Introduction

Human norovirus (hNoV) was discovered as the cause of an outbreak of acute gastroenteritis in an elementary school in 1968 in the city of Norwalk, Ohio, and in 1972, the gastroenteritis was confirmed by immunoelectron microscopy to have a viral etiology [47, 48]. Although gastroenteritis is caused by bacteria, protozoa and viruses, viral gastroenteritis is particularly problematic due to the lack of effective antiviral therapies. Acute gastroenteritis caused by hNoV is estimated to account for 90 % of cases of virus-mediated gastroenteritis. Approximately 800 fatalities among infants and the elderly due to hNoV infection are reported each year in the US alone, and 200,000 children under the age of 5 years in developing countries die annually from the disease [79]. Norovirus infection in healthy adults causes self-limiting acute disease including vomiting and diarrhea and typically resolves in 2-3 days. However, in immunocompromised patients, hNoV can establish chronic and potentially fatal infections [7].

Since the discovery of hNoV and the cloning of its genome [57, 111], much effort has been made to develop in vitro infection models for hNoV in cultured cell lines [19, 31, 58]. However, the lack of cell lines that can support hNoV infection poses a barrier to its in vitro culture. Therefore, hNoV stocks have been prepared from stool samples of human patients or volunteers for human infection trials. For these reasons, progress in the study of the pathological characteristics and mechanisms of viral replication and gene expression has been severely impeded for this virus compared to other positive-sense RNA viruses such as hepatitis C virus and poliovirus. The present review focuses on recent progress and challenges in the development of in vitro culture models and alternative models available to study hNoV.

In vitro infection and culture models of hNoV

De novo infection of established cell lines

A large number of established cell lines have been tested for in vitro infection with hNoV (Table 1). Many different animal cell lines as well as human epithelial cells from the gastrointestinal tract have been tested for susceptibility to hNoV infection, with no clear indications of infection. As simple and conventional infections have not been successful, various culture methods and manipulations of cellular phenotypes have also been attempted (Table 1). [18, 19, 23, 25, 37, 44, 56, 63, 80, 82, 90, 96]. However, there have been no clear indications of de novo hNoV infection observed [19]. A plausible explanation may include the inactivation of virus particles upon excretion in feces. This possibility can be tested if virus stocks can be prepared from sources other than feces. However, Schwab et al. [87] reported that hNoV forms very stable particles that can survive for an extended period outside the human body. In addition, other enteric viruses are not inactivated by similar preparations from feces. Therefore, noroviral inactivation upon excretion might not fully explain its inability to infect established cell lines [19, 88].

Table 1 Norovirus infections in established human cell lines
Full size table

As some caliciviruses require the presence of intestinal contents for infection [22, 78, 86], the hypothesis that hNoV needs to be modified in a manner similar to that in the gastrointestinal tract of the human body was put forth and tested. For example, pre-treatment with trypsin and intestinal contents from a gnotobiotic pig has been reported to be required for infection with and replication of feline and swine norovirus, respectively. However, similar pre-treatment with supplements did not result in hNoV infection in the culture. Another possibility is that hNoV infection is inhibited or promoted by the presence of virus-specific antibodies. If hNoV particles are coated with neutralizing antibodies in the feces, viruses might lose infectivity in the subsequent infections in cultured cells. However, in some viral infections, the presence of virus-specific antibodies helps virions infect their target cells. Examples include, but are not limited to, human cytomegalovirus [65], foot-and-mouth disease virus [68, 84], and dengue virus [28, 29, 42]. Therefore, it would be intriguing to explore antibody-dependent enhancement of viral infection to test if the binding of neutralizing antibodies to hNoV renders them infectious under in vitro conditions.

Attempts have been made to infect not only established cell lines but also human macrophages and dendritic cells [58]. Murine norovirus (MNV) was first isolated and identified in immune-compromised mice [49]. MNV was found to infect and replicate in macrophages and dendritic cells from STAT1-deficient mice, and the same types of cells have been reported to be infectible in vitro [14, 109]. As the genome of hNoV is detected in the sera of pediatric patients [98], human macrophages and dendritic cells were tested for their ability to be infected in vitro [58]. However, no indications of infection were observed. Interestingly, it has been reported that infection with porcine enteric calicivirus (PEC) requires the presence of bile acids, which in turn inhibit the function of STAT1 [11]. Considering that type I and II interferons inhibit viral replication and protein expression of MNV, which can infect STAT1-deficient mice [14, 64], it is possible that STAT1 knockdown with specific siRNA can render otherwise resistant cells susceptible. However, the inhibition of STAT1 expression did not promote viral infection (Mary K. Estes, personal communication).

