTTCCACTCCA–3′
Thy-1	NM_009382	6678347a1	F:5′–TGCTCTCAGTCTTGCAGGTG–3′
			R:5′–TGGATGGAGTTATCCTTGGTGTT–3′
HGF	X84046	4165287A1	F:5′–ATGTGGGGGACCAAACTTCTG–3′
			R:5′–GGATGGCGACATGAAGCAG–3′
c-met	NM_008591	6678868a1	F:5′–GTCAACATGAAGTATCAGCTCCC–3′
			R:5′–TGTAGTTTGTGGCTCCGAGAT–3′
TWEAK	NM_011614	33438584a1	F:5′–CCGCCAGATTGGGGAATTTAC–3′
			R:5′–AGTCCAAAGTAGGTTAGGAAGGG–3′
Fn14	NM_013749	7305059a1	F:5′–GTGTTGGGATTCGGCTTGGT–3′
			R:5′–GTCCATGCACTTGTCGAGGTC–3′
MCP1	NM_011333	NA	F:5′–AAGTCCCTGTCATGCTTCTG–3′
			R:5′–TCTGGACCCATTCCTTCTTG–3′

NA, indicates not applicable. 

^*, Primer Bank, http://pga.mgh.harvard.edu/primerbank/ 

Liver Homogenate Preparation, Electrophoresis, and Immunoblotting

Liver homogenates were prepared and protein concentrations determined for immunoblotting.30 Thirty-five micrograms of protein were resolved on 15% polyacrylamide gels and transferred to Polyvinylidene fluoride membranes. Membranes were probed with antibodies specific for CYP2E1 and Hsc 70 (loading control).

α Smooth Muscle Actin Immunolocalization

Formalin-fixed paraffin embedded liver sections (5 μm) were deparaffinized and hydrated consecutively in 100% twice, 95% and 70% ethanol followed by 2 washes in distilled water. Protein epitopes were unmasked using citrate buffer incubation at 95–100°C for 20 minutes in an 1100 watt microwave using a pressure cooker (Nordic Ware, Minneapolis, MN). After heat treatment, the sections were allowed to cool at room temperature for an additional 20 minutes. Endogenous peroxidase activity and avidin/biotin were quenched using 3% hydrogen peroxide (H₂O₂) and appropriate avidin and biotin blocking reagents (Vector Laboratories, Inc., Burlingame, CA), respectively. Serum-containing blocking reagents from the Mouse on Mouse kit (Vector Labs) were used per the manufacturer’s instructions before incubation with the mouse αSMA antibody (1:1600). The horseradish peroxidase–conjugated secondary antibody was used as outlined in the Mouse on Mouse kit. After washing away unbound secondary antibody, sections were incubated with Avidin-Biotin Complex reagents followed by incubation with diaminobenzamidine for 1 minute. Sections were then washed in cold running tap water for 5 minutes and then immersed in hematoxylin (DAKO, Carpinteria, CA) for 5 minutes before another 5-minute wash in cold running water. Finally, sections were dehydrated and preserved using Permount (Sigma-Aldrich) and glass coverslips. No immunostaining was seen in sections incubated with blocking reagent alone in the absence of primary antibody.

A6 Immunolocalization and Quantification

A6 is a marker for mouse oval cells.31 Five-micrometer frozen sections from livers embedded in Tissue-Tek Optimal Cutting Temperature Compound (Sakura Finetek USA, Torrance, CA) were fixed in methanol:acetone (1:1) for 10 minutes and then air dried for 10 minutes. Endogenous peroxidase in liver sections was quenched by emersion in 3% H₂O₂ for 2 minutes. After blocking in 10% normal goat serum, sections were incubated with the anti-A6 antibody (1:40 dilution in 10% normal goat serum) for 60 minutes at room temperature in a humidified chamber. After two 5-minute washes in Tris-buffered saline (TBS), sections were incubated with an horseradish peroxidase–conjugated goat-anti-rabbit secondary antibody. After two additional washes in TBS, sections were incubated with diaminobenzamidine for 5 minutes, rinsed in TBS, and then distilled water and stained with hematoxylin for 5 minutes. After dehydration in a graded series of ethanols and xylene substitute, sections were preserved using Permount (Sigma-Aldrich) and glass coverslips. A6-positive cells were counted in four ×400 fields per slide by blinded individuals. The A6 antibody recognizes biliary epithelial cells found in bile caniculi of the portal tracts as well as oval cells.31 Therefore, only A6-positive cells found in the parenchyma in areas separate from portal tracts were enumerated.

