Secretory Immunoglobulin A Antibodies against the σ1 Outer Capsid Protein of Reovirus Type 1 Lang Prevent Infection of Mouse Peyer's Patches

Abstract

Reovirus type 1 Lang (T1L) adheres to M cells in the follicle-associated epithelium of mouse intestine and exploits the transport activity of M cells to enter and infect the Peyer's patch mucosa. Adult mice that have previously cleared a reovirus T1L infection have virus-specific immunoglobulin G (IgG) in serum and IgA in secretions and are protected against reinfection. Our aim in this study was to determine whether secretory IgA is sufficient for protection of Peyer's patches against oral reovirus challenge and, if so, against which reovirus antigen(s) the IgA may be directed. Monoclonal antibodies (MAbs) of the IgA isotype, directed against the σ1 protein of reovirus T1L, the viral adhesin, were produced and tested along with other, existing IgA and IgG MAbs against reovirus T1L outer capsid proteins. Anti-σ1 IgA and IgG MAbs neutralized reovirus T1L in L cell plaque reduction assays and inhibited T1L adherence to L cells and Caco-2_BBe intestinal epithelial cells in vitro, but MAbs against other proteins did not. Passive oral administration of anti-σ1 IgA and IgG MAbs prevented Peyer's patch infection in adult mice, but other MAbs did not. When anti-σ1 IgA and IgG MAbs were produced in mice from hybridoma backpack tumors, however, the IgA prevented Peyer's patch infection, but the IgG did not. The results provide evidence that neutralizing IgA antibodies specific for the σ1 protein are protective in vitro and in vivo and that the presence of these antibodies in intestinal secretions is sufficient for protection against entry of reovirus T1L into Peyer's patches.

Reovirus type 1 Lang (T1L) adheres selectively to the apical surfaces of M cells in the follicle-associated epithelium of mouse intestine and exploits the transepithelial transport activity of M cells to enter Peyer's patch mucosa and initiate infection (3, 46, 53). Adherent viruses that are transcytosed by M cells subsequently are taken up by phagocytic cells of the Peyer's patch mucosa (26, 52) or infect epithelial cells from the basolateral side (9, 56). Adult mice respond to a mucosal reovirus infection with a vigorous immune response, including virus-specific cytotoxic T lymphocytes, serum immunoglobulin G (IgG) antibodies, and secretory IgA (S-IgA) antibodies (38, 39, 57, 60). Both cytotoxic T lymphocytes and serum antibodies have been shown to contribute to clearance of the mucosal infection (6, 64, 65), and in normal adult mice, the infection is cleared within about 10 days (39).

Silvey et al. have recently demonstrated that adult mice that had previously cleared a reovirus T1L infection and were orally rechallenged were completely protected against Peyer's patch reinfection (60). At the time of challenge, the protected mice had antireovirus IgG in serum and IgA in secretions. In contrast, IgA-deficient mice effectively cleared the initial infection, but when orally rechallenged their Peyer's patches became reinfected despite high levels of antireovirus IgG in serum (60). These results suggested that S-IgA is required for complete mucosal protection, but they failed to demonstrate directly the protective capacity of S-IgA in the absence of other immune protection mechanisms. Furthermore, these studies did not prove that secretion of antibodies is essential for prevention of Peyer's patch infection, since IgA as well as IgG antibodies are normally present within mucosal tissues (42), where they would likely neutralize reovirus that had entered the mucosa.

S-IgA is the most abundant immunoglobulin on the intestinal mucosal surface, and S-IgA antibodies are known to play an important role as a first line of defense against adherence and invasion by enteric pathogens (42). The exact mechanisms through which IgA exerts its protective function are only partly understood. There is evidence that S-IgA prevents contact of pathogens with mucosal surfaces by facilitating entrapment in mucus and subsequent peristaltic or ciliary clearance (22, 36, 59). IgA may also sterically hinder the microbial surface proteins that mediate epithelial attachment (61), intercept incoming pathogens within epithelial cell vesicular compartments (13, 14, 36, 40), or mediate export of pathogens back into the lumen (35, 42). Numerous studies (reviewed in reference 44) have demonstrated that protection against mucosal infections by viruses is associated with the presence of virus-specific IgA in secretions.

On the other hand, there is evidence that mucosal protection can be provided by serum IgG and that S-IgA is not essential (11, 21, 23, 50, 53). For example, in an IgA-deficient mouse model (33), the presence of virus-specific IgG antibodies was correlated with protection against influenza virus infection of respiratory epithelium (41), herpes simplex virus infection of vaginal epithelium (54), and rotavirus infection of intestinal epithelium (51). Although these studies appear to conflict with the observation that serum IgG was not sufficient for protection against reovirus T1L infection of Peyer's patches (60), the differing results are likely to reflect the widely differing tropisms and pathogenetic strategies of the viruses tested. Reovirus T1L infection in the adult mouse is a particularly informative model, because entry of virus into Peyer's patches can be detected and quantitated within hours after oral challenge (60). This allows one to test the effects of specific IgA and IgG antibodies on very early events in M-cell-mediated mucosal entry and infection.

The capacity of antibodies to protect against mucosal infection is also dependent on their antigen specificities. Evidence from bacterial infection models has suggested that IgA directed against any antigen exposed on the microbial surface can protect the intestinal mucosa by entrapping pathogens in secretions and preventing mucosal contact (5, 43, 73). Three proteins in the outer capsid of native reovirus virions are exposed on the particle surface and subject to structural changes as the particles enter the mouse intestine: an outer icosahedral layer of σ3 protein monomers, an underlying icosahedral layer of μ1 protein trimers, and σ1 protein trimers at the icosahedral vertices (25, 27, 37, 46, 49). Upon contact with proteases in the intestinal lumen, σ3 is removed, μ1 is internally cleaved but remains on the viral particle, and σ1 undergoes a conformational change and probably extends up to 40 nm from the particle surface (12, 25, 27, 37, 45, 46, 49). In mice, this conversion to intermediate subviral particles (ISVPs) is required for adherence to M-cell surfaces (3) and for initiation of Peyer's patch infection (8, 12). The σ1 protein mediates viral adherence to several types of cells in culture (7, 18, 37, 48, 58, 70), and a recent report by Helander et al. has demonstrated that the σ1 protein of reovirus T1L is also responsible for adherence of T1L ISVPs to apical surfaces of M cells (34).

