Activation of NF-κB and STAT3 in rat oval cells during 2-acetylaminofluorene/partial hepatectomy-induced liver regeneration


  • Aránzazu Sánchez,

    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Valentina M. Factor,

    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Insa S. Schroeder,

    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Peter Nagy,

    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD
    2. 1st Institute of Pathology and Experimental Cancer Research, Semmelweis University of Medicine, Budapest, Hungary
    Search for more papers by this author
  • Snorri S. Thorgeirsson

    Corresponding author
    1. Laboratory of Experimental Carcinogenesis, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD
    • National Cancer Institute, 37 Convent Drive MSC 4262, Building 37, Room 4146A, Bethesda, MD 20892-4262
    Search for more papers by this author
    • fax: 301-496-0734


Proliferation and differentiation of hepatic stem cell progenies (i.e., oval cells) sustain liver regeneration when the replicative and functional capacity of hepatocytes is impaired. The signaling pathways that control stem cell activation remain poorly understood. In this study, we investigated the involvement of nuclear factor-kappa B (NF-κB) and signal transducer and activator of transcription 3 (STAT3) in oval cell-mediated liver regeneration induced by 2-acetylaminofluorene/partial hepatectomy (AAF/PH) protocol. Using OV1 as a marker for identification and sorting of oval cells, we established that both NF-κB and STAT3 were highly activated in the OV1+ cell population. Three distinct subpopulations of oval cells were defined as OV1low, OV1medium, and OV1high, based on the intensity of OV1 staining. Quantitative polymerase chain reaction analysis revealed that they represent different stages of oval cell differentiation along hepatocyte lineage. OV1low cells displayed the least differentiated phenotype as judged by high expression of c-kit and lack of hepatocytic differentiation markers, whereas OV1high cells lost c-kit expression, were more proliferative, and acquired more mature hepatocytic phenotype. Notably, NF-κB was activated uniformly in all three subpopulations of oval cells. In contrast, phosphorylation of STAT3 was detected only in OV1high cells. In conclusion, transcriptional activity supported by NF-κB and STAT3 is required for oval cell activation, expansion, and differentiation. The differential induction of NF-κB and STAT3 point to a distinct role for these transcription factors at different stages of hepatic stem cell differentiation. (HEPATOLOGY 2004;39:376–385.)

Although hepatocytes possess a remarkable proliferative potential, well illustrated in a series of recent transplantation experiments,1, 2 the adult liver contains a stem cell compartment that can be activated under conditions of severe liver injury.3, 4 It is generally thought that the distal part of the biliary tree, the so-called canals of Hering, constitute the stem cell niche in the adult liver. The stem cell progenies, commonly referred to as oval cells, possess traits that are phenotypically similar to both fetal hepatocytes and bile epithelial cells. Thus, oval cells express α-fetoprotein, a marker of fetal hepatocytes as well as isozymes of aldolase and pyruvate kinases found in fetal and adult hepatocytes. Similarly, oval cells express cytokeratins 7 and 19, specific for biliary epithelial cells.5 In an attempt to establish lineage relationships between oval cells and other epithelial cell types in the liver, a variety of antigens recognizing oval cells were generated, most of which also were shared by bile duct cells, including OV1 and OV6.6, 7 In addition, oval cells express unique markers not found on hepatocytes and biliary epithelial cells, such as stem cell factor/c-kit,8 bcl-2,9 and cytokeratin 14.10