Statistical Analysis

Values reported are means ± SEM. Because of the limited number of age-matched egr-1^−/− mice, data were collected from several different experiments. The data were analyzed by general linear models procedure (SAS, Carey, NC) followed by least square means analysis of differences between groups, blocking for experiment effects when data from more than one experiment were used in any given data set. Data were log transformed to obtain a normal distribution, if necessary. Comparisons between data sets with P values of less than 0.05 were considered significant.

Results

Egr-1 Deficiency Was Associated with Enhanced Hepatic Stellate Cell Activation after Acute CCl₄ Exposure

A single CCl₄ exposure in mice and rats activates HSCs, the cell type largely responsible for excessive ECM deposition and fibrosis after repeated CCl₄ exposure.32 Indices of HSC activation and hepatic fibrogenesis after acute exposure to CCl₄ can be used to predict the severity of fibrosis after chronic exposure to CCl₄. Hepatic αSMA and desmin, markers associated with HSC activation, were increased in livers from wild-type and egr-1^−/− mice 72 hours after acute CCl₄ exposure; this increase was greater in livers from mice deficient in Egr-1 (Figure 1, A and B). αSMA protein expression paralleled the mRNA data; immunoreactive αSMA was greater in livers from egr-1^−/− mice compared with wild-type mice (Figure 1C). Activated HSCs produce type I collagen molecules, the predominate collagen associated with liver fibrosis.3 Seventy-two hours after acute CCl₄ exposure, hepatic mRNA accumulation of collagen 1α1 (COL1A1) and collagen 1α2 (COL1A2) was increased in both strains of mice; however, this increase was fivefold greater in livers from egr-1^−/− mice (Figure 1, D and E, respectively). Associated with increased activation of HSC and production of type I collagen molecules in egr-1^−/− mice was increased expression of connective tissue growth factor (CTGF), a pro-fibrotic molecule, and tissue inhibitor of matrix metalloproteinase−1, an inhibitor of matrix degradation (Figure 1, F and G, respectively) compared with wild-type mice. Expression of cytochrome P450 2E1 (CYP2E1), the enzyme required for CCl₄ bioactivation and CCl₄-induced liver injury and fibrosis,33 was not different between wild-type and egr-1^−/− mice after acute CCl₄ exposure (unpublished data).

Figure 1.

HSC activation after acute CCl₄ administration. Seventy-two hours after a single olive oil or CCl₄ injection, biomarkers of HSC activation were measured in wild-type and egr-1^−/− mice. Hepatic (A) αSMA, (B) desmin, (D) COL1A1, (E) COL1A2, (F) CTGF, and (G) tissue inhibitor of matrix metalloproteinase−1 mRNA accumulation was determined by real-time PCR using gene specific primers (Table 1). Data are expressed as fold change over olive oil controls for each genotype after normalization to 18S (housekeeping gene). The data are means ± SEM, n = 4 to 6 per experimental condition. There is a significant change in gene expression after CCl₄ when compared with genotype-specific olive oil control (*P < 0.05). There is a significant change in gene expression after CCl₄ compared with genotype-specific olive oil control and between genotypes at the indicated time point (**P < 0.05). C: Immunolocalization of hepatic αSMA protein from olive oil or CCl₄-treated mice, ×200 magnification. Images are representative of n = 4 to 6 per experimental group. Asterisk in image denotes the central vein; PT denotes portal tract.

Exacerbated Fibrosis after Chronic CCl₄ Exposure in egr-1^−/− Mice

In animal models of acute injury, Egr-1 expression is robust, but transient.8,34 After chronic CCl₄ exposure, hepatic Egr-1 mRNA was increased over baseline at 3 and 7 days post CCl₄ exposure in wild-type mice (Figure 2). However, the level of expression was lower than after acute CCl₄ exposure (unpublished data). These data suggested that low, but sustained, Egr-1 expression was characteristic of chronic CCl₄-induced liver injury.

Figure 2.

Egr-1 expression was prolonged after chronic CCl₄ exposure in wild-type mice. Wild-type mice were treated with olive oil or CCl₄ for five weeks. Mice were euthanized at the time points indicated, their livers harvested and analyzed for hepatic Egr-1 mRNA accumulation by real-time PCR. Data are means ± SEM of n = 5 to 10 per experimental group. There is a significant difference compared with olive oil controls (*P < 0.05).