The μ1 and σ1 proteins are highly exposed on the surfaces of ISVPs (25) and, thus, should be highly accessible to antibodies in the intestinal lumen. If IgA coating is sufficient to entrap and clear reovirus ISVPs, then S-IgA antibodies against either μ1 or σ1 should be protective. Silvey et al. previously reported that IgA or IgG monoclonal antibodies (MAbs) specific for μ1 or σ3, when fed intragastrically to adult mice along with reovirus T1L, were incapable of preventing viral entry and Peyer's patch infection, whereas the neutralizing IgG MAb 5C6 specific for T1L σ1 (64, 66) was protective (60). The capacity of passively fed anti-σ1 IgG to protect against oral reovirus challenge may not be relevant to natural reovirus infection, however, because IgG is not normally present in significant amounts in the adult mouse intestine (30). An anti-σ1 IgA MAb was not available at the time of the previous study, so the investigators could not determine whether S-IgA against the σ1 protein may be sufficient for protection against reovirus T1L infection of Peyer's patches.

In this study we produced IgA MAb 1E1 directed against the σ1 protein of reovirus T1L. This new reagent, along with other, existing IgA and IgG MAbs against reovirus T1L outer capsid proteins, were tested for their capacities to block viral adherence to cells in vitro and to prevent Peyer's patch infection in vivo. Both anti-σ1 IgA MAb 1E1 and anti-σ1 IgG MAb 5C6 showed neutralizing activity, blocked adherence of T1L ISVPs to L cells and Caco-2_BBe intestinal epithelial cells in culture, and prevented Peyer's patch infection in adult mice when administered perorally along with virus. Anti-σ1 IgA MAb 1E1 also prevented Peyer's patch infection when produced in mice bearing hybridoma backpack tumors, but anti-σ1 IgG MAb 5C6 did not. The results provide evidence that the presence of neutralizing IgA antibodies specific for the σ1 protein in intestinal secretions is sufficient for protection against entry of reovirus T1L into Peyer's patches.

MATERIALS AND METHODS

Virus production, purification, and biotinylation.

Reoviruses T1L and type 3 Dearing (T3D) were laboratory stocks derived from ones from the Bernard N. Fields laboratory. Type 2 SV59 (T2S59) was a laboratory stock derived from one from the Terence S. Dermody laboratory. Reassortant virus strains containing the T1L S1 genome segment on a T3D background (3HA1) or the T3D S1 on a T1L background (1HA3) were also derived from stocks from the Fields laboratory (70). Recoated viral cores were prepared from purified reovirus T1L cores and insect cell lysates containing recombinant T1L μ1 and T1L σ3, plus or minus T1L σ1 or T3D σ1, as described previously (17, 18). Concentrations of the recoated particles were determined by densitometry of Coomassie brilliant blue-stained gels (17).

Virus was grown in mouse L929 fibroblast cells in suspension culture (25). Purified virions were prepared using second-passage L929 cell lysates, and ISVPs were generated by digestion with tosyl lysine chloromethyl ketone-treated α-chymotrypsin (Sigma, St. Louis, Mo.) as previously described (29). Concentrations of virions, ISVPs, and cores were calculated from the absorbance at 260 nm, and infectious titers were measured by plaque assay (63). The purity of virus preparations and conversion of native virus particles to ISVPs was verified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 10% reducing gels (12). For biotinylation of purified virions or ISVPs, particle suspensions were diluted to 0.5 mg of protein per ml in phosphate-buffered saline (PBS; pH 8.4), and 240 μl of a 5-mg/ml solution of EZ-Link sulfo-NHS-biotin (Pierce, Rockford, Ill.) was added to each milliliter of virus suspension. The mixture was incubated at room temperature for 4 h and then dialyzed at 4°C overnight.

Production of an IgA MAb directed against the reovirus σ1 protein.

Adult female BALB/c mice were obtained from Charles River Laboratories (Wilmington, Mass.) and maintained in the animal resource facility at Children's Hospital. All animal procedures were conducted in strict compliance with the Guidelines for Animal Experimentation established by Harvard Medical School, the Children's Hospital, and the National Institutes of Health. For production of monoclonal antireovirus IgA antibodies, 4- to 6-week old BALB/c mice were immunized intragastrically with 2 × 10⁷ PFU of reovirus T1L in 500 μl of PBS, by peroral inoculation using a feeding needle. On day 10, mice were anesthetized by intraperitoneal administration of Avertin, 250-mg/kg (2,2,2-tribromoethanol) (Aldrich, Milwaukee, Wis.), and sacrificed by cervical dislocation. Small intestines were removed and placed in incomplete Dulbecco's minimal essential medium (DMEM; Cellgro; MediaTech, Herndon, Va.) on ice, and Peyer's patches were excised. Peyer's patch cells were isolated, pooled, and fused with P3X63/Ag8U.1 mouse myeloma cells as previously described (71). Fusion products were seeded in 96-well flat-bottom plates with a feeder layer of thymocytes isolated from DBA/2 mice.

Clones were screened for reovirus-specific antibodies using enzyme-linked immunosorbent assay (ELISA) plates coated with whole reovirus virions as described previously (66). Undiluted samples of hybridoma supernatants were added to the wells and allowed to react overnight at 4°C. After washing in PBS containing 0.05% Tween (PBS-Tween), secondary antibodies (biotinylated goat anti-mouse IgA, IgG, or IgM; Southern Biotechnology Associates, Birmingham, Ala.) were added at a 1:3,000 dilution in PBS-Tween containing 5% goat serum (blocking buffer). Bound antibody was detected with streptavidin-horseradish peroxidase (HRP; Pierce) diluted 1:1,000, using a one-component peroxidase substrate system (Kirkegaard & Perry Laboratories, Gaithersburg, Md.). Plates were read at 650 nm in a SpectraMax 250 plate reader using the Softmax ELISA analysis program (Molecular Devices, Sunnyvale, Calif.). Reovirus-specific clones were expanded and cloned three times by limiting dilution. Cloned, stable hybridoma cell lines were grown in T-162 vent-cap tissue culture flasks (Costar, Corning, N.Y.) in RPMI 1640 medium (Gibco BRL, Gaithersburg, Md.) containing 10 mM HEPES, 24 mM NaHCO₃, and 1 mM sodium pyruvate and supplemented with 10% fetal calf serum (FCS; HyClone, Logan, Utah) as well as 2 mM glutamine, 100 U of penicillin/ml, and 100 μg of streptomycin (Gibco)/ml.

Dot blot and Western blot assays.

To identify IgA hybridoma clones producing antibodies against the σ1 protein, supernatants from reovirus-specific IgA clones were applied to paired dot blots of T1L virions or recoated cores containing the μ1 and σ3 proteins but lacking σ1 (17). Separate aliquots (3 μl) of each particle type (10¹³ particles/ml) were applied to individual squares of nitrocellulose paper and allowed to air dry. After nonspecific protein binding sites were blocked with PBS-Tween containing 5% FCS, squares were incubated with selected hybridoma supernatants, washed in blocking buffer, and then incubated with biotinylated goat anti-mouse IgA or anti-mouse IgG (1:3,000; Southern Biotechnology Associates) followed by streptavidin-HRP (1:1,000; Pierce). After washing, blots were developed with the Opti-4CN kit (Bio-Rad, Hercules, Calif.).

