Human serum leads to differentiation of human hepatoma cells, restoration of very-low-density lipoprotein secretion, and a 1000-fold increase in HCV Japanese fulminant hepatitis type 1 titers

Authors

  • Rineke H.G. Steenbergen,

    1. Department ofMedical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada
    2. Li Ka Shing Institute of Virology, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author
  • Michael A. Joyce,

    1. Department ofMedical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada
    2. Li Ka Shing Institute of Virology, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author
  • Bradley S. Thomas,

    1. Department ofMedical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada
    2. Li Ka Shing Institute of Virology, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author
  • Daniel Jones,

    1. Division of BioMedical Sciences, Memorial University of Newfoundland, St. John's, Newfoundland, Canada
    Search for more papers by this author
  • John Law,

    1. Department ofMedical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada
    2. Li Ka Shing Institute of Virology, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author
  • Rodney Russell,

    1. Division of BioMedical Sciences, Memorial University of Newfoundland, St. John's, Newfoundland, Canada
    Search for more papers by this author
  • Michael Houghton,

    1. Department ofMedical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada
    2. Li Ka Shing Institute of Virology, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author
  • D. Lorne Tyrrell

    Corresponding author
    1. Department ofMedical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada
    2. Li Ka Shing Institute of Virology, University of Alberta, Edmonton, Alberta, Canada
    • Address reprint requests to: Lorne Tyrrell, M.D., Ph.D., Department of Medical Microbiology and Immunology, Li Ka Shing Institute of Virology, University of Alberta, 6-010 Katz Center for Pharmacy and Health Research, Edmonton, Alberta, T6G 2S2, Canada. E-mail: lorne.tyrrell@ualberta.ca; fax: +-1-780-492-5304.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to declare.

  • The authors thank Dr. Rice for the kind gift of Huh7.5 cells and Dr. Wakita for the JFH-1 construct. This research was funded by the Li Ka Shing Foundation, Canadian Institutes for Health Research, and Alberta Advanced Education and Technology (AAET).

Abstract

In this study, we differentiated the human hepatoma cell line Huh7.5 by supplementing tissue culture media with human serum (HS) and examined the production of hepatitis C virus (HCV) by these cells. We compared the standard tissue culture protocol, using media supplemented with 10% fetal bovine serum (FBS), to media supplemented with 2% HS. Cells cultured in HS undergo rapid growth arrest, have a hepatocyte-like morphology, and increase the expression of hepatocyte differentiation markers. In addition, expression of cell adhesion proteins claudin-1, occludin, and e-cadherin are also increased. The lipid droplet content of these cells is highly increased, as are key lipid metabolism regulators liver X receptor alpha, peroxisome proliferator-activated receptor (PPAR)-α, and PPAR-γ. Very-low-density lipoprotein secretion, which is absent in FBS-grown cells, is restored in Huh7.5 cells that are cultured in HS. All these factors have been implicated in the life cycle of HCV. We show that viral production of Japanese fulminant hepatitis type 1 increases 1,000-fold when cells are grown in HS, compared to standard FBS culture conditions. The virus produced under these conditions is associated with apolipoprotein B, has a lower density, higher specific infectivity, and has a longer half-life than virus produced in media supplemented with FBS. Conclusion: We describe a convenient, cost-effective method to produce hepatocyte-like cells, which produce large amounts of virus that more closely resemble HCV present in serum of infected patients. (Hepatology 2013; 58:1907–1917)

Abbreviations
ABS

adult bovine serum

ALB

albumin

α1AT

alpha-1-antitrypsin

ApoB

apolipoprotein B

DMEM

Dulbecco's modified Eagle's medium

DMSO

dimethyl sulfoxide

ELISA

enzyme-linked immunosorbent assay

FBS

fetal bovine serum

HCV

hepatitis C virus

HCVcc

HCV in cell culture

HS

human serum

JFH-1

Japanese fulminant hepatitis type 1

HCV

hepatitis C virus

IP

immunoprecipitation

LDL-R

low-density lipoprotein receptor

LXR-α

liver X receptor alpha

mRNA

messenger RNA

NPC1L1

Niemann-Pick C1-like 1

PPAR

peroxisome proliferator-activated receptor

qRT-PCR

quantitative reverse-transcriptase polymerase chain reaction

SR-B1

scavenger receptor class BI

TCID50

50% tissue culture infectious dose

TG

triglyceride

VLDL

very-low-density lipoprotein

Hepatitis C virus (HCV) is an enveloped, positive-strand RNA virus of the family of Flaviviridae that causes acute and chronic hepatitis. HCV can cause cirrhosis, hepatocellular carcinoma, and steatosis in infected individuals.

