A robust model of natural hepatitis C infection using hepatocyte-like cells derived from human induced pluripotent stem cells as a long-term host

The mononuclear cells were separated from bone marrow using IsoPrep (Robbins Scientific, Canada) density gradient centrifugation. Isolated cells were seeded at 2 × 10⁶ cells on T-75 cell culture flask in Minimum Essential Medium (MEM) α (Gibco Invitrogen, NY), 10 % FBS (HyClone, GE Healthcare Life Sciences, Singapore), 100 units/mL penicillin, 100 μg/mL streptomycin at 37 °C in 5 % CO₂. Lentiviral particles were produced by co-transfection of packaging psPAX2 packaging plasmid, pMD2.G vesicular stomatitis virus G envelope, and the polycistronic plasmid encoding human OCT4, SOX2, KLF4, c-MYC and dTomato [32] using X-tremeGENE HP DNA Transfection Reagent (Roche Diagnostics, USA). hMSCs were transduced with concentrated lentivirus (MOI = 0.5). Production of dTomato was observed 48 h post-transduction. The transduced cells were seeded on MEF in hES medium to promote reprogramming process and colony formation. Reprogrammed colonies were identified with StainAlive TRA-1-60 Antibody (Stemgent, MA). Positive colonies were manually picked and passaged until stable iPS cells established.

Human iPS cells were maintained in a feeder free condition on Geltrex-coated dish in E8 medium at 37 °C in 5 % CO₂. The E8 medium consisted of DMEM/F12 supplemented with 64 mg/L L-ascorbic acid-2-phosphate, 14 μg/L sodium selenite, 100 μg/L FGF2, 19.4 mg/L insulin, 543 mg/L NaHCO₃, 10.7 mg/L transferrin and 2 μg/L TGF-β1. The medium was changed daily. Human iPS cell colonies were split at a ratio of 1:8 every 5 - 7 d and passaged onto a new Geltrex-coated dish. The iPSC colonies after the tenth passage were characterized for pluripotent markers, e.g., alkaline phosphatase, OCT4, SOX2, NANOG, TRA-1-60, TRA-1-81, and SSEA4 using alkaline phosphatase ES characterization kit (MERCK Millipore, USA), fluorescent antibody staining and quantitative RT-PCR analysis.

Approximately 2 × 10⁶ iPS cells were subcutaneously injected into nude mice (BALB/cMlac-nu). The tumor-like tissue was collected within 8–12 weeks after injection. The tissue was fixed in 10 % formalin and underwent histological processes. The tissue sections were stained with hematoxylin and eosin and examined under a light microscope for the presence of the specialized cells derived from three germ layers. All experiments performed on laboratory animals were reviewed and approved by the Animal Care and Use Committee, Institute of Molecular Biosciences, Mahidol University.

The characterized iPS cells were taken for hepatic induction. They were seeded on Geltrex-coated 6 well plate in E8 medium until 80 % confluence and differentiated using a modified hepatic lineage development protocol [31]. Briefly, the cultured cells were maintained in endoderm differentiation medium DMEM/F12 : IMDM (1:1) (Gibco Invitrogen, NY), 100 ng/mL activin A (PeproTech, US), 10 ng/mL bFGF, 10 ng/mL BMP-4, 1 μM LY294002 and 3 μM CHIR99021 for 24 h. Cells were maintained in endoderm commitment medium DMEM/F12 : IMDM (1:1), 100 ng/mL activin A, 100 ng/mL bFGF, 10 ng/mL BMP-4 and 1 μM LY294002 for 24 h. The endoderm-like cells were further maintained in anterior definitive endoderm differentiation medium RPMI/B27 with 50 ng/mL activin A for 3 d with a daily medium replacement. The differentiated cells were maintained in hepatic lineage specification medium RPMI/B27, 20 ng/mL BMP-4 and 10 ng/mL FGF10 for 4 d with daily medium replacement. The differentiated cells were cultured in hepatocyte basal medium (HBM, Lonza, UK), 30 ng/mL oncostatin M and 50 ng/mL HGF for 10 d with medium replacement every 2 d. For functional hepatocyte induction, the cells were cultured in Williams' Media E (Invitrogen, MD, 1 % DMSO for 3 d.

