Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Dec 1.
Published in final edited form as: Toxicol Pathol. 2015 Aug 18;43(8):1114–1126. doi: 10.1177/0192623315599853

Genomic Profiling Reveals Unique Molecular Alterations in Hepatoblastomas and Adjacent Hepatocellular Carcinomas in B6C3F1 Mice

Sachin Bhusari 1, Arun R Pandiri 1,6, Hiroaki Nagai 1, Yu Wang 1, Julie Foley 1, Hue-Hua L Hong 1, Thai-Vu Ton 1, Michael DeVito 2, Keith R Shockley 3, Shyamal D Peddada 3, Kevin E Gerrish 4, David E Malarkey 1, Michelle J Hooth 5, Robert C Sills 1, Mark J Hoenerhoff 1,*
PMCID: PMC4670276  NIHMSID: NIHMS709786  PMID: 26289556

Abstract

The cell of origin of hepatoblastoma in humans and mice is unknown; it is hypothesized to be a transformed hepatocyte, oval cell, or hepatic progenitor cell. In mice, current dogma is that hepatoblastomas arise from pre-existing hepatocellular neoplasms as a result of further neoplastic transformation. However, there is little evidence supporting this direct relationship. To better understand the relationship between hepatocellular carcinoma (HCC) and hepatoblastoma, and determine molecular similarities between mouse and human hepatoblastoma, global gene expression analysis and targeted mutation analysis were performed using hepatoblastoma, HCC, and adjacent liver from the same animals in a recent National Toxicology Program bioassay. There were significant differences in Hras and Ctnnb1 mutation spectra, and by microarray, hepatoblastomas showed dysregulation of embryonic development, stem cell pluripotency, and genomic imprinting compared to HCC. Meta-analysis showed similarities between hepatoblastoma, early mouse embryonic liver, and hepatocyte derived stem/progenitor cells compared to HCC. Our data shows that there are striking differences between hepatoblastoma and HCC, and suggests that hepatoblastoma is a significantly different entity that may arise from a hepatic precursor cell. Furthermore, mouse hepatoblastoma is similar to the human disease at the pathway level, and therefore is likely a relevant model for evaluating human cancer hazard.

Keywords: Carcinogenesis, Liver, Genomics, Microarray, Molecular pathology, Rodent pathology, Toxicologic pathology

Introduction

In mice, hepatoblastoma (HB) is a poorly differentiated, primitive, embryonal hepatocellular neoplasm that is often associated with preexisting hepatocellular adenomas (HCA) or carcinomas (HCC). They have a relatively late onset, occurring mainly in aged mice, and an increased incidence of HB has been associated with exposure to a number of chemicals in the National Toxicology Program (NTP) carcinogenesis bioassay, including compounds such as oxazepam, diethanolamine, methyleugenol, and Gingko biloba extract, among others (National Toxicology Program 1998, 1999, 2000, 2013). In B6C3F1 mice, the background rate of spontaneous HB in vehicle controls is low, with an average incidence of 4.86% (34/700) in males and 0.57% (4/698) in females (NTP historical controls, all routes and vehicles (National Toxicology Program 2014)). The cell of origin of this tumor remains unknown, but since a large number of HB appear to arise in association with preexisting hepatocellular adenoma (HCA) or carcinoma (HCC), it has historically been thought that mouse HBs are a more malignant variant of a hepatocellular tumor, and represent a later stage in the progression of hepatocellular malignancy (Diwan et al. 1989; Turusov et al. 2002). However, there is little scientific evidence to support a direct relationship between these two very different tumor types. While HBs do arise later in the course of disease, significant differences in biologic behavior to support HB as a more malignant form of hepatocellular tumors have not been shown; for example, the rate of lung metastases is similar between HB and HCC (23.6% and 23%, HB and HCC respectively) (Turusov et al. 2002). In addition, HBs are morphologically markedly different from their hepatocellular counterparts (Turusov et al. 2002), which does not support a progression of malignancy but rather suggests that these tumors are a distinct and separate entity. HBs have a more primitive phenotype with morphologic features of palisading of poorly differentiated cells with elongated deeply basophilic, hyperchromatin nuclei, and scant eosinophilic cytoplasm; many of which may be considered reminiscent of a progenitor cell phenotype, as opposed to an a distinct hepatocellular phenotype in HCC. Therefore, it is possible that HBs represent a distinct tumor type, which arises late in the course of the disease as a result of additional molecular events within the hepatic milieu that may result in transformation of a stem or multipotent progenitor cell in the liver.

The relevance of mouse HB in human hazard identification has also been a topic of great debate. Due to differences in biology between mice and humans, the use of mouse HB and other liver tumors to inform on human cancer hazard and risk has been controversial (Grisham 1997; Maronpot et al. 2004). In humans, HBs are predominantly observed in young children below age of 5 years, and are exceptionally rare in adults (Cruz Jr et al. 2013; Spector and Birch 2012). In contrast, HB in mice occurs in aged animals, most often late in the course of hepatic neoplasia (Chhabra et al. 1992). Hepatoblastomas in the mouse most often arise in association with other hepatocellular tumors (Turusov et al. 2002); in contrast, human HBs are always solitary tumors unassociated with pre-existing hepatocellular neoplasia, and are thought to arise primarily de novo from hepatoblast remnants of the embryonic liver (Lack et al. 1982; Turusov et al. 2002; von Schweinitz 2012). These tumors are commonly associated with chemical exposure in mice, and have been reported in children from mothers exposed to certain chemicals (Buckley et al. 1989). Hepatoblastoma in mice and humans share primitive, poorly differentiated morphologic features, they are predominant in males of both species, and have a similar metastatic rate (Turusov et al. 2002); however, while tumors in both species most commonly metastasize to the lungs (Herzog et al. 2000; Turusov et al. 2002), other less common metastatic sites include the CNS and bone in humans (Herzog et al. 2000) and lymph nodes in the mouse (Turusov et al. 2002). In humans, HBs are thought to develop from the transformation of hepatic stem cells (hepatoblasts) through impairment of the normal hepatic differentiation program, including alterations in oncogenic Wnt/β-catenin activation and genetic alterations such as 2p and 8q chromosomal gains (Armengol et al. 2011). Molecularly, human and mouse HBs both harbor mutations in Ctnnb1 (Anna et al. 2000; Cairo et al. 2008), suggesting that underlying molecular changes are at least in some ways similar.

