Background
Hepatitis C virus (HCV), a liver tropic positive-stranded RNA flavivirus, infects ~170 million people worldwide, causing acute and chronic hepatitis and hepatocellular carcinoma [
1]. However, since its discovery in 1989, a major obstacle impeding HCV research has been the lack of robust cell culture and small animal infection models. Notably significant advancement has been made with the identification of a genotype 2a HCV consensus clone (Japanese Fulminant Hepatitis, JFH-1) that can replicate and produce infectious HCV in vitro in the Huh7 human hepatoma-derived cell line [
2‐
4], allowing for the study of the entire viral life cycle. This system, however, is limited in that it makes use of a non-differentiated cell line that does not recapitulate the cellular conditions encountered by HCV in vivo [
5,
6]. In particular, hepatocyte polarity is likely relevant to HCV entry as growing evidence suggests interplay between HCV and tight junction (TJ) proteins claudin-1 (CLDN1) [
7] and occludin [
8,
9] is essential for viral uptake. In fact, recent reports surprisingly suggests that hepatocyte polarity may restricts HCV entry [
10,
11]. While an inverse relationship between viral entry and hepatocyte polarity would potentially represent a unique determinant of HCV entry, to date attempts to dissect this relationship have been difficult and inconclusive due to the inability of cell culture grown hepatocyte-derived cell lines, such as Huh7 cells, to mimic the complex polarized phenotype of hepatocytes in vivo. To circumvent these restriction, studies investigating HCV entry into Caco-2 cells [
10] and HepG2 cells [
11] have been performed as these cells can polarize to differing degrees in vitro, however, neither Caco-2 or HepG2 cells supports efficient HCV infection limiting their utility. As such, a more physiologically relevant hepatocyte tissue culture model is still needed to assess if cell polarity negatively affects HCV infection.
The NASA-engineered RWV is a horizontally rotating cylindrical culture vessel which reduces shear and turbulence associated with conventional stirred bioreactors; therefore, it simulates aspects of microgravity similar to the environment encountered during fetal development [
12‐
14]. In contrast to conventional static tissue culture systems, cells grown in the RWV are cultured in "suspended animation" where they are continuously free-falling [
12,
15]. Thus, while the 2-D environment of plastic substrates may alter gene expression and prevent cellular differentiation [
12,
16‐
21], the fluid dynamics of the RWV culture system allow cells to co-localize into three-dimensional (3-D) aggregates, promoting in vivo-like exchange of growth factors and efficient cell-to-cell interactions [
12‐
14,
20,
21]. This in vivo-like environment thus can promote transformed and primary cell lines to become more structurally and functionally similar to their in vivo counterparts [
13,
15,
20‐
24].
In the current study we demonstrate that RWV-cultured Huh7 cells formed complex, multilayered, 3-D aggregates that exhibited up-regulation of metabolic and hepatocyte-specific transcripts as well as increased expression and re-localization of tight junction, cell adhesion, and polarity markers. Importantly, these aggregates remained highly permissive for HCV infection suggesting that hepatic polarity does not limit HCV entry in 3-D-cultured Huh7 cells. As such, RWV-cultured Huh7 cells may represent a more appropriate physiologically relevant system for further in vitro studies of HCV entry and infection dynamics.
Discussion
Here we demonstrate that Huh7 cells cultured in RWV bioreactors form multi-layered tissue-like aggregates that are phenotypically distinct from traditional Huh7 2-D monolayers (Fig.
1 and
2). Specifically, the RWV-environment promoted increases in hepatocyte-specific, as well as Phase I and II metabolic gene transcripts in 3-D Huh7 aggregates relative to Huh7 monolayers (Fig.
2). Additionally, we observed increased expression and organization of cellular HCV receptors, cell adhesion, tight junction and polarity-specific proteins, and the loss of cancer-associated nuclear localization of β-Catenin, in RWV 3-D Huh7 aggregates as compared to 2-D monolayers (Fig.
3). These data therefore suggest that the RWV environment promotes differentiation of Huh7 cells down a more hepatocyte-like route. Importantly, since these 3-D Huh7 cultures remain highly permissive for HCVcc infection, this system represents a new in vitro cell culture system for the study of HCV infection and antiviral drug studies in more polarized, xenobiotically-competent cells.