HLCs and hMSCs were cultured on 35-mm dish for 3d. The cells were fixed in 5 % formaldehyde, 95 % ethanol for 1 min, rinsed for 1 min under running tap water, immersed in periodic acid solution for 5 min, rinsed thrice with dH₂O, immersed in Schiff’s reagent for 15 min, and rinsed with running tap water for 5–10 min. Samples were counterstained with Mayer’s hematoxylin for 1 min, rinsed with water, and assessed under light microscope.

The total RNA was extracted from iPS cells, HLCs, or HepaRG (HPRGC10, Life Technologies) by illustra RNAspin Mini RNA Isolation Kit (GE Healthcare, Asia Pacific) following the manufacturer instruction. The total RNA was immediately used for reverse transcription to construct cDNA using the ImProm-II reverse transcription system (Promega, WI) following the manufacturer instruction. The hepatocyte markers and cytochrome P450 markers included albumin, α-fetoprotein, cytokeratin18, G-6-PD, HNF-4α, tyrosine aminotransferase and major 7 CYPs isozymes. All gene specific primers (Additional file 1) were designed using vector NTI version 11.5 (Invitrogen, MD). They were amplified using KAPA SYBR^® FAST qPCR Kits (Kapa Biosystems). Each sample was measured in triplicate. PCR amplicons were confirmed using size and melting curve analysis. Expression levels were calculated using the ΔΔCt method and normalized to the endogenous GAPDH. ΔΔCt was transformed into fold change using the formula: fold change = 2^{- ΔΔCt}.

HLCs on 24-well plate were fixed with 4 % paraformaldehyde for 20 min at room temperature. The fixed cells were washed thrice with PBS, blocked and permeabilized with 0.2 % Triton X-100, 3 % BSA and 2 % normal goat serum for 1 h. Primary antibodies against pluripotent markers were TRA-1-60 (1:250, SC21705, Santa Cruz Biotechnology), SSEA4 (1:250, ab16287, Abcam), NANOG (1:250, SC33759, Santa Cruz Biotechnology), OCT4 (1:250, SC5279, Santa Cruz Biotechnology). Primary antibodies against HCV cell receptors were claudin-1 (1:100, ab15098), occluding (1:100, ab31721), CD81 (1:100, ab79559), SR-BI (1:200, ab106572) and hepatitis C core antigen (1:200, ab2740) ordered from Abcam. Primary antibodies against non-structural proteins of HCV were NS3 (1:200, SC69938), NS5A (1:200, SC52417) and NS5B (1:200, SC58146) ordered from Santa Cruz Biotechnology. The cells were incubated with the primary antibody for overnight h at 4 °C in moist chamber. After washing thrice, samples were incubated with the secondary antibody; Alexa fluor 488 conjugated goat anti-mouse IgG (1:1000, A11001) or Alexa fluor 488 conjugated goat anti-rabbit IgG (1: 1000, A11003s) or Alexa fluor 488 conjugated rabbit anti-goat IgG (1: 1000, A1178) (Life Technologies) for 1 h at 37 °C. Fibroblasts or untreated hMSCs served as negative controls. After washing thrice, the cells were counterstained with DAPI and mounted with anti-fade mounting medium in coverslip and examined under a fluorescent microscope and photographed.

The expression of CYP450 levels of hepatocyte-like cells responded to enzyme inducers was evaluated after incubate with cocktail of classical drugs [33]. HepaRG and hepatocyte-like cell derived from iPS cells were seeded on 6 well-plates for 3 d. These cells were maintained in Williams’ Media E, 10 % FBS before exposure to the inducers. Both of hepatocyte cells were incubated for 72 h with the cocktail agent: 20 μM rifampicin, 50 μM omeprazole, 1 mM phenobarbital and 88 μM ethanol. The treated cells were washed with 2–3 mL PBS, detached using 0.025 % trypsin-EDTA, and neutralized with 10 % FBS in DMEM/F12. The cell pellets from hepatocyte-like cell and HepaRG were stored at -80 °C until analysis for CYP450 gene expression.