There are a number of critical questions that need to be answered regarding HBs in mice and their relationship to HCC. First, as a common model for exposure-related carcinogenicity, it is important to understand the relevance of HB in the B6C3F1 mouse in the context of HCC resulting from carcinogen exposure. If these tumors are genetically related, then an increased incidence of HB associated with HCC would suggest a progression of malignancy. If unrelated genetically, then this would suggest that HB is an independent neoplastic population resulting from ongoing genetic selection within the hepatic microenvironment, potentially targeting a multipotent hepatic stem cell. Secondly, it is critical to further understand the relevance of HB in the B6C3F1 mouse to humans in hazard identification. If these tumors are similar in humans and mice in terms of their global gene expression, then despite their differences in clinical course, an increased incidence of HB as a result of chemical exposure would provide stronger evidence that such an exposure may pose a cancer risk in humans. Given these unanswered questions, the aim of this study was to compare concurrent HB and HCC to identify relationships in global gene expression and mutation spectra of common hepatic cancer genes (Ctnnb1, Hras) that would 1) provide an insight on the origin of HB tumors, and 2) examine the relevance of B6C3F1 HB to human HB.

Materials and Methods

Sample selection, laser capture microdissection, RNA isolation and amplification

All tissues used in this study were from animals in a single NTP chronic bioassay (bromodichloroacetic acid (BDCA), TR-583). The NTP conducts its studies in compliance with its laboratory health and safety guidelines and FDA Good Laboratory Practice Regulations and must meet or exceed all applicable federal, state, and local health and safety regulations. Animal care and use in this bioassay were in accordance with the Public Health Service Policy on Humane Care and Use of Animals. Samples containing HB, adjacent HCC, and associated normal (AN) liver in the same section were identified histologically in formalin-fixed, paraffin-embedded (FFPE) sections (Figure 1). Frozen samples collected from the liver of animals that corresponded to these FFPE tissues were used for RNA isolation. These frozen tissues were collected from the NTP Frozen Tissue Repository, embedded in OCT (optimum cutting temperature) freezing media, and cryosectioned for laser capture microdissection (LCM) in order to obtain each cell population (HB, HCC, AN) for analysis. HB (n=6), adjacent HCC (n=6), and associated AN liver (n=6) were laser microdissected from 2–5 serial 8 micron cryosections for microarray analysis (Figure 1) using MMI CellCut Plus® (Molecular Machines and Industries Inc, Zurich, Switzerland). Total RNA from microdissected tissue was extracted with the Arcturus Picopure® RNA isolation kit (Life Technologies, Carlsbad, CA) by using the manufacturer’s protocol. RNA integrity was analyzed with an Agilent 2100 Bioanalyzer RNA Pico Assay (Agilent Technologies, Santa Clara, CA).

Figure 1.

Figure 1

Open in a new tab

Tissue collection protocol for obtaining hepatoblastoma (HB), hepatocellular carcinoma (HCC), and associated normal liver (AN) for each sample, using laser capture microdissection.

Mutation analysis

Regions of the mouse Ctnnb1 (exon 2 and 3) and Hras (codon 61) gene sequences prone to mutation (“hot-spot” regions) that are orthologous in humans were analyzed in all samples. Analysis was performed using cDNA generated from the above LCM samples as well as FFPE HB (n=30) and adjacent HCC (n=30) samples from the same bioassay (BDCA, TR583). The PCR primers and procedures used to amplify the hot spot regions of Ctnnb1 and Hras were described previously (Hoenerhoff et al. 2013). HB and adjacent HCC samples were obtained by laser capture microdissection on deparaffinized, stained, and dehydrated FFPE sections. DNA was extracted from microdissected tissue with Arcturus Picopure® DNA extraction kit(Life Technologies, Carlsbad, CA) and DNA concentration were measured using Qubit® 2.0 Fluorometer and Qubit® dsDNA HS (High Sensitivity) assay (Life Technologies, Carlsbad, CA). Gene amplification reactions were performed by regular PCR; controls lacking template DNA were run with all sets of reactions. PCR amplified products were purified using a QIAquick Gel Extraction Kit (Qiagen, Valencia, CA). The purified PCR products were cycled with Terminal Ready Reaction Mix-Big Dye (Perkin Elmer, Foster City, CA), and the extension products were purified with DyeEx 2.0 Spin Kit (Qiagen). The lyophilized PCR products were sequenced with an automatic sequencer (Perkin-Elmer ABI Model 3100).

Microarray hybridizations

Gene expression analysis was conducted on laser capture microdissected HB (n=6), HCC (n=6), and adjacent non-tumor liver (AN; n=6) samples using Affymetrix Mouse Genome 430 2.0 GeneChip® arrays (Affymetrix, Santa Clara, CA). Total RNA (10ng) was amplified as directed in the WT-Ovation Pico RNA Amplification System protocol, and labeling with biotin following the Encore Biotin Module. Amplified biotin-aRNAs (5μg) were fragmented and hybridized to each array for 18 hours at 45°C in a rotating hybridization. Array slides were stained with streptavidin/phycoerythrin utilizing a double-antibody staining procedure and then washed for antibody amplification following manufacturer’s protocol. Arrays were scanned in an Affymetrix Scanner 3000 and data was obtained using the GeneChip® Command Console Software (AGCC; Version 1.1) using the MAS5 algorithm to generate .CHP files.

Data processing and identification of differentially expressed genes (DEGs)

Data processing and identification of differentially expressed genes was done as previously described (Hoenerhoff et al. 2011; Pandiri et al. 2012) Briefly, array fluorescent pixel intensity measurements were acquired and gene expression data were normalized across all samples using the robust multiarray analysis (RMA) methodology (Irizarry et al. 2003). Using RMA-normalized data, for each probe set, pairwise comparisons (AN vs. HB and AN vs. HCC) were made using a bootstrap t-test while controlling the mixed directional false discovery rate (FDR) (Guo et al. 2010). This methodology controls for the overall FDRs for multiple comparisons as well as directional errors when declaring a gene to be upregulated or downregulated. The mdFDR threshold was set at 5%. Microarray data files (.cel) and associated annotations have been submitted to the Gene Expression Omnibus (GEO) database: GSE67316.

Bioinformatics analyses

Partek Genomics Suite, version 6.6 (release date June, 2012) (Partek, St. Louis, MO) was used to perform principal component analysis (PCA) on the normalized data and to generate heat maps to compare samples for differentially expressed probe sets. PCA uses a linear transformation to find orthogonal variables (principal components, PC) that describe the variability in the data. The first three PC that capture the majority of the variation in the data were used to visualize the spatial relationship of the HB, HCC and AN samples.

Ingenuity Pathway Analysis (IPA) 9.0 (www.ingenuity.com), application build-220217, version-16542223, was used to evaluate the most statistically significant overrepresented canonical pathways. These canonical pathways are based on the Ingenuity knowledge base. The significant biological canonical pathways were derived from IPA and the statistical significance was set at p < 0.05 (Fisher’s exact test).