For Western blotting, gradient-purified reovirus was boiled in sample buffer (0.5 M Tris-HCl, 2% β-mercaptoethanol, 0.1% bromophenol blue, 20% glycerol, 4% SDS). Viral proteins were separated by electrophoresis on 10% polyacrylamide gels and transferred to nitrocellulose (Bio-Rad). Strips were blocked in PBS containing 5% FCS and 0.1% Tween, incubated with undiluted hybridoma supernatants from antireovirus IgA clones for 2 h at room temperature, washed in PBS-Tween, and then incubated with biotinylated goat anti-mouse IgA, IgG, or IgM (1:3,000; Southern Biotechnology Associates). Strips were washed, incubated with streptavidin-HRP (1:500), and developed with the Opti-4CN kit.

Radioimmunoprecipitation.

The specificity of the 1E1 antireovirus IgA hybridoma clone was tested by radioimmunoprecipitation from radiolabeled reovirus T1L-infected L-cell lysates. L-cell monolayers grown in T-25 flasks (Costar) were infected with reovirus T1L at 1 PFU/cell and incubated for 18 h at 37°C. The culture medium was removed, cells were rinsed with PBS, and cells were incubated in DMEM containing 2% dialyzed FCS and 50 μCi of [³⁵S]Met-Cys (Sigma)/ml for 2 h at 37°C. Cells were then rinsed with PBS and lysed by freeze-thaw in lysis buffer (1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate, 50 mM HEPES, 300 mM NaCl) with protease inhibitors (1% [vol/vol] aprotinin, 1 mM phenylmethylsulfonyl fluoride, 5 μg of leupeptin/ml, and 40 μg of bestatin [Sigma]/ml) and pelleted (10,000 × g; 4°C for 2 min). Radiolabeled lysates were precleared by incubation with protein G-coupled Sepharose 6MB (Pharmacia, Piscataway, N.J.), and aliquots were then incubated with hybridoma supernatants and protein G-Sepharose coated with either goat anti-mouse IgA or goat anti-mouse IgG for 2 h at 4°C. Beads were repeatedly washed and boiled in sample buffer, and released proteins were analyzed by SDS-PAGE on 10% gels run under reducing conditions. Gels were fixed, dried, and autoradiographed using X-ray film (Kodak, Rochester, N.Y.).

Other MAbs used in this study.

Existing IgA and IgG MAbs used in this study are listed in Table 1. These included antireovirus T1L IgA MAbs specific for σ3 (RB3) and μ1 (RB8) and IgG MAbs specific for σ3 (10G10), μ1 (4A3), and σ1 (5C6) of T1L (66). The two antireovirus IgA MAbs were previously generated by fusion of Peyer's patch cells after mucosal immunization of BALB/c mice with T1L (71). Control MAbs included IgA MAb 1E10 and IgG MAb P1B8 directed against the cholera toxin B subunit (CTB) (4) and IgA MAb CrA-2 directed against 15-kDa and 60-kDa proteins of Cryptosporidium parvum sporozoites (15). All of the IgA MAbs used in this study were produced by cloned hybridoma cells as a mixture of monomers, dimers, and higher polymers, determined by nonreducing SDS-PAGE as previously described (71).

TABLE 1.

Mouse MAbs used in this study

Clone	Isotype	Specificity	Reference
1E1	IgA	Reovirus T1L σ1	This study
5C6	IgG 2a	Reovirus T1L σ1	66
RB8	IgA	Reovirus T1L μ1	71
4A3	IgG 2b	Reovirus T1L μ1	66
RB3	IgA	Reovirus T1L σ3	71
10G10	IgG 2a	Reovirus T1L σ3	66
P1E10	IgA	CTB subunit	4
P1B8	IgG	CTB subunit	4
CrA-2	IgA	C. parvum	15

Plaque reduction assays.

Virus in a second- or third-passage cell lysate stock was diluted in PBS supplemented with 2 mM MgCl₂ and 0.4% (wt/vol) bovine serum albumin (Sigma) (PBS⁺-BSA) to an infectious concentration of 150 or 200 PFU per 100 μl. Clarified hybridoma supernatant of IgA MAb 1E1 was diluted to an approximate concentration of 2 mg of IgA/ml. Aliquots of this MAb 1E1 solution, or PBS⁺-BSA (no-MAb control), were then added to the diluted virus stock as 1/20 of the final volume, and the mixtures were incubated for 1 h at room temperature. The medium was removed from L929 monolayers in the wells of a six-well cluster plate, each monolayer was inoculated with 100 μl of the virus-MAb or control mixture, and virus-cell absorption was allowed to proceed for 1 h at room temperature. For each mixture, three wells were inoculated. After absorption, the wells were overlaid with a 1:1 (vol/vol) mixture of completed 2× 199 medium and 2% agar (Difco) and analyzed by standard plaque assay (29). Plaques were counted after staining with neutral red (Sigma). The mean value from the three wells of each mixture was used in calculating the relative plaque number in the matched MAb and no-MAb samples.

Viral adherence to L929 cells and inhibition by reovirus-specific MAbs.

L929 cells were harvested from T-75 flasks and seeded onto 12-mm glass coverslips in 24-well plates at approximately 5 × 10⁴ cells per well and cultured overnight at 37°C in a 5% CO₂ atmosphere in DMEM containing 25 mM glucose supplemented with 10% FCS, 2 mM glutamine, 100 U of penicillin/ml, and 100 μg of streptomycin/ml. Cells were then fixed in PBS containing 3% paraformaldehyde. Reactive aldehydes were quenched with PBS containing 50 mM NH₄Cl for 10 min, and nonspecific protein binding sites were blocked in buffered gelatin-saline (137 mM NaCl, 19 mM H₃BO₃, 0.87 mM MgCl₂, 0.27 mM CaCl₂, 0.13 mM Na₂B₄O₇, pH 7.2, with 0.3% [wt/vol] type A porcine gelatin [Sigma]). Aliquots of biotinylated T1L virions or ISVPs were suspended at 3 × 10¹¹ particles/ml in gelatin-saline and preincubated with reovirus-specific IgG or IgA MAbs at 10, 50, or 100 μg/ml for 2 h at 4°C. Coverslips with L929 cells were inverted onto 50-μl droplets of the ISVP-MAb mixtures and incubated for 2 h. Coverslips were then thoroughly washed in PBS and again immersed in fixative solution for 15 min, quenched, and blocked as above. Bound virus or ISVPs were visualized by incubating the coverslips with streptavidin-fluorescein isothiocyanate (FITC; Pierce), diluted 1:1,000 in gelatin-saline containing 1% goat serum, for 1 h. Coverslips were then washed and mounted onto microscope slides with Moviol (Calbiochem, San Diego, Calif.) containing 2.5% (wt/vol) 1,4-diazabicyclo[2.2.2]octane (Sigma). Slides were examined and photographed using a Zeiss Axiophot microscope (Carl Zeiss, Thornwood, N.Y.) equipped for epifluorescence and phase contrast.

