Wnt/β-Catenin Signaling Contributes to Activation of Normal and Tumorigenic Liver Progenitor Cells

Introduction

In mammals, preexisting hepatocytes and cholangiocytes may replace liver tissue lost as a consequence of liver injury ( 1). However, when this pathway is impaired in response to liver damage, the facultative hepatic progenitor cells initially expand from a niche near the portal triad known as the Canal of Herring and can participate in the liver regeneration ( 1). These cells are viewed as a progenitor cell type because of their coexpression of hepatocellular (ALB, AFP) and biliary (CK19) markers and their known ability to differentiate into either hepatocytes or biliary epithelial cells ( 2). Several antibodies specific for putative hepatic progenitor cell markers, such as OV6, A6, EpCAM, have been developed to aid in their identification ( 3– 5). To date, anti-OV6, a monoclonal antibody raised against cells isolated from carcinogen-treated rat liver ( 6), although also reacts with bile duct epithelium in the rat and in humans ( 7), remains the best available marker of hepatic stem cells ( 8).

With the development of improved techniques for detecting oval cells in liver sections, it has become apparent that hepatic progenitor cells participate in a range of injuries to human liver and is implicated in hepatic cirrhosis and carcinogenesis ( 9). In addition, hepatocellular carcinoma (HCC) individuals with a similar gene expression profile to hepatic stem/progenitor cells have a poor prognosis ( 10). So it is of paramount importance to elucidate the signaling mechanisms regulating the activation, expansion, and differentiation of hepatic progenitor cells, as well as their involvement in human hepatocarcinogenesis.

β-Catenin is an essential component of both intercellular junctions and the canonical Wnt signaling, which plays critical roles in the development of multiple tissues through the regulation of cell proliferation, differentiation, and movement ( 11). Activation of β-catenin in progenitor/stem cells seems to enhance their self-renewal activity and carcinogenesis potential in many tissues ( 12). In liver, there is growing evidence of the involvement of the Wnt/β-catenin pathway in various aspects of liver biology ( 13). Recent evidence indicates that β-catenin is crucial for cell division during embryonic liver development ( 14, 15) and liver regeneration after partial hepatectomy (PH; ref. 16). Cellular accumulation of β-catenin is primarily involved in the pathogenesis of hepatic tumors, especially HCC ( 13). More recently, β-catenin is also shown to regulate the proliferative response of hepatic oval cells in rodent models ( 17, 18); however, its role in expansion of hepatic progenitor cells, especially during liver carcinogenesis, still remains unclear.

In the current study, we show that Wnt/β-catenin signaling promotes expansion of the hepatic progenitor cell population both when overexpressed in transplanted rat oval cells and when transiently expressed in adult mice. More importantly, we define a subpopulation of less differentiated human HCC cells that are characterized by expression of the hepatic progenitor marker OV6 and are enriched by excessive activation of Wnt/β-catenin signaling, as well as cisplatin treatment. Indeed, elimination of β-catenin virtually abrogated the chemoresistant cell population endowed with progenitor-like features in HCC cells. Targeting Wnt/β-catenin signaling in cancer stem/progenitor-like cells may overcome this chemoresistance and provide a therapeutic model for malignant hepatic tumors.

Materials and Methods

Animals and regeneration protocol. Male Fischer 344 rats (160–180 g) and male C57BL/6 mice (6 wk of age) were purchased from Shanghai Experimental Center (Chinese Science Academy) and maintained in the barrier facility under pathogen-free conditions. In rats, oval cells were induced by the method of the 2-acetylaminofluorene (AAF)/PH model ( 18). To induce oval cell proliferation in C57BL/6 mice, regular standard chow mixed with 0.1% 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC; Sigma) was used. All procedures regarding animals were conducted according to Guide for the Care and Use of Laboratory Animals prepared by National Academy of Sciences and published by NIH (publication 86-23 revised 1985).

Collection of human tissue specimen. Liver specimens were obtained from patients with hepatocellular carcinoma after hepatectomy or liver transplantation in Eastern Hepatobiliary Surgery Hospital. All human sample collection procedures were approved by China Ethical Review Committee.