The NextBio tool (www.nextbio.com) was used to compare the transcriptomic data sets of mouse HB tumor to most similar human and rodent transcriptomic datasets and evaluate pathway enrichment in both systems. NextBio is a curated and correlated repository of experimental data derived from an extensive set of public sources (e.g., ArrayExpress and GEO) that allows the user to compare patterns of gene expression in their experiment to thousands of published transcriptomic data sets. The statistical approach used by NextBio is analagous to gene set enrichment analysis (Kupershmidt et al. 2010; Subramanian et al. 2005). NextBio software was utilized to compare the differential transcriptomic changes (biogroups) common to human tumors curated by NextBio and mouse HB tumors.

Quantitative Real-Time PCR

RNA extraction and amplification was performed using Ovation Pico WTA System V2 (NuGEN, San Carlos, CA) following manufacturer’s recommendation. Total RNA (10ng) was amplified using Ovation Pico WTA System and resulted in 6 μg of amplified cDNA. Relative quantitative gene expression levels were detected using real-time PCR with the ABI PRISM 7900HT Sequence Detection System (Life Technologies, Grand Island, NY) using SYBR (Synergy Brands) green methodology. Primers were designed using Primer3Plus software (Rozen and Skaletsky 2000) to span exon-exon junctions with an annealing temperature of 60°C and amplification size of less than 150 bp. Briefly, 25 ng of cDNA were added to a 25 μl PCR reaction to get a final concentration of 1.00 ng/μl of cDNA. Forward and reverse primer final concentrations were 100nM in the SYBR green assay. The reactions were performed using the Power SYBR® Green PCR Master Mix (Life Technologies, Grand Island, NY). 18S was chosen as the endogenous control gene in our qPCR experiments. Relative quantification of gene expression changes was recorded after normalizing for 18S expression, computed by using the 2−ΔΔCT method (user manual #2, ABI Prism 7700 SDS).

Immunohistochemistry (IHC)

Immunohistochemistry was performed on unstained formalin fixed, paraffin embedded sections of three samples containing HCC and HB using an avidin-biotin peroxidase system (Vectastain Elite ABC kit; Vector Laboratories, Burlingame, CA) according to the manufacturer’s protocol. Proteins for immunohistochemical analysis were selected based on differential gene expression between HB and HCC from the microarray data, as well as the role of these proteins in HCC and HB development. The 5um sections were collected on charged glass slides and were deparaffinized in xylene and rehydrated through a decreasing graded ethanol series. Following antigen retrieval and blocking of endogenous peroxidases, primary antibody was applied. These included hepatocyte nuclear factor-4a (HNF4a (H-171), 1:250, rabbit polyclonal, Santa Cruz Biotechnology Inc., Santa Cruz, CA), cMYC (cMYC (AB32072), 1:75, rabbit polyclonal, Abcam Inc., Cambridge, MA), β-catenin (CTNNB1 (H-102), 1:750, rabbit polyclonal, Santa Cruz Biotechnology Inc., Santa Cruz, CA) and CYP2E1 ((AB1252) 1:1000, rabbit polyclonal, Millipore, Billerica, MA). Sections were incubated with avidin–biotin–peroxidase complex (Vector Laboratories), and 3,3-diaminobenzidine (DAB) was used to visualize all immune reactions. Positive controls included tissues known to exhibit positive expression of proteins of interest. Negative controls received the antisera from the same animal species as the source of the secondary antibody.

Results

HB and HCC from B6C3F1 mice do not share common mutations in Hras and Ctnnb1

Regions of the mouse Ctnnb1 and Hras gene sequences prone to mutation (“hot-spot” regions) which correspond to orthologous regions in humans were analyzed in all samples. We hypothesized that if HCC is a precursor lesion of HB, then these mutations found HCC should be preserved in HB during the course of neoplastic transformation and progression. Overall, the rates of Hras (5.5 and 11%) and Ctnnb1 (19 and 8.3%) mutation were relatively comparable respectively, between HB and HCC in these samples (Table 1). Of the 36 tumor samples containing both HB and adjacent HCC (6 frozen, 30 FFPE), 11 samples harbored one or more mutations in at least one tumor type (Table 2). Two of these 11 samples harbored mutations in both HB and HCC, but only one sample contained a shared mutation. A TCT → TTT transition in codon 37 of the Ctnnb1 gene was observed in both HCC and HB from one sample. Mutations were not observed in AN liver.

Table 1.

Incidence of Hras and Ctnnb1 mutations in HB and HCC from B6C3F1/N mice

Hras
Ctnnb1
Hepatoblastoma
Hepatocellular carcinoma
Hepatoblastoma
Hepatocellular carcinoma
Spontaneousa –b 260/473c (55%) – 1/59d (2%)
BDCA-exposed 2/36 (5.5%) 4/36 (11%) 7/36 (19%) 3/36 (8.3%)
Open in a new tab
a

Hepatocellular tumors arising in vehicle control animals.

b

Historical incidence unavailable.

c

Historical incidence of Hras mutations in spontaneous hepatocellular carcinomas in B6C3F1/N mice in NTP studies (Sills et al., 1999; Maronpot et al., 1995).

d

Historical incidence of Ctnnb1 mutations in spontaneous hepatocellular carcinomas in B6C3F1/N mice in NTP studies (Hayashi et al., 2003).

Table 2.

Hras and Ctnnb1 mutation spectrum in hepatoblastoma and hepatocellular carcinoma in BDCA-exposed B6C3F1/N mice

Animal no. Hras
Ctnnb1
Hepatoblastoma Hepatocellular carcinoma Hepatoblastoma Hepatocellular carcinoma
LM101 –a 61 CAA → CGA – –
LM122 – 61 CAA → CGA 37 TCT → TTTb 34 GGA → GTA
37 TCT → TTT
HM343 – 61 CAA → CGA 32 GAT → GTT –
LM106 – 61 CAA → CTA – –
MM214 61 CAA → CTA – – –
HM333 61 CAA → CTA – 52 CCT → CAT –
LM129 – – 32 GAT → AAT –
MM236 – – 32 GAT → GGT
33 TCT → TTT		 41 ACC → GCC
MM244 – – 5 GCT → TCT –
HM360 – – 35 ATC → AGC –
HM364 – – – 19–46 deletion
Open in a new tab
a

No mutation detected.

b

Single shared mutation in Ctnnb1 between HB and HCC in animal LM122.