CYP1A1, CYP2B6, CYP2C9 and CYP3A4 enzyme activities were evaluated directly in all cultured cells (HepaRG, HLCs and primary human hepatocytes) attaching to the collagen type IV-coated 6-cm dish at a density of 10⁶ cell. All cultured cells were treated with a cocktail of 20 μM rifampicin, 50 μM omeprazole, 1 mM phenobarbital and 88 μM ethanol. All treated cells were incubated for 72 h with daily medium change. CYP450 activities were assessed based on luciferase activity using the P450-glo 1A1, 2B6, 2C9 and 3A4 assays (V8751, V8321, V8791, V9001; Promega, WI). After 3-d incubation period, cells were incubated with Williams’ Media E supplemented with 100 μM Luciferin-CEE, Luciferin-2B6, Luciferin-H for 3–4 h or 3 μM Luciferin-IPA for 30–60 min. An aliquot (50 μL) of the medium was transferred to 96-well opaque white luminometer plate (Nunc, Denmark) and luciferin detection reagent was added to each well. After sitting at room temperature for 20 min, luminescence was measured with a SpectraMax M5 spectrofluorometer.

HCV infection requires several types of cellular receptors to permit viral entry [34]. Cellular receptors essential for HCV infection such as Claudin-1, Occludin, SR-BI, CD81 were detected using quantitative RT-PCR and immunofluorescent technique. Mature HLCs were seeded on 24-well plate and maintained until 80 % confluence in Williams’ Media E, 10 % FBS. Cells were fixed and stained with fluorescent-conjugated antibodies raised against Claudin-1, Occludin, SR-BI or CD81.

Cell culture based hepatitis C virus (HCVcc genotype 2a) was prepared from JFH-1 system [35]. The JFH-1 plasmid was propagated in E Coli and extracted using NucleoBond Xtra Midi plasmid (MN, Germany). The plasmids were purified and linearized by a restriction enzyme XbaI, and used as a template for JFH-1 RNA synthesis. The full length HCV RNA was synthesized by TranscriptAid T7 High Yield Transcription Kit (Thermo Fisher Scientific, MA) following the manufacturer’s instruction. HCVcc was produced by transient transfection of JFH-1 RNA into Huh7 cell using either lipofectamine 3000 (Invitrogen, MA) or electroporation (Gene Pulser Xcell Electroporation Systems, Bio-Rad, CA). HCV in Huh7 supernatant was passed through 0.45 μm syringe filter, concentrated by sucrose gradient ultracentrifugation and stored at −80 °C for future use.

HCV-positive sera were collected from patients with chronic HCV infection at Ramathibodi Hospital, Mahidol University. The collection of leftover blood specimen was approved by the Ethics Committee on Research Involving Human Subjects (2556/250). All subjects are adults and provided written informed consent. Serum from individual patient with HCV load > 10⁶ IU/mL was selected to infect HLCs. These sera carried HCV genotypes 1a, 1b, 3a, 3b, 6f and 6n that were prevalent in Thailand. For HCVser and HCVcc infection, HLCs and Huh7 cells were seeded on 6-well plate in Williams' Media E, 10 % FBS until 80 % confluence. Some host cells were pretreated with 1 μM or 5 μM α-tocopherol for 24 h prior to HCVser infection. For HCVser infection, 50 μL of infected serum in 2 mL of medium was added to host cells. For HCVcc infection, concentrated JFH-1 HCV load at 10⁷ IU/mL in 2 mL of medium was added to host cells. After 24 h incubation, the infected cells were washed 6–10 times with 0.1 % BSA in PBS and reconstituted with 2 mL complete growth medium.

