Background
Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide and the third leading cause of cancer mortality [
1,
2]. The genetic and physiological complexity of HCC is a major obstacle to the study of tumorigenesis and identification of therapeutic targets. HCC patients typically have poor tolerance of systemic chemotherapy due to underlying liver dysfunction and are resistant to conventional chemotherapeutic agents. Nevertheless, effective chemotherapy and other treatment strategies for patients with HCC are still lacking. Clinicians attempt to select the most effective anti-cancer treatment in order to improve the quality of lifer of HCC patients. To this end, personalized medicine, the potential to tailor therapy with the best response and highest safety margin to ensure better patient care, is optimal to reduce unnecessary exposure to chemotherapy, which has limited efficacy and/or acute toxicity in HCC patients. However, the development of “personalized” treatments for HCC patients is plagued by limitations, such as a lack of fresh biopsy from advanced HCC, the low incidence of genetic driver mutations in HCC, and the tumor heterogeneity. Therefore, suitable experimental models are needed to overcome these problems and to explore the mechanism of HCC development, progression, and metastasis. In this study, we established and characterized two new HCC cell lines, AMC-H1 and AMC-H2, from HBV-infected HCC patients as experimental model of personalized medicine. Both cell lines displayed common hepatocytes characteristics but differed in other way from each other. Also, these secreted HBV DNA stably after serial culture for more than 100 passages.
The 3D culture models have widely been used for the past two decades in an effort to mimic the in vivo behavior of normal or transformed cells under conditions more amenable to experimental investigation. Currently, these methods have been routinely applied to the analysis of malignant transformation and tumor progression. These 3D cultures have potential to greatly improve cell-based drug screening and identify toxic and ineffective substances at an earlier stage of the drug discovery pipeline than animal or clinical trials. Additionally, spheroids composed of patient-derived tissues maintained their own characteristics, such as high levels of glucose consumption, lactate production, and HIF1a levels as well as lower levels of radiation-induced oxidative stress, for a long time in vitro, and studies have suggested that these models serve as proper culture methods for personalized therapy in colorectal cancer [
3], glioma [
4], and head and neck cancer [
5]. Moreover, a 3D system with homogenous and automated methods in the extracellular matrix (ECM) was developed and used to visualize and analyze the migration and invasion ability of cancer cells using mouse breast and human sarcoma biopsy samples [
6] as well as melanoma cells [
7]. In our previous study, we probed the crosstalk between HCC and tumor microenvironments (TME) that promoted HCC chemoresistance and migration in multicellular tumor spheroids (MCTS). Our results suggested that targeting TME components may offer a promising therapeutic strategy for HCC liver cancer therapy. Thus, in this study, we propose to apply the 3D co-culture model to a system of patient-derived HCC cells and stromal cells representing human hepatic stellate cells (HSCs), human fibroblasts, and human umbilical vein endothelial cells (HUVECs), to screen for personalized cancer therapy. Ultimately, we aim to provide more informative results for clinical drug application through convergence of patient-derived liver cancer cells and TME components in MCTS.
Methods
Cell lines and culture conditions
The stromal cells WI38 (human fibroblasts), LX2 (human hepatic stellate cells), and HUVECs (human umbilical vein endothelial cells) were purchased from ATCC (Manassas, VA, USA), Merck Millipore, (Darmstadt, Germany), and PromoCells (Heidelberg, Germany), respectively. All cells were maintained at 37 °C in a humidified atmosphere of 5% CO2.
WI38 cells were cultured in Minimum Essential Media (MEM; Welgene, Daegu, Korea) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gibco, Grand Island, NY, USA) and 1× penicillin-streptomycin (P/S; Gibco). LX2 cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM; Welgene) supplemented with 2% heat-inactivated FBS and 1× P/S. HUVECs were cultured in endothelial basal medium (EBM) purchased from PromoCells.