The signaling pathways that control oval cell proliferation and differentiation remain poorly understood. Elucidating the mechanism(s) that regulate the activation and expansion of the hepatic stem cell compartment is of fundamental importance because it is a prerequisite for understanding the stem cell-driven liver regeneration and the potential application of stem cells in regenerative medicine. Previous work from our laboratory has demonstrated a unique activation of the stem cell factor/c-kit system and the interferon-γ network during liver regeneration from oval cells.8, 11 In addition, the primary growth factors involved in normal hepatic regeneration, including transforming growth factor (TGF)-α, hepatocyte growth factor (HGF), acidic fibroblast growth factor (aFGF), and TGF-β, have been shown to play important roles in proliferation and differentiation of oval cells.12, 13 Similarly, suppression of interleukin (IL)-6 and tumor necrosis factor (alpha) (TNF) production after Dexamethasone administration inhibited not only replication of hepatocytes, but oval cell proliferation as well.14 These data are consistent with the observation that oval cell proliferation is impaired in TNFR1 knockout mice.15 Furthermore, Kirillova et al.16 demonstrated that TNF was capable of stimulating proliferation of an oval cell-derived cell line in vitro, an effect that was associated with activation of nuclear factor-kappa B (NF-κB) and signal transducer and activator of transcription 3 (STAT3). These data support the role of TNF-driven signaling in oval cell proliferation. The importance of IL-6 in oval cell-mediated regeneration is less clear because an increased proliferation of oval cells was found in IL-6 knockout mice after cocaine-induced periportal injury.17 Together, these observations suggest that growth regulation of hepatocytes and liver stem cells may share a number of common features, but distinct features for each of these cell lineages also may exist.

To determine if NF-κB and STAT3 are involved in the activation or expansion of oval cells during liver regeneration, or both, we examined the activities of these transcription factors in oval cells induced by the 2-acetylaminofluorene/partial hepatectomy (AAF/PH) protocol. Three distinct subpopulations of oval cells were isolated, phenotypically characterized, and used to assess the requirements for each transcription factor in oval cell activation, proliferation, and differentiation. We show here that NF-κB is activated in all three subpopulations, whereas activation of STAT3 is linked to differentiation and propagation of oval cells. These studies suggest critical but distinct roles for NF-κB and STAT3 in stem cell-mediated liver regeneration.


TGF, transforming growth factor; HGF, hepatocyte growth factor; aFGF, acidic fibroblast growth factor; NF-κB, nuclear factor-kappa B; STAT3, signal transducer and activator of transcription 3; AAF, 2-acetylaminofluorene; PH, partial hepatectomy; FACS, fluorescence-activated cell sorting; EMSA, electrophoretic mobility shift assay; HNF4, hepatocyte nuclear factor 4.

Materials and Methods

Animal Model.

Oval cell proliferation was induced in male Fischer F344 rats (weight, 140–180 g) using the AAF/PH protocol.18 Briefly, 10 mg/kg of AAF was administered daily by gavage for 8 days. A standard 2/3 partial hepatectomy was performed 4 days after starting the AAF administration. Animal study protocols were conducted according to National Institutes of Health guidelines for animal care.

Hepatocyte Isolation.

Hepatocytes were isolated by two-step collagenase perfusion of the liver followed by isodensity centrifugation in Percoll as described.19 Viability was determined by trypan blue exclusion and was >90%.

Flow Cytometry and Cell Sorting.

Nonparenchymal cell (NPC) populations were isolated at 1 to 14 days after PH as described20 and were stained with a mouse monoclonal OV1 antibody recognizing an unknown surface antigen specific for oval and biliary epithelial cells.7 The staining was performed in Hanks balanced salt solution (HBSS)/0.1% BSA for 1 h at 4°C at a 1:10 dilution, followed by incubation with a Phycoerythrin (PE)-conjugated secondary antibody (goat F(ab′)2 anti-mouse IgG3), 1:10 dilution (SouthernBiotech, Birmingham, AL). After two washes in HBSS/0.1%BSA, cells were used either for magnetic cell sorting with anti-PE beads (Miltenyi Biotec Inc., Auburn, CA) or for fluorescence-activated cell sorting (FACS) using the FACS Vantage (BD Bioscience, San Jose, CA). The FACS Vantage was equipped with Turbo Sort for high speed sorting and Digital Electronics and Diva Software (BD Bioscience) that allow sorting of four populations at once. Sorting settings were as follows: pressure, 22 psi; speed, 15,000 cells/second; nozzle, 90 μm. After sorting, viability was determined for each of the cell populations by trypan blue exclusion and was typically >90%.