HB is markedly different from HCC based on global gene expression

Comparisons between global gene expression of laser capture microdissected hepatoblastoma (HB), hepatocellular carcinoma (HCC), and adjacent-normal liver (AN) from the same individual animals were made in order to isolate gene expression changes common and different between HB and HCC in the context of associated normal liver in exposed animals. This comparison was designed to minimize background chemical effect on the adjacent normal liver and tumor samples, in order to identify alterations more specific to tumorigenesis in the absence of chemical effect. While some chemical effects on tumor biology may remain at play after this comparison, this method was the most stringent to control for chemical effect while identifying significantly altered genes associated with tumorigenesis between HB and HCC. Results indicated 10,346 differentially expressed genes in HB compared to AN liver (mdFDR < 5%). In HCC, there were 1,087 genes differentially expressed (mdFDR < 5%) compared to AN liver. Principal component analysis (PCA) based on all interrogated probe sets showed distinct tight clustering of HBs from HCC and AN liver (Figure 2), suggesting that there is marked variance in global gene expression that segregates HB from HCC and AN. Similarly, unsupervised hierarchical cluster analysis of differentially expressed genes segregated HB from HCC and AN (Figure 3).

Figure 2.

Figure 2

Open in a new tab

Principal components analysis (PCA) comparing gene expression profiles of HB (black), HCC (green), and AN liver (red) for differentially expressed probesets. The plot illustrates marked differences in global gene expression in HB compared to both HCC and AN liver; overlap of HCC and AN liver samples suggests that these samples have less variance in their gene expression compared to HB, which is distinctly and markedly different.

Figure 3.

Figure 3

Open in a new tab

Unsupervsed hierarchical cluster analysis (HCA) illustrating significant differences in global gene expression of HB tumors compared to HCC and AN liver (red = upregulated genes, green = downregulated genes). Similar to PCA analysis, hierarchichal clustering shows that gene expression alterations in HCC are more similar to AN liver than HB. Alterations in significantly differentially expressed genes in HB show a marked difference when compared to HCC.

Overrepresented pathways in mouse hepatoblastoma include embryonic development and genomic imprinting compared to normal liver

IPA analysis was performed to identify altered pathways in HB compared to AN liver. Genes targeted by Wnt/Ctnnb1 pathway signaling were altered in HB, including upregulation of various Wnt signaling genes (Wnt9a, Wnt10a, Wnt 7a), genes involved in Wnt feedback-regulation (Axin2, Nkd1), positive effectors (Lef1, Dvl3), and Wnt antagonists (Dkk2, Dkk3, Wif1). Secondly, there was significant upregulation of a number of genes involved in genomic imprinting (Igf2, Peg1, Peg10, Bex1, Meg3, H19, Ndn), which are normally expressed only in fetal liver. Finally, genes involved in embryonic stem cell pluripotency (T, Bmp4, Fzd6, Fzd10, Nog) and stem cell-related target genes (Sox9, Tbx3, Suz12, Zfp42, Gata4, Fzd4 and Fzd8) were differentially expressed, consistent with expression of hepatic stem/progenitor cell pathways (Table 3).

Table 3.

Ingenuity Pathway Analysis of Select Differentially Expressed genes from overrepresented pathways in mouse hepatoblastomas.

Fold Change
Gene Symbol Gene Name Hepatoblastoma Hepatocellular carcinoma
Wnt/Ctnnb1 target genes
Axin2 Axin2 (conductin) 20.3 –a
Dkk1 Dickkopf homolog 1 33 –
Lef1 Lymphiod enhancer-binding factor 1 43.1 –
Bmp4 Bone morphogenetic protein 4 58.8 –
Dkk2 Dickkopf 2 homolog 8.6 –
Dvl3 Dishevelled, dsh homolog 3 (Drosophila) 2 –
Wnt 5, 6, 7, 9, 10 Wingless-type MMTV integration site family 10.0 to 90.0 –
Wif1 WNT inhibitory factor 1 475 –
Stem cell-related targets
Tbx1 T-box1 37.3 –
Sox9 SRY (sex determining region Y)-box 9 3.2 –
Suz12 Suppressor of zeste 12 homolog (Drosophila) 3 –
T T, brachyury homolog (mouse) 268 –
Hepatic Targets
Cyp2e1 Cytochrome P450, family 2, subfamily E, polypeptide 1 −233.2 –
Cyp1a1 Cytochrome P450, family 1, subfamily A, polypeptide 1 −42 –
Hdac2 Histone deacetylase 2 3.2 –
Genomic Imprinting Genes
Igf2 Insulin-like growth factor 2 67 3.3
Peg1 Paternally expressed 1 5.7 –
Peg10 Paternally expressed 10 3.5 –
Bex1 Brain expressed, X-linked 1 166 44
Meg3 Maternally expressed 3 26 –
Ndn Necdin homolog (mouse) 6 −1.23
H19 H19, imprinted maternally expressed transcript 40.6 –
Open in a new tab
a

Not significantly differentiatially expressed.

Hepatoblastomas show overrepresentation of pathways associated with carcinogenesis and stem/progenitor cell signaling compared to HCC

Results of IPA analysis indicated overrepresentation of a number of biologic functions shared between HCC and HB, including cancer and cellular movement, and also identified differentially altered pathways involving protein synthesis, metabolic disease, and developmental disorders (Figure 4A). Overrepresented canonical pathways related to carcinogenesis in HB compared to HCC included EIF2 signaling, cell cycle control of chromosome replication, NRF2-mediated oxidative stress response, basal cell carcinoma signaling, and mTOR signaling (Figure 4B). Finally, there were a number of stem cell/developmental regulation pathways represented in HB compared to HCC, including Ephrin B signaling, mouse and human embryonic stem cell pluripotency, DNA methylation and transcriptional repression signaling, and sonic hedgehog and NANOG signaling (Figure 4C). Select genes that were differentially expressed between HB and HCC were validated using qPCR (Figure 5A-B) and immunohistochemistry (Figure 6). Targets for qPCR analysis included those genes observed as significantly differentially expressed between HB and HCC on microarray analysis, observed as related to genomic imprinting (Igf2, Meg3, Bex1), Ctnnb1/Wnt mediators (Axin2, Lef1, Wnt6, Wif1, Wnt10a), hepatic genes (Cyp2E1, Cyp1A2, Hnf4a), and stem cell-related genes (cMyc, Lrg5, Tbx3, Id1, T (brachury)). Those proteins selected for immunohistochemistry validation of microarray results were chosen based on differential gene expression on microarray, and roles in hepatic tumorigenesis (CTNNB1, cMYC, HNF4A, CYP2E1). Results of immunohistochemistry (labeling intensity, localization, and distribution) were consistent between the three different samples for each antibody; all three samples showed strong nuclear and cytoplasmic immunoreactivity to CTTNB1 in mouse HB compared to faint membrane labeling in HCC, increased nuclear immunoreactivity to cMYC in mouse HB compared to HCC, multifocal nuclear expression of HNF4A in HCC and no observable expression in HB, and decreased nuclear expression of CYP2E1 in mouse HB compared to HCC.

