INT-407 cells were obtained from the American Type Culture Collection (Manassas, VA, USA) (CCL-6) and grown in GTSF-2 medium (Hyclone, Logan, UT, USA) containing penicillin/streptomycin and amphotericin B. The human colonic adenocarcinoma epithelial cell line (Caco-2; HTB-37) was obtained from the American Type Culture Collection and was cultured in Earle minimum essential medium, supplemented with 10% fetal bovine serum. All cell cultures were grown at 37°C in 5% CO_2.

NV was purified from the stool specimens of infected volunteers by using a CsCl isopycnic ultracentrifugation gradient as described (5). Endotoxin levels were determined by using the Pyrotell Gel-Clot method (Associates of Cape Cod Inc., East Falmouth, MA, USA), performed according to the manufacturer’s protocol (31). NV was inactivated by γ-irradiation at doses of 4.25 kGy at room temperature by using a Gammacell-1000 Irradiator (Best Theratronics Ltd., Ottawa, Ontario, Canada) with a cesium-137 source, at a dose rate of 0.5 kGy/h. The titer of NV RNA in each fraction was determined by immunomagnetic capture reverse transcription (RT) PCR as described (31,34).

The 3-D INT-407 model used in these studies was established by using procedures that were as close as possible to those previously described (30). In brief, the INT-407 cells were initially grown as monolayers in GTSF-2 medium (Hyclone) containing penicillin/streptomycin and amphotericin B at 37°C in 5% CO₂ in preparation for seeding into the rotating wall vessel (RWV) bioreactor (35). A total of ≈2 × 10⁶ cells were subsequently added to the RWV containing 0.25 mg/mL porous Cytodex-3 microcarrier beads (Sigma-Aldrich, St. Louis, MO, USA) at a ratio of ≈10 cells/bead. Cells were then cultured in the RWV at 18–20 rotations/min as described (30,35). Fresh medium was replenished every 24–72 h, depending on metabolic activity of cultures. Intestinal aggregates were harvested for use in all studies 28–35 days after they were seeded into the RWV.

By using a wide-bore pipette (10-mL wide), mature aggregates in 250 μL of GTSF-2 medium were seeded into a 24-well plate (≈1×10⁶ cells/well) and infected with NV at a multiplicity of infection (MOI) of 25 (live or inactivated) for 1 h at 37°C in 5% CO₂. Plates were rocked every 15 min during the 1-h absorption phase and then overlaid with 750 μL of fresh medium. Wells were incubated for 0-, 6-, 24-, 48-, 72-, or 96-h following infection, imaged, and harvested for RNA extraction or immunofluorescence analysis.

RNA was extracted from samples by using the QIAamp Viral RNA Mini Kit (QIAGEN, Valencia, CA, USA). RNA was treated with DNase (Ambion, Foster City, CA, USA) and converted to cDNA by RT by using the iScript cDNA Synthesis Kit (BioRad, Hercules, CA, USA) according to the manufacturer’s instructions. cDNA was stored at −20°C until further analysis.

Quantitative RT-PCR (qRT-PCR) was carried out in 25 μL of a reaction mixture containing 2.5 μL of cDNA, 12.5 μL of Supermix (BioRad), a 400 nmol/L concentration of each primer designed to bind to the open reading frame 1 and 2 junction (36), and 15 pmol of RING1(a)-TaqMan probe and 5 pmol of RING1(b)-TaqMan probe fluorogenic probes for NV genogroup I (GI) detection. PCR amplification was performed with an iQ5 cycler (BioRad) as described (36). For each PCR, an NV GI-specific standard curve was generated by a 10-fold serial dilution (10⁹–10³ copies) of purified NV GI cDNA plasmids.

Phase-contrast images were produced on a Zeiss Axiovert 40 CFL microscope (Carl Zeiss, Thornwood, NY, USA) by using Axiovision 4 software (Carl Zeiss) for image processing. For analysis by confocal microscopy, confluent Caco-2 monolayers, INT-407 monolayers, and 3-D INT-407 aggregates were fixed in 4% paraformaldehyde for 30 min at room temperature. Cells were washed and permeabilized with 0.1% Triton X-100 in Dulbecco phosphate-buffered saline (PBS) and blocked at 37°C for 1 h in 4% bovine serum albumin. Cells were stained with anti-H type 1 antibodies (BG-4 [#SIG-313]; Covance, Princeton, NJ, USA; and UEA-1 lectin [L#8146]; Sigma-Aldrich), anti-H type 2–specific monoclonal antibody (Biogenesis; Poole, UK), anti-Le^a monoclonal antibody (#4861; Immucor Gamma, Atlanta, GA, USA), and anti-Le^b monoclonal antibody (BG-6 [#SIG-3315]; Covance) for 2 h at 37°C at 1:25 dilution. Antiviral protein (VP) 1 and VPg antibodies were used as described (5). Unstained control slides and PBS control samples were also used to distinguish background and establish appropriate laser intensity for visualizing experimental slides. Samples were analyzed by using the Zeiss LSM 510 inverted confocal microscope and a 63× objective and Zeiss LSM software.