Hoechst Staining and Cell Cycle Analysis.

5x106NPC were stained with 5 μg/mL Hoechst 33342 (Sigma, St. Louis, MO) in HBSS with 2% fetal bovine serum (FBS) for 90 minutes at 37°C. 50 μM Verapamil (Sigma) was added to prevent dye efflux. Cells were centrifuged and stained with OV1 antibody as described. Double staining analysis for Hoechst and OV1 (blue and red channels respectively) was performed in a FACS Vantage cytometer. DNA content in each gated cell subpopulation was quantified by using ModFit software (Verity Software House, Inc., Topsham, ME).

Nuclear Extracts and Electrophoretic Mobility Shift Assay (EMSA).

Preparation of nuclear extracts from livers and sorted cell populations was performed as described.21, 22 Protein concentration was measured by the Bio-Rad protein reagent following the recommendations of the supplier. For EMSA, 5 μg of nuclear protein were mixed with 0.5 ng of radiolabeled oligonucleotide and 1μg poly (dI-dC) as nonspecific DNA competitor and incubated for 30 minutes at 4°C in binding buffer containing 10 mM HEPES, pH 7.9, 50 mM NaCl, 1 mM ethylenediaminetetraacetic acid, and 10% glycerol. Supershift experiments were performed by preincubating 2 μL of antibody targeted to NF-κB p65 with the nuclear extracts for 2h at 4°C before addition of the radiolabeled probe. DNA binding complexes were resolved by gel electrophoresis on 6% polyacrylamide/0.5x tris borate EDTA (TBE). The gels were dried and exposed to Kodak X-AR film (Eastman Kodak Co., Rochester, NY) at −80°C for 3 to 12 hours. NF-κB consensus oligonucleotides were obtained from Promega and end labeled using T4 polynucleotide kinase (Invitrogen, Carlsbad, CA) in the presence of γ-32P-adenosine triphosphate (ATP) (ICN Biomedicals Inc., Irvine, CA). Anti-NF-κB p65 (F-6, X) was from Santa Cruz Biotechnologies, Santa Cruz, CA.

Protein Preparation and Western Blot Analysis.

Fresh sorted cell populations were lysed with mammalian protein extraction reagent (M-PER™) buffer (Pierce Chemical Co., Rockford, IL) supplemented with 1 mM phenylmethylsulfonylfluoride (PMSF), 0.2 mM Na3VO4, 20 mM NaF and protease inhibitor cocktail (Roche). Fifty micrograms of protein were separated by electrophoresis on 4% to 20% gradient gels (Novex) and blotted on polyvinylidene fluoride (PVDF) membranes (Invitrolon, Invitrogen). Primary antibodies used were: anti-STAT3 antibody (BD Transduction Laboratories, San Diego, CA), anti-Phospho-STAT3 (Tyr705; Cell Signaling Technology, Beverly, MA), anti-OV6 (kindly provided by Dr. Douglas Hixson), anti-cdc2 (H-297), anti-cyclin D1 (C-20) from Santa Cruz Biotech (Santa Cruz, CA), and anti-GAPDH (Trevigen, Gaithersburg, MD). Secondary antibodies were conjugated with horseradish peroxidase (Amersham Biosciences Corp., Piscataway, NJ) and proteins were visualized by using Super Signal-West Pico Chemiluminescent Substrate (Pierce).

RNA Isolation.

Total cellular RNA was prepared using the RNeasy Kit (Qiagen, Valentia, CA) according to the manufacturer's instructions with on-column DNase treatment (RNase-free DNase Set, Qiagen) for 45 minutes to avoid genomic DNA contamination. RNA yields and purity were analyzed using a spectrophotometer (DU 640; Beckman Coulter, Fullerton, CA).

Quantitative Polymerase Chain Reaction (PCR) Analysis.