Most recently, a breakthrough has finally been made by Jones et al. in the development of an in vitro hNoV culture model [43] using B cells in the presence of a commensal bacterium, Enterobacter cloacae. In fact, it has been reported that B cells are required for replication of MNV [5, 73, 113]. Jones et al. [43] showed that while hNoV infection of B cells required the presence of E. cloacae, MNV could infect B cells in vitro even in the absence of the bacteria and that oral antibiotic administration reduced MNV replication in vivo. It has been established that norovirus infection in humans is correlated with the histo-blood group antigen (HBGA) expression profile [38, 66]; however, it remains unclear how the HBGA-expressing bacteria promote hNoV infection in B cells, especially considering the fact that HBGA-like molecules expressed on commensal bacteria would compete for hNoV binding with HBGA on hNoV target cells. The requirement for enteric bacteria for hNoV infection of B cells is a groundbreaking finding in the norovirus field, but it might not be surprising, as it is known that some enteric viruses, including mouse mammary tumor virus [46], poliovirus [55, 83], and reovirus [55] require bacteria for infection and replication.

Three-dimensional (3-D) cell culture methods

3-D culture techniques have been developed, helping to promote infections and cultures of various pathogenic bacteria and viruses. Infection of cells in a monolayer is very different from that in an in vivo environment. Proper differentiation of epithelial cells requires apical and basolateral polarization, and two-dimensional (2-D) culture techniques might not be able to support the same cellular differentiation found in vivo. 3-D organoid culture techniques were first developed and used to investigate the infection and pathogenicity of Salmonella enteritica in INT-407 human intestinal epithelial cells [32, 74, 75]. Similar culture techniques have been exploited for establishing infections with various bacteria (Escherichia coli [32, 74] and Pseudomonas species [9]) and viruses (Epstein-Barr virus [62], Kaposi’s sarcoma-associated herpesvirus [16], rotavirus [21] and hNoV [93]). This culture technique was first developed by NASA [75], and the 3-D organoid culture is basically a bioreactor with a rotating cylinder. Cell culture medium is added to the cylinder along with collagen I-coated porous microcarrier beads. With the addition of cells, the cylinder continues to rotate to prevent the cells from binding to the walls of the cylinder. Cells grow in and on the porous beads to form 3-D structures that closely resemble physiological tissues or organs of the body. INT-407 cells in 3-D cultures have been shown to differentiate into various cell types, thus enabling ‘co-cultures’ in a bioreactor [74, 93]. The co-culture of various cell types in 3-D has been reported to allow noroviral infection [19, 93], which is not possible in the 2-D culture of the same cell type. In addition, Straub et al. [93] also reported that a type of histo-blood group antigen (i.e., Lewis antigen A), a cellular attachment receptor for hNoV, was expressed at the apical tip of the 3-D culture. When the 3-D culture of INT-407 was infected with hNoV, the authors observed a cytopathic effect, and an increase in the number of viral genome copies was detected by reverse transcription polymerase chain reaction (RT-PCR) and fluorescence in situ hybridization (FISH). Based on these findings, Straub et al. claimed that the 3-D culture of INT-407 was susceptible to hNoV. However, in the same report [93], Caco-2 and HT-29 cells were not susceptible to hNoV, even in 3-D organoid cultures. In 2011, the same group of researchers reported [94] that a subclone of Caco-2 cells (C2bbe1) was susceptible. They claimed that C2bbe1 cells in the 3-D culture were able to fully differentiate in the 3-D culture, while the parental Caco-2 cells were not.

These findings, however, have been disputed by the other research groups: norovirus infection was not detected in 3-D culture [31]. A joint study conducted by the Nickerson and Estes groups reported that norovirus infection was not detected in a 3-D culture by real-time PCR or immunofluorescent assay. In addition, histo-blood group antigens such as Lewis antigen were not expressed. Based on these and other findings, the joint group concluded that the CPE in the 3-D culture that was observed upon inoculation with norovirus was likely due to the toxicity of contaminating lipopolysaccharides in the virus stock prepared from the fecal samples. The claims of the papers by Straub et al. [93, 94] could also not be verified by Takanashi et al. [99]. This discrepancy will only be resolved when virus stocks can be prepared from samples other than fecal samples.