Replicon systems, both subgenomic and full length, and the Japanese fulminant hepatitis type 1 (JFH-1) tissue culture infection models have yielded important insight into the HCV life cycle. Most of these models make use of HuH-7 or HuH-7-derived cells, such as Huh7.5. HuH-7 or HuH-7-derived cells have many advantages for the in vitro study of HCV: they are readily available and rapidly dividing, and therefore enable large-scale experiments. However, these systems do not necessarily accurately represent events that occur during a natural HCV infection in vivo, because hepatocytes are normally nondividing and fully differentiated. To circumvent this, dimethyl sulfoxide (DMSO) has been used to differentiate HuH-7 cells,[1] which resulted in increased expression of hepatocyte-specific genes. These differentiated, growth arrested cells can be infected using HCV JFH-1 and produce viral titers that are comparable to those in dividing cells.[1]

Freshly isolated primary human hepatocytes are a more representative in vitro model to study HCV infectivity. However, viral titers produced in these cells are low, and experiments longer than a few days require coculture with other cell types.[2, 3] Additionally, primary hepatocytes exhibit large interdonor variability, are often cost prohibitive, and have limited availability. Thus, they are generally not suitable for large-scale experiments.

We have previously shown that infection in chimeric mice is not reliably achieved until the humanization of the liver is nearly complete.[4] We postulated that infection with HCV was not only dependent on the presence of human hepatocytes, but also on human factors in serum of mice that have to reach critical levels to support HCV infection. Indeed, we found a good correlation between successful infection and the humanization of lipoprotein profiles in mouse serum.[4] In this study, we extended this observation to Huh7.5 cells in culture and investigated whether the presence of HS in tissue culture is advantageous to infection and viral production. To this end, we compared the “standard” tissue culture protocol, using media containing 10% fetal bovine serum (FBS), to the use of 2% human serum (HS). Surprisingly, cells cultured in HS differentiated into a cell type with increased hepatocyte functionality: they underwent growth arrest, adopted a primary hepatocyte-like morphology, and we showed increased expression of hepatocyte differentiation markers, tight junction proteins, and key regulators of lipid metabolism. Importantly, a hepatocyte-specific function, very-low-density lipoprotrein (VLDL) secretion, which is nearly absent in cells cultured in FBS media, is restored in HS-containing media. The benefits of growing these cells in HS go beyond differentiation alone: viral replication increases over 1,000-fold when cells are grown in HS, compared to standard FBS culture conditions. Additionally, virus produced under these conditions more closely resembles virus isolated from patient serum, with respect to infectivity, viral density and apolipoprotein B (ApoB) association, and has a much longer half-life.

We present an easy, cost-effective method to produce large amounts of hepatocyte-like cells, which produce large amounts of virus that more closely resembles HCV present in serum of infected patients.

Materials and Methods

Standard Cell Culture Conditions for Proliferation

Huh7.5 cells were a kind gift of Dr. C. Rice and were maintained according to the protocols provided. In short, cells were maintained in Dulbecco's modified Eagle's medium (DMEM; D5796; Sigma-Aldrich, St. Louis, MO), 10% FBS (F1051; lot nos.: 11M369, 080M8403, and 11D025; Sigma-Aldrich), penicillin, and streptomycin and discarded after 25-35 passages.

Maintenance of Cells in HS

Because the use of HS (34005-100; pooled human AB serum, lot nos.: 1274112, 1189296, and 1127343; Invitrogen, Carlsbad, CA) results in growth arrest, cell cultures were normally maintained in FBS-containing media, as described above. At the time of transfer to HS, cells were trypsinized, trypsin was inactivated with DMEM, and cells were centrifuged at 300×g. Cell pellets were then resuspended in DMEM/2% HS/penicillin/streptomycin and plated at a density of 30%-50%. At confluency, cells were trypsinized again, plated at a density of 50%, and left to form confluent layers of undividing cells. Cells can be subcultured for approximately 7-10 days; after that, cells appear to loose their ability to reattach to untreated cell culture plastic.

Production of Viral Stocks

JFH-1 was electroporated into FBS-cultured cells, as described previously,[5] and each cell suspension was split in two and maintained in either FBS- or HS-containing media. Viral production (RNA/mL and 50% tissue culture infectious dose [TCID50]/mL) was further monitored for up to 65 days. Four days after electroporation, culture supernatants were collected and these viral stocks were used for infection experiments described below. Virus produced by cells maintained in FBS and HS media is referred to as “JFH-FBS” and “JFH-HS,” respectively.