DNA was extracted from 200 μL of INT-407 cell pellets by using the QIAamp DNA Blood Mini Kit (QIAGEN). A region of the FUT2 gene was amplified, by using the following primers: forward 5′-CCCATCTTCAGAATCACCCTGCCGGTGCTG-3′, and reverse 5′-TCGGCCGGCCCGTGGAAACATCCCCAGGTA-3′, which anneal at positions 280–309 and 535–564, respectively, as described (18). Results were analyzed as described; in brief, homozygous amino acid at position 428 or TT at position 385 of the FUT2 gene defines a secretor-negative genotype (18,31).

Escherichia coli O111:B4 lipopolysaccharide (LPS) (10⁶ EU/mg) was purchased from InvivoGen (LPS-EB; InvivoGen, San Diego, CA, USA). Aggregates were treated with PBS (negative control) or LPS doses ranging from 0.01 to 100.0 μg/mL and incubated for 24 h at 37°C.

Prism software (GraphPad Software, La Jolla, CA, USA) was used to graph all data and to perform statistical analyses. One-way analysis of variance, followed by Bonferroni post hoc analysis, was used to make statistical comparisons for qRT-PCR analysis. For statistical significance, p<0.05 was considered significant and p<0.01 was considered highly significant.

Straub et al. (30) previously reported the use of RWV-derived 3-D INT-407 models of small intestinal epithelium for successful NV infection, demonstrated in part by cellular vacuolization and detachment after a 24-h challenge with unpurified extracts obtained from human stool specimens that contained genogroup I or genogroup II NoVs (30). To test whether exposure to NV led to gross morphologic changes in 3-D INT-407 aggregates, we prepared a highly purified and concentrated (1.63×10⁹ RNA genomic copies/μL) stock of NV. (Notably, Straub et al. used virus stocks that were clarified by using a 10,000 molecular weight cutoff filter [30]). A γ-irradiated NV stock was used as a negative control for NV infectivity and replication because this treatment inactivates virus but preserves the integrity of the virus and the remaining components of the sample (31). Three-dimensional INT-407 aggregates were inoculated with live NV or γ-irradiated NV or mock-inoculated with media alone (uninfected) at an MOI of 25 for 6, 12, and 24 h and monitored for potential CPE by light microscopy (Figure 1). Twelve hours post inoculation (hpi), morphologic changes initially attributed to CPE were observed in the 3-D aggregates. These observed morphologic changes included changes in size, shape, and clumping of cells, and detachment of cells from the beads on which the cells had been cultured. In Figure 1, the arrows indicate dissociated cells. Larger distinct circular beads without attached intestinal cells are more abundant in the samples showing damage. These CPE-like morphologic changes were not observed in the uninfected or inactivated samples, which suggests that the NV stock was able to induce gross morphologic changes in the 3-D intestinal aggregates.

To determine whether the CPE-like morphologic changes observed in 3-D INT-407 aggregates were a direct result of NV replication, we performed a kinetic analysis of NV RNA levels by qRT-PCR following NV challenge (Figure 2). The 3-D aggregates were challenged at an MOI of 25 (≈4 × 10⁹ copies of NV total) with live or γ-inactivated NV or mock-challenged with media alone. At each time point (0, 6, 24, 48, 72, and 96 hpi), cellular pellets (containing possible intracellular and cell-associated virus) and supernatants (containing input and released virus) were harvested for RNA extraction, and viral RNA titers were quantified by qRT-PCR. NV RNA copy numbers (indicative of viral replication) did not significantly (p>0.05) increase relative to input copies at any of the time points measured in both cell pellets and culture supernatants (Figure 2). The input virus served as a positive control and for establishing baseline RNA levels for comparison with those of the experimental samples. NV RNA consistently decreased over time in this experiment; however, the differences between these decreases and levels of input virus NV RNA copies over time were not statistically significant for either live or γ-irradiated NV inoculations. These data clearly illustrate that viral RNA levels were not increasing over time as would be expected during a productive viral infection.