Primary culture of HCCs
Immediately after surgery, a portion of the excised tumor was immersed in Hanks balanced salt solution (HBSS; Gibco) and transported from the operating room at 0 °C to the laboratory. The specimens were collected under sterile conditions and rinsed 2-3 times with HBSS free of calcium and magnesium to remove blood. After removal of blood, the liver sample was cut into small fragments, gently dispersed, and placed in HBSS containing 0.03% pronase (Gibco), 0.05% type IV collagenase (Gibco), and 0.01% deoxyribonuclease (DNase, from bovine pancreas, Gibco) for 20 min at 37 °C. The resultant suspension was filtered through a 100-μm-nylon filter (BD Falcon, Franklin Lakes, NJ, USA) and centrifuged at 50 x g for 2 min at 4 °C to obtain hepatocytes. The pellet was washed twice in HBSS containing 0.005% DNase. The final cell suspensions were cultured in collagen-coated T25 flasks (BD Falcon) in hepatocyte basal medium (Lonza, Basel, Switzerland) supplemented with 10% heat-inactivated FBS, 1 ng/ml hepatocyte growth factor (HGF, Prospec, Rehovot, Israel), and 1× antibiotic-antimycotic (Gibco) as HBM media at 37 °C in a humidified incubator with 5% CO2. The medium was changed 24 h after seeding to remove dead cells and debris. When cells reached 70-80% confluence, the cells were re-plated in HBM medium with supplements. Confluent cells were trypsinized, counted, and diluted 1:3-1:5 at every passage. Once cell lines were maintained for more than 30 passages, the cells were collected and stored in liquid nitrogen.
Ethics approval and consent to participate
The study was conducted in accordance with the Declaration of Helsinki principles. The study was approved by the Human Research Ethics Committee of ASAN Medical Center (Permit Number: 2007-0332). The institutional review board at ASAN Medical Center complies with all applicable guidelines, including the ICH, KGCP, and bioethics and safety act. Written informed consent for the use of tissues for research was obtained from patients at the time of procurement of tumor specimens. One line named AMC-H1 was acquired from a 55-year-old female patient, and another, AMC-H2, was from a 51-year-old male patient. The etiology of HCC was HBV infection in both patients.
Immunocytochemistry
To validate the primary cells, cells were fixed with 4% paraformaldehyde (PFA; Sigma, St Louis, MO, USA) for 10 min at room temperature, permeabilized with 0.1% Triton X-100 (Sigma) in Dulbecco’s phosphate-buffered saline (DPBS; Welgene) for 30 min at room temperature, and then washed three times with DPBS. The following primary antibodies were used: mouse monoclonal anti-human serum albumin (ALB; 15C7, ab10241, 1:500; Abcam, Cambridge, UK), mouse monoclonal anti-human Hep Par-1 (Clone OCH1E5, M7158, 1:100; Agilent Technologies, Santa Clara, CA, USA), and mouse monoclonal anti-human CD133/1 (AC133, 130-090-422, 1:100; Miltenyi Biotec, Bergisch Gladbach, Germany). Samples were incubated with the primary antibodies for 16 h at 4 °C and then washed for 10 min three times with DPBS. The secondary antibodies used for staining were goat anti-mouse IgG conjugated with Alexa® Fluor 488 and goat anti-rabbit IgG conjugated with Alexa® Fluor 488 (Invitrogen, Eugene, OR, USA). Samples were then incubated with secondary antibodies for 1 h at room temperature in the dark and washed for 10 min five times with DPBS. For nuclei staining, cells were incubated with Hoechst 33,342 (Invitrogen) for 10 min at room temperature in the dark and washed with PBS twice quickly. All fluorescence images were obtained using the Operetta® High Content Screening (HCS) System (Perkin Elmer, Waltham, MA, USA).
Growth properties of AMC-H1 and AMC-H2
The AMC-H1 and AMC-H2 cell lines were seeded at 1 × 103 cells per well in 96-well plates and allowed to grow 1-7 d. Cell viability was assessed by the MTS assay (Promega, Madison, WI, USA) at the indicated times.
AMC-H1 and AMC-H2 primary cells were seeded in 6-well ultra-low attachment (ULA) plates (Corning Life Sciences, Amsterdam, The Netherlands) in DMEM/F12 (Gibco) supplemented with 1× B27 (Invitrogen), 20 ng/ml basic fibroblast growth factor (bFGF; Invitrogen), 20 ng/ml epidermal growth factor (EGF; Invitrogen), 25 μg/ml insulin (Sigma-Aldrich) as LCSC media. Cells were cultivated for 5-7 d without changing the LCSC medium.
Wound healing assay
To examine metastatic potential, a wound healing assay was performed. Cells were plated in a 6-well plate and grown to 80-90% confluence. After reaching the appropriate confluence, a scratch to generate a wound edge was made through the monolayer using a pipette tip. The migration of cells from the edge of the wound was monitored. Time-lapse images were captured using a microscope equipped with a Nikon 3 camera.