Repeated exposure of mice to CCl₄ over several weeks induces excessive ECM deposition in liver which associated with fibrosis.26 To determine the effect Egr-1 deficiency had on development of hepatic fibrosis, we exposed wild-type and egr-1^−/− mice to CCl₄ for a total of 5 weeks. Sirius red staining of ECM demonstrated established fibrosis in both genotypes. The quantity of ECM was greater in Egr-1–deficient mice compared with wild-type mice both 3 and 7 days after the final CCl₄ exposure and was predominantly bridging in morphology (Figure 3, A–H, quantification in I). In addition, the pattern of ECM deposition differed between wild-type and egr-1^−/− mice. Fibrotic septa in wild-type mice were tight distinct bands between hepatic chords, minimizing the total area of the liver affected by ECM deposition (Figure 3, A and E, C and G). In contrast, fibrotic septa in egr-1^−/− mice were diffuse and punctuated by areas of Sirius red-negative tissue (Figure 3, B and F, D and H). Differences in hepatic CYP2E1 expression were not observed between genotypes at any time point examined after chronic CCl₄ exposure (see supplemental Figure S2 at http://ajp.amjpathol.org).

Figure 3.

Hepatic fibrosis after chronic CCl₄ exposure. Wild-type and egr-1^−/− mice were treated with olive oil (not shown) or CCl₄ for five weeks. Three (A and B) and seven days (C and D) after final injection, mice were sacrificed and their livers removed, processed, and stained with Sirius red to visualize collagen fiber/ECM deposition. Outlined areas in A–D are enlarged to illustrate differences in the pattern of ECM deposition and are denoted in E–H, respectively. Asterisks in enlarged images denote central veins. Images are representative of n = 4 to 5 mice per experimental group, and two ×100 fields per mouse. I: Morphometric assessment of total Sirius red density in livers from mice three and seven days after cessation of chronic CCl₄ exposure. Data represent means ± SEM, n = 4 to 5 mice per experimental group and two ×100 fields per mouse. There is a significant difference when compared with olive oil control for same genotype (*P < 0.05). There is a significant difference when compared to wild-type at same time point (**P < 0.05).

Enhanced Activation of the Oval Cell Compartment in egr-1^−/− Mice after Chronic CCl₄

Masson’s trichrome staining was used to visualize localization of ECM components (in bright blue) within the context of liver architecture (in shades of dark purple and pink) to further investigate the morphology of the diffuse ECM deposition in livers from egr-1^−/− mice after chronic CCl₄ exposure (Figure 4). The diffuse nature of ECM staining in egr-1^−/− mice revealed by Sirius red staining was attributable, at least in part, to the interruption of the fibrotic septa by large numbers of small cells situated singly or in small groups within the collagen fibers at both time points after chronic CCl₄ (Figure 4, A and B); these cells were not present in the livers of wild-type mice. The nuclear morphology of a portion of the cells was consistent with the presence of a hepatic progenitor cell type called oval cells (Figure 4, enlarged insets). The putative oval cells were not active HSC as they were not labeled by immunostaining for αSMA (Figure 5). αSMA-positive cells were limited to uninterrupted areas between hepatic chords in wild-type mice, whereas diffuse and extensive αSMA staining found in livers from egr-1^−/− mice was localized predominately to the perimeter of hypercellular regions that contained cells with oval cell morphology (Figure 5).

Figure 4.

Pronounced cellular hyperplasia in egr-1^−/− mice was associated with diffuse ECM deposition after chronic CCl₄ exposure. Wild-type and egr-1^−/− mice were treated with olive oil (not shown) or CCl₄ for five weeks. Mice were then sacrificed and their livers removed for histological analysis and ECM deposition analysis by hematoxylin and eosin (H&E) and Masson’s trichrome staining, respectively in serial sections (A) 3 days or (B) seven days after the end of chronic CCl₄ exposure. Outlined areas in ×400 H&E images are enlarged for clarity and reproduced to the right of each ×400 image. Black arrows denote cells whose nuclear morphology is consistent with hepatic oval cells. Images are representative of n = 4 to 5 mice per experimental group.

Figure 5.

