Background
Hepatocellular carcinoma (HCC) is a leading cause of cancer death worldwide [
1,
2], and it is usually associated with specific risk factors including hepatitis B or C virus infection, high alcohol intake, hemochromatosis, and nonalcoholic fatty liver disease [
3]. A typical cancer can harbor thousands of somatic mutations, of which 10–100 might occur in the coding region of genes [
4‐
8]. With the advent of next generation sequencing, recent studies have shown that the HCC genome can contain various somatic mutations, intrachromosomal rearrangements, gene fusions, and focal copy number alterations [
9‐
11]. These studies have also indicated that genes related to two pathways, the p53 and Wnt/β-catenin signaling pathways, are most frequently mutated in HCC. Furthermore, whole genome analysis of the HCC genome has indicated that one of the most frequent mutations identified in HCC is TERT promoter region mutation [
12].
HCC is a heterogeneous disease in terms of its morphology, biological behavior, response to treatment, and clinical outcome. Traditionally, this heterogeneity has been explained by cancer cell clonal evolution, and the step-wise acquisition of genetic mutations [
13]. However, recent evidence has suggested that HCC may conform to the cancer stem cell (CSC) hypothesis; this hypothesis proposes that a subset of cells with stem cell-like features divide asymmetrically to generate a heterogeneous cell population, and these stem cell-like cells play a fundamental role in tumor maintenance, chemoresistance, and metastasis [
14].
In HCC, several CSC markers, including CD133, CD90, CD44, CD24, and CD13 have been identified [
15‐
19]. We have previously demonstrated that HCC subtypes can be defined by the expression of the hepatic stem/progenitor cell markers epithelial cell adhesion molecule (EpCAM) and α-fetoprotein, and that these subtypes correlate with distinct gene expression signatures and patient prognoses [
20,
21]. Our previous data also suggested that EpCAM is a marker of liver CSCs, and might be used to enrich a highly tumorigenic and chemoresistant cell population.
In the present study, we sorted EpCAM+ and EpCAM− cell populations from fresh HCC specimens and performed whole exome sequencing on the sorted cell populations, to identify the somatic mutations that may explain the intratumor heterogeneity of cells (EpCAM+ CSCs and EpCAM− non-CSCs in the same tumor). We further evaluated the identified somatic mutations in independent 57 HCC tissues and EpCAM expression status, to identify the somatic mutations that may explain the intertumor heterogeneity of HCCs (EpCAM-positive and -negative HCCs). Our aim was to examine whether EpCAM expression is associated with specific genetic mutations in EpCAM+ CSCs (intratumor heterogeneity) or EpCAM-positive HCCs (intertumor heterogeneity), and to determine whether HCC conforms to the clonal evolution or CSC model.
Methods
Cell culture
HuH1, HuH7, HLE, HLF, Hep3B, HEP-G2, SK-Hep-1, and PLC/PRL/5 human liver cancer cell lines were obtained from the Japanese Collection of Research Bioresources (JCRB; Osaka, Japan) or American Type Culture Collection (ATCC; Manassas, VA). Cells were routinely cultured in DMEM supplemented with 10% FBS. Two fresh HCC specimens (HCC1 and HCC2) were obtained and were used for xenotransplantation and to prepare single-cell suspensions for analysis. Primary HCC tissues were dissected and digested in 1 mg/mL type 4 collagenase (Sigma-Aldrich Japan K.K., Tokyo, Japan) solution at 37 °C for 15–30 min. Contaminated red blood cells were lysed with ammonium chloride solution (STEMCELL Technologies, Vancouver, BC, Canada) on ice for 5 min.
Fluorescence activated cell sorting (FACS)
Cultured cells were trypsinized, washed, and resuspended in Hank’s Balanced Salt Solution (Lonza, Basel, Switzerland) supplemented with 1% HEPES and 2% FBS. Cells were then incubated with antibodies on ice for 30 min. Labeled cells were analyzed by FACS using a FACSCalibur (BD Biosciences, San Jose, CA). The antibodies used were: a FITC-conjugated anti-EpCAM monoclonal antibody (Clone Ber-EP4; DAKO, Carpinteria, CA); an APC-conjugated anti-CD326 (EpCAM) antibody (Miltenyi Biotec K.K., Tokyo, Japan); an APC-conjugated anti-CD90 monoclonal antibody (Clone 5E10; eBioscience, San Diego, CA); an APC-conjugated anti-CD133/2 antibody (Clone 293C3; Miltenyi Biotec K.K.); an APC-conjugated anti-CD44 mouse monoclonal antibody (eBioscience); an APC-conjugated anti-CD13 antibody (eBioscience); and a PE-conjugated anti-CD24 antibody (Miltenyi Biotec K.K.).