Challenges to the development of in vitro cultivation of hNoV

HNoV is known to bind to HBGA to infect the host [38, 39, 50, 60, 66, 67, 77]. Interestingly, the same is true of rotavirus [40], one of the two major gastroenteritis-causing viruses in humans. When functional α (1,2) fucosyltransferase 2 (FUT2) was not expressed, norovirus infection was not detected in experimental infections in human volunteers. FUT2 is an enzyme required for the expression of HBGA on the surface of epithelial cells, some of which is secreted in bodily fluids, including saliva [54, 76]. HBGA type 1 (e.g., Lewis b [Leb]) is mainly expressed on the epithelial cells at the junction of the stomach and duodenum, and HBGA type 2 is expressed at the glandular level [66, 71, 85, 92]. In an experiment using virus-like particles (VLPs), VLPs bind to HBGA, inducing their internalization [66]. In addition, VLPs seem to specifically bind to A, H1, and Lewis b antigens [17, 30, 35, 66]. Therefore, functional receptor expression appears to be critical for successful infection with hNoV in culture. The overexpression of FUT2 in Huh-7 cells, a human hepatoma cell line, resulted in strong binding of viruses to the cells; however, it did not influence viral internalization or replication. In fact, many other studies have led to the same conclusion: overexpression of FUT2 has a negligible effect on virus internalization, uncoating and viral genome replication [1, 11, 27, 31, 108]. These results imply that HBGA expression alone may not be sufficient for viral infection [31, 100]. Furthermore, the presence of other proteins may be required for norovirus infection. One interesting hypothesis is that not only HBGA but also a co-receptor may be required for norovirus infection [100], the discovery of which would lead to the development of an in vitro culture model of hNoV. In fact, this hypothesis has been proposed by many researchers in the field [31, 105]. Identification and characterization of co-receptor(s) would provide a breakthrough in the development of a convenient and reliable hNoV culture model. In addition, to develop a successful hNoV in vitro culture model, more experimental infection systems need to be investigated, including primary human intestinal cells or tissue explants and 3-D co-cultures of different cell types. An interesting alternative that has shown promising results in a recent study is the development of intestinal organoids using pluripotent stem cells [21].

One of the major barriers to hNoV study is the inability to prepare purified virus stocks. Virus stocks prepared from human feces are often contaminated with LPS or other enteric viruses (e.g., rotavirus), which makes it difficult to interpret the CPE observed in experimental infections [31]. Therefore, the development of a norovirus producer cell line that enables the preparation of a large quantity of purified hNoV will certainly provide a breakthrough in the field of hNoV research.

Research methods for hNoV

Utilization of norovirus replicons or infectious cDNA clones

Due to the lack of susceptible cell lines, the development of preventive vaccines and virus-specific therapies has been hampered. However, the use of virus replicons containing part of the viral genome enables screening of antiviral drugs and efficacy testing. The first of such replicons expresses neomycin in place of the VP1 capsid protein in Huh-7 and BHK21 cells (Fig. 1B) [12]. This replicon was stably maintained over extended passages, and viral protein expression was detected in those cells. Using the hNoV replicon, Chang et al. identified interferon alpha, interferon gamma, ribavirin [13], and peptide-conjugated phosphorodiamidate morpholino oligomers (PPMOs) [6] as antiviral agents. Viral-replicon-containing cells are useful for identifying antiviral agents against non-capsid proteins but might not be adequate for studying the full life cycle of the virus due to the lack of VP1. To overcome these limitations, three independent groups of researchers have developed systems that harbor the full cDNA of hNoV [2, 45, 52]. The overall experimental designs of two of the systems are similar. First, the T7 promoter was added to the 5’ end of the full viral cDNA with a poly(A) tract of 26 to 30 nucleotides at the 3’ end (Fig. 1B and C). To regulate transcription by T7 polymerase, the T7 terminator sequence was inserted at the far end of the 3’ end of the construct (Fig. 1C and D). Between the poly(A) and T7 terminator, the ribozyme sequence was added so that the full viral genome was precisely processed (Fig. 1C and D). When the T7 polymerase was expressed using vaccinia virus, virus particle production was detected by electron microscopy at a density of 1.318 g/cm3 as determined by gradient ultracentrifugation. These data suggest that complete virions were formed by the binding of the viral genome and capsid proteins. However, the infectivity of the virions could not be determined due to the lack of susceptible cell lines. Furthermore, these systems require the presence of a helper virus to express functional T7 polymerase. To overcome this inconvenience, Katayama et al. [53] developed a plasmid-based hNoV reverse genetics system (Fig. 1D), which was successfully exploited to produce GFP-expressing recombinant hNoV. Development of the long-awaited recombinant hNoV will certainly help identify susceptible cells in vivo and in vitro. For example, virus stocks prepared from cell culture without contamination with endotoxin, as is seen in stocks prepared from patient stool samples, will effectively help resolve the debate over whether contaminating endotoxin in hNoV stocks was the primary cause of the cytopathic effect observed in the 3-D organoid culture model. In addition, the use of GFP-expressing recombinant hNoV will enable investigators to detect low-level hNoV infections in both primary and established cells, allowing sensitive and high-throughput viral detection using fluorescence. Furthermore, with the availability of an in vitro B cell infection model and recombinant hNoV viruses, the requirement(s) for viral entry in cultured cell lines can be effectively analyzed and probed.

