Main

Rotavirus is the most frequent etiologic agent of infantile gastroenteritis worldwide, and it is responsible for approximately 1 million deaths per year in developing countries(1). Treatment ofRotavirus gastroenteritis is based upon oral or parenteral rehydration. No specific therapy is currently available againstRotavirus infection.

We have previously reported that enteral administration of human serum immunoglobulin in two children with severe and protracted rotaviral diarrhea was associated with full recovery(2). Subsequently, a double blind, placebo-controlled clinical trial was performed in children with acute rotaviral gastroenteritis to establish the efficacy of oral administration of immunoglobulin. This treatment proved to be effective on all the parameters considered, including the duration and severity of diarrhea, the duration of viral excretion, and the time needed for hospital care(3).

It has been previously shown that Rotavirus is able to infect the human intestinal epithelial cells Caco-2(4). Caco-2 cells derive from a moderately differentiated colon carcinoma(5) and can be grown on a filter, where they form a single layer and produce TEER typical of polarized epithelial cells(6). When confluent, Caco-2 cells undergo a functional enterocytic differentiation, which is confirmed by the presence of microvilli with high levels of alkaline phosphatase and sucraseisomaltase activities(7). Caco-2 infected with Rotavirus showed a decrease in TEER(4). This model appeared suitable to study the Rotavirus-enterocyte interaction and the effects on this of immunoglobulin because the time course events related to the infection may be easily monitored by measuring the TEER(4). We have therefore used this model to establish the kinetics of Rotavirus infection in Caco-2 cells and to investigate the enterocyte-Rotavirus-immunoglobulin interplay.

METHODS

Cell growth and culture. Caco-2 cells were grown in DMEM, with high glucose concentration (4.5 g/L), at 37 °C in 5% CO2 atmosphere as previously described(8). The volume of the apical and basal bathing media was 2.5 and 3.0 mL, respectively, and the medium was supplemented with 10% FCS, 1% nonessential amino acids, penicillin (50 mU/mL), and streptomycin (50 μg/mL). The medium was changed daily. Cells were grown on nontransparent polycarbonate Transwell filters of 24.5-mm diameter with 0.4-μm pores (Costar Packaging Corp. Cambridge, MA) for TEER measurement. Each filter was seeded with 2 × 106 cells. Caco-2 were also grown on 24-well tissue culture plates (Falcon Labware, Becton Dickinson, Oxnard, CA) for microscopic evaluation. Finally Caco-2 cells were grown on glass coverslips for immunofluorescence studies. All the experiments were performed on cells at 15 d postconfluence.

Virus strain and infection protocol. A simianRotavirus strain, SA11, was used, because it is a well characterized strain(9) that grows well in Caco-2 cells(10) and belongs to the third serotype defined for human rotaviruses(11). The strain was kindly supplied by Prof. Donelli (Laboratory of Ultrastructure, Istituto Superiore di Sanità, Rome, Italy). The virus was activated with 20 μg/mL trypsin for 30 min at 37 °C(12). Confluent monolayers of Caco-2 cells were washed twice and incubated overnight in FCS-free medium before virus infection. Viral suspension was added to the apical side of the cell monolayer. After 90 min of incubation at 37 °C, the cells were rinsed three times with DMEM and incubated in serum-free medium for the established times p.i. The timing of infection was started after having removed the excess of viral particles. Viral load was determined by PFU assay using MA104 cells as previously described(11).

TEER measurements. The TEER of cell monolayers grown on filters was measured by Millicell-ERS resistance monitoring apparatus (Millipore S.p.A., Milan, Italy). The background was the value obtained in the absence of cells. The net resistance was calculated by subtracting the background from the actual value and multiplying the value obtained by the area of the filter(4.9 cm2). The resistance is expressed in ohms/cm2. Data are means ± SD.

Cell morphology and viability. Virus-induced cytopathic damage was evaluated by inspection of cell monolayer at subsequent times p.i. through an inverted microscope (Diaphot 2 ELWD 0.3, Nikon Tokyo, Japan).

Cell viability was determined by staining the monolayer with 0.4% trypan blue for 10 min, after washing twice with PBS, and the percent of stained cells was determined as reported previously(13). Data are means ± SD.