Infection of Cells With HCV JFH-1

Cells were replated at 30% density and infected 2 days later with either JFH-FBS or JFH-HS (multiplicity of infection: 1 RNA per 5 cells). After 4 hours of infection, cells were washed to remove remaining virus and placed in either DMEM/10% FBS/penicillin/streptomycin or DMEM/2% HS/penicillin/streptomycin for the remainder of the experiment.

Determination of TCID50 Value and Viral RNA in Cell Culture Supernatants

TCID50 value was determined as described previously.[5] Viral RNA from tissue culture supernatant samples was extracted using the Roche Pure Viral Nucleic Acid Kit (Roche Diagnostics, Indianapolis, IN), according to the manufacturer's instructions. RNA was quantitated as described previously.[4]

Quantitative Reverse-Transcriptase Polymerase Chain Reaction

RNA was isolated using Trizol (Invitrogen), according to the manufacturer's instructions. Complementary DNA was produced from RNA using the QuantiTect Reverse Transcription kit (Qiagen, Hilden, Germany). Gene-specific primer-probe sets were designed by Applied Biosystems (Foster City, CA). We used an Applied Biosystems 7900HT Fast Real-Time polymerase chain reaction (PCR) system for quantitation of gene products. Gene expression was calculated, relative to hypoxanthine phosphoribosyltransferase, according to Pfaffl[6] and depicted as fold increase compared to FBS.

Quantitation of Albumin Secretion

Huh7.5 cells were washed extensively with OptiMEM (Gibco, Grand Island, NY) to remove albumin (ALB) present in serum. The last wash was collected to determine background levels of ALB. Cells were then kept in OptiMEM at 37°C for 6 hours, and samples were taken every 2 hours. The amount of secreted ALB was determined using quantitative enzyme-linked immunosorbent assay (ELISA), as described previously.[4] ALB secretion (calculated as ng albumin/hour/10 × 106 cells) was normalized to FBS between experiments and expressed as fold-increase compared to FBS.

Visualization and Quantitation of Lipid Droplets by Immunofluorescence

Cells were grown on poly-L-lysine-coated coverslips and cultured in either FBS or HS. Lipid droplets were stained with Bodipy 493/503 (Invitrogen), according to the supplier's instructions. Quantity of neutral lipid staining was visualized using a conventional fluorescence microscope (Zeiss Axiovert200; Carl Zeiss, Göttingen, Germany) and quantitated using ImageJ software (National Institutes of Health, Bethesda, MD). Images were taken using identical microscope and exposure settings. Data were collected in three independent experiments, with four to eight microscopic fields per condition.

Fast Protein Liquid Chromatography Analysis of Secreted Lipoproteins

Lipoprotein analysis was performed as described previously,[7] using size-exclusion chromatography (large particles elute first) combined with in-line triglyceride (TG) and cholesterol measurements.

Sucrose Gradients/Density Centrifugation

Sucrose density-gradient ultracentrifugation was performed as previously described.[8] Fractions of 0.5 mL each were collected from the top of the gradient, and the RNA titer in each fraction was determined by quantitative reverse-transcriptase PCR (qRT-PCR).

Immunoprecipitation of ApoB-Containing Particles

Immunoprecipitation (IP) experiments were performed as previously described.[4]

Determination of Half-Life of Viral Particles

Freshly collected tissue culture supernatants from infected cells were filtered through a 22-μm filter and placed in clean tissue culture plates (without cells) and kept at 37°C or 4°C. Samples were taken at the start of the incubations and then at 4- and 12-hour intervals. Viral RNA was extracted and quantitated as described above.

Statistical Analysis

For calculation of significance, all experiments consisted of a minimum of three independent replicates. Statistical significance was calculated using the Student t test using the Prism statistics package. P values of 0.05 or smaller were considered significant.

Results

HS Supplementation Leads to Contact Inhibition of Huh7.5 Cells, Changes in Morphology, and Differentiation

We noticed that cells grown in media supplemented with HS, compared to FBS, had reduced growth rates, and show contact inhibition after approximately 7 days. From this point on, cells could be kept in confluent monolayers without further subculturing. HS cells did not pile up or detach from the culture plate. Cells in HS media could be maintained in monolayers for at least 2 months with regular media changes. Determination of cell numbers over a 3-week period confirmed that cell numbers did not increase after approximately 7 days in HS media (Fig. 1A). Morphology of cells cultured in HS-containing media changed dramatically during the first 3 weeks (Fig. 1B-D). After approximately 21 days, these cells had many morphological features of cultured primary hepatocytes. They formed tightly packed monolayers that were strongly attached to the tissue culture substrate, with a pavement-like organization. Similar to primary hepatocytes in culture (Fig. 1D), HS cells were mono- or binucleated and had a granular appearance (Fig. 1C). The size of the cells also increased. These changes were most obvious after more than 21 days of culturing in HS media.

