Lack of Norovirus Replication and Histo-Blood Group Antigen Expression in 3-Dimensional Intestinal Epithelial Cells

Melissa M. Herbst-Kralovetz

, Andrea L. Radtke, Margarita K. Lay, Brooke E. Hjelm, Alice N. Bolick, Shameema S. Sarker, Robert L. Atmar, David H. Kingsley, Charles J. Arntzen, Mary K. Estes, and Cheryl A. Nickerson

Author affiliations: Author affiliations: University of Arizona, Phoenix, Arizona, USA (M.M. Herbst-Kralovetz, A.L. Radtke); Arizona State University, Tempe, Arizona (M.M. Herbst-Kralovetz, A.L. Radtke, B.E. Hjelm, A.N. Bolick, S.S. Sarker, C.J. Arntzen, C.A. Nickerson); Baylor College of Medicine, Houston, Texas, USA (M.K. Lay, R.L. Atmar, M.K. Estes); Pontificia Universidad, Catolica de Chile, Santiago, Chile (M.K. Lay); Translational Genomics Research Institute, Phoenix (B.E. Hjelm); US Department of Agriculture, Dover, Delaware (D.H. Kingsley)

Abstract

Noroviruses (NoVs) are a leading cause of gastroenteritis worldwide. An in vitro model for NoV replication remains elusive, making study of the virus difficult. A previous study, which used a 3-dimensional (3-D) intestinal model derived from INT-407 cells reported NoV replication and extensive cytopathic effects (CPE). Using the same 3-D model, but with highly purified Norwalk virus (NV), we attempted to replicate this study. Our results showed no evidence of NV replication by real-time PCR of viral RNA or by immunocytochemical detection of viral structural and nonstructural proteins. Immunocytochemical analysis of the 3-D cultures also showed no detectable presence of histo-blood group antigens that participate in NV binding and host tropism. To determine the potential cause of CPE observed in the previous study, we exposed 3-D cultures to lipopolysaccharide concentrations consistent with contaminated stool samples and observed morphologic features similar to CPE. We conclude that the 3-D INT-407 model does not support NV replication.

Norovirus (NoV) has been identified as the primary etiologic agent of acute epidemic viral gastroenteritis in industrialized countries (1,2). Norwalk virus (NV) is the human prototype GI.1 NoV strain; it belongs to the Caliciviridae family of positive-sense, single-stranded RNA, nonenveloped viruses (3). Human-to-human transmission of NoV occurs primarily through the fecal-oral route, with the small intestine being the initial site of viral replication (1,4). The lack of an efficient cell culture system in which to study NoV infections has hindered development of antiviral drugs to control or limit NoV outbreaks (5).
Currently, animal models are being used to increase our understanding of NoV infections (6–12); however, no small animal model for NoV mimics the disease manifestations observed in humans (7,11–13). As a result, human outbreaks and volunteer studies have been the primary source for existing knowledge of NV epidemiology and pathogenesis, respectively (1,14,15). A major finding from human NV volunteer studies is that persons with strains containing 2 mutated alleles of the α(1,2) fucosyltransferase (FUT2) gene were resistant to NV infection (16–20). FUT2 encodes an enzyme that produces histo-blood group antigens (HBGA) on the surface of epithelial cells and in mucosal secretions (21,22). Persons who lack a functional FUT2 gene cannot generate ABH antigens in secretions and, thus, are termed nonsecretors (23,24). HBGA are complex carbohydrates distinguished by different monosaccharides added to a precursor oligosaccharide by fucosyltransferase enzymes (23,24). In the human gastroduodenal junction where NV has been shown to bind HBGA type 1 (Lewis b [Le^b]), is found exclusively on epithelial surfaces, whereas HBGA type 2 is primarily found at the glandular level (19,24–26). In vitro experiments using NV virus-like particles (VLPs) directly showed NV VLP attachment to HBGA, resulting in VLP internalization into the cell (19). In addition, NV VLPs were found to preferentially bind to A and H type 1 and Le^b carbohydrates (19,27–29). From these critical studies, putative NV receptors were identified, and thus it was hypothesized that a successful in vitro cell culture system would most likely possess these receptors to support NV replication.
Although most of the understanding of NoV infections in humans has been derived from volunteer studies and authentic sporadic outbreaks, extensive volunteer studies have limitations (variability, cost, and institutional review board considerations). As a result, many attempts have been made to develop a reproducible in vitro model system to culture NoV, but none has yet proved successful (13,30–33). However, one publication has reported establishing a productive NoV infection in a 3-dimensional (3-D) organotypic model of human intestinal epithelium derived from the human embryonic intestinal epithelial cell line INT-407 (30). We report here an unsuccessful attempt to replicate these findings using NV.