Western blot analysis
Molecular profiles were examined by western blot analysis. Cells were collected and centrifuged at 1000 rpm for 3 min. The cell pellets were collected and lysed using RIPA buffer (InTron, Jungwon-gu, South Korea) for 30 min at 4 °C with vigorous vortexing. The resultant lysate was centrifuged at 12,500 rpm for 20 min at 4 °C, and the supernatants were collected. The protein concentration was measured with the BCA assay (Promega). Following SDS-PAGE and transfer, membranes were blotted with antibodies to the epidermal growth factor receptor (EGFR, Cell Signaling, Danvers, MA, USA), β-catenin (Cell Signaling), N-cadherin (Cell Signaling), p53 (Santa Cruz Biotechnology, Dallas, TX, USA), PTEN (Cell Signaling), pERK (Cell Signaling), ERK (Cell Signaling), pAkt (Cell Signaling), and Akt (Cell Signaling). Following binding to the appropriate secondary antibodies, signals were detected using ECL reagents (GE, Fairfield, CT, USA).
Single nucleotide polymorphism (SNP) microarray
To examine the genomic aberrations in AMC-H1 and AMC-H2 cells, we performed SNP array analysis using Affymetrix CytoScanTM 750 K (Affymetrix, Santa Clara, CA, USA). DNA was isolated from AMC-H1 and AMC-H2 using the AxyPrep Multisource Genomic DNA Kit (Axygen, Union City, CA, USA). The SNP microarray was performed according to the standardized protocol provided by the manufacturer. The raw data were analyzed using the Chromosome Analysis Suite 2.0 software (Affymetrix) to detect copy number (CN) alterations and loss of heterozygosity (LOH).
Tumor spheroid and multicellular tumor spheroid (MCTS) formation and drug treatment
To generate tumor spheroids, cells suspended in HBM media were seeded at a density of 6 × 103 cells/well in 96-well round bottom ULA microplates (Corning Life Sciences). The plates were incubated for 3 d at 37 °C in a humidified atmosphere of 5% CO2. To produce MCTSs, four kinds of cells (primary HCC cells, LX2 cells, WI38 cells, and HUVECs) suspended in HBM media were seeded at a density of 6 × 103 cells/well in 96-well round bottom ULA microplates. The plates were incubated for 3 d at 37 °C in a humidified atmosphere of 5% CO2. After 3 d, 5-fluorouracil (5-FU; Sigma-Aldrich), cisplatin (Sigma-Aldrich), and sorafenib (Santa Cruz Biotechnology) were added and incubated for an additional 7 d. A solution of 0.5% dimethylsulfoxide (DMSO; Sigma-Aldrich) was used as a negative control.
Drug sensitivity in monolayer culture conditions
Primary HCC cells were seeded at a density of 2.5 × 103 cells/well in 384-well plates (Greiner Bio-One, Monroe, NC, USA). After a 16-h incubation, 5-FU, cisplatin, and sorafenib were added for 48 h. After 48 h, cells were fixed with 4% PFA for 10 min at room temperature and washed twice with DPBS. Hoechst 33,342 was used for nuclear staining. To capture enough cells (> 1000) for analysis, five image fields were collected from each well, starting at the center of the well. All of the image analysis was performed using the HCS system and Harmony software. Cell counts were calculated and normalized to the control (0.5% DMSO).
Cell death detection in spheroids
Spheroid cell death was detected using the cell-impermeant viability indicator ethidium homodimer-1 (EthD-1; Invitrogen). EthD-1 is a high-affinity nucleic acid stain that fluoresces weakly until binding to DNA and then emits red fluorescence (excitation/emission maxima ~ 528/617). Spheroids were incubated in 4 μM EthD-1 in HBM medium for 30 min at 37 °C, and images and the intensity of EthD-1 were obtained using the HCS system.