Viral adherence to Caco-2_BBe cell monolayers and inhibition by reovirus-specific MAbs.

The Caco-2_BBe intestinal adenocarcinoma cell clone (55) was obtained from the American Type Culture Collection (Manassas, Va.). Cells were grown in T-75 flasks (Costar) at 37°C in a 5% CO₂ atmosphere in DMEM containing 10% FCS, 2 mM glutamine, 100 U of penicillin/ml, 100 μg of streptomycin/ml, and 10 μg of holo-transferrin (Sigma)/ml. Approximately 10⁵ cells were seeded onto 12-mm glass coverslips (Bellco, Vineland, N.J.), precoated with 1% (wt/vol) rat tail collagen in 60% (vol/vol) ethanol, in 24-well tissue culture plates (Costar). The medium was replaced 8 h after seeding and then every second day. At 2 and 6 days postconfluence, Caco-2_BBe cell monolayers were fixed by immersion in 3% paraformaldehyde (Electron Microscopy Sciences, Fort Washington, Pa.) in PBS for 30 min and rinsed with PBS. Free aldehydes were quenched, and nonspecific protein binding sites were blocked as described above.

To test the capacity of various antireovirus T1L MAbs to inhibit binding of T1L, aliquots of biotinylated T1L ISVPs were suspended at 3 × 10¹¹ particles/ml in gelatin-saline and preincubated with antireovirus IgA or IgG MAbs at 1, 10, 50, 150, or 300 μg/ml for 2 h at 4°C. For determination of the serotype specificity of MAb 1E1, aliquots of biotinylated ISVPs derived from various reovirus strains, reassortants, or recoated cores were preincubated with reovirus T1L σ1-specific MAbs at 10 or 50 μg/ml as above. In all cases, the coverslips with Caco-2_BBe cells were inverted onto 50-μl droplets of the ISVP-MAb mixtures and incubated for 2 h. Coverslips were then washed, immersed in fixative solution for 15 min, and blocked as described above. Bound ISVPs were visualized with streptavidin-FITC and photographed as described above.

Passive oral protection assay.

Anti-σ1 IgA MAb IE1, anti-σ1 IgG MAb 5C6, and anti-μ1 IgA MAb RB8 were tested by passive peroral feeding using a protocol established previously (60). Aliquots of virus were preincubated with antibodies by mixing 50 μg of antireovirus IgA or IgG MAb and 2 × 10⁷ PFU of reovirus T1L in 500 μl of PBS for 30 min at 37°C. This corresponds to a ratio of approximately 6 × 10⁴ IgA dimers or 1.2 × 10⁵ IgA monomers per viral particle and 1.2 × 10⁵ molecules of IgG per viral particle. Unanesthetized adult BALB/c mice were inoculated intragastrically with the mixtures, and 8 h later Peyer's patches were collected for plaque assay as described below.

Backpack tumor protection assay.

Hybridoma cells producing MAbs specific for σ1 (IgA clone 1E1 and IgG clone 5C6) and control MAbs specific for CTB (IgA clone P1E10 and IgG clone P1B8) were used. Hybridoma cells grown in tissue culture flasks were washed in sterile PBS, and 10⁶ cells in 0.5 ml of PBS were injected subcutaneously on the upper back of each adult BALB/c mouse. In most (but not all) mice, tumors were visible at the injection site on days 10 to 14. Mice that had tumors approximately 1 to 3 cm in diameter were selected for challenge. Mice were inoculated intragastrically with 2 × 10⁷ PFU of reovirus T1L in 500 μl of PBS, and Peyer's patches were collected 8 h later. Serum and feces were collected from all mice prior to injection of tumor cells and on the day of challenge. Feces were placed in preweighed microcentrifuge tubes containing 1 ml of PBS containing 0.5% nonfat dry milk (wt/vol) and protease inhibitors [aprotinin, 1 μg/ml; leupeptin, 5 μg/ml; 4-(2-aminoethyl)benzenesulfonyl fluoride, 48 μg/ml (Calbiochem); and bestatin, 1 μg/ml]. Fecal pellets were disrupted by vortexing, and supernatants were obtained by centrifugation in an Eppendorf Microfuge at maximum speed for 20 min at 4°C. Aliquots of fecal supernatants were stored at −20°C.

Quantitation of virus in Peyer's patch tissue.

Mice were anesthetized by intraperitoneal administration of Avertin and sacrificed by cervical dislocation. Small intestines were removed, and the lumens were rinsed with cold DMEM. Peyer's patches were excised and collected in preweighed microcentrifuge tubes containing 1 ml of gelatin-saline with 2% Fungibact (Irvine Scientific, Irvine, Calif.), and tissue weight was determined. Tissues were disrupted by freeze-thawing twice followed by probe sonication. Tissue plaque assays were done as previously described (60, 68), and the concentration of infectious virus was expressed as PFU per milligram of original tissue.

Evaluation of total immunoglobulins and reovirus-specific antibodies in serum and feces.

For evaluation of total IgA and IgG, 96-well flat-bottom plates (Nunc Maxisorb) were coated with goat anti-mouse IgA (Southern Biotechnology Associates) or goat anti-mouse IgG (Cappel, Durham, N.C.). Plates were washed in PBS-Tween and blocked with PBS-Tween containing 5% goat serum (blocking buffer). Serial twofold dilutions of fecal supernatants or serum were applied in duplicate to the plates, along with standards. Standards were purified mouse monoclonal IgA (Southern Biotechnology Associates) or purified mouse serum IgG (Sigma). After washing with PBS-Tween, secondary biotinylated goat anti-mouse IgA or IgG was added at 1:3,000 dilution in blocking buffer. Bound antibody was detected with a 1:5,000 dilution of streptavidin-HRP and the one-component peroxidase substrate system. Plates were read at 650 nm in a SpectraMax 250 plate reader using the Softmax ELISA analysis program.

For measurement of antireovirus IgG and IgA, 96-well flat-bottom plates were coated overnight with 2 × 10¹¹ reovirus T1L viral particles/ml in PBS at 4°C in a humidified chamber. Plates were washed in PBS-Tween followed by blocking buffer. Serial twofold dilutions of serum and fecal supernatants in blocking buffer were applied in duplicate. Known concentrations of MAbs 1E1 (anti-σ1 IgA) and 5C6 (anti-σ1 IgG) were used as standards for antireovirus IgA and IgG ELISAs, respectively. After washing in PBS-Tween, secondary biotinylated goat anti-mouse IgA or IgG was added at a 1:3,000 dilution in blocking buffer, and bound antibody was detected and measured as described above.