Cell culture. The diploid WB-F344 rat liver epithelial cell line and the HCC cell lines (Huh7, HepG2, PLC/PRF/5, Hep3B, and SMMC 7721) were obtained from Shanghai Cell Bank and Department of Pathology of Second Military Medical University (Shanghai). Cells were routinely cultured in DMEM (Invitrogen) supplemented with 10% fetal bovine serum (FBS; Life Technologies) within a humidified incubator containing 5% CO₂ at 37°C and passed every 2 to 3 d to maintain logarithmic growth.

WB-F344 cell transplantation. Recipient F344 rats were given 15 mg/kg AAF (Sigma) by garbage for 4 d. On the fifth day, 70% PH was performed on the rats, and 6 × 10⁵ SP-DiI–labeled WB cells [infected with Ad-LacZ or Ad–β-catenin (S37Y), respectively] were transplanted into these AAF/PH rats via the portal vein. Six days after transplantation, transplanted rats were sacrificed and all livers were examined to analyze the transplanted cells.

Plasmid injection. The recombinant β-catenin (S37Y) expression plasmid [pCMV–β-catenin (S37Y)] and GFP expression plasmid (pCMV-GFP) were cloned, as described previously ( 19). Plasmid DNA was given to mice by a hydrodynamics-based gene transfer technique via rapid injection of a large volume of DNA solution through the tail vein, as described by Liu and colleagues ( 20).

Magnetic sorting and culture of OV6⁺ HCC cells. Cells were labeled with primary OV6 antibody (mouse IgG1, R&D Systems), subsequently magnetically labeled with rat anti-mouse IgG1 microbeads and separated on MACS LS column (Miltenyi Biotec). All the procedures were carried out according to manufacturer's instructions. The purity of sorted cells was evaluated by flow cytometry.

Flow cytometry. The flow cytometry was carried out with a Becton Dickinson FACS-Vantage Flow Cytometer, using OV6 primary antibody (R&D Systems), CD133 antibody (Miltenyi Biotec), and FITC-conjugated or PE-conjugated secondary antibody (Sigma).

In vivo tumorigenicity experiments. Various numbers of magnetic sorting OV6⁺ or OV6⁻ cells were resuspended in 100 μL PBS and injected s.c. into 6-wk-old nonobese diabetic (NOD)/severe-combined immunodeficient (SCID) female mice (Chinese Science Academy). Mice were killed between 3 and 6 mo postinjection, at which time tumors were harvested for further examination.

Colony formation assay and drug treatments. Cells were plated in DMEM/10% FBS on 96-well plates precoated with 100 μL Matrigel (Becton Dickinson). Cisplatin was added at the beginning of the culture period for 14 days. Control cells were treated with PBS. The number of the colony formation was assessed by counting under microscope. Representative views were photographed.

Statistics. Values presented are expressed as mean ± SD. After acquiring all data for histologic variables and in vitro assays, Student's t test was applied to determine statistical significance. A value of P < 0.05 was considered significant. Data analysis was performed by Microsoft Excel software (Microsoft).

Results

Activated β-catenin signaling in hepatic regeneration from rat oval cells. To examine the role of β-catenin signaling in oval cell–mediated liver regeneration, an AAF/PH model was established to activate the oval cell compartment in rat liver. The identity of the oval cells was determined both by morphologic criteria and by the expression of OV6, a known oval cell marker ( Fig. 1A ). Western blot and quantitative PCR analyses revealed that the expression of β-catenin and its putative target genes c-Myc, cyclin D1, and Axin2 ( Fig. 1B and Supplementary Fig. S1A) were markedly increased and remained elevated up to days 5 to 7 post-PH, thus paralleling the induction of oval cells.