For long-term HCV infection assay, HLCs and Huh7 were infected with HCVcc or HCVser. After 24 h incubation, the culture medium was renewed with 10 % FBS, DMEM/F12. The HCV-containing supernatant was collected ever 4–6 days and HCV RNA was detected using RT-PCR. During long-term cultivation process, HLCs and Huh7 cell were maintained in complete medium with 1 % concentrated lipids (Life Technologies, USA) and 1 μM (α-tocopherol). The HCV production was confirmed using the Abbott real-time HCV assay (Abbott Diagnostics, Illinois, USA).

HCV RNA in the supernatants was extracted with NucleoSpin RNA Virus isolation kit (MN, USA); while the intracellular RNA was extracted with illustra RNAspin Mini RNA Isolation Kit (GE Healthcare, Asia Pacific). The stand specific primers were designed using Vector NTI version 11.5 (Invitrogen, USA). The primers for positive strand are: 5′-CCCTGTGAGGAACTACTGTCTTCACGCA-3′ and 5′-CTCGCAAGCACCCTATCAGG-CAGTAC-3′. The primers for the negative strand are: 5′-GATGTACCCCATGAGGTCGG-3′ and 5′-GCGCGACAAG-GAAGACTTCG-3′ [36]. The viral RNA and total RNA from infected host cells were converted to cDNA with the ImProm-II reverse transcription system (Promega, WI). For HCV (-) stand, one microgram of isolated RNA was incubated with 12.5 μM HCV ‘reverse’ primers in a total volume of 5 μL at 70 °C for 5 min and chilled on ice-water immediately for at least 5 min. The reverse transcription mix (15 μL of 5× reaction buffer, 25 mM MgCl₂, 2 mM dNTP Mix, 40 U/μL RNasin ribonuclease inhibitor, and 200 U/μL Improm-II™ RT) was added to the RNA-primer mix to make a total volume of 20 μL. The mixture was incubated at 25 °C for 5 min, and 42 °C for further 1 h. The RT reaction was terminated by heating at 70 °C for 15 min followed by chilling on ice. The HCV cDNA was amplified by KAPA SYBR® FAST qPCR Kits (Kapa Biosystems, UK) and 2.5 μM HCV ‘forward’ primer at 95 °C for 30 s, 66 °C for 45 s and 70 °C for 90 s. Hepatitis C virus (+) stand was quantified by real-time RT-PCR using the m2000sp and m2000rt instruments and RealTime HCV kit (Abbott Diagnostics, Illinois, USA).

All results of experiments were performed in triplicated. Data were expressed as means ± SD. Data from quantitative real-time RT-PCR were evaluated for statistical significance using either students’ t test or ANOVA for difference followed by post-hoc tests (Tukey’s HSD) between the treatments and their respective controls. * and ** represented statistically different data with p < 0.05 and p < 0.01 respectively.

hMSC isolated from aspirated bone marrow exhibited spindle-shaped morphology at 90–95 % confluence (Fig. 1a). After infecting hMSCs with polycistronic OSKM-dTomato lentivirus particles for 48 h, the transduced hMSCs expressing OCT4, SOX2, KLF4 and C-MYC were determined by positive dTomato (Fig. 1b and c). The dTomato-positive cells were detached and seeded on mitomycin c-treated MEF feeder at 2 × 10⁵ cells/10-cm dish. Cell morphology of dTomato positive population transformed from spindle-shape to epithelial morphology around day 5 post-transduction (Fig. 1d). These cell populations divided and formed a small cluster of dTomato after 2-3 d (Fig. 1e). The transduced colonies expanded and transformed to a tightly packed colony with a high nucleocytoplasmic ratio (Fig. 1f). Upon reaching full reprogramming, dTomato expression was silenced and reprogramming colonies were negative for dTomato on day 13 as monitored under a fluorescence microscope (Fig. 1g). The transduced cells reached iPS as determined by their positive pluripotent surface marker TRA-1-60 using a live staining technique (Fig. 1h and i). The TRA-1-60 positive colonies were selected and picked into a new MEF-seeded plates for expansion and characterizations.