Figure 1.

HS supplementation leads to contact inhibition and changes in morphology. (A) Cells were cultured in FBS- or HS-containing media and cell numbers were monitored over a period of up to 30 days. Cell numbers in FBS increased exponentially, whereas cells in HS-supplemented media did not increase after approximately 7 days. (B-D) Morphology of Huh7.5 cells cultured in FBS (B), cells cultured in HS for 21 days (C), and cultured primary hepatocytes (PH) plated onto collagen plates (D).

To further investigate whether cells that were cultured in HS-supplemented media underwent differentiation to a hepatocyte-like cell, we first examined transcript levels of hepatocyte differentiation markers alpha-1-antitrypsin (α1AT), ALB, and low-density lipoprotein receptor (LDL-R). No significant changes were observed after culturing in HS for 7 days. However, after 21 days, messenger RNA (mRNA) levels of both α1AT (Fig. 2A) and ALB (Fig. 2C) were significantly higher than in FBS-cultured cells and were comparable to those in cultured human primary hepatocytes. We did not find an increase in LDL-R mRNA as a result of culturing in HS (Fig. 2B). We used quantitative ALB ELISA to confirm that the increase in ALB mRNA resulted in increased ALB secretion (Fig. 2D). In line with mRNA levels, after 7 days in HS, no significant changes in ALB secretion were observed; however, after 21 days in HS, ALB secretion had increased approximately 6-fold.

Figure 2.

Transcript levels of hepatocyte differentiation markers. Cells were cultured in FBS-containing media or in HS-containing media for 7 or 21 days. (A-C) mRNA levels of hepatocyte differentiation markers α1AT (A), LDL-R (B), and ALB (C) were measured. For comparison, transcript levels of these markers in cultured primary hepatocytes were also included (prim. hep). Fold change compared to FBS was calculated. (D) ALB secretion rates (ng ALB/hour/10 million cells) were determined by quantitative ALB ELISA and expressed as fold increase compared to FBS. n ≥ 3; *P < 0.05; ***P < 0.0005.

HS Supplementation Leads to Increased Expression of Cell–Cell Contact Components

The presence of tight and adherens junctions are well-recognized features of hepatocytes in vivo and linked to increased liver-specific functionality in vitro[9]; loss of cell-junction components is commonly associated with metastatic cell types.[10]

Cells that were grown in HS-supplemented media for 14 days or more became very strongly attached to the plate and to each other. They were difficult to release by trypsinization. Cells that were eventually released remained organized in large clumps, indicating strong cell–cell contacts. We determined the mRNA levels of the two main tight junction components, claudin-1 and occludin, and the chief component of cell adherens junctions, e-cadherin. In HS-supplemented media, Huh7.5 cells had a 4-fold increase of claudin-1 and occludin mRNA (Fig. 3A,B), which was similar to the levels of expression noted in cultured primary hepatocytes. We also found that e-cadherin is up-regulated approximately 6-fold in cells cultured in HS media (Fig. 3C).

Figure 3.

Transcript levels of cell–cell contact proteins and known entry receptors of HCV. Cells were cultured in HS for 21 days, and transcript levels of claudin1 (A), occludin (B), and E-cadherin (C) were determined and compared to levels in FBS-cultured cells. All cell–cell contact components increased significantly. We also measured transcript levels of known HCV entry receptors CD81 (D), SR-B1 (E), and NPC1L1 (F); no significant changes were measured. Fold change compared to FBS was calculated. n ≥ 3 ; **P < 0.005; ****P < 0.0001. n.s., not significant; significance was calculated compared to FBS value.

Other Entry Receptors

LDL-R, claudin-1, and occludin have also been recognized as factors involved in HCV entry. To investigate alterations in some of the other factors involved in entry of HCV, we also determined mRNA levels of CD81, scavenger receptor class BI (SR-B1), and Niemann-Pick C1-like 1 (NPC1L1). No changes were observed in mRNA levels of any of these entry factors as a result of culturing in HS-supplemented media (Fig. 3D-F).