In addition, a strong higher expression of β-catenin and one of its target gene, cyclin D1, were observed in nonparenchymal cells isolated from the livers at day 5 post-PH than that in parenchymal cells (Supplementary Fig. S1B and C). As it has been shown that oval cells are present in the nonparenchymal cell fraction, but not in the parenchymal cell faction, our data indicated that β-catenin signaling may be active in the oval cell compartment. Supportingly, immunohistochemical staining revealed membranous distribution of β-catenin in hepatocytes and biliary cells but marked cytoplasmic/nuclear expression of β-catenin in small cells around periportal regions (Supplementary Fig. S1D). By immunofluorescent double staining, we confirmed in the AAF/PH model that cells with cytoplasmic/nuclear staining of β-catenin were indeed oval cells, as shown by its colocalization with OV6 ( Fig. 1C). These data suggest that the activation of Wnt/β-catenin signaling occurs during rat oval cell induction, and these cells were the sites with higher level of β-catenin activity.

Propagation effect of β-catenin in rat oval cells. We then studied the effects of activated β-catenin signaling on the growth characteristics of oval cells. The diploid WB-F344 cells, a rat liver oval cell line, were adenovirally transduced with a constitutively active β-catenin (S37Y), which significantly increased pGL-OT reporter activity in WB-F344 cells (Supplementary Fig. S2A). Growth curves showed a significant difference between WB-F344 cells expressing mutant β-catenin and control vector (Supplementary Fig. S2B), suggesting that the active β-catenin signaling favors the accumulation of WB-F344 cells in culture.

To determine the contribution of Wnt/β-catenin pathway to oval cell–mediated liver repopulation in vivo, WB-F344 cells were infected with Ad-LacZ or Ad–mutant β-catenin and then transplanted into the liver of AAF/PH rats via portal vein. At 6 days after cell transplantation, X-gal staining revealed numerous clusters of engrafted WB-F344 cells throughout the control recipient livers (data not shown). To distinguish between donor and recipient cells, transplanted WB-F344 cells were labeled with a fluorescent vital dye SP-DiI. The location of the SP-DiI–labeled WB-F344 cells did not differ between the two groups, with all being scattered in the liver lobule. However, there were much larger and more fluorescent clusters in liver transplanted with cells expressing mutant β-catenin than that for LacZ control ( Fig. 1D). These data suggest that the activation of Wnt/β-catenin signaling in oval cells facilitates liver repopulation when hepatocyte replication is impaired.

Role of β-catenin signaling in mouse oval cells. We next asked whether the effect of Wnt/β-catenin pathway activation observed in rat oval cells is conserved in mouse oval cells. A mouse liver injury model was established, in which oval cell proliferation was induced chemically by DDC feeding ( Fig. 2A ). As observed in the rat model, cells positive for A6 antibody (an oval cell marker for mouse) were also found to have a stronger cytoplasmic/nuclear accumulation of β-catenin ( Fig. 2C). Furthermore, a chromatin immunoprecipitation (ChIP) assay revealed that β-catenin was indeed recruited more to the promoter regions of c-myc gene during oval cell induction than in normal control liver ( Fig. 2B), showing higher β-catenin signaling activity in oval cell–mediated mouse liver reproduction.

Then, we sought to determine whether activating Wnt/β-catenin signaling would lead to increased oval cell expansion in mice liver. The active form of β-catenin (S37Y) was transiently overexpressed in DDC-treated mouse liver using the systemic administration of naked plasmid DNA, a hydrodynamic-based transfection method with high levels of foreign gene expression in mouse liver. Control mice received an equivalent number of plasmid DNA containing GFP cDNA, and transduction was confirmed by GFP expression in the liver (data not shown). As shown in Fig. 2D, expression of mutant β-catenin resulted in significantly more A6-positive oval cells 7 days after plasmids administration in DDC-fed mice. In contrast, TUNEL staining of the liver sections showed a similar number of apoptotic cells in two groups of mice, indicating no significant apoptosis related to β-catenin overexpression (data not shown). Together, these data suggest that stabilization of β-catenin in mouse liver progenitors increases oval cell proliferation, leading to expansion of A6-positive oval cells.