Figure 4.

Figure 4

Open in a new tab

Ingenuity pathway comparison analysis of differentially expressed genes in HB compared to HCC in BDCA-exposed B6C3F1 mice, including (A) top biological functions, (B) cancer-associated pathways, and (C) stem-cell and developmental pathways overrepresented in HB.

Figure 5.

Figure 5

Open in a new tab

Quantitative PCR validation of microarray targets involved in (A) genomic imprinting, Wnt/Ctnnb1 pathway, hepatic function, and (B) stem cell genes in HB and HCC from BDCA-exposed B6C3F1 mice.

Figure 6.

Figure 6

Open in a new tab

Immunohistochemistry validation of microarray targets differentially expressed between HB and HCC. (A) Diffuse cytoplasmic and membrane immunoreactivity to CTNNB1 antibody in HB compared to mild membrane staining in adjacent HCC (anti-CTNNB1 antibody, hematoxylin counterstain, 40×). (B) Diffuse nuclear immunoreactivity to cMYC antibody in HB compared to multifocal nuclear immunoreactivity in adjacent HCC (anti-cMYC antibody, hematoxylin counterstain, 20×). (C) Negative immunoreactivity to HNF4A antibody in HB compared to strong and diffuse nuclear immunoreactivity in adjacent HCC (anti-HNF4A antibody, hematoxylin counterstain, 40×). (D) Moderate to strong cytoplasmic immunoreactivity to CYP2E in HCC compared to weak cytoplasmic labeling in adjacent HB (anti-CYP2E antibody, hematoxylin counterstain, 20×).

Gene expression of mouse HB is highly concordant with early mouse embryonic liver, human pluripotent hepatic stem cells, and human hepatoblastoma

NextBio meta-analysis software (Kupershmidt et al. 2010) was used to compare mouse HB to other mouse and human datasets in the literature in order to identify similarities in datasets based on the total number of shared genes, and the number of genes similarly changed based on directionality of fold change in both datasets. Meta-analysis showed that, in contrast to HCC, mouse HB is transcriptomically very similar to early (E10.5) embryonic mouse liver (Chaignat et al. 2011; Otu et al. 2007) and highly concordant with hepatocyte-derived pluripotent stem cells (Lee et al. 2012) (Figure 7).

Figure 7.

Figure 7

Open in a new tab

NextBio meta-analysis comparing HB and HCC in BDCA-exposed B6C3F1 mice with mouse embryonic liver (E10.5, top) and hepatocyte derived pluripotent stem cells (bottom). There is a higher degree of concordance between hepatoblastoma in BDCA-exposed mice and mouse embryonic liver (3830 total shared genes, p=9.4E−183) with 1972 concordant upregulated (p=1.8E−59) and downregulated genes (p=4.2E−306), compared to HCC, which shared 694 total genes with mouse embryonic liver (p=9.6E−37), with only 294 concordant upregulated (p=5.6E−20) and 288 concordant downregulated genes (p=2.8E−54). Similarly, there is increased concordance between mouse HB and hepatocyte derived pluripotent stem cells, with 5223 total shared genes (p=7.3E−130), 1790 of which were upregulated (p=2.1E−21) and 2466 of which were downregulated (p=2.6E−238). In contrast, comparison of mouse HCC with this bioset showed 1226 shared genes (p=8.3E−29) and only 399 concordant upregulated (p=1.4E−11) and 582 concordant downregulated genes (p=2.1E−46).

Discussion

The primary objectives of the current study were to better understand the underlying molecular changes that characterize the relationship between HB and HCC in B6C3F1 mice, and how these changes relate to human HB in order to better understand the relevance of the mouse HB to human health in chemical hazard identification studies. Using global gene expression profiling and mutation analysis, we have shown that HB is markedly different and quite distinct from HCC in mice. First, mutation analysis of Ctnnb1 and Hras did not support a progression of HCC to HB through mutation conservation, and therefore do not provide evidence of neoplastic progression. The somatic mutation theory suggests that cancer begins with a genetic change in a single cell that passes it on to its progeny thereby generating a clone of malignant cells that accumulate further genetic changes in a step-wise manner (Berenblum and Shubik 1949). As such, a mutation that occurs during the transformation of a precursor lesion or early tumor should remain fixed within the genome throughout the progression to later stages of malignancy or metastasis. Rather, the mutation spectra between these tumors was quite different, they did not share common mutations, with the exception of only one sample. These data are consistent with previous studies (Anna et al. 2000), which show that the mutational spectrum is different in HB than that of HCC, suggesting that they may be distinct entities. In terms of global gene expression, it is quite remarkable to note that the distribution of samples observed on the PCA plot suggests that there was less variance between HCC and associated AN normal liver than the variance between HCC and HB. These findings strongly suggest that HB and HCC are distinct entities based on their global gene expression and mutation analysis. Overall, the marked differences in HB gene expression are consistent with the distinct primitive morphologic features of these tumors compared to HCC, and sheds light into possible origins of this tumor in the mouse.

In the current study, numerous pathways and their respective genes involved in both mouse and human embryonic stem cell pluripotency, embryonic development, and genomic imprinting were prominent in mouse HB, but not significantly represented in mouse HCC (Cairo et al. 2008; Clevers 2006; Monga 2011). Overrepresentation of genes from the hedgehog signaling (Hh) pathway, including strong upregulation of sonic hedgehog (Shh), its receptor patched 1 (Ptch1), and the Hh ligand smoothened (Smo), was observed in HB and not HCC. This pathway is responsible for critical cell fate decisions including proliferation, apoptosis, migration, and differentiation (McMillan and Matsui 2012). It plays vital roles in tissue morphogenesis during fetal development including embryonic development of the liver and hepatic regeneration in the adult; constitutive activation of Hh signaling can lead to transformation of progenitor cells when continuously activated, and lead to various cancers in humans (Beachy et al. 2004; Omenetti et al. 2011). Activation of the Hh pathway is also important in cancer stem cell regulation and enhanced tumor initiating and self-renewal potential (McMillan and Matsui 2012), and overexpression of this pathway has been reported in about two-thirds of human HB (Eichenmüller et al. 2009; Oue et al. 2010). Significant upregulation of genomic imprinting genes primarily expressed in fetal liver was observed in mouse HB from the current study; this upregulation is similar to that observed in human HB (Cairo et al. 2008; Tomlinson and Kappler 2012). Genomic imprinting is an epigenetic mechanism in which individual alleles are silenced based on the parent of origin, and loss of imprinting (LOI) is recognized in various childhood developmental disorders and in various cancers (Damaschke et al. 2013).