TRA-1-60-positive colonies were passaged and maintained on MEF in ES medium until reprogramming cells achieved a typical human ESC morphology. The cell morphology consisted of a sharp border, flat, dense colony, high nucleocytoplasmic ratio and prominent nucleoli (Fig. 1h). The karyotype analysis confirmed no chromosomal alteration (Fig. 1j). The establishment of human iPS cell was confirmed using alkaline phosphatase (AP) activity, immunofluorescent staining of pluripotent markers and the expression of pluripotent genes (Fig. 1k). The reprogramming cells were maintained in E8 medium with feeder-free condition on Geltrex-coated dishes (Invitrogen, USA) prior to the analysis (Fig. 1l). The colony of reprogramming cells highly expressed essential pluripotent marker proteins, OCT4 and SOX2, required for maintaining pluripotent status. The expressions of OCT4 and SOX2 targeted proteins such as the self-renewal transcription factor NANOG, transmembrane proteins TRA-1-60 and TRA-1-81, and a cell surface glycoprotein SSEA4 were also expressed in the reprogramming cells. Human fibroblasts were used as negative controls for non-iPS cells (Fig. 2a). Exogenous OCT4 and SOX2 produced from the lentivirus were active at the early stage of cell reprogramming. To confirm the maintenance of endogenous OCT4 and SOX2 expression, specific OCT4 and SOX2 primers were designed to detect only the mRNA from endogenous source. The RT-PCR analysis confirmed the transduced cells carried endogenous OCT4 and SOX2 (Fig. 2b). Other pluripotent markers such as REX1, NANOG, GDF3, DNM3TB and UFT1 were expressed in the transduced cells and compared with those in human ES cell (BG01V) and non-transduced hMSCs (Fig. 2b). GAPDH and NTC served as positive and negative controls for PCR products.

To determine whether these newly established iPS cells could differentiate into all three germ layers, the iPS cells were evaluated for ex vivo and in vivo differentiation. For spontaneous ex vivo differentiation, iPS cells were maintained in suspension culture with embryoid bodies formation within a week. After 2 weeks, embryoid bodies were seeded on gelatin-coated culture plates to allow cell adhesion and migration. The adherent cells displayed various morphologies (Fig. 3a). Immunofluorescent staining revealed differentiated cell positive for βIII-tubulin (ectoderm marker), vimentin (mesoderm marker) and α-fetoprotein (AFP, endoderm marker) (Fig. 3a). These indicated the differentiation to several lineages. For teratoma formation, one million iPS cells were resuspended in DMEM 10 % FBS and injected subcutaneously into nude mice. Four weeks after injection, teratoma was formed at the injection site. Histological examinations revealed that the teratoma contained several tissues (Fig. 3b), e.g., neural rosette (ectoderm) (Fig. 3b, left panel), mesoderm derived smooth muscle, adipose tissue, cartilage (mesoderm) (Fig. 3b, middle) and endoderm-derived gut-like epithelial (endoderm) (Fig. 3b, right). Taken together, iPS cells derived from hMSCs not only support the spontaneous ex vivo differentiation but they can form teratoma in vivo.