Expression of β-catenin in OV6⁺ progenitor cells in human liver diseases. Based on these data in animal models, we asked whether β-catenin was activated in human liver diseases involving progenitor cell proliferation, such as liver cirrhosis and HCCs. In the nondiseased human liver, we found a few OV6⁺ cells residing in the bile ductules and the periportal parenchyma ( Fig. 3A ). Interestingly, the oval-like cells were markedly increased in hepatocarcinogenic conditions, such as liver cirrhosis ( Fig. 3B). In accordance with the above observation, immunostaining of serial sections showed that these cells also expressed high levels of β-catenin, and the cytoplasmic or nuclear staining was consistently detected ( Fig. 3B). Significantly, a subset of intermediate hepatocyte-like cells found in HCCs was also positive for OV6 and nuclear β-catenin ( Fig. 3C). In addition, these cells displayed increased proliferation index (Ki67 staining) compared with adjacent OV6⁻ HCC cells ( Fig. 3D). Together, these data implied the existence of a distinct cancer cell subpopulation in HCC with active Wnt/β-catenin signaling.

Identification and characterization of OV6⁺ subpopulation in cultured HCC cell lines. To test this hypothesis, we investigated whether human HCC cell lines also contain OV6⁺ subset recapitulating the features of primary cells detected in HCC specimen. As expected, flow cytometry indeed showed presence of a rare OV6⁺ population ranged from 0.2% to 3% in various HCC cell lines, including Huh7, PLC, SMMC7721, Hep3B, and HepG2 ( Fig. 4A ). Recently, CD133, which is expressed by normal and malignant stem cells of the neural, hematopoietic, epithelial, and endothelial lineages, was also used to identify cancer stem cells population in HCC ( 21). Double staining of OV6 and CD133 antigen showed that, in Huh7 cells which expressed CD133 in nearly three-fourths of the cells, the percentage of CD133⁺ cells was 95.5% in OV6⁺ subpopulation versus 76.6% in total cells. Furthermore, in SMMC7721 cells which expressed CD133 only in 0.12% of the cells, nearly all CD133⁺ cells were detected in the OV6⁺ subpopulation. Therefore, CD133⁺ population was significantly enriched in cells positive for OV6 ( Fig. 4B), indicating that OV6⁺ cells may represent a potential stem/progenitor-like cell population. When injected into the NOD/SCID mice, OV6⁺ SMMC7721 cells could initiate significantly larger tumors (n = 4), whereas OV6⁻ cells showed small or no tumors ( Fig. 4C). Furthermore, as few as 5 × 10³ OV6⁺ SMMC7721 cells were sufficient for consistent tumor development in NOD-SCID mice, whereas at least 50 to 100 times as many OV6⁻ cells were necessary to consistently generate a similar tumor in the same model (data not shown). These data indicate that OV6⁺ cells possess a greater tumorigenic ability in vivo.

We then examined the expressions of hepatic progenitor cell markers and transcription factors involved in hepatic embryogenesis in OV6⁺ and OV6⁻ cells by quantitative reverse transcription–PCR (RT-PCR). As shown in Fig. 4D, progenitor cell markers, such as ABCG2, EpCAM, c-kit, AFP, and ALB, as well as transcription factors in the earlier phase of hepatic development, including GATA6, CAAT/enhancer binding protein α (C/EBPα) and C/EBPβ, were more intensely expressed in OV6⁺ cells compared with OV6⁻ cells. These findings indicate that OV6⁺ cells may therefore represent a less differentiated subpopulation.

Then we asked whether the OV6⁺ cells have some intrinsic properties of stem cells, such as preferential expression of “stemness” genes, including Notch-1, Bmi1, Nanog, and Oct-4. These genes are known important for stem cell self-renewal, proliferative capacity or fate determination. Quantitative RT-PCR analysis showed that OV6⁺ Huh7 and SMMC7721 cells indeed had increased mRNA levels of these genes compared with levels in OV6⁻ cells ( Fig. 4D). Furthermore, concordantly, a luciferase reporter assay of β-catenin–mediated transcriptional activation revealed an increase in β-catenin activity in OV6⁺ cell population ( Fig. 5A ), suggesting a role for β-catenin signaling in this progenitor-like subpopulation of HCC cells.