Genes associated with Ctnnb1/Wnt signaling and normal hepatic function were significantly differentially altered in mouse HB in this study compared to HCC. The Wnt/Ctnnb1 pathway is a master regulator of cell fate and proliferation during embryonic development, and it is essential for stem cell maintenance in a wide variety of tissues (Clevers 2006; Haegebarth and Clevers 2009). In this study, we observed dysregulation in a number of genes associated with a variety of Wnt/Ctnnb1 functions, including feedback-regulation, positive effectors and antagonists of the Wnt pathway, and target genes in carcinogenesis. Many of these genes expressed at high levels in HB tumors are targets of Lef1, a known Wnt/Ctnnb1 pathway transcription factor. Dysregulation of the Wnt/Ctnnb1/Lef1 pathway has also been observed in human cancer, and plays an important role in cancer stem cell biology (Polyak and Weinberg 2009), including maintenance of self-renewal and differentiation of cancer stem cells (Bisson and Prowse 2009). In this study, immunohistochemistry for CTNNB1 showed strong nuclear and cytoplasmic immunoreactivity in mouse HB compared to faint membrane labeling in HCC. CTNNB1 is very often overexpressed in HB, and translocation from the cell membrane to the cytoplasm and nucleus is commonly observed, in contrast to membrane staining in HCC (Anna et al. 2000). A number of downstream oncogene targets of the Wnt pathway associated with human HB were also altered in this study (cMyc, Ccnd1). cMyc was overexpressed in HB compared to HCC on microarray and qPCR, and immunohistochemistry for cMYC showed increased nuclear expression in HB compared to HCC. Others have shown that MYC oncogene overexpression is often associated with excessive CTNNB1/WNT signaling in human HB, and is also associated with aggressive and poorly-differentiated HB in humans (Cairo et al. 2008; Takayasu et al. 2001). Finally, a number of genes associated with normal mature hepatocyte metabolism and function were significantly downregulated in HB compared to HCC on microarray and qPCR in this study. HNF4A, which is necessary for epithelial transformation of progenitor cells in the fetal liver, and has an important role in maintaining mature hepatocyte function (Si-Tayeb et al. 2010), was significantly downregulated in HB on microarray and qPCR analysis compared to HCC. Consistent with this, there was multifocal nuclear expression of HNF4A in HCC by immunohistochemistry, and no observable expression in HB. This would suggest a more primitive, fetal phenotype of HB cells lacking mature hepatocyte function, which is consistent with the morphologic and transcriptomic phenotype of HB in this study. Similarly, CYP2E1 expression was decreased in HB compared to HCC by microarray analysis, qPCR, and immunohistochemistry. CYP2E1 plays an important role in detoxification in the mature liver, and loss or downregulation of this protein would suggest a phenotype lacking in adult hepatic detoxification systems, such as is observed in fetal liver. In fact, the expression of CYP2E1 has been shown in both human and mouse fetal liver to be decreased compared to that of adult liver (Choudhary et al. 2005).

Meta-analysis of mouse HB indicated that the key proceses of embryonic development and stem/multipotent cell regulatory pathways as the major characteristics of these tumors. There was significantly higher concondance not only with mouse embryonic liver, but also hepatic pleuripotent stem cell biosets in HB, compared to mouse HCC. This supports an embryonic phenotype of HB tumors as seen histologically, and further indicates that these are distinct tumors likely arising from a primordial hepatic precursor cell. Although there are several differences between human and mouse HB, our results suggest that mouse and human HB share some important similiarities in their molecular landscape, and therefore mouse HB is a relevant lesion in toxicology and carcinogenicity bioassays, and is an important indicator of potential cancer risk in hazard identification studies. Due to its marked genomic and phenotypic differences, this tumor may be considered a distinct entity from hepatocellular carcinoma.

Conclusions

Our data shows that mouse HBs are profoundly different from HCC in terms of their global gene expression and mutation spectra. In addition, mouse and human HB share significant similarities in the global gene expression, including dysregulation of embryonic developmental pathways, embryonic/stem cell pluripotency pathways, and genomic imprinting. In addition, meta-analysis supports that mouse HB is highly concordant with mouse embryonic liver and hepatic pluripotent stem cells. These findings suggest that HB and HCC are very distinct and separate entities, although possibly arising from a common hepatic stem or progenitor cell. It is possible that the development of HB may require a particular microenvironment which is provided by an adjacent HCC, including one with increased genomic instability. Further studies with greater sample sizes from several bioassays as well as spontaneously occurring tumors would be warranted for further understanding of these molecular changes. Additionally, since each tumor was sampled as a whole, factors such as tumor heterogeneity may possibly have played a role in these results. Future studies sampling multiple sites within each tumor may provide more information about variable mutation spectra and differential gene expression between HB and adjacent HCC. Using newer technologies such as exome sequencing at high read depths may provide more evidence on the conservation of mutations between HB and HCC, and cell lineage tracing experiments may provide more definitive evidence on the origin of HB.

Acknowledgments

The authors would like to thank the NIEHS histology and immunohistochemistry laboratories and the NTP Archives for their technical expertise, and the CMPB pathologists and NTP toxicologists for helpful discussions.