Changes in mRNA levels of specific genes were quantified via TaqMan methodology in an ABI PRISM 7900HT Sequence Detection System (PE Applied Biosystems, Foster City, CA). All materials, primers, and probes were purchased from PE Applied Biosystems. A set of specific TaqMan oligonucleotide primers and probes were designed using Primer Express software 2.0 (PE Applied Biosystems). All probes were 5′-6-carboxyfluorescein (FAM) (6-carboxyfluorescein)-labeled at their 3′ end with 6-carboxytetramethylrhodamine (TAMRA) as a quencher at the 5′ end. Differences in the threshold cycle number were used to quantify the relative amount of PCR targets. The amplification mix was prepared following the TaqMan Gold reverse transcription PCR core kit protocol. Primers and probes for lineage marker genes were used in concentrations of 100 nM and 50 nM, respectively, and the sequences are given in Table 1. Rodent glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a reference gene (TaqMan Rodent GAPDH Reagents kit, VIC-labeled). Primer and probe concentrations for GAPDH were 200 nM and 100 nM. Each reaction was performed in triplicate in 96-well optical reaction plates under the following conditions: 50°C for 2 minutes, 95°C for 10 minutes, followed by 45 to 50 cycles of 95°C for 15 seconds (denaturation) and 60°C for 1 minute (annealing and extension). Relative expression of hepatic marker genes was calculated using the comparative CT method after confirming that reference and marker genes were amplified with nearly identical efficiencies.

Table 1. Sequences of Primer and Probes Used in Real-Time Polymerase Chain Reaction
GeneAmplicon Size (bp)Forward PrimerReverse PrimerProbe*
  • Anneling and extension temperature, 60°C.

  • *

    Reporter: FAM/Quencher: TAMRA.

  • Primers used are specific for the 2.1-kb transcript.


cDNA Arrays.

GEArray Q Series mouse apoptosis gene array and mouse cell cycle array (Superarray, Inc., Frederick, MD) were used to hybridize cDNA from sorted cells. Five micrograms total RNA were used for cDNA probe synthesis. Both probe labeling and hybridization were performed according to the manufacturer's instructions following the radioactive detection method.


Proliferation of Oval Cells Is Associated With Activation of NF-κB and STAT-3.

To evaluate the involvement of NF-κB in the regulation of oval cell proliferation, we performed a time-course analysis of NF-κB activation in animals subjected to AAF/PH treatment. EMSA of liver nuclear extracts revealed a biphasic pattern of NF-κB activation after PH (Fig. 1A). The first, smaller peak was observed from 30 minutes to 2 hours after surgery, consistent with previous findings in regenerating livers.23, 24 However, a more prominent and persistent increase in NF-κB activity was found at 4 to 14 days after PH, coincident with proliferation and expansion of oval cells.18 The specificity of the reaction was confirmed by supershift experiments with antibodies against the p65 subunit of NF-κB (Fig. 1B). This result strongly suggested that oval cell proliferation and expansion was associated with an activation of NF-κB. To support the hypothesis that NF-κB activity resided in the oval cells, we purified oval cells from the NPC population isolated at 8 and 9 days after PH using magnetic sorting with an OV1 monoclonal antibody that recognizes a surface-specific epitope on oval cells.7 The purity of the sorted OV1 and OV1+ populations was 88.23% ± 1.65% (SD) and 87.91% ± 4.09%, respectively, as evidenced by flow cytometric analysis (Fig. 2A). Nuclear extracts prepared from these highly enriched cell populations showed strong NF-κB DNA binding in OV1+ cells, whereas the OV1 subpopulation displayed very low or no binding (Fig. 2B). Similarly, high levels of active STAT3 were found only in OV1+ cells. These results demonstrate a strong and prolonged activation of NF-κB and STAT3 in oval cells during stem cell-mediated liver regeneration.

Figure 1.