Role of β-catenin signaling in self-renewal of OV6⁺ HCC cells. As the active β-catenin signaling was detected in OV6⁺ HCC cells, we then evaluated the potential involvement of Wnt/β-catenin pathway in the self-renewal of OV6⁺ cell population. The canonical Wnt/β-catenin signaling can be activated by direct, intracellular inhibition of GSK-3 function using specific inhibitors. The indirubins BIO, a particularly selective and potent inhibitor of GSK-3, was used to activate the Wnt/β-catenin pathway in cultured HCC cell lines. We first assessed whether Huh7 and SMMC7721 cells were capable of transducing Wnt signaling when treated with BIO. As shown in Supplementary Fig. S3, cells treated with BIO showed a substantial, dose-dependent increase in pGL-OT reporter activity, whereas its kinase-inactive analogue, MeBIO-treated cells, did not show any change in activity, indicating efficient activation of the canonical Wnt/β-catenin pathway by BIO.

To uncover a potential role of β-catenin signaling in regulating self-renewal in OV6⁺ cells, we thereafter examined the change in percentage of OV6⁺ subpopulation in the cultivated SMMC7721 and Huh7 cells after BIO treatment. Flow cytometry revealed that BIO treatment of short-term cultures enriched the OV6⁺ subpopulation 2-fold to 11-fold relative to MeBIO-treated group ( Fig. 5B). In addition, BIO treatment of MACS-sorted OV6⁺ cells also led to the preservation of much higher percentage of OV6⁺ cell population than MeBIO treatment (51.7% versus 9.04%; Fig. 5C), whereas lentiviral transduction of OV6⁺ cells with microRNA (miRNA) targeting β-catenin (Supplementary Fig. S4) substantially reduced the subpopulation percentage ( Fig. 5C and Supplementary Fig. S5). These results indicate that the increased OV6⁺ fractions were caused by enrichment of original OV6⁺ subpopulations. To investigate whether OV6⁺ cell population gained higher proliferation potential as a result of β-catenin activation, the same amount of SP-DiI–labeled OV6⁺ and nonlabeled OV6⁻ SMMC7721 cells was mixed and cultured in the presence of BIO or MeBIO control. Two days later, the relative frequencies of SP-DiI⁺ tumor cells of OV6⁺ origin were evidently higher than nonlabeled OV6⁻ tumor cells, suggesting faster growth rate of OV6⁺ subset in a competitive culture condition. Interestingly, BIO treatment not only further increased the frequency of fluorescent tumor cells but also resulted in greater abundance of OV6⁺ cells in SP-DiI⁺ cell mass (5.73% versus 13.64%). These data therefore indicate that activation of the canonical Wnt/β-catenin pathway by BIO can markedly enrich the OV6⁺ subpopulations and influence the self-renewal of this subset.

Resistance of OV6⁺ HCC cells to standard chemotherapy. HCCs are commonly resilient to treatment because the malignant cells survive chemotherapy. It has been speculated that stem or progenitor cells in many tissues are more resistant to conventional cancer therapies. To explore the possible role of OV6⁺ cells in the development of chemoresistance, cultured Huh7 and SMMC7721 cells were treated with cisplatin and the percentage of OV6⁺ progenitor-like cells in the total population after treatment was analyzed by flow cytometry. As shown in Fig. 6A , treatment with cisplatin significantly enriched the OV6⁺ subpopulation in a dose-dependent and time-dependent manner, suggesting that OV6⁺ tumor cells have greater chemoresistance and repopulation potential than OV6⁻ cells in vitro.

To identify further the cell subpopulations that contribute to HCC chemoresistance, we studied the chemosensitivity of OV6⁺ and OV6⁻ tumor cell subpopulations. Because chemotherapy-induced cell killing is generally attributed to mitotic cell death, clonogenic assays to assess the replicative competence of cisplatin-treated subpopulations were performed. Although no significant difference was observed between the two subpopulations in the absence of cisplatin, OV6⁺ cells isolated from Huh7 and SMMC7721 cells were more resistant to cisplatin treatment than were corresponding OV6⁻ cells ( Fig. 6B and data not shown). To determine the role of Wnt/β-catenin signaling in mediating chemoresistance of HCC subpopulation, purified OV6⁺ cells were infected with lentiviral miRNA against β-catenin and allowed to grow in Matrigel for 2 weeks in the presence or absence of cisplatin. As expected, transduction of the cells with interfering miRNA significantly reduced β-catenin expression (Supplementary Fig. S4) and, importantly, diminished chemoresistant colonies ( Fig. 6C). These results indicate that β-catenin signaling was required for protection of OV6⁺ progenitor-like cells from chemotherapeutics-induced cytotoxicity.