Cell sorting
Cells were trypsinized, washed, and resuspended in Hank’s Balanced Salt Solution supplemented with 1% HEPES and 2% FBS. Cells were then incubated with an APC-conjugated anti-CD326 (EpCAM) on ice for 30 min, and EpCAM positive and negative cells were isolated using a BD FACSAria II cell sorting system (BD Biosciences). In addition, EpCAM+ and EpCAM− cells were also sorted for functional studies using an autoMACS pro cell separator and CD326 (EpCAM) microbeads (Miltenyi Biotec K.K.).
Immunohistochemistry (IHC) analyses
HCC tissue samples were obtained from patients who had undergone radical resection at the Center for Liver Diseases in Kanazawa University Hospital, Kanazawa, Japan. All patients provided informed consent and the tissue acquisition procedures were approved by the Ethics Committee of Kanazawa University. In total, 57 formalin-fixed paraffin-embedded HCC samples, resected between 2008 and 2011, were used for the IHC analyses. IHC was performed using EnVision+ Kits (DAKO, Carpinteria, CA), according to the manufacturer’s instructions. An anti-EpCAM monoclonal antibody (VU-1D9; Oncogene Research Products, San Diego, CA) was used to detect EpCAM expression.
For spheroid formation assays, single cell suspensions from HCC1 and HCC2 were generated using FACS and 1.5 × 104 HCC1 cells or 1.0 × 104 HCC2 cells were seeded in 6-well Ultra-Low Attachment Microplates (Corning, Corning, NY). The number of spheroids was determined 21 days after seeding.
Tumorigenicity in NOD/SCID mice
The protocol for animal procedures was approved by the Kanazawa University Animal Care and Use Committee. Cells were suspended in 200 µL of 1:1 DMEM:Matrigel (BD Biosciences) and were subcutaneously injected into 6-week-old NOD/SCID mice (NOD/NCrCRl-Prkdcscid) purchased from Charles River Laboratories, Inc. (Wilmington, MA). The size and incidence of subcutaneous tumors was recorded. For histological evaluation, tumors were formalin-fixed and paraffin-embedded prior to storage.
Quantitative reverse transcription-PCR analysis
Total RNA was extracted using High Pure RNA Isolation Kit (Roche Diagnostics K.K., Tokyo, Japan) according to the manufacturer’s instructions. The expression of selected genes was determined in triplicate using the 7900 Sequence Detection System (Applied Biosystems, Foster City, CA). Each sample was normalized relative to 18s rRNA expression. The probes used were PCDH18, Hs01556217_m1; CYP2R1, Hs01379776_m1; TECTA, Hs00193706_m1; ITGB8, Hs00174456_m1; CSMD1, Hs00899130_m1; PER1, Hs01092603_m1; ALKBH3, Hs00286731_m1; OSCP1, Hs00376771_m1; and 18s rRNA, Hs99999901_s1 (Applied Biosystems).
RNA interference
Small interfering RNAs (siRNAs) specific to PCDH18 (#1, HSS122980: #2, HSS122982) and a negative control (12935200) siRNA were designed and synthesized by Invitrogen (Invitrogen, Carlsbad, CA). A total of 2 × 105 cells were seeded into 6-well plates 24 h before transfection. Transfection was performed using Lipofectamine RNAiMAX Transfection Reagent (Invitrogen), according to the manufacturer’s instructions. A total of 20, 40, 60, and 100 pmol/L of siRNAs was used for each transfection in SK-Hep-1, HCC2, HLE, and HLF, respectively.