Activation of NF-κB after AAF/PH. (A) Time course of NF-κB DNA binding activity by EMSA. Five micrograms of nuclear protein were used to analyze DNA binding activity by EMSA as described in Materials and Methods. (B) Representative supershift analysis performed with extracts obtained from liver 30 minutes after PH. The bracket indicates the NF-κB–DNA supershifted complexes.

Figure 2.

Magnetic sorting of the OV1+ and OV1 oval cell populations. (A) FACS dot plots (FL2-H vs. side scatter, SSC-H) corresponding to the nonparenchymal cell population stained with OV1 antibody before sorting (left plot) and after sorting (middle and right plots). Gates define the negative (R1) and positive (R2) populations. Note the significant enrichment in the OV1+ population after sorting. (B) EMSA of NF-κB DNA binding activity in sorted cells from livers at 8 and 9 days after PH. Five micrograms of protein were used for the assay. (C) Immunoblot analysis of STAT3 in sorted cells. Detection of P-STAT3 and STAT3 was carried out as described in Materials and Methods. A representative experiment of three is shown.

Phenotypic Characterization of the OV1+ Cell Population.

Work from our laboratory as well as that of others has established that the composition of the oval cell compartment is very heterogeneous with regard to both phenotype and gene expression profile.10, 25, 26 Because OV1 is thought to recognize an early population of oval cells,7, 26 we reasoned that the OV1+ cells should preferentially harbor an early hepatic stem cell progeny. Interestingly, the intensity of OV1 staining was extremely heterogeneous. FACS analysis revealed that the range of OV1 fluorescence varied within three orders of magnitude in the OV1+ population. To rule out the possibility that the extent of OV1 expression is related to protease digestion during NPC isolation, we verified that the pattern of OV1 staining did not depend on the duration of pronase treatment (10, 20, 40, or 60 minutes; data not shown). In all cases, OV1+ cells could be subdivided into three distinct subpopulations: OV1low, OV1medium, and OV1high cells. As shown in Fig. 3, a total of four distinct cell populations were gated and selected for FACS sorting. Cells within the R1 gate corresponded to OV1 cells, representing mixed nonparenchymal cells of different origin, whereas the R2, R3, and R4 gates defined three oval cell subpopulations based on the intensity of OV1 fluorescence. The frequency of OV1low, OV1medium, and OV1high cells was 6.48% ± 1.49% (SD), 8.16% ± 2.14%, and 8.46% ± 1.54%, respectively. Dot plots for each of the sorted populations confirmed the high purity of isolated cell subpopulations, which was 96.38% ± 5.24%, 96.48% ± 1.10%, 94.82% ± 2.03%, and 94.79% ± 1.71% for R1, R2, R3, and R4, respectively (Fig. 3C). To understand the functional significance of the distinct OV1+ subpopulations, we next prepared whole cell lysates from all sorted cell fractions and examined the expression of OV6, another oval cell-specific marker.6 Notably, only OV1high but not OV1low or OV1medium cells expressed OV6 antigen (Fig. 3D), consistent with previous findings that OV1 and OV6 antibodies recognize different subpopulations of oval cells.7, 27 As expected, no OV6 expression was detected in OV1 cells. These results supported our hypothesis that phenotypic heterogeneity defined by the intensity of OV1 staining may be related to the differentiation status of oval cells. To address this issue directly, we analyzed the expression of lineage specific markers in all sorted cell populations by quantitative PCR. Remarkably, the expression of c-kit, a well-known stem-cell marker, was readily detected in OV1low and OV1medium cells, but it was markedly lower in the OV1high population (Fig. 4). In contrast, mRNA levels of transthyretin, albumin, α-fetoprotein, connexin 32, and hepatocyte nuclear factor-4 progressively increased from OV1low to OV1medium and reached the maximum in OV1high cells. Nonetheless, the expression levels of hepatocyte differentiation markers in oval cells were still lower than in adult hepatocytes. Taken together, these results clearly show that the OV1+ population contains oval cells at different stages of hepatic differentiation, ranging from the primitive OV1low/OV6 cells to more mature OV1high/OV6+ cells.