Discussion

The canonical Wnt/β-catenin signaling pathway plays a prominent role in modulating stemness, proliferation, and differentiation in several adult stem cell niches, such as the hair follicles in the skin, the mammary gland, the intestinal crypt, and the hematopoietic tissues ( 12). Our data support such a role for Wnt/β-catenin signaling in activation and proliferation of the hepatic progenitor cells in the liver. First, β-catenin is highly expressed and mainly localized to the proliferating oval cells in both rat and mouse models, showing extensive cytoplasmic and nucleic staining. Second, endogenous β-catenin and its target genes were colocalized in oval cell fraction and cycled with similar kinetics during oval cell response in rat AAF/PH model. Furthermore, in DDC-fed mice, the ChIP assay showed an enhanced β-catenin transcriptional activity in oval cell–mediated liver reproduction. Thus, the canonical Wnt/β-catenin pathway was activated during the regenerative response. Third, constitutive activation of Wnt/β-catenin signaling stimulated oval cell proliferation in vitro and caused a substantial clonal expansion of the transplanted rat oval cells in vivo, indicating that the activated Wnt/β-catenin signaling is required for oval cell expansion. Finally, the hyperplastic phenotype was consistently observed in the livers of DDC-fed mice that transiently overexpressed the constitutively active β-catenin.

During the course of the present study, Sylvester and colleagues ( 17) and Monga and colleagues ( 18) independently reported the essential role of Wnt/β-catenin signaling in oval cell response in rodents. Our data provide more functional evidences that confirm the positive regulatory function of Wnt/β-catenin signaling on oval cell activation, but go further in showing that Wnt/β-catenin signaling also plays an essential role in chronic human liver diseases in which progenitor cell activation is featured. Human liver progenitor cells have mainly been studied in regeneration after severe hepatocellular necrosis ( 9). Recently, these cells have also been identified in human liver diseases associated with a high incidence of HCC or cholangiocarcinoma (such as chronic viral and alcoholic hepatitis, sclerosing cholangitis; refs. 22, 23). This indicates that human liver progenitor cells represent a potential target cell population for hepatocarcinogens. However, there is hitherto no direct evidence in support of the hypothesis that human HCCs can originate from liver progenitor cells. In this study, we observed the presence of OV6⁺ progenitor-like and intermediate hepatocyte-like cells in half of the studied small cell dysplastic foci and HCC ( Fig. 3), which strongly suggests that these lesions arise from liver progenitor cells that proliferate and differentiate toward the hepatocytic lineage. Most strikingly, we have identified a subpopulation of highly tumorigenic cells with stem/progenitor cell properties from human HCC cells by using an oval cell marker OV6. These OV6⁺ tumor cells have features typical of stem/progenitor cells with highly expressed stem/progenitor cell markers ( 4, 24). They are capable of self-renewal as shown by clonogenic expansion in vitro and tumor formation in vivo and are able to differentiate into OV6⁻ tumor cells ( Fig. 5C). They also express certain markers of both hepatocytic and biliary lineages, as well as transcription factors involved in hepatic embryogenesis. Several very recent studies in HCCs have used a putative stem cell marker of CD133 for the identification of cancer stem-like cells, which are capable of producing outgrowths of liver tumor after transplantation into immunodeficient mice ( 21). By flow cytometry analysis, we found that the percentage of CD133⁺ tumor cells varied from 0.1% to 75% in different HCC cell lines, whereas the proportion for OV6 is relatively constant (ranging from 0.2% to 3%). Interestingly, CD133⁺ cells are enriched in cells positive for OV6. This further indicates that OV6⁺ cells may represent a potential stem/progenitor-like cell population in HCC that has a higher tumor-initiating capability. Our findings thus support the hypothesis that a considerable proportion of human HCCs arise from hepatic progenitor cells. As to the origin of these hepatic progenitor cells, numerous studies conducted on animal models and humans have suggested that bone marrow cells may give rise to hepatic oval cells in certain circumstances ( 25). Wang and colleagues also revealed that epithelial cancers can originate from marrow-derived sources ( 26). So the source of these hepatic progenitor cells remains to be determined.