References

  1. Anna CH, Sills RC, Foley JF, Stockton PS, Ton TV, Devereux TR. Beta-catenin mutations and protein accumulation in all hepatoblastomas examined from B6C3F1 mice treated with anthraquinone or oxazepam. Cancer Res. 2000;60:2864–8. [PubMed] [Google Scholar]
  2. Armengol C, Cairo S, Fabre M, Buendia MA. Wnt signaling and hepatocarcinogenesis: The hepatoblastoma model. The international journal of biochemistry & cell biology. 2011;43:265–270. doi: 10.1016/j.biocel.2009.07.012. [DOI] [PubMed] [Google Scholar]
  3. Beachy PA, Karhadkar SS, Berman DM. Tissue repair and stem cell renewal in carcinogenesis. Nature. 2004;432:324–331. doi: 10.1038/nature03100. [DOI] [PubMed] [Google Scholar]
  4. Berenblum I, Shubik P. An experimental study of the initiating state of carcinogenesis, and a re-examination of the somatic cell mutation theory of cancer. Br J Cancer. 1949;3:109–18. doi: 10.1038/bjc.1949.13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bisson I, Prowse DM. WNT signaling regulates self-renewal and differentiation of prostate cancer cells with stem cell characteristics. Cell research. 2009;19:683–697. doi: 10.1038/cr.2009.43. [DOI] [PubMed] [Google Scholar]
  6. Buckley JD, Sather H, Ruccione K, Rogers PC, Haas JE, Henderson BE, Hammond GD. A case-control study of risk factors for hepatoblastoma. A report from the Childrens Cancer Study Group. Cancer. 1989;64:1169–76. doi: 10.1002/1097-0142(19890901)64:5<1169::aid-cncr2820640534>3.0.co;2-i. [DOI] [PubMed] [Google Scholar]
  7. Cairo S, Armengol C, De Reynies A, Wei Y, Thomas E, Renard CA, Goga A, Balakrishnan A, Semeraro M, Gresh L, Pontoglio M, Strick-Marchand H, Levillayer F, Nouet Y, Rickman D, Gauthier F, Branchereau S, Brugieres L, Laithier V, Bouvier R, Boman F, Basso G, Michiels JF, Hofman P, Arbez-Gindre F, Jouan H, Rousselet-Chapeau MC, Berrebi D, Marcellin L, Plenat F, Zachar D, Joubert M, Selves J, Pasquier D, Bioulac-Sage P, Grotzer M, Childs M, Fabre M, Buendia MA. Hepatic stem-like phenotype and interplay of Wnt/beta-catenin and Myc signaling in aggressive childhood liver cancer. Cancer Cell. 2008;14:471–84. doi: 10.1016/j.ccr.2008.11.002. [DOI] [PubMed] [Google Scholar]
  8. Chaignat E, Yahya-Graison EA, Henrichsen CN, Chrast J, Schütz F, Pradervand S, Reymond A. Copy number variation modifies expression time courses. Genome research. 2011;21:106–113. doi: 10.1101/gr.112748.110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chhabra R, Eustis S, Haseman J, Kurtz P, Carlton B. Comparative carcinogenicity of ethylene thiourea with or without perinatal exposure in rats and mice. Fundamental and Applied Toxicology. 1992;18:405–417. doi: 10.1016/0272-0590(92)90139-9. [DOI] [PubMed] [Google Scholar]
  10. Choudhary D, Jansson I, Stoilov I, Sarfarazi M, Schenkman JB. Expression patterns of mouse and human CYP orthologs (families 1–4) during development and in different adult tissues. Arch Biochem Biophys. 2005;436:50–61. doi: 10.1016/j.abb.2005.02.001. [DOI] [PubMed] [Google Scholar]
  11. Clevers H. Wnt/β-catenin signaling in development and disease. Cell. 2006;127:469–480. doi: 10.1016/j.cell.2006.10.018. [DOI] [PubMed] [Google Scholar]
  12. Cruz RJ, Jr, Ranganathan S, Mazariegos G, Soltys K, Nayyar N, Sun Q, Bond G, Shaw PH, Haberman K, Krishnamurti L, Marsh JW, Humar A, Sindhi R. Analysis of national and single-center incidence and survival after liver transplantation for hepatoblastoma: New trends and future opportunities. Surgery. 2013;153:150–159. doi: 10.1016/j.surg.2012.11.006. [DOI] [PubMed] [Google Scholar]
  13. Damaschke NA, Yang B, Bhusari S, Svaren JP, Jarrard DF. Epigenetic susceptibility factors for prostate cancer with aging. The Prostate. 2013 doi: 10.1002/pros.22716. n/a-n/a. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Diwan BA, Ward JM, Rice JM. Promotion of malignant ‘embryonal’ liver tumors by phenobarbital: increased incidence and shortened latency of hepatoblastomas in (DBA/2 × C57BL/6)F1 mice initiated with N-nitrosodiethylamine. Carcinogenesis. 1989;10:1345–8. doi: 10.1093/carcin/10.7.1345. [DOI] [PubMed] [Google Scholar]
  15. Eichenmüller M, Gruner I, Hagl B, Häberle B, Müller-Höcker J, von Schweinitz D, Kappler R. Blocking the hedgehog pathway inhibits hepatoblastoma growth. Hepatology. 2009;49:482–490. doi: 10.1002/hep.22649. [DOI] [PubMed] [Google Scholar]
  16. Grisham JW. Interspecies comparison of liver carcinogenesis: implications for cancer risk assessment. Carcinogenesis. 1997;18:59–81. doi: 10.1093/carcin/18.1.59. [DOI] [PubMed] [Google Scholar]
  17. Guo W, Sarkar SK, Peddada SD. Controlling False Discoveries in Multidimensional Directional Decisions, with Applications to Gene Expression Data on Ordered Categories. Biometrics. 2010;66:485–492. doi: 10.1111/j.1541-0420.2009.01292.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Haegebarth A, Clevers H. Wnt Signaling, Lgr5, and Stem Cells in the Intestine and Skin. The American Journal of Pathology. 2009;174:715–721. doi: 10.2353/ajpath.2009.080758. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Herzog CE, Andrassy RJ, Eftekhari F. Childhood cancers: hepatoblastoma. Oncologist. 2000;5:445–53. doi: 10.1634/theoncologist.5-6-445. [DOI] [PubMed] [Google Scholar]
  20. Hoenerhoff MJ, Pandiri AR, Lahousse SA, Hong HH, Ton TV, Masinde T, Auerbach SS, Gerrish K, Bushel PR, Shockley KR, Peddada SD, Sills RC. Global gene profiling of spontaneous hepatocellular carcinoma in B6C3F1 mice: similarities in the molecular landscape with human liver cancer. Toxicologic Pathology. 2011;39:678–99. doi: 10.1177/0192623311407213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hoenerhoff MJ, Pandiri AR, Snyder SA, Hong HHL, Ton TV, Peddada S, Shockley K, Witt K, Chan P, Rider C, Kooistra L, Nyska A, Sills RC. Hepatocellular Carcinomas in B6C3F1 Mice Treated with Ginkgo biloba Extract for Two Years Differ from Spontaneous Liver Tumors in Cancer Gene Mutations and Genomic Pathways. Toxicologic Pathology. 2013;41:826–841. doi: 10.1177/0192623312467520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003;4:249–264. doi: 10.1093/biostatistics/4.2.249. [DOI] [PubMed] [Google Scholar]