Figure 3.

FACS of the OV1+ and OV1 oval cell populations. (A) Dot plot representing the OV1 staining profile before sorting. Gates R1, R2, R3, and R4 define the OV1, OV1low, OV1medium, and OV1high populations, respectively, separated by the signal intensity of OV1 fluorescence and selected for the cell sorting. (B) Dot plot corresponding to the background signal from cells incubated with PE-conjugated IgG (secondary antibody alone), used to define the position of the negative (R1) and positive (R2, R3, and R4) cell populations. (C, D, E, F) Dot plots of the gated populations after sorting. Note the high purity of all sorted populations. A representative experiment of five is shown. (G) Example of Western blot analysis of OV6 expression after sorting.

Figure 4.

Comparative analysis of expression of hepatic differentiation markers in sorted cell populations by quantitative PCR. Total RNA isolated from fresh sorted cell populations at 8 days after PH was used for quantitative PCR analysis as described in Materials and Methods. Values are mean ± SD of three separate experiments made in triplicate and expressed as fold difference relative to the OV1 population. Hepatocytes values are shown as a reference.

NF-κB and STAT-3 Are Differentially Activated Within OV1+ Subpopulations.

In further experiments, we examined possible roles for NF-κB and STAT3 at different stages of oval cell differentiation. For this purpose, we isolated OV1low and OV1high cells at 8 days after PH and performed gel shift and Western blotting analyses. Both cell fractions demonstrated high DNA binding activity of NF-κB regardless of their differentiation status (Fig. 5A). In contrast, phosphorylation of STAT3 was found only in the more differentiated OV1high cells (Fig. 5B). OV1 cells showed very low or undetectable levels of NF-κB and STAT3, consistent with our previous findings (Fig. 2B). These results demonstrate that oval cell differentiation toward hepatocytic lineage is driven by differential activation of NF-κB and STAT3. The fact that STAT3 is activated specifically in the more differentiated oval cells provides strong evidence for the importance of the STAT3-mediated signaling pathway in the differentiation process.

Figure 5.

Analysis of NF-κB and STAT3 activation in sorted cell subpopulations. (A) EMSA of NF-κB DNA binding activity in OV1 (N), OV1low (Low), and OV1high(Hi) sorted populations. Five micrograms of nuclear extracts isolated from sorted cells at 8 days after PH were used. (B) Western blot analysis of STAT3 in sorted cells. Fifty micrograms of total protein isolated from the same cell populations were used for SDS-PAGE electrophoresis and detection of both P-STAT3 and STAT3. The experiments were repeated three times using independent animals.

Activation of STAT3 But Not NF-κB Correlates With a Higher Proliferative Activity in Oval Cells.

To understand the possible roles for NF-κB and STAT3 in regulation of oval cell response, the proliferative competence of OV1+ subpopulations was measured by three different assays: FACS analysis, cDNA microarray, and Western blotting. OV1low cells exhibited a low rate of proliferation as judged by Hoechst staining, similar to that found in OV1 cells (Fig. 6A). The number of cycling cells doubled in the OV1high subpopulation throughout the whole process. To explore the differences in proliferation further between OV1lowand OV1high cells, we performed analysis of cell cycle regulated genes using total RNA isolated from OV1+ cells sorted at 8 days after PH. The gene array analysis showed that several genes involved in cell cycle control, including cyclins A2, B, B2, D1, and the cdc2 kinase, required for both G1/S and G2/M transitions, were upregulated two- to ninefold in the OV1high cells as compared with either OV1 or OV1low cells, supporting the results of FACS analysis (Fig. 6B). In addition, the expression of a cyclin-dependent kinase inhibitor, p57/Kip2, was induced in both OV1low and OV1high cells. This is not totally surprising because cyclin kinase inhibitors, including p27/Kip1 and p57/Kip2, are known to be involved in the regulation of progenitor cell proliferation, cell fate specification, and differentiation.28, 29 Finally, we have shown that OV1high cells express higher levels of cyclin D1 and cdc2 proteins than OV1low cells in agreement with microarray data (Fig. 6C). These results indicate that OV1high cells represent the most amplifying cell compartment within the OV1+ population. The preferential activation of STAT3 in the OV1high cells suggests that this transcription factor may play a dual role in supporting both their proliferation and differentiation.