Human iPS cells were cultured in feeder free condition until reaching 70 % confluence before hepatic induction (Fig. 4a). During hepatic differentiation, iPS cells were monitored daily for morphological changes. Initially, iPS cells were differentiated into definitive endoderm (DE) cells (Fig. 4b) that shared the same precursor with all endodermal organs (pancreas, liver, lung, thyroid and gut) [37]. The differentiated cell transformed into round-shaped morphology (Fig. 4b) with positive staining for SOX17, an endoderm differentiation marker (Fig. 4c). The second stage, definitive endoderm cells were developed to anterior definitive endoderm (ADE) with cell morphology underwent epithelial to mesenchymal transition (EMT) [38]. Cells displayed the mesenchymal-like morphology (Fig. 4d). At the third stage of differentiation, ADE cells became hepatic progenitors (Fig. 4e). The differentiated cell at this stage was confirmed with the expression of alpha-fetoprotein (AFP) (Fig. 4f), a characteristic of hepatoblast [31]. At maturation, hepatoblasts developed into functional hepatocyte-like cells (HLCs). HLCs shared both fetal and adult hepatocyte markers, e.g., albumin and minimal level of alpha-fetoprotein (Fig. 4g). The expression of HNF4-α, found only at the end of maturation, revealed that the cells fully developed into mature hepatocytes (Fig. 4h). After 4-week differentiation, human iPS cell developed into a near homogenous population with more than 95 % pure hepatocyte-like cell. HLCs exhibited typical hepatocyte morphology, e.g., a polygonal shape, a dark cytoplasm with large nuclei containing nucleoli and bi-nucleated (Fig. 4i). Mature HLCs maintained in Williams' Media E exhibited homogeneous population (Fig. 4j), hepatic triad-like structure (Fig. 4k) and some populations produced bile pigment (Fig. 4l). HLCs exhibited polygonal morphology during passages (Fig. 4m). HLCs showed the capability to synthesize and store glycogen greater than did hMSCs as detected by Periodic acid Schiff staining (Fig. 4n and o). The HLCs were also positive for gene expressions of α-fetoprotein, albumin and HNF4-α using quantitative RT-PCR (Fig. 4p). For major metabolic functions in hepatocyte involving carbohydrate, lipid and amino acid metabolism, HLCs expressed both glucose-6-phosphate dehydrogenase (G-6-PD) and tyrosine aminotransferase (TAT). The expression levels of basic hepatocyte markers (i.e., ALB, AFP, HNF4-α, and TAT) were increased between 10–20 folds over those from undifferentiated iPS cells.

To achieve fully functional hepatocyte, mature HLCs were maintained in 10 % FBS Williams’ Media E for 3 d and the cells were screened for CYP expression. HLCs expressed a number of CYP isozymes as illustrated by immunostaining and quantitative real-time PCR. The basal mRNA levels of CYP2B6, CYP2D6, and CYP2C9 in HLCs were 22-, 25-, and 20-fold that of HepaRG cells respectively; while CYP3A4, CYP2C19, CYP1A2 were around 10 folds. CYP2E1 in HLCs was up-regulated to 40-fold that of the undifferentiated iPS cells (Fig. 5a). Other isozymes achieved comparable expression levels to those of HepaRG and the primary human hepatocytes. To evaluate the inducibility of these CYP isozymes, the HLCs were incubated with the cocktail of chemical inducers (20 μM rifampicin, 50 μM omeprazole, 1 mM phenobarbital and 88 μM ethanol). The expressions of CYP2B6, CYP2C9, CYP2C19 and CYP3A4 in HLCs were increased to 2, 2, 4, 3.5 folds respectively after the induction with the cocktail. In comparison, the expressions of CYP2B6, CYP2C9 and CYP3A4 in HepaRG were increased to 2, 2, 3.5 folds and comparable to those in HLCs (Fig. 5b). In particular, some CYP isozymes, e.g. CYP2B6, CYP2C9, CYP2C19 and CYP2E1 were more inducible in HLCs than in HepaRG cells. The expression levels of CYPs in primary human hepatocytes were dramatically decreased post-induction by 72 h. In addition to CYP expressions, HLCs contained high levels of organic anion transporting polypeptide 2 (OATP2). The elevated OATP2 expression could be extensively up regulated to 3 folds. For protein determination, more than 50 % of HLC populations were positive for intracellular CYP1A1, CYP3A4, CYP2B6 and CYP2E1 proteins (Fig. 5c–g). The immunofluorescent straining demonstrated the increasing expression of major CYP proteins (Fig. 5c–g), e.g., CYP3A4, CYP2B6, CYP1A1, CYP2C9 and CYP2E1. HLCs expressed multidrug resistance-associated protein 2 (MRP2) that represented the canalicular (apical) membrane of the hepatocyte as a biliary transport (Fig. 5h). The hMSC served as a control of non-hepatic cell (Fig. 5i). The increasing of CYP450 activities was investigated in hepatocytes (Fig. 5j). Mature HLCs population could proliferate, but still maintained hepatic morphology without losing basal levels of hepatic related gene expression.