To further investigate whether the 3-D aggregates were productively infected, we used confocal immunofluorescence microscopy to analyze 3-D INT-407 aggregates challenged with live NV, γ-irradiated NV, or media alone (mock) and monitored for the presence of NV structural (VP1) and nonstructural (VPg) proteins (Figure 3). During productive NV replication, both structural and nonstructural proteins are detected within cells (5). After 24 hpi, VP1 and VPg proteins were detected in the live NV-inoculated aggregates (Figure 3, panels A and B); however, by 48 hpi, evidence for these proteins had markedly diminished and the images resembled the NV-inactivated images (Figure 3, panels C–H). We conclude that the VP1 signal at 24 hpi is most likely a result of input virus and not viral replication, consistent with the qRT-PCR data (Figure 2). The faint VPg signal observed may be nonspecific staining or low-level initial gene expression that does not lead to permissive replication.

To better understand why these 3-D intestinal aggregates were not able to productively support NV replication, we characterized them for HBGA expression. NV has been demonstrated to bind to H-type antigens on the surface of mammalian cells, and this binding has been hypothesized to be the mechanism for NV to enter human cells (5,16,17,19,21). To determine whether 3-D intestinal aggregates contain the FUT2 gene that converts precursor molecules to H-type antigens (21), we performed genotyping analysis on the INT-407 intestinal cells used to produce the organotypic cell culture model. The intestinal cell line was shown to contain the genetic determinants to express a functional FUT2 (data not shown); therefore, the INT-407 cell line (used to make the 3-D intestinal model) is likely a secretor-positive cell line with the potential to express the H-type receptors necessary for NV binding and, presumably, for infection. However, after performing H-type antigen phenotyping by confocal immunofluorescence microscopy on monolayers and 3-D differentiated intestinal cells, we were unable to detect H type 1 on the surface of the cells by using 2 antibodies specific for different regions of the antigen (Figure 4) (5). Further confocal analysis of the INT-407 cells also showed no expression of the other carbohydrate antigens involved in binding of other genogroups of NoV, H type 2, Le^b, although minimal antigen expression of Le^a was observed. Caco-2 cells are shown as a positive control for staining of the surface glycans (5). These data demonstrate that both monolayer and 3-D intestinal models lack sufficient expression of the cell surface receptors believed to be required for susceptibility to NV infection.

Our data clearly demonstrate that the 3-D intestinal aggregates do not support a productive NV replication; however, we did observe CPE-like morphologic changes after treatment with highly purified NV (Figure 1). One potential cause of these morphologic changes in the 3-D intestinal aggregates could be the presence of contamination of the NV stocks with a bacterial endotoxin, LPS, or other potentially cytotoxic molecules commonly found in human stool samples (the source of virus used in our studies and those of Straub et al. [30]). The highly purified virus preparations used in this study were determined to contain low LPS levels, resulting in 10 ng/mL and 1 ng/mL of LPS for live NV and γ-irradiated NV stocks, respectively, added to the aggregates on NV inoculation. LPS is a known cytotoxic inflammation-inducing product; we therefore hypothesized that the contaminating LPS could be a potential cause of the CPE-like morphologic changes observed after NV inoculation.

To test whether LPS alone induces comparable CPE-like effects at relevant concentrations in the 3-D intestinal aggregates, we treated the aggregates with increasing amounts of LPS (InvivoGen) and monitored for alterations in morphology indicative of cytotoxicity (Figure 5). After 24 h, aggregates treated with a range of LPS concentrations (0.01–100.0 µg/mL) exhibited increasing signs of CPE, as shown by cellular dissociation from the support beads and a partial loss of cell viability (Technical Appendix Figure 1). In comparison with the morphologic changes shown in the technical appendix figure (observed with NV addition to the aggregates), we found that LPS at levels of ≈1 µg/mL was able to induce the CPE-like morphologic changes equivalent to those induced by inoculation of nonpurified NV. Although these levels are higher than the measured level of LPS in the purified NV samples we used, there might also be additional cytotoxic contaminants that were carried from the infected patients’ stool specimens. When these data are taken together, we interpret them as indicating the CPE-like morphologic changes were caused by stool-derived NV samples (Technical Appendix Figure and [30]) contaminated with LPS, and perhaps other cytotoxins, and were not a result of virus replication.