DNA extraction and whole exome sequencing
DNA extraction was performed using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). The SureSelect Human All Exon V4 Kit (Agilent Technologies, Santa Clara, CA) was used for whole exome capture, and the HiSeq 2000 Sequencing System (Illumina Inc., San Diego, CA) was used for massive parallel sequencing. The sequence reads were mapped against the University of California, Santa Cruz hg19 Genome Browser (
http://hgdownload.cse.ucsc.edu/goldenPath/hg19/chromosomes/). Sequence variations, including single nucleotide polymorphisms and insertion/deletions were detected using the Genome Analysis Toolkit software (GATK; Broad Institute, Cambridge, MA). All of the whole exome sequencing and analysis was performed at Riken Genesis (Riken Genesis, Tokyo, Japan). To predict the effect of nonsynonymous single nucleotide substitutions on protein structure, function, and phenotype, we used tools available online, such as SIFT (
http://sift.jcvi.org/) [
22] and Polyphen2 (
http://genetics.bwh.harvard.edu/pph2/) [
23].
DNA extraction and Sanger sequencing
DNA extraction was performed using the QIAamp DNA Mini Kit (Qiagen). PCR Primers were designed by Primer-BLAST (
http://www.ncbi.nlm.nih.gov/tools/primer-blast/). Primers are listed in Tables
1 and
2. PCR amplifications were performed using Takara Taq Hot Start Version (Takara, Shiga, Japan), or PrimeSTAR GXL DNA Polymerase (Takara) using a standard application protocol and the manufacturer’s instructions. PCR cleanup was performed using the QIAquick PCR Purification Kit (Qiagen). The cleaned PCR products were sequenced using a BigDyeTerminator v3.1 cycle sequencing KIT (Applied Biosystems). Sequenced products were run on the Life Technologies 3130xl Genetic Analyzer (Applied Biosystems). Electropherograms were visualized and analyzed using Sequence Scanner v2.0 software (Applied Biosystems).
Table 1
Primers used for PCR amplification
PCDH18
| Exon 1 | Forward 5′-TAATCTGGGAAGCAAGGGGAC-3′ |
| | Reverse 5′-ACGACCAAACAAACGCAAGG-3′ |
| Exon 2 | Forward 5′-CACTGTCTCCTGCCTCACTG-3′ |
| | Reverse 5′-ATAGTTGGTAGCTGGCTGCG-3′ |
| Exon 3 | Forward 5′-GGCTGTATCGGATGAGGTGG-3′ |
| | Reverse 5′-CCAGCAGGTCTCTCAGCTTC-3′ |
| Exon 4 | Forward 5′-CAGTGGCTAGTTTCTACACGAC3′ |
| | Reverse 5′-TCACACCTAGTTCTTCCCACG-3′ |
Table 2
Primers used for Sanger sequencing
PCDH18
| Exon 1 #1 | 5′-GCTAAAGTGTGCATCTTTGCTAC-3′ |
| Exon 1 #2 | 5′-CAGCAACACTGCACAAATTGC-3′ |
| Exon 1 #3 | 5′-CTTCGGGCTTCCTCCATCTC-3′ |
| Exon 1 #4 | 5′-TCAGCCCAGAATCCTTGTCC-3′ |
| Exon 1 #5 | 5′-TCTGAGGCAGTGAGCTGAAG-3′ |
| Exon 2 | 5′-CACTGTCTCCTGCCTCACTG-3′ |
| Exon 3 | 5′-GGCTGTATCGGATGAGGTGG-3′ |
| Exon 4 #1 | 5′-CACACTTGCATTGTGTACATACG-3′ |
| Exon 4 #2 | 5′-GAAGGCGGTAAGAGACGCTG-3′ |
Cell proliferation assays
For cell proliferation assays, single cell suspensions of 2 × 103 cells were seeded in 96-well plates, and cell density was evaluated 48 h after seeding using the Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan) according to the manufacturer’s instructions.
Statistical analysis
Different test groups were compared using GraphPad Prism software 6.0 (GraphPad Software, San Diego, CA).
Discussion
It remains to be determined whether intratumor heterogeneity originates from the clonal evolution of tumor cells, with the step-wise acquisition of genetic changes (clonal evolution model), or a balance of self-renewal and differentiation by CSCs, which could potentially be regulated by the microenvironmental niche (CSC model). It is also possible that both models are true to a greater or lesser extent [
41‐
44]. We have postulated that CSCs have a greater potential to acquire genetic mutations than non-CSCs because they are reported to be more resistant to chemo/radiation therapy, and are highly tumorigenic and metastatic. However, in the current study, we found that our two primary HCC cells that follow the CSC model had similar somatic mutation patterns in EpCAM
+ CSCs and EpCAM
− non-CSCs. This suggests that at certain points and conditions in the process of tumorigenesis, CSCs and non-CSCs are genetically similar, and that differences in their tumorigenic/metastatic ability may be conferred by signaling pathways rather than genetic alterations. However, because our data only reflect the exome status of two HCC nodules following the CSC hypothesis, it is possible that HCC CSCs acquire more genetic mutations at different organ sites or after chemo/radiation treatments that may confer a treatment-resistant phenotype. Further studies are required to evaluate the relationship between cancer cell evolution, CSCs, and treatment resistance.