Figure 6.

Cell cycle analysis in oval cell subpopulations. (A) Plot representing percentage of cells in S+G2/M phases of the cell cycle at different times after PH. Double staining with Hoechst 33352 and OV1 was performed as described in Materials and Methods. OV1, OV1low, and OV1high cell subpopulations were gated based on the flow cytometric profiles for OV1 staining using CellQuest. Cell cycle analysis was performed in each individual population by measuring DNA content using ModFit software. Each data point represents mean ± SD of at least three separate experiments. (B) Microarray analysis of cell cycle-related genes. Five micrograms of total RNA were used for microarray analysis (GEArray Q series mouse cell cycle gene arrays) following manufacturer's instructions. A fragment of the microarray membranes containing spots of interest is shown. Genes showing differential expression among the cell subpopulations are marked by the numbered circles (the cutoff was twofold difference in expression). Densitometric analysis of the signals is summarized in the table. Values are presented as arbitrary units after normalization with cyclophilin A (four spots inside the dotted rectangle). (C) Western blot of cdc2 and cyclin D1 in FACS sorted cells. Fifty micrograms of total protein were used for the assay. The experiment was repeated twice with independent animals.


In this paper, we examine the involvement of NF-κB and STAT3 in oval cell-mediated liver regeneration. We show herein that oval cells, the hepatic stem cell progenies, display high NF-κB and STAT3 activities. Furthermore, we demonstrate a differential activation of these transcription factors that depends on the stage of the oval cell development. NF-κB activity appears very early during hepatic stem cell activation and persists throughout the differentiation process along hepatocytic lineage. In contrast, phosphorylation of STAT3 occurs later, coincident with increased oval cell proliferation and differentiation.

Similarly to other stem cell systems, the hepatic stem cell compartment contains only a small fraction of dormant stem cells among a multitude of transit cells at various stages of differentiation.25, 30, 31 Oval cells are believed to represent these transitional cells. Our data clearly demonstrate that OV1 and OV6 are acquired in a sequential order during the oval cell maturation process, in agreement with previous observations.7, 32 Moreover, we were able to isolate and to characterize several distinct subpopulations of oval cells within the OV1+ compartment. The OV1low/OV6 cells seemed to be the early progeny of the dormant stem cells, whereas OV1high/OV6+ cells constituted the amplifying compartment containing highly proliferative and more mature oval cells along the hepatocytic lineage. The conclusion that OV1low/OV6 cells represent an early committed hepatic progenitor cell population is based on the following observations: (1) a high expression of c-kit, (2) a low proliferation rate, and (3) a lack of expression of hepatocytic differentiation markers. These findings and previous studies8, 33 show that the stem cell factor and its receptor c-kit are important during the early stages of hepatic stem cell commitment, although the precise role for this ligand and receptor system is still unknown. The identification of a population of the most primitive hepatic progenitor cells committed to hepatocyte differentiation has allowed us to gain insight into the molecular mechanisms controlling activation and expansion of oval cells. Thus, our work has demonstrated that both NF-κB and STAT3 are activated during oval cell-mediated liver regeneration. The kinetics of NF-κB and STAT3 activation in the AAF/PH model differs substantially from that observed after normal PH. Both transcription factors are considered to be key molecules during the priming step of the regenerative response, providing proliferative competence to hepatocytes.34 NF-κB is rapidly and transiently activated within 0.5 to 2 hours after the operation, whereas the activation of STAT3 occurs later and reaches maximum levels at 2 to 4 hours after PH.23, 24 However, after AAF/PH treatment, we observed a prolonged activation of NF-κB and STAT3, suggesting the functional relevance of these transcription factors during the entire process of oval cell activation, expansion, and differentiation. In the liver, two major roles have been assigned to NF-κB. Besides being a key component of the proliferative response after PH, accumulating evidence suggests that NF-κB constitutes one of the major survival pathways in the liver, not only during liver development but also in the adult organ. Mice lacking RelA/p65, IKKβ, or NEMO/IKKγ die at midgestation as a result of massive fetal liver apoptosis.35–37 NF-κB protects hepatocytes and other hepatic cell lines against apoptosisin vitro.38, 39 Finally, inhibition of NF-κB during liver regeneration after partial hepatectomy either by adenovirus-based delivery of a superrepresor form of IKBα or by administration of gliotoxin leads to a decrease in the mitotic index and massive apoptosis.40, 41 Apoptosis is also of fundamental importance for stem cell biology. In fact, early experiments conducted in our laboratory have demonstrated that 2-AAF treatment induced not only proliferation of oval cells but also apoptosis, which contributed to maintaining an equilibrium in the cell number and to retaining liver morphologic features during the oval cell response.42 Given the prolonged activation of NF-κB and the lack of a specific association either with proliferation or differentiation of oval cells, it seems likely that sustained NF-κB signaling has a critical role in protecting oval cells against apoptosis during stem cell-mediated liver regeneration.