Statistics.

The Statview 5.0.1 computer system program (Abacus Concepts, Berkeley, Calif.) was used for all calculations and statistical analyses. Results were logarithmically transformed to obtain geometric means. Between-group comparisons were performed by unpaired two-tailed t test at the 99% confidence level. Differences were considered significant only if P values were 0.01 or less.

RESULTS

Production of an IgA MAb against the reovirus T1L σ1 protein.

To obtain monoclonal, dimeric IgA antibodies against the σ1 protein of reovirus T1L, hybridomas were generated by fusion of Peyer's patch cells after oral inoculation of adult BALB/c mice with reovirus T1L. Screening of hybridoma culture supernatants by ELISA with purified T1L virions, followed by further cloning and expansion, resulted in seven stable antireovirus IgA hybridoma cell lines. When tested on Western blots of reovirus proteins separated by reducing SDS-PAGE, supernatants from all seven clones failed to recognize viral proteins, suggesting that they recognize discontinuous epitopes. When tested on dot blots of intact, unreduced viral particles, the supernatant from IgA hybridoma clone 1E1 reacted with T1L virions, but not with T1L recoated cores lacking σ1 (17) (Fig. 1), suggesting the specificity of this MAb for σ1. Supernatant from IgG hybridoma clone 5C6, which is known to be specific for T1L σ1 (66), showed a binding pattern identical to that of 1E1. The other six IgA hybridoma supernatants failed to recognize viral proteins in this assay (Fig. 1).

FIG. 1.

Binding of antireovirus MAbs to whole reovirus T1L or recoated T1L viral cores lacking σ1. Reovirus T1L particles (A) or T1L cores recoated with σ3 and μ1 but lacking σ1 (B) were dotted onto nitrocellulose paper and incubated with hybridoma supernatants (rows 1 to 3), mouse fecal extract (row 4), or buffer (rows 5 and 6) followed by biotinylated anti-IgA or -IgG secondary antibodies and HRP-streptavidin. Row 1, antireovirus IgA MAb 2F3 (shown here) and five other IgA MAbs (data not shown) failed to recognize either whole virus or recoated cores. Row 2, IgA MAb 1E1 recognized whole virus but failed to recognize recoated cores lacking σ1. Row 3, IgG MAb 5C6, specific for the T1L σ1 protein, recognized whole virus but not recoated cores. Row 4, pooled fecal extracts from BALB/c mice orally immunized with reovirus T1L contained IgA that recognized both whole virus and recoated cores. Rows 5 and 6, control dots exposed to anti-mouse IgA (row 5) or IgG (row 6) secondary antibodies.

The specificity of IgA MAb 1E1 for σ1 was investigated further by immunoprecipitation of radiolabeled, reovirus T1L-infected L-cell lysates and separation of radiolabeled precipitates by SDS-PAGE. In this assay the anti-σ1 IgG MAb 5C6 served as positive control, and anti-μ1 IgA MAb RB8 and anti-C. parvum IgA MAb CrA-2 served as negative controls. The precipitates obtained with CrA-2 contained a faint band of radiolabeled protein corresponding in mobility to reovirus protein σ2 (Fig. 2, lane 2). This band was present at comparable intensity in all lanes, suggesting that the σ2 protein interacted nonspecifically with mouse IgA and/or the Sepharose beads under the conditions of this experiment. The CrA-2 precipitates also showed a faint, nonspecific σ3 band but did not contain detectable amounts of σ1 protein. Anti-μ1 IgA MAb RB8 precipitates showed no detectable σ1 but contained bands corresponding in mobility to μ1 and σ3 (Fig. 2, lane 4). This is consistent with previous evidence that these two major components of the outer capsid are largely associated in infected cell lysates (37, 71). Precipitation by IgA MAb 1E1 resulted in the same pattern of bands as the anti-σ1 IgG MAb 5C6, including a band corresponding in mobility to σ1 (Fig. 2, lanes 3 and 5). Both 5C6 and 1E1 precipitates contained a band corresponding to σ3, suggesting that under the conditions used here, σ1 and σ3 failed to dissociate completely. Taken together, the immunoprecipitation data were consistent with the hypothesis that IgA MAb 1E1 recognizes the σ1 protein.

FIG. 2.

Immunoprecipitation of radiolabeled reovirus T1L proteins by antireovirus and control MAbs. Lysates of radiolabeled L cells infected with reovirus T1L were incubated with hybridoma supernatants and protein G-Sepharose coated with goat anti-mouse IgA or IgG, and precipitates were analyzed by SDS-PAGE. Lane 1, total proteins from infected L-cell lysates show major reovirus bands. Lane 2, control IgA MAb CrA-2 (anti-C. parvum) did not specifically precipitate reovirus proteins. Faint bands corresponding to σ2 (present in all lanes) and σ3 (in this lane) indicate low-level, nonspecific interactions of these proteins with mouse IgA. Lane 3, MAb 5C6 (anti-σ1 IgG) immunoprecipitates contained a protein corresponding in mobility to σ1. Since this MAb is known to be specific for the head region of σ1, the σ3 band in this lane indicates that in the presence of anti-σ1 antibody, the two proteins remained associated in cell lysates. Lane 4, MAb RB8 (anti-μ1 IgA) coprecipitated σ3 and μ1c (band not shown) but did not precipitate σ1. Lane 5, IgA MAb 1E1 precipitates were identical to those of 5C6 in that they contained a σ1 band along with coprecipitated σ3.

IgA MAb 1E1 neutralizes reovirus T1L in a strain-dependent and S1-determined manner.

Plaque reduction assay in cell culture is the classical method for demonstrating antibody-mediated neutralization of viral particles (64, 69). We used this assay to determine the neutralizing capacity of IgA MAb 1E1 with several different reovirus strains in L929 cells. The results demonstrated that 1E1 is neutralizing for reovirus T1L, against which the antibody was initially raised, but not for reoviruses T2S59 or T3D (Fig. 3). Moreover, 1E1 was shown to be neutralizing for reassortant virus 3HA1, which contains the T1L S1 genome segment and all nine other segments from T3D, but not for reassortant 1HA3, which contains the T3D S1 and all nine other segments from T1L (Fig. 3) (70). We conclude that the capacity of reovirus T1L to be neutralized by 1E1 is determined by the σ1-encoding S1 genome segment, consistent with preceding evidence that 1E1 binds to σ1. Neutralization of T1L by σ1-specific IgG MAb 5C6 (66) was also confirmed in this study (data not shown).

FIG. 3.