Real-time PCR
Total RNA was isolated from cells using TRIzol® (Invitrogen) according to the manufacturer’s instructions. The reaction mixture contained RT buffer (Bio Basic, Amherst, NY, USA), dNTP solution (Bio Basic), RNasin® inhibitor (Promega), oligo (dT)15 primer (Bioneer, Daejeon, Korea), total RNA, and M-MLV reverse transcriptase (Invitrogen). The reaction mixtures were incubated at 37 °C for 1 h, and the transcription reaction was terminated by heating the mixture to 95 °C for 5 min and then rapidly cooling it on ice. PCR reactions were performed in 96-well plates in a mixture composed of cDNA, SYBR Green master mix (Applied Biosystems, Waltham, MA, USA), primers, and DEPC using a StepOnePlus real-time PCR system (Applied Biosystems). The reaction conditions were 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. The threshold cycle (CT) was defined as the fractional cycle number at which the fluorescence passes the fixed threshold. CT values were calculated according to the mathematical model R = 2 − ΔΔCT, where ΔCT = CTtarget gene − CTGAPDH, and ΔΔCT = ΔCTtest − ΔCTcontrol. CT values were normalized to GAPDH expression. All real-time PCR was performed in triplicates, and the data are presented as the mean ± standard deviation (SD). All primers were designed and purchased from Bioneer.
Statistical analysis
All experiments were performed at least three times. The results are expressed as the mean ± SD. Statistical analyses were performed using the Student’s t-test.
Discussion
HCC is a highly heterogeneous disease in terms of its molecular profiles, and such heterogeneity has been a major obstacle to the study of tumorigenesis and identification of therapeutic targets. Therefore, the establishment of cell lines derived from HCC patients, especially from HBV-related HCC is required for the in vitro study of HCC, as HBV is the most common etiology of HCC worldwide. In this study, we established two new HBV-associated, patient-derived HCC cell lines (AMC-H1 and AMC-H2) and characterized the morphology; HBV DNA secretion; cellular properties, such as invasiveness; cytogenetic; and molecular features of these cell lines (Fig.
1).
AMC-H1 and AMC-H2 both expressed AFP and more importantly secreted HBV DNA stably. AFP expression levels were higher in AMC-H2 than in AMC-H1; yet, the AMC-H1 donor showed much higher serum AFP levels at the time of tumor resection than the AMC-H2 donor. We cannot explain this discrepancy, but we assume that both cell lines were altered with respect to their capacity to express AFP during the establishment of the cell lines (Fig.
1b-c). In terms of HBV DNA, we were unable to detect the whole genome of HBV in cell lysates; however, stable HBV DNA expression in these cells lines must occur due to HBV DNA incorporation, at least in part, into the chromosomes of both cell lines (Additional file
1: Table S1, Additional file
3: Table S2, Additional file
4: Table S3). Furthermore, we did not detect HBsAg in either AMC-H1 or AMC-H2 supernatants. One explanation for this could be that incorporation of only partial HBV DNA into the host genome precluded expression of this antigen. Another explanation is that HBsAg escape mutations [
12‐
14] occurred during the in vitro establishment period, although this possibility is quite low. Collectively, both AMC-H1 and AMC-H2 expressed AFP and HBV DNA stably after a long period of serial culture, but infectious intact HBV virion was not produced. Further virological study is needed to assess the mechanism of stable HBV DNA secretion from our cell lines and its usefulness in the study of HBV-related HCC studies.
We also examined the tumor-related genes and signaling pathways in these cell lines. EGFR [
15] and β-catenin [
16] are well-known factors involved in HCC tumorigenesis, but neither was overexpressed in the cell lines. In HCC, it has been reported that EGFR mutation is quite rare, but β-catenin mutation is more frequent. Thus, specific examination of β-catenin mutation in AMC-H1 and AMC-H2 will be required in future studies (Fig.
1d). The representative tumor suppressors, p53 and PTEN, were also examined in our new cell lines, and both tumor suppressors were expressed at a low level compared to expression in other HCC cell lines. Among receptor-mediated signaling pathways, the Ras/Raf/MEK/ERK pathway was not activated in either cell line, but the PI3K/Akt pathway was activated in the AMC-H2 cell line. These results suggest that other signaling pathways such as Wnt/β-catenin [
17,
18], VEGF [
19], FGF, or IGF [
3] may play a more important role in the proliferation of our cell lines, and as such, further study is warranted.