Immunolocalization of αSMA-positive cells in liver after chronic CCl₄ exposure. Wild-type and egr-1^−/− mice were treated with olive oil (not shown) or CCl₄ for five weeks. Mice were sacrificed and their livers removed and processed for αSMA immunohistochemistry to visualize activated HSC three days after the final injection. Outlined areas in ×100 images were photographed again at ×200 to better illustrate localization of αSMA-positive cells in relation to hypercellular areas observed in livers from egr-1^−/− mice, and the lack of this pattern in livers from wild-type mice. Images are representative of n = 4 to 5 mice per experimental group.

Hepatic oval cell activation is associated with the expression of a set of genes including α fetoprotein (AFP)35 and thymocyte differentiation antigen (Thy) −1,36 and oval cell growth factors including CTGF, hepatocyte growth factor (HGF), and its receptor c-met, as well as TWEAK and its receptor, Fn14.37,38 AFP expression was induced in livers of both wild-type and egr-1^−/− mice 3 days after final CCl₄ exposure, but this increase was sevenfold greater in egr-1^−/− mice (Figure 6A). Although AFP levels returned to baseline in wild-type mice at 7 days, AFP remained eightfold over baseline in egr-1^−/− mice (Figure 6A). Hepatic Thy-1 and HGF expression were induced in egr-1^−/− mice, but not in wild-type mice, 3 and 7 days after CCl₄ cessation (Figure 6, B and C). Along with its profibrotic roles,39 CTGF is also an important mediator of oval cell activation and migration.40 CTGF is expressed in the livers of both wild-type and egr-1^−/− mice 3 days after CCl₄ (Figure 6D). CTGF expression returns to baseline in wild-type mice 7 days after CCl₄ while its expression is maintained in egr-1^−/− mice (Figure 6D). There were no detectable changes in expression of HGF receptor c-met or differences in TWEAK and Fn14 expression between wild-type or egr-1 ^−/− mice (see supplemental Figure S3 at http://ajp.amjpathol.org).

Figure 6.

Expression of mRNAs associated with oval cell activation after chronic CCl₄ exposure. Wild-type and egr-1^−/− mice were treated with olive oil or CCl₄ for five weeks. Mice were sacrificed and hepatic expression of (A) AFP, (B) Thy-1, (C) HGF, and (D) CTGF was determined by real-time PCR at the indicated time points. Relative amounts of gene-specific mRNA were calculated using the 2^−ΔΔCt method and results expressed as fold change over olive oil controls for each genotype after normalization to 18S (housekeeping gene). The data are means ± SEM, n = 5 to 9 mice per experimental group. There is a significant difference in gene expression after CCl₄ when compared with genotype-specific olive oil control (*P < 0.05). There is a significant difference in gene expression after CCl₄ when compared with genotype-specific olive oil control and between genotypes at the indicated time point (**P < 0.05).

To specifically identify oval cells in the livers of mice after chronic CCl₄, we performed immunostaining using the well-characterized A6 antibody, which recognizes an antigen expressed in mouse oval cells.31 Livers from olive oil–treated wild-type and egr-1^−/− mice did not exhibit positive A6 staining apart from normal A6 immunolocalization in the portal tract (data not shown). After chronic CCl₄, a few small A6-positive cells were found in clumps outside the portal tracts of wild-type mice (Figure 7A). In contrast, egr-1^−/− mice had robust A6 immunostaining throughout the liver both 3 and 7 days post cessation of chronic CCl₄ exposure (Figure 7A). These cells were larger than those in wild-type mice and were found migrating through the liver parenchyma in chords, as well as in clumps and rings (Figure 7A). Three days after cessation of chronic CCl₄, there were fourfold more A6-positive cells per 100 × field in livers from egr-1^−/− mice compared with wild-type mice (Figure 7B). The A6 response in livers of egr-1^−/− mice was sustained 7 days after cessation of CCl₄; sevenfold more A6-positive cells were found in livers of egr-1^−/− mice compared with wild-type mice at this time point (Figure 7B).

Figure 7.

Localization of A6-positive cells in fibrotic livers from wild-type and egr-1^−/− mice. A: Both three and seven days after cessation of chronic CCl₄ exposure, mice were sacrificed and frozen liver sections probed with an antibody to the oval cell–specific antigen, A6, by immunohistochemistry. B: Quantification of A6-positive cells in olive oil–treated mice and three and seven days post chronic CCl₄. Data represent means ± SEM, n = 4 to 5 mice per experimental group, four ×100 fields per mouse. There is a significant difference in A6 expression after CCl₄ when compared with genotype-specific olive oil control (*P < 0.05). There is a significant difference in A6 expression after CCl₄ when compared with genotype-specific olive oil control and between genotypes at the indicated time point (**P < 0.05).