Although we did not detect unique mutations that were enriched in EpCAM+ CSCs in our two primary HCC samples, we did identify a number of novel somatic mutations. One of these somatic mutations, PCDH18 (HCC1), was detected in 3/57 HCCs, and was significantly associated with EpCAM-positive HCC. The total PCDH18 mutation frequency was 5.3%, but in EpCAM-positive HCCs the PCDH18 mutation frequency was 15.8%. Furthermore, although we did not detect PCDH18 mutations in HCC cell lines, we did find that PCDH18 gene expression was suppressed in EpCAM-positive HCC cell lines compared with EpCAM-negative cell lines. These data suggest that a functional loss of PCDH18, by genetic mutation or other mechanisms such as epigenetic gene silencing, may be associated with the generation of EpCAM-positive HCCs. Indeed, PCDH18 knockdown in EpCAM-positive HCC2 cells resulted in a slightly enhanced rate of proliferation, indicating that the requirement for PCDH18 expression may have been bypassed in EpCAM-positive HCCs; in EpCAM-negative HCC cell lines, PCDH18 knockdown instead inhibited cell proliferation. Taken together, these data suggest that PCDH18 may play different roles in EpCAM-positive and EpCAM-negative HCCs.
Protocadherins (PCDHs) are members of the nonclassic subfamily of calcium-dependent cell–cell adhesion molecules, which is part of the cadherin superfamily [
45]. The cadherin superfamily is classified into classical cadherins, desmosomal cadherins, and PCDHs. The PCDH family is largely divided into two groups based on their genomic structure: clustered PCDHs and non-clustered PCDHs. Non-clustered PCDHs are further classified into three subgroups: δ1, δ2, and ε.
PCDH18 belongs to the δ2-PCDH subgroup. Other δ2-PCDHs include
PCDH8,
PCDH10,
PCDH17, and
PCDH19 [
46].
PCDH18 is reported to be expressed in the central nervous system and pharyngeal arches of zebrafish embryos [
47], and plays a role in cell adhesion, behavior, and migration in zebrafish development [
48]. Although the function of
PCDH18 in humans is unclear, some studies have shown that
PCDH18 deletion may be associated with altered brain development, intellectual disability, and multiple malformations with pulmonary hypertension [
26‐
28,
49,
50]. Significantly, several δ2-PCDH members have been reported to function as tumor suppressor genes. For example,
PCDH8 is genetically or epigenetically silenced in breast cancer [
51] and mantle cell lymphoma [
52]. The
PCDH10 and
PCDH17 promoter regions are reported to be hypermethylated in uterine cervical cancer [
53], head and neck cancer, and some gastrointestinal cancers [
54‐
56]. And our study has shown that a loss of
PCDH18 gene expression may be related to the development of EpCAM-positive HCCs. In contrast, our data indicated the requirement of
PCDH18 expression in EpCAM-negative HCC cell lines. We previously demonstrated that EpCAM
+ CSCs show epithelial cell feature with highly tumorigenic capacity with activation of Wnt signaling, whereas CD90
+ CSCs show mesenchymal cell feature with highly metastatic capacity with activation of c-Kit signaling. Furthermore, CD90
+ CSCs were detected in all EpCAM-negative HCC cell lines [
57]. Therefore, it is plausible that
PCDH18 may have a role on maintenance of mesenchymal features of CD90
+ CSCs. The different role of
PCDH18 gene in different cellular contexts needs to be further evaluated in the future.
Authors’ contributions
TH, TY, MH, and SK contributed to the study design. TH, TY, and HO contributed to data acquisition and analysis. KN, YH, YN, TH, YA, MY, NO, HS, and HT discussed the interpretation of the data. TH and TY wrote the manuscript. All authors read and approved the final manuscript.