Chromosome instability is a hallmark of solid tumors and is a characteristic genetic alteration of HCC [
4]. It has been reported that HBV-related HCC have common chromosomal abnormalities, including gains of 1q, 6p, 8q and 17q and losses of 1p, 4q, 6q, 8p, 9p, 13q, 16q, and 17p [
5,
6]. As shown in Tables
2,
3, these common aberrations were also found in AMC-H1 and AMC-H2 and chromosomal aberrations observed in AMC-H1 and AMC-H2 were markedly divergent even at the same chromosome loci. In addition to common chromosomal abnormalities, analysis of an SNP array showed that AMC-H1 and AMC-H2 have numerous genomic alterations on most chromosomes. Furthermore, chromosomal alterations varied between the two cell lines; for instance, AMC-H2 mainly had chromosome losses while AMC-H1 did not. In general, chromosome gains frequently occur by duplication of genetic regions; however, chromosome gains by duplication have little oncogenic potential, as oncogenic drivers are typically the result of chromosome gains by focal amplification (CN state = 6-7). AMC-H1 and AMC-H2 have highly duplicated genetic loci but no focal amplification. Chromosome gains in AMC-H1 were mainly related to proliferation, while AMC-H2 showed amplification in regions containing survival genes. Chromosome amplification may play important roles in the survival of both cell lines by different signaling pathways. LOH is the common chromosomal alteration in HCC, and both cell lines have LOH on most of their chromosomes. Common LOHs observed in AMC-H1 and AMC-H2 are IGF2R and p53, which have clinical significance in HCC, as both are known to predict poor clinical outcomes in surgically resected primary HCC [
7,
20]. In addition to these common LOH, AMC-H1 and AMC-H2 have unique and divergent LOH, which may confer distinct properties to each of these cell lines.
Because of the molecular heterogeneity of HCC, proper in vitro systems are required to study HCC pathogenesis [
21]. Many research groups reported the establishment of HCC cell lines from HBV-infected patients. Although some of these groups observed HBV DNA integration into the host genome [
22‐
26], only a few HCC lines secrete HBV DNA or HBsAg stably [
27]. In this study, we established two new HCC cell lines from HBV-infected Korean HCC patients and characterized their properties in detail. The cell lines, designated as AMC-H1 and AMC-H2, were very different from one another in morphology, signaling pathways, and chromosomal aberrations but expressed AFP and albumin and more importantly secreted HBV DNA continuously in their supernatant.
In recent years, a paradigm shift from 2D to 3D cell culture techniques has occurred because culturing cells in 2D result in unnatural cell attachment, whereas culturing cells in 3D results in formation of natural cell-to-cell attachments. Accordingly, 3D cell cultures have been shown to have dramatic effects on cell polarity and differentiation as well as signaling cascades and gene expression profiles compared with that in monolayer culture. Moreover, cells grown in 3D are more valid for discovering new drugs to treat cancer, in part, because cells grown in 3D cultures form multilayers of cells, whereas cells in 2D form a monolayer of cells spread thin on a plastic surface. When testing a drug in 2D, the drug needs only to diffuse a short distance across the cell membrane to reach its target. In 3D, however, the situation is more realistic, and a drug needs to diffuse across multi layers of cells to reach its target.
Particularly, we examined MCTS models, which were necessary in order to establish an in vitro model that includes additional cell types to resemble the heterogeneity and complexity of in vivo tissue. We sought to develop this model to provide a physically relevant source for drug discovery and personalized medicine [
11,
28,
29] Therefore, we compared the drug sensitivities of sorafenib, cisplatin and 5-FU in monolayer cultures, tumor spheroids, and MCTSs from AMC-H1 and AMC-H2 cell lines and primary HCC cells from liver cancer patients. The size of spheroids were diverse depend on the cells, but average volume of spheroids were 300-400 μm. Although each primary HCC cell culture showed different drug sensitivities depending on the culture method, we concluded that simplifying the conditions for a monolayer-based chemosensitivity system is not be beneficial in terms of optimized treatment for patients with hepatocellular carcinoma therapy of HCC. The key of optimized therapy of HCC stems from the difference in results of chemosensitivity between homogeneous HCC spheroids and heterogeneous MCTS-HCC models. Because we found that MCTS-HCC models exhibited clear selective response to sorafenib, cisplatin, and 5-FU as opposed to the homogeneous HCC spheroids (Fig.
3d and Table
3), we suggest that MCTS-HCC models are the best methodology to screen for optimized therapy for patient with HCC. We, however, still cannot confidently assert that MCTS-HCC models represent identical in vivo tumor microenvironments for personalized medicine of HCC, because we do not yet have consistent evidence for clinical treatment. Therefore, we are currently securing abundant patient samples in conjunction with their history of clinical treatment with patient consent in order to inform future investigations.