  23. Kupershmidt I, Su QJ, Grewal A, Sundaresh S, Halperin I, Flynn J, Shekar M, Wang H, Park J, Cui W, Wall GD, Wisotzkey R, Alag S, Akhtari S, Ronaghi M. Ontology-based meta-analysis of global collections of high-throughput public data. PLoS One. 2010;5 doi: 10.1371/journal.pone.0013066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lack EE, Neave C, Vawter GF. Hepatoblastoma: A clinical and pathologic study of 54 cases. The American Journal of Surgical Pathology. 1982;6:693–705. [PubMed] [Google Scholar]
  25. Lee SB, Seo D, Choi D, Park KY, Holczbauer A, Marquardt JU, Conner EA, Factor VM, Thorgeirsson SS. Contribution of hepatic lineage stage-specific donor memory to the differential potential of induced mouse pluripotent stem cells. Stem cells (Dayton, Ohio) 2012;30:997–1007. doi: 10.1002/stem.1074. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Maronpot RR, Flake G, Huff J. Relevance of animal carcinogenesis findings to human cancer predictions and prevention. Toxicol Pathol. 2004;32(Suppl 1):40–8. doi: 10.1080/01926230490425003. [DOI] [PubMed] [Google Scholar]
  27. McMillan R, Matsui W. Molecular Pathways: The Hedgehog Signaling Pathway in Cancer. Clinical Cancer Research. 2012;18:4883–4888. doi: 10.1158/1078-0432.CCR-11-2509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Monga SPS. Role of Wnt/β-catenin signaling in liver metabolism and cancer. The international journal of biochemistry & cell biology. 2011;43:1021–1029. doi: 10.1016/j.biocel.2009.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. National Toxicology Program N. NTP Toxicology and Carcinogenesis Studies of Oxazepam (CAS No. 604-75-1) in F344/N Rats (Feed Studies) Natl Toxicol Program Tech Rep Ser. 1998;468:1–189. [PubMed] [Google Scholar]
  30. National Toxicology Program, N. NTP Toxicology and Carcinogenesis Studies of Diethanolamine (CAS No. 111-42-2) in F344/N Rats and B6C3F1 Mice (Dermal Studies) Natl Toxicol Program Tech Rep Ser. 1999;478:1–212. [PubMed] [Google Scholar]
  31. National Toxicology Program, N. NTP Toxicology and Carcinogenesis Studies of Methyleugenol (CAS NO. 93-15-2) in F344/N Rats and B6C3F1 Mice (Gavage Studies) Natl Toxicol Program Tech Rep Ser. 2000;491:1–412. [PubMed] [Google Scholar]
  32. National Toxicology, Program N. NTP Technical Report on the Toxicology and Carcinogenesis Studies of Ginkgo biloba Extract (CAS NO. 90045-36-6) in F344/N Rats and B6C3F1/N Mice (Gavage Studies) Natl Toxicol Program Tech Rep Ser. 2013:1–184. [PubMed] [Google Scholar]
  33. National Toxicology Program, N. NTP Historical Controls Report: All Routes and Vehicles, B6C3F1 mice. 2014. November, 2014. [Google Scholar]
  34. Omenetti A, Choi S, Michelotti G, Diehl AM. Hedgehog signaling in the liver. Journal of Hepatology. 2011;54:366–373. doi: 10.1016/j.jhep.2010.10.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Otu HH, Naxerova K, Ho K, Can H, Nesbitt N, Libermann TA, Karp SJ. Restoration of Liver Mass after Injury Requires Proliferative and Not Embryonic Transcriptional Patterns. Journal of Biological Chemistry. 2007;282:11197–11204. doi: 10.1074/jbc.M608441200. [DOI] [PubMed] [Google Scholar]
  36. Oue T, Yoneda A, Uehara S, Yamanaka H, Fukuzawa M. Increased expression of the hedgehog signaling pathway in pediatric solid malignancies. Journal of Pediatric Surgery. 2010;45:387–392. doi: 10.1016/j.jpedsurg.2009.10.081. [DOI] [PubMed] [Google Scholar]
  37. Pandiri AR, Sills RC, Ziglioli V, Ton TV, Hong HH, Lahousse SA, Gerrish KE, Auerbach SS, Shockley KR, Bushel PR, Peddada SD, Hoenerhoff MJ. Differential Transcriptomic Analysis of Spontaneous Lung Tumors in B6C3F1 Mice: Comparison to Human Non-Small Cell Lung Cancer. Toxicologic Pathology. 2012;40:1141–59. doi: 10.1177/0192623312447543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Polyak K, Weinberg RA. Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nature Reviews Cancer. 2009;9:265–273. doi: 10.1038/nrc2620. [DOI] [PubMed] [Google Scholar]
  39. Rozen S, Skaletsky HJ. Primer3 on the WWW for general users and for biologist programmers. Humana Press; Totowa, NJ: 2000. [DOI] [PubMed] [Google Scholar]
  40. Si-Tayeb K, Lemaigre FP, Duncan SA. Organogenesis and development of the liver. Developmental cell. 2010;18:175–89. doi: 10.1016/j.devcel.2010.01.011. [DOI] [PubMed] [Google Scholar]
  41. Spector LG, Birch J. The epidemiology of hepatoblastoma. Pediatric Blood & Cancer. 2012;59:776–779. doi: 10.1002/pbc.24215. [DOI] [PubMed] [Google Scholar]
  42. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–50. doi: 10.1073/pnas.0506580102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Takayasu H, Horie H, Hiyama E, Matsunaga T, Hayashi Y, Watanabe Y, Suita S, Kaneko M, Sasaki F, Hashizume K, Ozaki T, Furuuchi K, Tada M, Ohnuma N, Nakagawara A. Frequent deletions and mutations of the beta-catenin gene are associated with overexpression of cyclin D1 and fibronectin and poorly differentiated histology in childhood hepatoblastoma. Clin Cancer Res. 2001;7:901–8. [PubMed] [Google Scholar]
  44. Tomlinson GE, Kappler R. Genetics and epigenetics of hepatoblastoma. Pediatric blood & cancer. 2012;59:785–92. doi: 10.1002/pbc.24213. [DOI] [PubMed] [Google Scholar]
  45. Turusov VS, Torii M, Sills RC, Willson GA, Herbert RA, Hailey JR, Haseman JK, Boorman GA. Hepatoblastomas in mice in the US National Toxicology Program (NTP) studies. Toxicol Pathol. 2002;30:580–91. doi: 10.1080/01926230290105802. [DOI] [PubMed] [Google Scholar]
  46. von Schweinitz D. Hepatoblastoma: recent developments in research and treatment. Seminars in Pediatric Surgery. 2012;21:21–30. doi: 10.1053/j.sempedsurg.2011.10.011. [DOI] [PubMed] [Google Scholar]

RESOURCES