Indirect immunofluorescence. Cells grown on glass coverslip were washed three times with PBS and then fixed in 2% paraformaldeide in PBS for 20 min. After fixation, the cells were rinsed and permeabilized with 0.075% w/v of saponin in PBS for 30 min. All solutions used after fixation were supplemented with 0.075% of saponin. The cells were incubated for 30 min at room temperature with a commercially available rabbit polyclonal antibody(Dako B-218) raised against the viral inner capsidic antigen VP-6, or with commercially available human immunoglobulin from a single lot. After washing three times, a goat anti-rabbit rhodamine-conjugated antibody or, respectively, a fluorescein-conjugated rabbit anti-human immunoglobulin (Dako F202), was added, and incubation was carried on for 30 min at room temperature.

The cells were rinsed three times again, mounted in PBS-glycerol solution, and inspected by a fluorescence microscope (Zeiss Axiomat, Oberkochen, Germany). To minimize nonspecific immunoglobulin binding, the human immunoglobulin was preadsorbed on fixed and Triton-permeabilized uninfected Caco-2 cell monolayer.

Immunoglobulin-virus interaction studies. Increasing doses of human serum immunoglobulin were added before, and at various times after, virus-cells coincubation. Preincubation of immunoglobulin withRotavirus was performed in FCS-free DMEM at 37 °C for 2 h. The tools to investigate the effect of immunoglobulin on virus-Caco-2 cells interaction were the following: immunofluorescence, TEER, cell morphology, and viability.

Statistics. Each experiment was performed in duplicate or triplicate and was repeated at least three times. A t test for paired or unpaired data was applied where appropriate.

Chemicals. Human serum immunoglobulins used in our study belonged to a single batch of commercially available human serum immunoglobulin for i.v. use. We have used a preparation with the following specific neutralizing titers: Rotavirus strain Wa (serotype 1), 1:800; strain DS-1 (serotype 2), 1:1600; strain Price (serotype 3), 1:3200; strain ST-3 (serotype 4), 1:1600. Specific titers were determined as reported previously(3). All chemicals were of reagent grade and were obtained from Sigma Chemical Co.-Aldrich srl (Milan, Italy). Media were purchased from Life Technologies, Inc. (Milan Italy). Anti-Rotavirus antibody and rabbit anti-human immunoglobulin serum were obtained from Dako(Glostrup, Denmark). Goat anti-rabbit rhodamine-conjugated antibody was purchased from Jakson (Tecno Genetics, Milan, Italy).

RESULTS

Kinetics of Rotavirus-Enterocyte Interaction

Immunofluorescence. The examination of cells infected with 5 PFU/cell revealed the presence of VP-6 antigen, which was evident from 3 h p.i. At 5 h p.i., the signal appeared to have a granular perinuclear distribution (Fig. 1,A andB). At that time, approximately 90% of cells were VP-6-positive. This pattern remained largely unchanged up to 18 h p.i. After 24 h, increasing cell detachment during washing procedures was observed, which prevented further reliable immunofluorescence evaluation.

Figure 1
figure 1

Indirect immunofluorescence with polyclonal antibody to VP-6 rotaviral antigen in: (A) uninfected Caco-2 cells and(B) cells infected with Rotavirus (5 PFU/cell). Anti-VP-6 polyclonal antibody shows a granular perinuclear distribution in about 90% of permeabilized cells at 5 h p.i. Bar = 10 μm.

Full size image

TEER. At 15 d postconfluence, Caco-2 cells showed a stable TEER value of approximately 600 ohms/cm2. The addition ofRotavirus (5 PFU/cell) to the apical side of Caco-2 cells induced a progressive decrease of TEER (Fig. 2). The effect was initially seen between 20 and 22 h p.i. TEER progressively decreased to reach 0 value in a time range of 60 to 72 h p.i.

Figure 2
figure 2

Effect of input virus multiplicity on TEER of Caco-2 cell monolayer. The TEER value decreased below a detectable level within approximately 36 h p.i. with 25 PFU/cell, 60-72 h p.i. with 5 PFU/cell, and 96 h p.i. with 1 PFU/cell.

Full size image

Increasing the viral load to 25 PFU/cell induced a faster decrease of TEER values, whereas decreasing the viral load to 1 PFU/cell induced a delayed and slower decrease of TEER compared with that observed with 5 PFU/cell(Fig. 2). When 0.5 PFU/cell was added, TEER was stable for at least 36 h p.i. and then started to drop (data not shown). Subsequent experiments were performed by the addition of a standard viral load of 5 PFU/cell to the apical side.