HS Supplementation Leads to Increased Lipid Droplets and Increased Expression of Lipid-Activated Nuclear Receptors

The cytoplasm of cells in HS media had a prominent granular appearance. To determine whether this change in morphology was the result of alterations in the amount of lipid droplets, cells were stained with Bodipy 493/503, a lipophilic fluorophore with a high affinity for lipid droplets. We found that Bodipy fluorescence intensity was approximately 4× higher in Huh7.5 cells in HS media than in Huh7.5 cells cultured in FBS (Fig. 4A-C). We next investigated the expression of three key lipid regulators: liver X receptor α (LXR-α) and peroxisome proliferator-activated receptors (PPAR-α and PPAR-γ). LXR-α is highly expressed in liver, is activated by cholesterol metabolites, and regulates genes involved in cholesterol processing and secretion.[11] Consistent with increased lipid droplet contents, we found that LXR-α expression is highly increased in cells cultured in HS, compared to FBS (Fig. 4D). Transcription of PPAR-α as well as PPAR-γ was up-regulated significantly in cells cultured in HS, compared to FBS (Fig. 4E,F). PPAR-α is highly expressed in liver and regulates mitochondrial function, fatty acid uptake, beta-oxidation, and TG metabolism, as well as lipoprotein assembly.[11] PPAR-γ also regulates genes involved in lipid metabolism and is activated by an array of ligands, including unsaturated fatty acids.[11]

Figure 4.

HS supplementation leads to increased lipid droplet content and increased expression of lipid-activated nuclear receptors. (A and B) Lipid droplets in FBS (A) and HS (B) cultured cells that were stained with Bodipy493/503. Images were taken with a conventional fluorescence microscope, using identical microscope and exposure settings. (C) Quantitation of Bodipy493/503 fluorescence, with fold change compared to FBS. (D-F) Transcript levels of lipid-activated nuclear receptors LXRα (D), PPARα (E), and PPARγ (F), as determined by qRT-PCR. Fold change compared to FBS was calculated. n ≥ 3; *P < 0.05; ***P < 0.0005; significance was calculated compared to FBS value.

HS Supplementation Leads to Restoration of VLDL Secretion

Importantly, we wanted to determine whether Huh7.5 cells cultured in HS media regain some of the complex functionality of primary hepatocytes that is considered lost in FBS-cultured Huh7.5 cells. The ability to secrete nascent VLDL particles is one example of such a complex process[12] because it depends on the integration of biogenesis, modification, and transportation processes. In line with previous observations,[7, 13] VLDL secretion is virtually absent in Huh7.5 cells that are grown in FBS-supplemented serum (Fig. 5). In cells cultured in HS media, VLDL secretion is gradually restored when cells are cultured in HS: After 5 days, minor changes can be noted on the triacylglyceride- and cholesterol-based lipoprotein profiles (Fig. 5A,B), and by 14 days, a prominent VLDL peak appears in HS-cultured cells. Also, the LDL peak increases in size and elutes earlier, indicating larger particles. Lipoprotein profiles secreted by HS-cultured hepatoma cells closely resemble those secreted by primary human hepatocytes in culture,[7] as well as the lipoprotein profiles of human blood (Fig. 5A[4]).

Figure 5.

VLDL secretion is restored in Huh7.5 cells cultured in HS-containing media. Cells were cultured in FBS- or HS-containing media for up to 30 days. Cells were washed with serum-free OptiMEM (Gibco, Grand Island, NY) to remove lipoproteins present in serum of tissue culture media, and the last wash was collected to serve as background (wash). Cells were then placed in serum-free OptiMEM overnight. Secreted lipoproteins were separated based on size using a size-exclusion column (large particles elute first). After separation, TG (A and B) or cholesterol (C) was measured in-line. (A) A typical lipoprotein profile of human blood (averaged profile from three different blood samples. (B and C) Triacylglycerol- and cholesterol-based lipoprotein profiles of cells cultured in FBS or in HS for 5-30 days. A gradual increase of the VLDL-sized peak can be seen. After 5 days, changes in TG-based lipoprotein profiles (B) are minimal, but a shift can be seen in the LDL-sized peak on the cholesterol profiles (C). From 14 days on, the VLDL peak becomes prominent on both types of profiles. An earlier elution the LDL peak is also observed, indicating larger LDL particles.

Comparison to DMSO and Adult Bovine Serum-Mediated Growth Arrest

We wanted to determine whether some of the same effects noted in HS (adult human serum) could be achieved by switching FBS to ABS (adult bovine serum) or to media supplemented with 1% DMSO, as reported previously.[1] The results of these experiments are presented in the Supporting Materials. Summarizing, both culturing in DMSO and ABS induces growth arrest and results in some of the morphological and transcriptional changes noted in HS. However, neither method induces all changes nor at similar levels as HS supplementation does.