A potential application for HLCs is for the study of pathogen-host cell interactions. The HLCs derived from iPS cells were evaluated for the capability to host HCV replication. The expression of HCV receptors is essential for virus hepatitis C viral entry. Mature HLCs expressed all major receptors for HCV entry (Fig. 6a–c), e.g., specific tight junctions: claudin-1, occludin; cell surface receptor: TAPA-1 (CD81); scavenger receptor B1 (SR-BI), ApoB and ApoE. The expression of specific HCV receptors was increased together with the maturation stage of HLCs. On the cell surface of cultured HLCs, the protein products of these receptors were also detected (Fig. 6d–i), e.g., claudin-1, occludin, SR-BI, CD81, LDL-R and ApoE. Human fibroblast served as a negative control (Fig. 6j).

To determine if HLCs could replace the Huh7 cells to host HCV, HLCs were transfected with JFH-1 RNA using liposomal transfection [14]. Seven days after transfection, we found cytopathic effect with the aberration of cytoplasm in HLCs, but not with mock transfection (Fig. 7a and b). These cells had lower growth rate, aggregated and formed clusters of syncytial cells (Fig 7c). These findings suggested that HCV might re-infect the HLCs in high titer. We could not observe the CPE nor the aberration of cytoplasm in mock transfection or JFH-1 transfected Huh7 cell on day 7 and day 14 (Fig. 7d‐f). Conditioned medium from HLCs and Huh7 was harvested every 3 d and filtered through a 0.45 μM syringe filter. The suitable multiplicity of infection (MOI) was optimized using 10-fold HCV dilution. Cell viability was determined with MTT assay (Fig. 7g and h). The HCV viral load was determined and compared between different cell lines (Fig. 7i). The kinetics of viral production varied depending on the hosting cells. The maximal HCV production was achieved between 12-15 d post-transfection. HCV viral load from Huh7 was about 10⁵ IU/mL but HLCs could provide up to 10⁷ IU/mL (Fig. 7i), comparable to those offered (10⁸ IU/mL) by the transfected Huh7.5 [17]. The detection of HCV in HLCs supernatant was maintained beyond 1 month. In transfected Huh7 cells, HCV viral load was decreased to 10⁴ IU/mL within 2 weeks and completely vanished in two months (Fig. 7i and j).

The naïve HLCs and naïve Huh7 cells were infected with either HCVcc (JFH-1 genotype 2a) or with HCVser of various genotypes including 1a, 1b, 3a, 3b, 6f and 6n. HCVcc could be obtained from the earlier transfection of HLCs or Huh7 with JFH-1 RNA. HCVser was collected from chronic hepatitis C patient. The cytopathic effect (CPE) in the infected cells was observed on both day 7 and 14 after infection (Fig. 8a–c). The cytoplasmic localization of non-structural proteins: NS3A, NS5A, NS5B and structural protein: HCV core antigen was clearly detected. Therefore, HLCs allowed both HCV entry and the transcription of viral RNA (Fig 8d–h). To ensure that the HLCs support entire HCV life cycle, the total RNA prepared from infected HLCs and Huh7 cell was investigated for the presence of HCV negative-strand RNA using specific RT-PCR primers. The viral RNA was maintained in infected cells up to 4 weeks in HLCs 2 weeks in Huh7 cells. The presence of positive-strand of viral RNA was measured in the supernatant from the infected cells using RT-PCR analysis (Fig. 8i and j). The HCVcc viral loads were 10⁷ IU/mL and 10⁴ IU/mL from the supernatants of HLCs and Huh7, respectively. The levels of intracellular HCV negative–strand RNA and positive-strand RNA were determined by reverse transcription PCR every week after transfection until 4 weeks. The negative-strand RNA was merely detectable on day 7 and increased on day 14 (Fig. 8i and j). For HCVser, HLCs allowed viral propagation as determined by the increasing HCV particles in culture medium (Fig. 8k) and the detection intracellular HCV RNA (Fig. 8l). Huh7 cell, in contrast, could barely hosted HCV derived from patient serum. The HCVser viral loads were 10⁴ IU/mL from the supernatants of HLCs. The kinetics of viral replication post-infection with either HCVcc (JFH-1) or HCVser (genotype 1a) in HLCs were determined using HCV titer and intracellular HCV RNA. The kinetics of HCV production suggested that HLCs supported complete HCV replication of both HCVcc and HCVser (Fig. 8m and n). However, HLCs did allow HCVser replication beyond 4 weeks (Fig. 8n). The immunostaining and RT-PCR analysis confirmed that the entire replication cycle and formation of HCV progeny took place in HLC and the supernatants from the infected cultures could efficiently infect naïve HLCs and Huh7 cells.