Fig. 1
figure 1

HNoV replicons and infectious cDNAs

Full size image

Utilization of virus-like particles (VLPs)

VLPs are suitable study materials for investigating the immunological aspects of hNoV infection. VLPs are particles made of self-assembled viral capsid proteins containing no viral RNA genome. The norovirus genome encodes two capsid proteins: VP1 and VP2. Of the two, VP1 alone can be assembled to form VLPs, and the function of VP2 has only begun to be revealed. Interestingly, VP2 is not required for VLP assembly but seems to enhance the expression of VP1 in cell culture and associate with VP1 within the shell domain, promoting the stability of VLPs [59, 106]. The outer structure of VLPs made of VP1 alone is known to be identical to that of complete virions containing the RNA genome. To date, many protein expression systems have been exploited, including insect cells [26, 41], human cells (293T [102], Caco-2 [4]), and plant cells (tomato [36], and potato [97]). In the absence of de novo infection systems for hNoV, VLPs have played a critical role in determining how norovirus interacts with host cells. VLPs have been shown to bind directly to HBGA molecules on the surface of host cells in vitro [66], including A, H type1, and Leb carbohydrates [30]. These data indicate that the cellular receptor for hNoV is HBGA, which attests to the usefulness of VLPs. In this regard, VLPs are invaluable for the study of virus-host interactions.

Utilization of other animal noroviruses

Due to the lack of cell culture or animal models for hNoV, other caliciviruses that infect experimental animals represent useful alternatives. In fact, a large portion of the known mechanisms of regulation of viral gene expression, gene function, and genome structure have been extrapolated from animal noroviruses. Animal noroviruses that can be cultivated in vitro are listed in Table 2. Among them, the best-characterized model is MNV (MNV) [49, 107, 110]. MNV infection occurs through the same fecal-oral route as hNoV. MNV is easy to manipulate experimentally, and it infects murine macrophages and dendritic cells as well as RAW264.7[109], a macrophage cell line. MNV belongs to genogroup V, while hNoV belongs to genogroup I, II, or IV. MNV is especially useful for studying virus-specific immune responses in a variety of knockout mice [89]. Using these models, primary and memory responses to norovirus infection have been extensively studied [3, 69, 73]. Furthermore, MNV has been successfully used for development of vaccines against norovirus infection [10, 61]. Details are reviewed elsewhere [51, 103, 110].

Table 2 Animal infection models for norovirus
Full size table

However, despite its many advantages, the MNV model has clear limitations as an alternative to hNoV infection. First, MNV-infected mice do not show symptoms such as diarrhea or vomiting. Second, mice are chronically infected [33, 34], which is in stark contrast to acute infection by hNoV. Third, number of MNV genotypes is limited, and it thus may not be suitable for the development of vaccines [104]. Lastly, hNoV does not seem to be able to infect monocyte-derived macrophages and dendritic cells in vitro [10]. The availability of a small-animal model that mimics the pathology of hNoV infections will undoubtedly play a key role in the development of antivirals and vaccines in the future.

Conclusion

Since the discovery of hNoV, little progress has been made regarding its mechanism of infection, replication, or host immune responses, and the most important reason for this dearth of understanding is the lack of cell culture or animal infection models. Thus, it is of paramount importance to develop a cell culture model to identify antiviral agents and vaccines against hNoV. As such, the development of a cell culture model is the key to an explosive expansion of research on this virus.