HS Supplementation Leads to Production of High Viral Titers of JFH-1

Next, we investigated the effect of HS supplementation and differentiation of Huh7.5 cells on HCV production. We first investigated viral production after electroporation. FBS-cultured cells were electroporated with JFH-1 RNA and each cell suspension was then split in two, with one half continuously cultured in FBS and the other half in HS. We followed both RNA titers and viral infectivity (TCID50/mL). After approximately 10-14 days postelectroporation, cells cultured in FBS underwent massive cell death, with a loss of RNA titers and infectivity (Fig. 6A,B). However, in HS, this cell death did not occur, and viral titers (RNA copies, TCID50/mL) continued to increase until approximately 20 days postelectroporation, then remained stable for at least 65 days (Fig. 6A,B).

Figure 6.

Comparison of viral production in HS- and FBS-cultured cells. (A) Viral RNA titers after electroporation of JFH-1 in FBS- or HS-cultured cells. In the early time points, some of the detected RNA may have been leftover RNA from the electroporations. (B) Viral infectivity of virus collected after electroporation was measured by determining TCID50/mL. (C) Comparison of the FBS-based method of viral production and infection to the HS-based method described here; viral titers exceed titers in FBS 1,000 times when cells become fully differentiated. (D) Infection of FBS-grown cells with JFH-HS and JFH-FBS. The same RNA titers were used at time of infection (E). Infection of HS- or FBS-grown cells with JFH-HS. Note that the viral titers represented by the open circles in this panel are the same as those represented by the open circles in (D). (F) Long-term production of JFH-1 in cells that were grown in HS. Although high viral titers are achieved sooner in cells that were already differentiated (closed squares), compared to cells that were infected at the time of transfer to FBS (open circles), eventually, cells produced similar titers. Viral production was followed for up to 105 days; also depicted are the viral infectivities (TCID50/mL) for some time points (bars). Depicted are the data obtained in a single experiment; however the experiments were repeated three times or more, with similar outcomes.

We next investigated the ability of JFH-FBS and JFH-HS to infect cells cultured in FBS or HS. We used virus isolated 4 days after electroporation to minimize effects of viral adaptation at time of infection.

First, we compared the traditional method of producing HCV in tissue culture (JFH-FBS variant in FBS-maintained cells) to the tissue culture method described here (JFH-HS variant in HS-maintained cells). In the first 5 days, there was no obvious benefit of using HS for virus production and maintenance of the cells, because viral titers were similar. However, the HS-based method resulted in 1,000-2,000 times more virus, when differentiation was complete (after 15-20 days; Fig. 6C).

To assess whether these changes could be attributed to changes in virus or in cells, we first infected FBS-cultured cells with either JFH-FBS or JFH-HS (same cells different virus). JFH-HS immediately produced higher viral RNA titers, exceeding viral titers after JFH-FBS infection ∼15×, indicating higher infectivity of JFH-HS. Approximately 15 days after JFH-HS infection of FBS cells, a plateau was reached (Fig. 6D).

We next measured viral RNA production after infection with JFH-HS in FBS- or in HS-cultured cells (Fig. 6E, “same virus, different cells”). During the first 10 days, there was no obvious benefit of culturing cells in HS. Viral titers of FBS-cultured cells plateaued approximately 10-15 days postinfection; however, viral RNA titers produced by HS-cultured cells rose rapidly 10-15 days after infection (Fig. 6E), and 20 days after infection, these cells produced approximately 100 times more virus than did FBS-cultured cells. Timing of rapid increase in viral titers coincides with differentiation of cells. Similar trends were observed when JFH-FBS was used instead of JFH-HS; however, initial viral titers were lower (not shown).

After 21 days, viral titers in HS-cultured cells reached a plateau. We were able to achieve continuous production of viral titers of ∼108 RNA copies/mL for at least 105 days using HCV JFH-1 (Fig. 6F). We were able to infect cells before differentiation, as well as cells that were fully differentiated. Eventually, similar titers were reached using either method (Fig. 6F).

We have also tried to infect differentiated cells with 35 different patient sera (genotypes 1-6), but infection was not detected in any of those cultures (using RNA titering).

Previously, only extensive adaptation of JFH-1 resulted in production of high viral titers, and this typically resulted in induction of cell death.[14] To examine whether we had also produced tissue culture adaptations, we have sequenced a JFH-HS viral variant after 24 days of culture and only could confirm a single mutation in NS2, at nucleotide position 2925 (A to G), resulting in a Q to R change. Additionally, we detected three mixed positions in NS5a, but none of these mutations resulted in an amino acid change.

HS Supplementation Changes the Physical Properties of HCV

Last, we investigated whether the biophysical properties of the virus produced by cells in HS media were different from virus produced by cells in FBS media. We investigated viral stability, viral density, ApoB association, and specific infectivity (Fig. 7).