The context of STAT3 activation in oval cells is different. Only the amplifying oval cell compartment exhibits high activity of STAT3. The capacity of STAT3 to promote proliferative responses has been demonstratedin a variety of systems, including liver.43, 44 Here, we provide strong evidence that STAT3 is important for driving the proliferation of the hepatic stem cell progenies. However, the known functions of this transcription factor are broader. Recent work with glycoprotein 130 conditional knock-out mice has shown that activation of STAT3 is glycoprotein 130 dependent and that IL-6/glycoprotein 130-mediated signaling may be more important for protecting the liver from injury than for cell cycle progression after PH.45 STAT3 also can be involved in regulation of differentiation. Interestingly, STAT3 is essential for self-renewal and maintenance of the stem cell phenotype in embryonic stem cells, suppressing their differentiation without affecting the capacity to propagate clonally.46, 47 The role of STAT3 in hepatic stem cell regulation seems to be more complex, because it supports both replication and differentiation. Work performed with murine fetal hepatic cells (E14) demonstrated that hepatocyte maturation and consequent downregulation of cyclin D expression induced by oncostatin M are mediated by STAT3,48, 49 correlating STAT3 signaling with both growth arrest and differentiation. These varied responses elicited by STAT3 in different cell systems could reflect a differential transcriptional regulation, depending on the cell stage. In fact, a maturation state-dependent activation of STAT3 resulting in induction of specific cellular responses has been found during granulocytic differentiation.50 Accordingly, we hypothesize that activation of STAT3 is necessary for the differentiation of oval cells along the hepatocyte lineage. However, because a link between the increased levels of phospho-STAT3 and the higher proliferative capacity of the cells also has been demonstrated, we cannot eliminate the possibility that STAT3 may drive proliferation of the expanding population of oval cells independently on their differentiation stage and may constitute a major force driving proliferation of the expanding population of oval cells.

In conclusion, we provide the first evidence for the involvement of NF-κB and STAT3 in adult hepatic stem cell-mediated liver regeneration. Although the nature of the signals driving NF-κB and STAT3 activation still awaits characterization, our results have opened new research opportunities in liver stem cell research by providing a better characterization of the early populations arising during hepatic stem cell differentiation. Nevertheless we do recognize that a complete understanding of the involvement of NF-κB and STAT3 in hepatic stem cell biology will require the application of novel tools such as specific inhibitors and adenoviral-based systems that will allow selective silencing of the transcription factors in vivo.


The authors thank Dr. Douglas Hixson for the kind gift of OV1 monoclonal antibody and Barbara Taylor for her excellent assistance in FACS sorting.