Hepatoblastoma: A Need for Cell Lines and Tissue Banks to Develop Targeted Drug Therapies

Limited research exists regarding the most aggressive forms of hepatoblastoma. Cell lines of the rare subtypes of hepatoblastoma with poor prognosis are not only difficult to attain but also challenging to characterize histologically. A community-driven approach to educating parents and families, regarding the need for donated tissue, is necessary for scientists to have access to resources for murine models and drug discovery. Herein, we describe the currently available resources, existing gaps in research, and the path to move forward for uniform cure of hepatoblastoma.

Limited research exists regarding the most aggressive forms of hepatoblastoma. Cell lines of the rare subtypes of hepatoblastoma with poor prognosis are not only difficult to attain but also challenging to characterize histologically. A community-driven approach to educating parents and families, regarding the need for donated tissue, is necessary for scientists to have access to resources for murine models and drug discovery. Herein, we describe the currently available resources, existing gaps in research, and the path to move forward for uniform cure of hepatoblastoma.
Keywords: hepatoblastoma, roadmap, xenograft models, cell lines, tissue procurement iNtrODUctiON Hepatoblastoma is the most common primary liver tumor diagnosed in childhood (1), with approximately 100 cases in the U.S. annually (2). Despite a high cure rate for those children whose tumor is resectable, there remains a group of children for whom a cure is out of reach.
The disease predominantly occurs in young children, from birth to 5 years of age (1). The histological subtypes of hepatoblastoma are fetal, embryonal, mixed epithelial-mesenchymal, and small cell undifferentiated (3). However, there is currently a lack of understanding regarding the origins and pathophysiology of these different subtypes of hepatoblastoma.
Clinically, the empirically driven advancements in postoperative chemotherapy and surgery, including the multidisciplinary approach set forth through the Pretreatment Extent of Disease guidelines (1), has improved outcomes for hepatoblastoma. These guidelines rely on a standardized staging system using imaging for detecting amount of tumor involvement (4). Despite these clinical advancements, the more aggressive forms of hepatoblastoma remain difficult to treat. Current treatments for aggressive forms of hepatoblastoma include doxorubicin, irinotecan (clinical trials), hepatic artery chemoembolization in addition to chemotherapy agents, as well as liver transplantation or partial resection with neoadjuvant chemotherapy (5).
Scientists and clinicians are now seeking non-chemotherapeutic treatments for patients with unresectable or metastatic tumor -treatments that directly target the molecular underpinnings of hepatoblastoma progression. For example, clinical trials regarding cixutumumab and pazopanib, monoclonal antibodies, and alisertib, a kinase inhibitor, have all been completed or are actively being investigated in phase 2 clinical trails for the treatment of refractory hepatoblastoma (6).
In order to find other targeted therapies, researchers need hepatoblastoma tissues and cell lines. There is an absence of diversity in hepatoblastoma cell lines for scientists and clinicians to use to better understand the disease. In this paper, we describe the need for more cell lines and murine models to advance the discovery of therapeutic targets for the more aggressive subtypes of hepatoblastoma.

MetHODs
An extensive literature review of hepatoblastoma via PubMed was conducted to obtain information on research with unique hepatoblastoma cell lines and murine models. First, the authors found the number of distinct hepatoblastoma cell lines published in the literature. The search terms used were "'hepatoblastoma' [All Fields] and 'cell line' [MeSH Terms]." The search provided approximately 450 publications, all reviewed by the authors of this paper. Any publication that had a focus on hepatoblastoma was mentioned in Table 1.
For each unique cell line, the authors attempted to find the primary article characterizing the cell line. Secondary publications using hepatoblastoma cell lines were also included in Table 1.
Manuscripts that had differing characterizations of certain cell lines were included in the "Unsure Hepatoblastoma" portion of Table 1.
Next, the authors reviewed literature via PubMed to find murine models of hepatoblastoma. The search terms used were "hepatoblastoma murine models. " This provided approximately 50 publications that the authors reviewed. All publications that studied murine models of hepatoblastoma were included in Table 1. The models were classified as chemically induced, transgenic, and cell-derived xenografts. Furthermore, the xenograft studies were sorted based on subcutaneous or orthotopic models.
Finally, the authors searched for patient-derived xenografts via PubMed and found no manuscript publications. The authors then searched the European Journal of Cancer using the term "hepatoblastoma xenograft" and found a published abstract using patient-derived xenografts, which is included in Table 1. resULts Data regarding histological characterization and experimental murine models from only a few hepatoblastoma cell lines exist. These cell lines tend to have a favorable histology, leading to an underrepresentation of the high-risk subtypes (41).

cell Lines
Fifteen hepatoblastoma cell lines are described in current literature (Table 1). Additionally, there are four cell lines that are potentially hepatoblastoma, but significant inconsistencies in the literature render the data obtained from these lines unreliable. Even among the confirmed hepatoblastoma cell lines, however, there are many documented instances in which cell lines were mistaken for hepatocellular carcinoma (7). A fibroblast cell line harvested from the liver of a Beckwith-Wiedemann syndrome patient with hepatoblastoma is described ( Table 1).
Most confirmed cell lines are of the mixed histology subtype. However, there are no cell lines of the small cell undifferentiated subtype, which carries the worst prognosis (7).