Cytopathic effects and cell viability. At 24 h p.i. the cells were not appreciably different from uninfected control (Fig. 3A). Starting from 36 h p.i., progressive abnormalities of cell morphology were observed. These consisted initially of cell rounding with granular appearance of cytoplasm, leading to a progressive dilation of intercellular spaces. Empty holes, corresponding to cell death and detachment, were subsequently observed up to a complete loss of monolayer pattern(Fig. 3,B andC).

Figure 3
figure 3

Cytopathic effects induced by Rotavirus (5 PFU/cell) in Caco-2 cell monolayer. (A) Uninfected Caco-2 cell monolayer; (B) Caco-2 cells at 48 h p.i.: cellular vacuolization, opening of intercellular junction and spotting cell detachment are observed;(C) Caco-2 cells at 96 h p.i.: extensive cellular detachment is observed with only few picnotic cells yet present. Bar = 10 μm.

Full size image

The results of cell viability experiments were consistent with those assessing the cyotpathic effects. Namely, a time-related increase of cells incorporating trypan blue was observed (Table 1). Less than 30% of cells were still viable after 72 h p.i..

Table 1 Effects on TEER and cell viability by immunoglobulin added at subsequent times before, and at various times after, Caco-2 cell infection with Rotavirus
Full size table

Effects of Immunoglobulin on Virus-Enterocyte Interaction

Immunofluorescence. All the experiments were evaluated at 5 h p.i. When Rotavirus was preincubated with 2.5 mg/mL of immunoglobulin, no viral antigen was detected inside the cells. The same result was obtained with the addition of immunoglobulin to the cells at 0, 5, and 10 min p.i. When immunoglobulin was added 15 min after viral addition, approximately 5-10% of the cells were infected, but when this was added at 20 min p.i., the number of cells showing a positive signal rose to 50%. When immunoglobulin was added to the cells 90 min after Rotavirus addition, most of the cells examined were positive, and the intracellular distribution of viral antigen was similar to that observed in immunoglobulin-free infected cells.

To determine whether serum immunoglobulin was able to directly bind the virus, a human immunoglobulin fluoresceinconjugated antibody was added to infected, fixed, and permeabilized cells. Fluorescence distribution of anti-VP-6 rhodamine-conjugated antibody (Fig. 4A) strongly resembled that seen with human immunoglobulin fluorescein-conjugated antibody (Fig. 4B).

Figure 4
figure 4

Immunofluorescence of Caco-2 cell monolayer infected with Rotavirus (5 PFU cell), incubated with: (A) anti-VP-6 plus rhodamine conjugated anti-rabbit antibody and (B) human immunoglobulin plus fluorescein-conjugated anti-human antibody. Bar = 3μm.

Full size image

TEER. Preincubation of cells with immunoglobulin prevented or hampered the decrease in TEER induced by the virus, in a dose-dependent manner(Fig. 5). When immunoglobulin, in a concentration range of 0.5 to 2.5 mg/mL, was preincubated with the standard viral load of 5 PFU/cell in a total volume of 0.5 mL, the TEER decrease induced byRotavirus was totally abolished (Fig. 5). The effect of preincubation with immunoglobulin on TEER was dependent on immunoglobulin concentration (Fig. 5).

Figure 5
figure 5

Dose dependency of the protective effect of immunoglobulin preincubated with Rotavirus as judged by the modifications of TEER. Increasing doses of immunoglobulin were incubated withRotavirus (5 PFU/cell) before adding the virus to Caco-2 cells and TEER values were monitored in comparison with immunoglobulin-free infected controls. *p < 0.001.

Full size image

In the next set of experiments, a standard immunoglobulin concentration of 2.5 mg/mL was added at various times p.i. The efficacy of immunoglobulin in preventing the fall of TEER was dependent on the times of its addition. Indeed, the early (within 1 h p.i.) addition of immunoglobulin to cells exposed to Rotavirus totally prevented the fall of TEER(Fig. 6). The addition of immunoglobulin at subsequent times p.i. significantly reduced the decrease of TEER, but the effect of immunoglobulin progressively decreased by increasing the times of its addition.