Neutralization of reoviruses by IgA MAb 1E1. Viral stocks were evaluated for neutralization by 100 μg of 1E1 per ml in parallel with matched samples containing no MAb. The number of plaques in each 1E1-treated sample was expressed as a percentage of that in the matched no-MAb sample. Each bar represents the mean of three determinations ± the standard error. Reoviruses T1L, T2S59, and T3D were evaluated along with reassortant viruses 3HA1 and 1HA3, which have the T1L and T3D σ1-encoding S1 genome segments, respectively, in a background of all other segments from the opposite strain (70).

IgA MAb 1E1 prevents binding of reovirus T1L to L cell membranes in vitro.

We then tested the hypothesis that the neutralizing activity of anti-σ1 MAbs is due to their capacity to inhibit binding of reovirus T1L to L cells. L cells were grown on glass coverslips to confluence and briefly fixed to stabilize cell membranes and prevent endocytosis. Biotinylated T1L virions and ISVPs were then applied to the cells in the presence or absence of MAbs 1E1 or 5C6, or antibodies against other outer capsid proteins, and bound virus was detected using streptavidin-FITC followed by fluorescence microscopy. The binding of both T1L virions and ISVPs was completely inhibited by preincubating the viral particles with anti-σ1 IgA MAb 1E1 or IgG MAb 5C6 at 10 μg/ml prior to overlay on the L cells (data not shown). In contrast, MAbs specific for σ3 or μ1 had no effect on viral binding even at 100 μg/ml, the highest concentration tested (data not shown).

IgA or IgG MAbs against the σ1 protein inhibit binding of reovirus T1L to apical membranes of Caco-2_BBe cells.

Reovirus T1L is known to adhere to confluent Caco-2 cells (2). Previous studies showed that at 2 days postconfluence, cloned Caco-2_BBe cells are polarized, but most have not yet assembled well-organized brush borders, and they lack the thick brush border glycocalyx typical of enterocytes in vivo (28, 30). We recently demonstrated that reovirus T1L adheres to apical membranes of polarized Caco-2_BBe cells at 2 days postconfluence via the σ1 protein but that viral binding is lost as the cells differentiate and develop brush borders (34). Binding of virions or ISVPs shows a mosaic pattern, with clusters of Caco-2_BBe cells virus positive and most cells negative, an observation consistent with the fact that these cells do not differentiate synchronously (10, 55). We therefore tested the ability of the anti-σ1 IgA MAb 1E1, as well as the five other IgG and IgA MAbs against reovirus T1L outer capsid proteins, to inhibit reovirus binding to apical surfaces of polarized Caco-2_BBe cell monolayers at 2 days postconfluence. To prevent access of virus to basolateral membrane domains of these cells, monolayers were fixed prior to apical application of virus. Biotinylated reovirus T1L ISVPs bound to clusters of Caco-2_BBe cells as previously observed (Fig. 4A). Biotinylated T1L ISVPs were then mixed with each of the antireovirus MAbs, and the mixtures were applied to the monolayers. Anti-σ1 IgG MAb 5C6 and anti-σ1 IgA MAb 1E1 completely abolished binding of ISVPs at 10 μg/ml or more (Fig. 4C and D). In contrast, IgG or IgA MAbs directed against T1L σ3 or μ1 proteins had no effect on the binding of ISVPs, even at 300 μg/ml (Fig. 4E to H). These results support the conclusion that the σ1 protein mediates T1L adherence to epithelial cells and are consistent with the σ1 specificity of IgA MAb 1E1.

FIG. 4.

Binding of reovirus T1L ISVPs to confluent Caco-2_BBe cell monolayers in the absence (A) or presence (C through H) of reovirus-specific IgG or IgA MAbs. Biotinylated ISVPs and MAbs were mixed and applied to fixed monolayers, and bound viral particles were visualized with streptavidin-FITC. (A) In the absence of MAbs, ISVPs adhered to clusters of cells in the monolayer. (B) Phase-contrast image of the monolayer shown in panel A, demonstrating that the Caco-2_BBe cells were confluent. In the presence of anti-σ1 IgG MAb 5C6 (C) or anti-σ1 IgA MAb 1E1 (D), ISVP binding was abolished. In the presence of anti-σ3 IgG MAb 10G10 (E), anti-σ3 IgA MAb RB3 (F), anti-μ1 IgG MAb 4A3 (G), or anti-μ1 IgA MAb RB8 (H), ISVP adherence was indistinguishable from control monolayers exposed to virus alone. Bar, 50 μm.

Inhibition of binding of reovirus T1L to Caco-2_BBe cells by IgA MAb 1E1 is strain dependent and S1/σ1 determined.

The IgG MAb 5C6 is known to recognize the σ1 protein of T1L in a strain-dependent manner commonly relating to viral serotype (64, 66). We thus sought to determine whether MAb 1E1, like 5C6, recognizes the σ1 protein in a strain-dependent manner. As a test, we preincubated biotinylated virions and ISVPs of reoviruses T1L, T2S59 (serotype 2), and T3D (serotype 3) with MAb 1E1 or 5C6 at concentrations shown to inhibit T1L binding in the experiments described above (10 and 50 μg/ml), prior to applying the particles to intact Caco-2_BBe cell monolayers (Table 2). Control monolayers were exposed to virus particles with no antibodies. Binding was scored as “inhibited” only when adherence of virus particles was completely abolished and as “unaffected” only when levels of adherence were indistinguishable from controls. MAbs 1E1 and 5C6 again inhibited the binding of T1L virions and ISVPs at both concentrations tested, but binding of T2S59 and T3D was unaffected (Table 2). Since serotype specificity is known to reside in the σ1 protein (7, 37, 46, 62, 70), this finding was consistent with the conclusion that the IgA MAb 1E1 is specific for σ1. We then tested the capacity of IgA MAb 1E1 to inhibit viral binding to Caco-2_BBe cells, using ISVPs derived from reassortant strains 3HA1 and 1HA3 (69), as well as recoated cores bearing either T1L or T3D σ1 protein (18). Binding of reassortants and recoated cores bearing T1L σ1 was inhibited by MAbs 1E1 and 5C6 at concentrations of 10 or 50 μg/ml, but binding of both types of particles bearing T3D σ1 was not affected (Table 2). These results confirmed the specificity of IgA MAb 1E1 for σ1 and reinforced the conclusion that the σ1 protein in T1L particles mediates epithelial cell binding.

TABLE 2.

Binding of reovirus ISVPs and ISVP-like particles derived from recoated cores to Caco-2_BBe cell monolayers and inhibition by reovirus T1L σ1-specific MAbs 1E1 and 5C6

Source of ISVPs	Binding activity in the presence of:
	No MAb	IgA MAb 1E1	IgG MAb 5C6
Viral isolates
T1L	+a	−	−
T2S59	+	+	+
T3D	+	+	+
Reassortants
3HA1	+	−	−
1HA3	+	+	+
Recoated cores
T1L σ1	+	−	−
T3D σ1	+	+	+

^a+, binding of virus at control levels; −, complete lack of binding.