Figure 7.

Biophysical properties of JFH-FBS and JFH-HS. (A) Viral half-life time of JFH-HS is increased, compared to JFH-FBS. (B) Density distribution of virus produced in HS- and FBS-maintained cells. Cells were infected with JFH-FBS and then maintained under either condition, and viral density distribution was determined in a sucrose gradient. Median density for FBS conditions and HS conditions is shown (1.16 and 1.09 g/mL, respectively). (C) Quantitation of viral density distribution. The fraction of cells with a density lower than 1.16 g/mL is depicted. (D) ApoB association of JFH-FBS and JFH-HS, as determined by IP with anti-ApoB antibodies. (E) Specific infectivity of JFH-HS and JFH-FBS at different time points postinfection (5-65 days). Specific infectivity of JFH-FBS did not change, whereas the specific infectivity of JFH-HS increased in the first 21 days, then reached a plateau. Values are expressed as TCID50/RNA copies. (F) Specific infectivity of JFH-FBS and JFH-HS produced by differentiated cells (21 days in HS and beyond). n ≥ 3; *P < 0.05; ***P < 0.0005.

We wanted to determine whether a change in viral half-life of JFH-HS, compared to JFH-FBS, could be a contributing factor to the increased viral titers. At 4°C, both viral variants were stable. However, at 37°C, the half-life of JFH-FBS was 10-14 hours, whereas the half-life of the JFH-HS variant was approximately 75 hours (Fig. 7A).

We found that virus produced by cells in HS media shifts toward a lower density on a sucrose gradient. Virus produced in FBS media had a median density of JFH of 1.16 g/mL, consistent with previous reports.[15] However, virus produced in HS media had a median density of 1.09 g/mL. In addition, a peak with a very low density appeared (Fig. 7B). Overall, 35% of the virus produced in FBS media had a density lower than 1.16 g/mL, whereas 75% of the virus produced in HS media had a density lower than 1.16g/mL (Fig. 7C). The low density of virus produced by cells in HS media is more consistent with the density of virus derived from patients and chimeric mice.[16]

HCV in patients has been consistently shown to be associated with ApoB[4, 17]; however, previous reports have shown that virus produced in culture is not associated with ApoB, but instead with apolipoprotein E.[18] We determined whether HCV produced in HS media was associated with ApoB (Fig. 7D). Consistent with previous reports, approximately 5% of the JFH-FBS virus variant was associated with ApoB. However approximately 80% of the JFH-HS viral variant is ApoB associated. For comparison, HCV from patient sera or chimeric mice was 30%-80% ApoB associated.[4]

A third difference between virus produced in tissue culture and virus produced in animal models is its specific infectivity. Virus produced in tissue culture (FL-J6/JFH variant; HCV in cell culture [HCVcc]) has a specific infectivity of approximately 1 TCID50/1,230 genomes, whereas the specific infectivity of virus produced in either mice or chimpanzees was 1 TCID50 per 10-150 viral genomes.[19] We determined specific infectivity of the two viral variants, expressed as TCID50/RNA copies at various time points (Fig. 7E). The specific infectivity of the JFH-HS increased gradually in the first 21 days, then reached a plateau, whereas the specific infectivity of JFH-FBS did not change. The average specific infectivity of virus produced by cells that were fully differentiated in HS was 1 TCID50 per 236 viral genomes, compared to an average specific infectivity of 1/2513 for JFH-FBS (Fig. 7F). The overall specific infectivity of JFH-HS is now similar to the highest specific infectivity that was reported previously for HCVcc,[5] corresponding to the small low-density peak in that study.

Discussion

In this study, we investigated the effects of HS on tissue culture of Huh7.5 cells. We compared the standard tissue culture protocol, using media supplemented with 10% FBS, to the use of media supplemented with 2% HS. Cells cultured in HS media undergo rapid growth arrest and show increased expression of hepatocyte differentiation markers (α1AT and ALB). In HS-supplemented media, the expression of cell-contact proteins claudin-1, occludin, and e-cadherin was also increased. These factors are indicative of differentiated epithelial cells. Because previous reports have shown that claudin-1 and occludin are entry factors and confer infection of nonpermissive cell types,[20, 21] the increase in claudin-1 and occludin likely plays a role in the increase in viral titers in HS media. The level of expression of other HCV-entry receptors (CD81, SR-B1, and NPC1L1) did not change when Huh7.5 cells were cultured in media with HS. Expression of key lipid metabolism regulators (LXR-α, PPAR-α, and PPAR-γ) was increased, and consistent with this, the lipid droplet content of these cells was highly increased. We showed that VLDL secretion was restored, a complex process that requires the integration of various biogenesis, modification, and transportation steps.[12]

All these factors have been implicated in the life cycle of HCV, and, in particular, HCV has been shown to hijack the VLDL secretion machinery for egress.[25] Consistent with this, we have shown that under these new tissue culture conditions, production of JFH-1 increased more than 1,000-fold. The virus produced under these conditions more closely emulates HCV that is found in serum of patients and animal models, was associated with ApoB, had a lower density, and was highly infectious. In addition, HCV JFH-1 that was produced in HS-containing media had a significantly longer half-life.