Discussion

The data presented in this article are the first to demonstrate that the transcription factor Egr-1 serves as a brake against fibrogenesis in the liver after chronic CCl₄ exposure in mice. Enhanced HSC activation and presence of a profibrotic milieu in livers from Egr-1–deficient mice after a single exposure of CCl₄ suggests that perturbations early in the hepatic wound healing response contribute to enhanced fibrosis after multiple exposures to CCl₄. Furthermore, this is the first description of robust oval cell activation in mice after chronic CCl₄ exposure in the absence of an exogenous antimitotic or other DNA damaging agent. These novel data provide evidence to suggest that Egr-1 is a critical negative regulator of fibrosis and oval cell activation in the liver after hepatotoxicant exposure.

Hepatic fibrosis is the result of repeated cycles of liver injury coupled to incomplete resolution of a normal wound healing response.3 Successful injury repair involves an orderly progression of coagulation, inflammation, angiogenesis, tissue formation, and ECM remodeling.1 However, during chronic injury, this cycle is interrupted mid-sequence and then re-initiated. This promotes excessive ECM deposition and reduced matrix metabolism and contributes to eventual hepatic dysfunction associated with fibrosis. In addition to fibrosis, prolonged exposure to liver injury is associated with impaired hepatocyte regenerative capacity. When this occurs, activation of the hepatic progenitor cell compartment and progenitor cell-mediated hepatocyte replacement ensues.41

Egr-1 is a transcription factor with known roles in the regulation of genes involved in the wound healing response, including those involved in inflammation, coagulation, angiogenesis, cell proliferation, and matrix remodeling.11,12,13,17,18,19,20,21,22,23,24,25 Egr-1 is induced by lipopolysaccharide, inflammatory cytokines, and reactive oxygen species. These activators are often increased in response to tissue injury5 and suggest that Egr-1 is a master regulator of gene expression during wound healing.

Making use of egr-1^−/− mice, we identified the transcription factor Egr-1 as a negative regulator of hepatic fibrosis. Enhanced activation of HSC and expression of profibrotic genes after acute CCl₄ exposure was associated with enhanced ECM deposition after chronic CCl₄ exposure in egr-1^−/− mice. Wild-type mice transiently lose weight after CCl₄ exposure. Body weight is recovered in these mice by 7 days after completion of chronic CCl₄ exposure, but not in egr-1^−/− mice (see supplemental Figure S1 at http://ajp.amjpathol.org). These data support the hypothesis that worsened fibrosis was associated with delayed recovery from CCl₄ exposure in egr-1^−/− mice. Enhanced hepatic fibrosis in egr-1^−/− mice in this study contrasts with the previously reported role that Egr-1 plays in bleomycin-induced fibrosis of the skin, lung, and unilateral ureteral obstruction-induced fibrosis of the kidney. In these reports, deficiency of Egr-1 attenuates fibrosis and fibrotic markers.16,42 Tissue specific differences between the liver and these other organs in the contribution of Egr-1 to chemically-induced fibrosis may stem from the liver’s predominant role in toxin metabolism. This is exemplified by the liver’s unique capacity to regenerate after toxin-mediated hepatocyte loss, as well as after other forms of liver injury.

Deficiency of Egr-1 is associated with elevated plasma liver function enzyme activity, necrosis, and apoptosis after acute CCl₄ exposure (unpublished data). In addition, expression of the Egr-1–regulated proinflammatory cytokine TNFα is reduced in livers from egr-1^−/− mice after acute CCl₄ exposure (unpublished data). This is associated with attenuated expression of hepatoprotective genes including inducible nitric oxide synthase, cyclooxygenase 2, and oncostatin M (unpublished data). Ergo, in the absence of appropriate hepatoprotection after acute CCl₄ exposure, liver injury is enhanced in mice deficient in Egr-1 and would promote worse fibrosis and delayed recovery from chronic CCl₄ exposure. Increased tissue debris in the egr-1^−/− mice could promote enhanced HSC activation3,43 and subsequent fibrosis after chronic CCl₄ exposure. Increased injury along with excess tissue debris and HSC activation suggest a delay in wound resolution in egr-1^−/− mice after a single exposure to CCl₄. Delayed wound resolution would predispose Egr-1–deficient animals to progressively worse CCl₄-induced liver injury and fibrosis with each dose in a chronic (fibrosis-inducing) CCl₄ exposure paradigm.