Anti-σ1 IgA MAb 1E1 protects mouse Peyer's patch mucosa from reovirus T1L infection in a passive protection protocol.

We previously established a passive protection assay in which reovirus T1L is administered intragastrically along with anti-T1L MAbs, and levels of reovirus infection in Peyer's patches are measured at 8 h after challenge. In this assay, IgA or IgG MAbs specific for the outer capsid proteins μ1 and σ3 failed to protect mice against Peyer's patch infection, whereas the anti-σ1 IgG MAb 5C6 did protect (60). To test the protective capacity of anti-σ1 IgA MAb 1E1, groups of mice were challenged intragastrically with 2 × 10⁷ PFU of reovirus T1L mixed with an excess (50 μg) of IgA MAb 1E1, IgG MAb 5C6, or IgA MAb RB8. Control mice received virus with no antibody. Plaque assays of Peyer's patch tissue showed that the anti-μ1 IgA MAb RB8 had no significant effect on levels of virus in the mucosa (Fig. 5), in agreement with a previous study (60). In contrast, both anti-σ1 IgG MAb 5C6 and anti-σ1 IgG MAb 1E1 prevented infection of Peyer's patches (Fig. 5). Although anti-σ1 antibodies of either isotype were effective in protection against oral challenge in this assay, the physiological significance of the IgG-mediated protection observed was not clear, because IgG levels are very low in the intestinal lumens of normal mice (31) and mice immunized with reovirus T1L (60).

FIG. 5.

Effects of passively fed antireovirus MAbs on viral entry into Peyer's patches of adult BALB/c mice. Aliquots of reovirus (2 × 10⁷ PFU) were mixed with MAbs (50 μg) and administered intragastrically to groups of six mice. Naïve mice were not inoculated, and positive control mice received virus only. Each symbol represents an individual mouse; bars indicate medians. Concentrations of virus in Peyer's patches of mice given anti-μ1 IgA were not significantly different from those in positive controls (P = 0.22), indicating that this MAb failed to prevent viral entry. In mice fed either anti-σ1 IgG MAb 5C6 or anti-σ1 IgA MAb 1E1, levels of virus in Peyer's patches were undetectable.

Production of serum and secretory antibodies in mice with antireovirus hybridoma backpack tumors.

To compare the protective capacities of endogenously produced serum IgG and secretory IgA MAbs directed against the σ1 protein, backpack hybridoma tumors were generated in groups of adult mice by subcutaneous injection of hybridoma cells. Control groups were injected with hybridoma cells producing IgG or IgA against an irrelevant antigen, CTB. Additional controls received no hybridoma cells. The IgA hybridoma cell lines used here (1E1and P1E10) produced a mixture of IgA monomers, dimers, and higher polymers in culture (reference 4 and data not shown). The IgG hybridoma cells (2D6 and P1B8) produced IgG monomers (4, 66).

Mice with visible tumors (about 1 to 3 cm in diameter) were selected for challenge. To confirm that the hybridoma tumors produced MAbs in these mice, concentrations of total IgA and IgG in serum and feces, collected prior to injection of hybridoma cells and again on the day of reovirus challenge when tumors were visible, were determined by isotype-specific ELISA. In sera of mice bearing IgG hybridoma tumors (5C6 or P1B8), total IgG levels were significantly elevated above normal, but levels of total IgG in feces remained unchanged (Table 3). In mice with IgA hybridoma tumors (1E1 or P1E10), total IgA levels increased significantly in both serum and feces (Table 3). Antireovirus IgA and IgG in serum and feces were also measured prior to hybridoma cell injection and on the day of challenge. Specific antibodies in serum and feces were at background levels prior to hybridoma cell injection, as expected (Fig. 6). On the day of challenge, serum levels of antireovirus IgG and IgA were high (Fig. 6A). In feces, however, levels of antireovirus IgG remained low while antireovirus IgA levels were high (Fig. 6B).

TABLE 3.

Total IgA or IgG in serum and feces of mice bearing hybridoma backpack tumors, determined by immunoglobulin isotype-specific capture ELISA

Hybridoma clone, antibody source, and isotype	Antibody level at indicated timea		P valueb
	Pretumor	Challenge day
1E1
Serum IgA	250 (151-344)	4,050 (3,530-6,450)	<0.0001
Fecal IgA	135 (120-170)	3,550 (1,850-5,250)	<0.0001
P1E10
Serum IgA	500 (105-770)	6,010 (2,640-152,000)	0.0032
Fecal IgA	140 (68-185)	1,550 (840-3,190)	0.0162
5C6
Serum IgG	150 (80-800)	5,810 (2,830-13,000)	<0.0001
Fecal IgG	<15 ng/ml	<15 ng/ml
P1B8
Serum IgG	310 (80-580)	3,660 (1,000-4,930)	0.0013
Fecal IgG	<15 ng/ml	<15 ng/ml

^aValues shown are median concentrations (and ranges), expressed as micrograms per gram of feces or micrograms per milliliter of serum unless otherwise indicated.

^bSignificance of differences between pretumor and challenge-day concentrations was determined by an unpaired two-tailed t test at the 99% confidence level.

FIG. 6.

Antireovirus IgA and IgG levels in serum and feces of adult BALB/c mice with hybridoma backpack tumors. Levels of reovirus-specific antibodies were measured by ELISA before injection of hybridoma cells (pretumor) and prior to intragastric inoculation of reovirus T1L (challenge day). Each symbol represents an individual mouse; bars indicate medians for each group. (A) Reovirus-specific antibodies in serum. Specific antibodies were not detected before injection of hybridoma cells, as expected. On the day of challenge, serum levels of anti-σ1 specific IgA or IgG of mice with anti-σ1 IgA or anti-σ1 IgG backpack tumors were significant (P < 0.001 for IgA; P < 0.001 for IgG). (B) Reovirus-specific antibodies in feces. Low background levels of IgA reactive with reovirus were measured in feces of pretumor naïve mice (64 to 353 μg/g of feces), as noted previously (60). On the day of challenge, mice with 1E1 IgA tumors had significantly elevated levels of specific IgA in feces (P < 0.0001). In mice with 5C6 IgG tumors, specific IgG was not detected in feces despite high levels of specific IgG in serum.

Anti-σ1 IgA but not IgG prevents reovirus T1L infection of Peyer's patches in mice with backpack hybridoma tumors.