Consistent with a switch to proper lipoprotein secretion in Huh7.5 cells in HS media, there was an increase in ApoB association with HCV. Though it is possible that the ApoB association of the virus occurs outside the cell, we do not think that this is the case. When we incubated JFH-FBS with human ApoB-containing lipoproteins, we found an increase in the fraction of the virus associated with ApoB; however, the vast majority (99%) of the virus was degraded. Therefore, it is possible that the ApoB-free virus is degraded, leading to an increase in the fraction of ApoB-associated virus. In support of this explanation, we found that the virus secreted by Huh7.5 cells cultured in HS media, which was ApoB associated, was exceptionally stable.

Higher viral infectivity has been linked to lower viral density,[5] presumably through lipoprotein association. We observed a gradual increase in viral infectivity, as well as a gradual increase of VLDL secretion, whereas the external environment remained the same (2% HS throughout), supporting the hypothesis that the virus associates to ApoB-containing lipoproteins intracellularly: We would expect an instant increase in infectivity if the virus associated with lipoproteins extracellularly, which was not observed. Further studies are needed to address this hypothesis and to investigate whether JFH-HS remains ApoE associated or now associated with ApoB instead.

We do not believe that the increase in viral titers can be attributed to a single factor. Rather, we have shown many changes, including cell–cell contacts, increased entry receptors, increased lipid droplets, increased infectivity, as well as increased viral stability. We envision a scenario where the JFH-HS viral variant, which is associated with ApoB, shows increased binding to heparan sulfate proteoglycans and, possibly, LDL-R and SR-B1. Eventually, the virus enters the cell at tight junctions through claudin-1 and occludin.[22] Increased cellular lipid droplet content allows the cells to establish the proper environment for HCV replication,[23, 24] and the viral assembly hijacks the VLDL secretion machinery,[25] which is now functional. Thus, the virus becomes associated with ApoB in the process, whereas proper VLDL secretion facilitates viral egress. Consistent with this, we detected far less core staining in HS-cultured cells than in FBS-cultured cells, even when secreted RNA titers were similar or higher. Similar observations have been presented previously in cells with elevated expression of carboxylesterase 1, an important factor in lipid loading of nascent ApoB particles.[26] This suggests that viral secretion is indeed more efficient in HS-cultured cells. It also may suggest that core accumulates in FBS-cultured cells, possibly leading to endoplasmic reticulum stress and apoptosis.

Although we currently do not fully understand the mechanism of HS-induced differentiation, the effect of HS appears to be at least 2-fold. Because the use of ABS also results in growth arrest, absence of fetal growth factors, and/or presence of differentiation-inducing factors in adult serum could partly explain the observed changes. We hypothesize that the absence of growth-stimulating factors allows for growth arrest and provides the opportunity to form cell–cell contacts and tight junctions. Cell–cell contacts, in turn, are important factors in facilitating intercellular communication and have been linked to increased hepatic functionality, including bile secretion, glycogenolysis, and ALB secretion.[9]

Our current data suggest that an important difference between cells cultured in ABS-supplemented (or DMSO-supplemented) media and cells cultured in HS-supplemented media is the intracellular lipid stores. Our study indicates that only in HS is the lipid droplet content increased. The increased lipid content can, in turn, facilitate activation of lipid-dependent nuclear receptors, such as LXR-α, PPAR-α, and PPAR-γ, enabling de novo synthesis of lipids and lipoprotein secretion.

The method we have presented in this study for culturing hepatoma cells provides a convenient, cost-effective model for the study of liver disease, lipoprotein secretion, and other liver-related processes. We have used this model to produce HCV strain JFH-1 at high titers. When cells are differentiated, JFH-1 production in HS media exceeded that in FBS media by 1,000 times or more. We have achieved production of viral titers of over 108 RNA copies/mL for extended periods of time. Besides functioning as a production platform for HCV, this model can also provide further insight into the cellular factors and processes essential for efficient production of HCV, resulting in virus that closely resembles HCV derived from patient sera.

Ancillary