Macrophages play a key role in response to injury, at least in part by acting to clear necrotic tissue. Expression of MCP-1, an Egr-1–regulated macrophage chemoattractant, was reduced in livers from egr-1^−/− mice relative to wild-type mice after acute CCl₄ exposure (see supplemental Figure S4 at http://ajp.amjpathol.org), consistent with reduced MCP-1 expression in egr-1^−/− mice in other animal models of acute injury.8,10 Reduced MCP-1 expression could negatively affect recruitment of macrophages to the liver and delay ‘wound sterilization’ and subsequent liver repair. In support of this hypothesis, liver injury is potentiated and repair is attenuated after acute CCl₄ exposure in mice depleted of Kupffer cells, the resident macrophage in the liver, as well as in mice deficient in the MCP-1 receptor, CCR2.44,45 Together, these data indicate that recruitment of macrophages is an integral component to wound healing in the liver after CCl₄ exposure and that Egr-1 deficiency impairs this response.

The hepatic oval cell response is activated only in situations where liver injury is so severe that proliferation of existing hepatocytes is insufficient to rapidly reinstate normal liver function and/or when proliferation of hepatocytes is inhibited.38 Despite its hepatotoxicity, CCl₄ exposure alone is not sufficient to promote oval cell activation.27 To study the oval cell response in mice and rats, investigators combine two different agents: a hepatotoxin or 2/3 partial hepatectomy to induce liver injury and an antimitotic or DNA-damaging agent to inhibit hepatocyte proliferation.41,46,47,48,49,50 Aside from a single report that suggested an inconsistent and rare presence of cells with oval cell-like morphology in liver parenchyma after CCl₄ exposure,51 the current study is the first to fully characterize a robust activation of the oval cell response in mice using CCl₄ in the absence of exogenous inhibitors of hepatocyte proliferation.

We hypothesize that more severe liver injury in the egr-1^−/− mice after acute CCl₄, (unpublished data), coupled with an attenuation or delay of hepatocyte proliferation, is responsible for activation of the oval cell response in egr-1^−/− mice observed in the current study. Expression of cytokines, TNFα, IL6, and the complement activation products C3a and C5a, are required to induce hepatoprotection and proliferative competence in hepatocytes after CCl₄ exposure or partial hepatectomy.52,53,54,55,56 TNFα expression was reduced in livers from egr-1^−/− mice after acute CCl₄, exposure (unpublished data). This is associated with attenuated bromodeoxyuridine incorporation (Pritchard and Nagy, manuscript in preparation) and suggests a negative impact of reduced TNFα expression on liver regeneration in egr-1^−/− mice. After partial hepatectomy, Liao et al demonstrated attenuated hepatocyte proliferation in egr-1^−/− mice and identified reduced expression of Cdc20, a protein required for the anaphase promoting complex and successful mitosis, as the explanation for reduced hepatocellular proliferation.20 In that study, Egr-1 deficiency delayed hepatocyte proliferation but did not prevent liver regeneration.20 Importantly, activation of the oval cell response was not noted. Unlike the single cycle of liver injury and repair induced after partial hepatectomy, the iterative injury and fibrosis induced by chronic CCl₄ is likely critical for oval cell activation in egr-1^−/− mice.

Collectively, the data in this report suggest that Egr-1 is an important negative regulator of hepatic fibrosis and oval cell activation; the specific mechanisms of which remain to be elucidated. The aberrant response to tissue injury in egr-1^−/− mice is likely initiated by reduced hepatic inflammation early in the response to CCl₄ exposure (unpublished data) and supplemental Figure S4 (http://ajp.amjpathol.org). Reduced inflammation is then associated with a cascade of events promoting more severe liver injury and reduced hepatic wound healing. The cumulative result of dysregulated hepatic wound healing in egr-1^−/− mice was enhanced fibrosis and activation of the oval cell response. In this genetic model, attenuated inflammation and wound healing promoted hepatic fibrosis. Therefore, caution should be applied when developing new strategies to eliminate early inflammatory processes, such as increased TNFα, in response to injury to attenuate fibrosis. Maintaining the normal inflammatory response, rather than eliminating it, is critical for normal wound-healing. Indeed, the current data, as well as our recently reported data (unpublished data), support the hypothesis that precisely orchestrated induction and resolution of hepatic inflammation is a prerequisite for normal execution of hepatic wound healing and that Egr-1 is a critical regulator of this process.