The effect of α-tocopherol to HCVser infection was investigated using HLCs and Huh7 cell line. After the pretreatment with 1 μM or 5 μM α-tocopherol for 24 h and infection with either HCVcc (JFH-1) or HCVser genotype 1a, the infected HLCs and Huh7 cells were further maintained for 14 days. The HCVser viral load was raised to 10⁶ IU/mL in HLCs but the HCVcc viral load was not affected by α-tocopherol. In Huh7 cell, α-tocopherol did not improve the HCV viral load in HCVser infection and HCVcc (JFH-1) infection (Fig. 9a and b).

To screen for the sensitivity of infected HLCs to anti-HCV agents, naïve HLCs were infected with 10⁶ IU/mL of either HCVcc (from conditioned medium of HLCs) or HCVser. Different anti-HCV agents or their combinations, e.g., ribavirin (20 μM), IFN-α (10 IU/mL), and sofosbuvir (PSI-7977, 200 nM) were incubated with the 24-h infected cells for 7 d. The HCV RNA expression in HCVcc and HCVser served as drug response. The intracellular HCV RNA was determined by real-time qPCR at 4, 8, 12, and 24 d after the infection initiation. Ribavirin, IFN-α and sofosbuvir were not cytotoxic up to 100 μM, 10 IU/mL and 100 μM respectively. Either ribavirin or INF-α decreased HCV RNA to 50 % in Huh7 and HLCs. The combination of INF-α and ribavirin decreased viral RNA to 20 %. Sofosbuvir decreased viral RNA below 20 %. (Fig. 9c and d). The IC₅₀ of INF-α toward HCV RNA level in HLCs and Huh7 cells were 5.3 and 2.4 IU/mL respectively (p < 0.01). The IC₅₀ of ribavirin in HLCs and Huh7 were 6.9 and 3.2 μg/mL respectively (p < 0.01). The IC₅₀ of sofosbuvir in HLCs and Huh7 were 96 and 92 nM respectively (p < 0.01) (data not shown).

Relative real-time qPCR on days 3 and 14 revealed that the expression of inflammatory markers, e.g. TNF-α, IL-28B and IL-29 were inducible during the infection. Particularly, IL-28B expression in response to HCV infection was observed earlier only in primary human hepatocyte, but also in our HLCs model (Fig. 9e). Long-term HCVcc production was evaluated by infecting naïve HLCs with viral load 10⁷ IU/mL for 72 h. The cell culture medium that had been replaced every week still carried HCV particles up to 6 months. The kinetics of HCV represented the new viral particles releasing from infected cell to the medium (Fig. 9f). At 4-week post-infection, the HCVcc viral load varied from 10⁴ to 10⁵ IU/mL and increased at 2-3 months later to 10⁵ IU/mL. The HCV viral production was still maintained at 10⁴-10⁵ IU/mL up to 5 months. Until 6 months, most of the HLCs died in accordance with the viral CPE.