Figure 6
figure 6

Effect of immunoglobulin (2.5 mg/mL) on TEER modifications induced by Rotavirus. Immunoglobulin was added at subsequent times post infection and values of TEER were monitored in comparison with time-matched values from immunoglobulin-free infected control.*p < 0.001.

Full size image

When immunoglobulin was added in a time range of 2 to 8 h p.i., after an initial drop, the values of TEER showed a later shift toward baseline values, suggesting restoration of the monolayer's integrity. On the contrary, when immunoglobulin was added at subsequent times p.i., only a minor effect on TEER was observed, consisting of an increased time for resistance to reach 0 value(Fig. 6).

Cytopathic effects and cell viability. Cells exposed toRotavirus, after its preincubation with immunoglobulin, were indistinguishable from uninfected control. The addition of immunoglobulin up to 8 h p.i. reduced the cytopathic effect of Rotavirus(Fig. 7,A-D). When immunoglobulin was added within 1 h p.i., the integrity of the monolayer was always preserved, in spite of modifications of cellular morphology; cells appeared enlarged and mytotic fuses were present, suggesting active replication (Fig. 7,A andB). Cells treated with immunoglobulin at 4 h p.i. and examined at 48 h p.i. showed only spotting disruption and dilation of intercellular spaces(Fig. 7C). However, junctional continuity was restored by expanded cells at 96 h p.i. (Fig. 7D). Cells treated with immunoglobulin at 12 h p.i. showed a similar cytopathic damage, but delayed in time, in comparison with immunoglobulin-free infected control cells (data not shown).

Figure 7
figure 7

Effect of immunoglobulin on cytopathic changes induced by Rotavirus. (A and B) Addition of immunoglobulin at 1 h p.i. (A) At 48 h p.i. immunoglobulin addition prevents the modifications observed in immunoglobulin-free control (seeFig. 3B for purpose of comparison). (B) At 96 h p.i., immunoglobulin addition prevents the modification observed in immunoglobulin-free control (see Fig. 3C for comparison).(C and D) Addition of immunoglobulin at 4 and 48 h p.i.(C) cell damage is similar to that observed in untreated infected cells, but at 96 h p.i. (D) monolayer integrity is almost completely restored by expanded cells; mitotic fuses are also present. Bar = 10 μm.

Full size image

Viability of cells treated with immunoglobulin at subsequent times p.i. showed a pattern similar to that described for cytopathic effects. Results of cell viability are reported in Table 1. A close similarity was also seen between electrical measurements and cell viability(see Fig. 6 andTable 1).

DISCUSSION

The protective role of immunity against Rotavirus is largely unknown. Rotavirus infection has an age-related pattern. Most children requiring hospitalization because of rotaviral diarrhea are younger than 24 mo(14). A majority of children have developed anti-Rotavirus antibodies within 2 y of age, and they are partially or totally immune to Rotavirus infection(15). Adult subjects can easily be infected by Rotavirus, but they are usually symptomless, and this is probably related to the protective role of specific antibody acquired earlier in life(16).

However, it is now becoming clear that the presence of antibody in the serum does not prevent Rotavirus infection(17). On the contrary, increasing evidence suggests that local rather than systemic immunity plays a crucial role againstRotavirus(18, 19). This is supported by the well known efficacy of breast milk for Rotavirus gastroenteritis(20).

The antirotaviral efficacy of local immunoglobulin resides mainly on clinical grounds(2, 3). However, experimental data obtained in mice suggested that passive immunization with MAb was effective against Rotavirus-induced diarrhea(21, 22).

We have used an in vitro model to investigate the human enterocyte-Rotavirus-immunoglobulin interplay. In this modelRotavirus was able to infect Caco-2 cells, as proved by immunofluorescence, and to induce cell damage which could be monitored either by the drop of the monolayer's TEER or by abnormalities in cell morphology and viability.

In a previous study, Svensson et al.(4) reported that rhesus Rotavirus was able to infect Caco-2 cells and to induce a decrease in TEER, due to an effect on the paracellular pathway limited by tight junctions, whereas only mild cytopathic changes were observed, and cell viability was relatively preserved. We had similar results with only minor quantitative differences. Namely we observed a slower decrease in TEER and more evident cytopathic effects, compared with that reported by Svensson et al.. Some modifications in the model, such as the age of cells or the different pore size of filters or even the different viral strain, may explain the quantitative differences in the results.