Relevant to the study of HCV, expression of the HCV receptors CD81 and SR-B1 were both diffuse and poorly organized in 2-D cultured Huh7s cells, while their expression was increased and localized to apical TJ regions and basolateral-sinusoidal surfaces in 3-D aggregates. Likewise, TJ proteins, which typically localize to the apical surface in polarized hepatocytes [
43], were more concentrated at the apical surface of 3-D Huh7 aggregates as compared to monolayer cultures. Notably however, the TJ protein CLDN1, a recently identified HCV receptor [
7], not only localized to TJs, but was also present at both apical and basolateral surfaces in 3-D aggregates. This localization pattern is in agreement with other studies [
47] and the model proposed by Reynolds et al., describing tight-junctional (apical) and nonjunctional (basolateral) forms of CLDN1 in polarized hepatocytes [
44]. As suggested by Mee et al, it may be that these non-junctional pools of CLDN1 have a direct role in HCV entry [
11]. Interestingly, Battle et al., have demonstrated a correlation between HNF4α and cell adhesion and TJ molecules expression and organization [
48]. Whether this is also the case in the 3-D Huh7 aggregates, which have increased HNF4α expression (Fig.
2A) remains to be determined. Nonetheless, the ability of 3-D cultured Huh7 cells to better organize cell adhesion and TJ proteins is a phenotype consistent with other RWV-cultured cell types [
14,
21,
23]. As such, RWV-cultured Huh7 cells provide an appropriate model for investigating HCV entry, particularly the interaction, organization, and stoichiometry of HCV receptors and TJ proteins. Additional analyses to determine the extent of differentiation and polarization of 3-D Huh7 aggregates is still warranted and a focus of ongoing studies.
To date, attempts to study HCV in polarized cells have been limited to colorectal adenocarcinoma Caco-2 cells [
10] or HepG2 cells [
11], neither of which support robust HCVcc infection. Although an inverse relationship between cell polarization and HCV entry into polarized Caco-2 [
10] and HepG2 [
11] cells has been observed no such phenotype was observed in 3-D Huh7 aggregates. Specifically, 3-D Huh7 aggregates, infected at various stages of differentiation (e.g. day 1, 7 or 14 post seeding), were equally permissive for HCVcc infection (Fig.
4B–E). Furthermore, 3-D aggregates treated with PMA, a known disruptor of TJ formation [
49], were no more permissive for HCV infection as compared to untreated parallel aggregates (data not shown), suggesting that the TJ barriers formed in 3-D Huh7 aggregates are not inhibitory for HCVcc infection.
Conclusion
Growing evidence suggests interplay between TJ protein expression, localization and function and HCV infection. Although, the current HCV infectious 2-D Huh7 cell culture system does not amend itself well to elucidating these dynamic relationships, the highly HCV-permissive 3-D Huh7 cell culture system described herein more closely mimics the differentiated and polarized state of hepatocytes. As such the RWV 3-D Huh7 cell culture system should prove useful for understanding the dynamic relationship between HCV and TJ protein expression as well as elucidating how HCV interacts with and disrupts key aspects of hepatocyte physiology.
Acknowledgements
We thank Drs. Heather L. LaMarca and Kerstin Hönzer zu Bentrup for helpful discussions, Dr. Francis Chisari for Huh7 cells, Dr. Takaji Wakita for the JFH-1 containing plasmid (pJFH-1), Dr. Dennis Burton for the monoclonal anti-HCV E2 human antibody (C1), Dr. Mei Ling Chen for assistance with confocal microscopy and Patricia A. Mavrogianis for paraffin embedding and sectioning of 3-D aggregates.
This work was supported by Public Health Service grant AI-070827 from the National Institute of Allergy and Infectious Diseases, Public Health Service grant CA-133266 from the National Cancer Institute and the University of Illinois Chicago Council to support Gastrointestinal and Liver Disease (UIC GILD). VTC was supported by an Institutional Ruth L. Kirchstein National Research Service Award (DK-007788-07) from the National Institute of Diabetes and Digestive and Kidney Diseases.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
BS and VT participated in the design of the study, performed the experiments and drafted the manuscript. SLU designed the study and participated in drafting the manuscript. All authors read and approved the final manuscript.