Aberrant activation of the Wnt/β-catenin pathway has been implicated in multiple hepatic tumors, especially including hepatoblastomas and HCCs ( 13). Nuclear and cytoplasmic localization of β-catenin were reported in 90% to 100% of all hepatoblastomas, which mainly originate from immature liver precursor cells and presents morphologic features recapitulating some of the developmental aspects of the liver ( 13). Consistently, expression of stabilized β-catenin promotes the self-renewal of hepatic stem/progenitor cells and leads to tumorigenesis in the liver ( 13, 27), as observed in central nervous system stem cells, keratinocyte stem cells, mammary stem cells, and chronic myelogenous leukemia (CML) granulocyte-macrophage progenitors ( 12). In HCC, however, heterogeneous intracellular distributions of β-catenin are observed within primary tumors and their metastases ( 13). Together with its colocalization with the OV6, the higher prevalence of β-catenin activation either in scattered cluster or in diffusely sporadic tumor cells strongly implicates deregulated Wnt/β-catenin signaling in malignant transformation of hepatic progenitor cells. This hypothesis was virtually validated by our observation that BIO treatment led to the enrichment of OV6⁺ cells, confirming that the activation of Wnt/β-catenin signaling confers a growth advantage or self-renewal capacity in a small HCC subpopulation endowed with progenitor-like features. In addition to Wnt/β-catenin pathway, several other signal pathways that regulate normal stem-cell self-renewal can also cause neoplastic proliferation when dysregulated by mutations. For example, the sonic hedgehog, Notch, and Bmi1 pathways ( 28) have all been shown to promote the self-renewal of somatic stem cells, as well as neoplastic proliferation in the same tissues when dysregulated. Recent studies have shown that expression of the Notch and Hedgehog family genes, as well as Bmi1 gene, are frequently increased in hepatoblastomas and HCCs ( 29). Interestingly, these stem cell markers were also highly expressed in OV6⁺ cells compared with OV6⁻ cells. In addition, the proliferation and self-renewal of human hepatic stem/progenitor cells was promoted by Bmi1 ( 27). Thus, it will be important to determine whether these pathways are also required and interact to regulate self-renewal of the progenitor-like subpopulation in HCC.

It is becoming increasingly evident that cancer stem/progenitor cells provides a compelling potential reason for why cancer chemotherapy may induce remission, yet rarely cures. Indeed, in cases in which bulk disease is eradicated and chemotherapy is given, only to be followed by a relapse, a plausible explanation is that the therapies may affect descendent cells that are irrelevant for the persistence and propagation of the disease, leaving the rarer, but more potent cancer stem/progenitor cell unperturbed ( 30). Understanding the molecular mechanisms of resistance in cancer stem/progenitor cells will be crucial for the development of new therapeutic strategies to cope with this challenge. Wnt/β-catenin pathway has been shown to mediate radiation resistance of mouse mammary progenitor cells and chemotherapy resistance of CML progenitors ( 31, 32). In the present study, we found substantial resistance of OV6⁺ cell population to standard chemotherapy. More importantly, inactivation of Wnt/β-catenin pathway can diminish the chemoresistance of OV6⁺ progenitor-like cells. Thus, the specific molecular mechanism for the resistance of the HCC progenitor-like cells to chemotherapy can be linked to the function of the Wnt/β-catenin signaling pathway. The Wnt/β-catenin signaling pathway may be an attractive target for directed anticancer stem/progenitor cell therapeutics. Most importantly, the demonstrated possibility of expanding HCC progenitor-like cells in vitro has considerable therapeutic implications for the evaluation of drug efficacy. This ability to maintain and even expand tumorigenic HCC cancer cells in vitro should be exploited for further preclinical studies. Therapies targeted to the Wnt/β-catenin signaling pathway in preclinical and clinical development may provide a specific method to disrupt this resistance mechanism to improve overall tumor control with chemotherapy.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.