Once we established the kinetics of Rotavirus infection in Caco-2 cells, we could study the effects of the immunoglobulin addition. In this model the immunoglobulin was effective in preventing or counteracting the modifications induced by Rotavirus. It did so in a time- and dose-dependent fashion.

When immunoglobulin was preincubated with Rotavirus before its addition to cell monolayers, no viral antigen was detectable by immunofluorescence. Furthermore, no TEER decrease nor cytopathic effect were observed, at least within the time frame of the experiments. These results indicate that cell infection was prevented by the addition of immunoglobulin to viral suspension, which is well in keeping with the reported efficacy of immunoglobulin in preventing rotavirus diarrhea in animal models(23, 24).

When immunoglobulin was added to infected cells, it reduced the modifications induced by Rotavirus. Interestingly, the protective efficacy was gradually lost with time, indicating that, in the absence of local immune response, the infection progressively leads to irreversible monolayer's damage. However, even when a large number of cells were infected, as judged by immunofluorescence, early addition of immunoglobulin was effective in reducing the severity of monolayer damage. When immunoglobulin was added to the cells within 1 h p.i., the monolayer integrity was preserved, as indicated by normal values of TEER, although the cells were infected withRotavirus, as shown by immunofluorescence. When immunoglobulin was added at subsequent times p.i., the TEER decrease was significantly reduced, compared with immunoglobulin-free infected controls, and subsequently, TEER values showed a progressive return toward uninfected control values, indicating restoration of the monolayer's integrity. These results are well in keeping with the reported efficacy of immunoglobulin treatment in children with Rotavirus diarrhea(3, 25).

One mechanism that can explain the prevention of Rotavirus infection by immunoglobulin is the inhibition of viral attachment and penetration, due to specific neutralizing antibodies in the immunoglobulin preparation. The presence of neutralizing antibodies againstRotavirus in the commercially available immunoglobulin preparations has been shown in previous works(3, 26). Neutralizing antibodies produced during Rotavirus infection are directed against the two major viral capsid antigens, VP-4 and VP-7(27). These proteins mediate the virus-cell attachment and internalization(28, 29). MAb to VP-4 and to VP-7 have been shown to be protective against Rotavirus-induced diarrhea in animal models(30), and antibodies to these antigens confer protection in children from natural rotaviral infection(31). It is of note that studies with mutants have shown that the VP-4 gene segment of SA11 is related to virulence(32).

The greater efficacy of immunoglobulin when it was preincubated withRotavirus rather than added after infection, supports the hypothesis that the immunoglobulin efficacy is related to its neutralizing activity against the virus. Whether the neutralization occurs only outside the cell or also inside is unclear at present. Intracellular neutralization activity against Influenzavirus by immunoglobulin A has been recently reported(33), and it could represent a general mechanism of antiviral effect by immunoglobulins.

Furthermore, it has been recently shown that specific IgA MAb that lack neutralizing activity had a protective effect against Rotavirus(34). That study was performed using a “backpack tumor” model, in mice injected with multiply cloned IgA hybridoma cell lines. It was also shown that IgA were inactive when given on the luminal side of the intestine. However, the preparation used in our work was IgA-free and furthermore it was active on the mucosal side.

It is quite possible that immunoglobulin counteracts viral infection also by neutralizing the viral progeny after exocytosis, thus preventing theRotavirus from reaching uninfected cells. This could explain the reported therapeutic efficacy in diarrheal children and also the reduction of viral excretion in treated children. It should be noted that other mechanisms could be active in vivo, including the inhibition of the protease activation, which is critical for efficient viral replication(35) and/or interaction with nonspecific ganglioside(36), or with other constituents of the immune response.

Taken together, these findings suggest that a complex network of humoral immune response operates against Rotavirus infection. The time course experiments had important clinical implications. It was clearly proved that the efficacy of immunoglobulin was related to its early administration. In our previous clinical study, immunoglobulin was administered to children early in the course of rotaviral disease(3). This could be a crucial point for obtaining the maximal efficacy of such treatment. To increase cost-effectiveness, immunoglobulin should be given immediately upon a positive response of rapid Rotavirus fecal antigen testing.

In conclusion, we have shown that human serum immunoglobulin is effective in preventing Rotavirus infection and in reducing virus-induced cell damage, in an in vitro model. These findings are in agreement with the anti-Rotavirus evidence of clinical efficacy of immunoglobulin and support its use in children with gastroenteritis.