Murine Models
Only one chemically induced murine model of hepatoblastoma has been reported ( Table 1). Although four different transgenic murine models are described in the literature, these murine models were not specifically developed for the purpose of modeling hepatoblastoma. The transgenic murine models phenotypically express both hepatocellular carcinoma and hepatoblastoma ( Table 1).
Ten unique cell line-derived subcutaneous xenografts and three cell line-derived orthotopic murine models of human hepatoblastoma exist ( Table 1). These models primarily utilize the Hep G2 and HuH6 cells lines. Twelve unique patient-derived xenografts exist ( Table 1).

Potential Genetic targets for Aggressive Hepatoblastoma
Many studies have noted genetic mutations specific to histological subtypes of hepatoblastoma (42). Hepatoblastoma cells have shown gain of 2q, 1q, Xp, and Xq; loss of 4q, 2q, and 1q; and loss of heterozygosity of insulin growth factor 2 (5). Subtypes with increased Notch expression are of the fetal subtype and tend to have a better prognosis. Those with overexpression of the Wingless-type MMTV Integration Site Family pathway are of the small cell undifferentiated subtype and carry a less favorable prognosis (3). Additionally, the more aggressive forms of hepatoblastoma have telomerase reverse transcriptase promoter mutations (43). Blocking the Wingless-type MMTV Integration Site Family pathway using NK1R antagonists has been shown to slow the progression of hepatoblastoma cell growth in vitro (44). Hepatoblastoma cells show an increase in activity of the hedgehog pathway, and abnormal signaling has been linked to more malignant potential (45). Forkhead Box G1 is overexpressed in hepatoblastoma, specifically the more aggressive subtypes, when compared to the fetal subtype (46).

DiscUssiON
In order to find targeted therapeutic options for hepatoblastoma, basic science studies need to be conducted. The few cell lines characterized and the inconsistencies in the literature on certain cell lines provide a major hurdle toward this goal. In addition, the availably of the cell lines is limited, which explains the narrow spectrum of cell lines used to derive xenografts from the already few hepatoblastoma cell lines. Additionally, diversity of histological subtypes is needed in order to find better treatment modalities for the more aggressive forms of hepatoblastoma. Interestingly, expression of fibroblasts enhances the growth of hepatoblastoma (47), which is why the hepatoblastoma-derived fibroblast cell line, GM08206 (Table 1), carries potential for more advanced studies. It is of important note that the majority of liver cells are aneuploid, which has been thought to protect the liver from chronic injury (48). Culturing surrounding normal liver tissue in addition to the tumor would provide insight into premalignant tissue field effect at the site of the tumor (46).
Certain repositories for hepatoblastoma are in the early stages of developing around the world, providing optimism for  In addition to the Children's Oncology Group, the Japanese Collection of Research Biosources hosts a cell bank that provided the cell lines for the majority of hepatoblastoma manuscripts in the literature review (9). However, currently only HUH6 is available for public distribution. Many published hepatoblastoma cell lines found in the literature review were not within the last decade, which could explain the difficulty in obtaining certain cell lines today.
Most importantly, many international groups, such as Childhood Liver Tumors Strategy Group and the Society for Pediatric Oncology and Hematology, have collaborated with Children's Oncology Group and the Japanese Collection of Research Biosources, which initially led to the Pretreatment Extent of Disease guidelines (50). It is this type of collaboration that can result in an increase in cell lines and tissue-banking repositories. One example is the Childhood Liver Tumors Strategy Group, which runs a tissue bank for childhood liver tumors (51).
Recently, further collaboration has allowed for the Children's Hepatic tumors International Collaboration, to obtain data on 1,605 hepatoblastoma patients, aimed at creating a database to identify prognostic factors for this rare pediatric cancer (52). One limitation to the database, mentioned by the authors, was the exclusion of histology due to the lack of international consensus in characterizing subtypes (52).
As more interaction among family members is made, newer registries, in addition to those previously mentioned, will continue to grow. The Macy Easom Foundation has committed to funding development of the Hepatoblastoma Registry, as well as the expense of administration, data compilation, and analysis (18).
Despite the current development of repositories, an increase in cell lines and murine models available for research purposes cannot progress unless methods are in place to increase awareness for tissue donation in hepatoblastoma. Both parents and treating physicians must be made aware of the need for hepatoblastoma tissue and the opportunity to support research via autopsy tissue donations. The decision, whether to make an autopsy tissue donation, is difficult, intensely personal, and unique for each family. The authors recognize the delicate balance between making parents aware of the need and opportunity while taking care to respect every family's response and perspective.
A parent who wishes to arrange for an autopsy donation should not be burdened with making the arrangements. With parents' consent, volunteers and professionals must be in place to make the necessary contacts and establish logistics of the donation. These arrangements may include contacting the treating physician, speaking with the local pathologist who will perform the autopsy, connecting the researcher who will receive the donated tissue with the pathologist, and arranging for transport of the body from the child's home to the hospital (and return to the funeral home).
Grassroots communication and interaction among family members, caregivers, and others affected by a particular diagnosis has significantly influenced progress in some areas of pediatric cancer research. As an example, interaction among families affected by diffuse intrinsic pontine glioma (DIPG) in an online discussion group is considered by some to be the first step in raising a tide turning awareness in that community. The result was a promising therapeutic drug, panobinostat, for treatment (53).
In summary, the greatest potential for the development of targeted therapy for aggressive forms of hepatoblastoma will come when scientists have access to hepatoblastoma cells lines and tissues with histological subtype diversity (Figure 1).

AUtHOr cONtriBUtiONs
RR, KS, RH, MSB, MB, and LH: contributed significantly to the acquisition, intellectual content, and final approval and are in agreement with all aspects of the work.