Groups of mice with hybridoma tumors producing MAb 1E1, P1E10, 5C6, or P1B8 or no tumors were orally challenged with 2 × 10⁷ PFU of reovirus T1L. At 8 h after challenge, the mice were sacrificed, Peyer's patches were removed and homogenized, and tissue concentrations of infectious virus were determined by plaque assay. Peyer's patches of control mice with no tumors or with P1B8 (anti-CTB IgG) tumors were consistently infected (Fig. 7). Peyer's patches of mice with either 5C6 (anti-σ1 IgG) or P1E10 (anti-CTB IgA) hybridoma tumors were also infected, but in both cases tissue levels of infectious virus were significantly reduced compared to controls. In contrast, the Peyer's patches of mice with 1E1 (anti-σ1 IgA) tumors contained little or no detectable virus (Fig. 7).

FIG. 7.

Reovirus T1L infection in Peyer's patches of adult BALB/c mice with hybridoma backpack tumors. Mice with hybridoma tumors producing anti-σ1 IgA or IgG, anti-CTB IgA or IgG, or no tumors were inoculated with 2 × 10⁷ PFU of reovirus T1L intragastrically, and at 8 h Peyer's patches were removed and viral entry was assessed by plaque assay. Each symbol represents an individual mouse; bars indicate median PFU per milligram of tissue. Control mice with no tumors showed infection of Peyer's patch tissue. Mice with anti-σ1 IgG hybridoma tumors were infected but at levels significantly lower than controls (P = 0.036). Mice with anti-σ1 IgA hybridoma tumors had viral levels that were undetectable or significantly lower than those in either control mice (P < 0.0001) or mice with anti-σ1 IgG hybridoma tumors (P < 0.0001). Mice with anti-CTB IgG hybridoma tumors were infected at control levels (P = 0.27). Mice with anti-CTB IgA hybridoma tumors were also infected but at lower levels than control mice without tumors (P = 0.004).

DISCUSSION

IgA MAb 1E1 directed against the reovirus T1L σ1 protein was newly produced in this study and used, along with other, existing antireovirus MAbs, to test and compare the protective capacities of specific IgA and IgG antibodies in vitro and in vivo. Both IgA and IgG MAbs against σ1 prevented adherence of T1L to apical membranes of cultured intestinal epithelial cells, but antibodies against μ1 or σ3 did not. Similarly, passive feeding of anti-σ1 IgA or IgG MAbs during oral T1L challenge protected mice against Peyer's patch infection, whereas feeding of antibodies against the other proteins did not. However, when anti-σ1 antibodies were produced in vivo from hybridoma backpack tumors, the anti-σ1 IgA, which was secreted into the intestinal lumen, prevented Peyer's patch infection, but the anti-σ1 IgG, which was not secreted, failed to protect.

The observation that antibodies directed against σ1 were capable of blocking T1L adherence to L cells and Caco-2_BBe cells in culture is consistent with previous evidence that the anti-σ1 IgG MAb 5C6 neutralized reovirus T1L in L-cell infectivity assays (64), and it is consistent with the neutralization of T1L by the anti-σ1 IgA MAb 1E1 demonstrated here. In separate studies, we observed that the neutralization activity of 5C6 in the standard L-cell plaque reduction assay is about 10-fold higher than that of 1E1 (data not shown). This difference was not apparent in our nonquantitative, fluorescence-based binding assays on cultured cells, in which anti-σ1 IgA and IgG appeared equally effective at blocking viral adherence. In any case, these results confirm previous ones identifying σ1 as the outer capsid protein that mediates binding of reovirus to nonpolarized L cells (7, 18, 19, 37, 45, 47, 48) and to apical membranes of polarized Caco-2_BBe cells (34). The results also suggest that the neutralizing activity of anti-σ1 antibodies at least partly reflects their capacity to prevent viral attachment to host cell surfaces. Neutralizing IgG MAb 5C6 recognizes the distal head region of T1L σ1 (19), the domain thought to mediate attachment to cellular protein receptors (20). The site recognized by neutralizing IgA MAb 1E1 was also recently identified in the distal head domain of σ1 (A. Helander, C. L. Miller, K. S. Myers, M. R. Neutra, and M. L. Nibert, submitted for publication), a location consistent with its capacity to block reovirus attachment to cultured cells.

IgA MAb 1E1 also showed functional resemblance to IgG MAb 5C6 in that it recognized reovirus in a strain-dependent manner. The fact that 1E1 blocked reovirus T1L adherence, but failed to block that of reoviruses T2S59 or T3D, opened the way for more specific antibody inhibition assays using T1L/T3D reassortants and viral cores recoated with either T1L or T3D σ1. In these assays, inhibition of adherence was consistently associated with antibody recognition of the T1L σ1 protein. The results support the hypothesis that the σ1 protein mediates attachment of T1L to epithelial cell apical membranes (34) and confirm that MAb 1E1 is indeed specific for T1L σ1.

The passive protection assay in adult mice, in which virus was mixed with specific IgA or IgG MAbs and then fed, served as a preliminary test of the protective capacities of antibodies directed against particular reovirus outer capsid proteins in the gastrointestinal tract. Some anti-σ3 and anti-μ1 IgG MAbs were previously shown to inhibit reovirus T3D infection of cultured L cells by mechanisms independent of effects on cell surface binding (67). In the mouse intestine in vivo, however, conversion of native virions to ISVPs involves removal of the σ3 protein (12, 25, 27, 37, 45, 46, 49). Our previous observation that passive feeding of anti-σ3 IgA or IgG MAbs failed to protect mice against reovirus T1L challenge (60) was expected, since anti-σ3 antibody would likely be removed along with σ3 in the intestinal lumen. A cleaved form of μ1 remains on the surface of ISVPs in the intestine, and IgA MAb RB8 binds to ISVPs in ELISAs (70). Thus, the lack of passive protection by anti-μ1 IgA or IgG MAbs shown in our previous study (60) and confirmed here was somewhat unexpected. If we assume that anti-μ1 IgA remained associated with ISVPs in the intestines of our mice, then these dimeric IgA molecules failed to entrap the virus in mucus gels, prevent mucosal contact, or sterically hinder the interaction of σ1 with M-cell surfaces. The apparent incapacity of IgA MAb RB8 to prevent M-cell attachment may be explained by the fact that dimeric IgA would measure at most 30 nm in length, whereas the σ1 protein can extend 40 nm from the surface of the ISVP (25, 29, 37). Another potential explanation is that the particular anti-μ1 MAbs used in these studies in fact do not bind well to ISVPs in the intestinal lumen, as recently suggested for anti-μ1 IgG MAbs 4A3 and 10F6 both in vitro and in cell culture (16). Further studies of the capacity of anti-μ1 MAbs to mediate protection in the mouse intestine may therefore be warranted.

In contrast, Peyer's patch infection was consistently prevented by feeding of anti-σ1 IgG MAb 5C6 in the previous study (60) and anti-σ1 IgA MAb 1E1 in this study. The σ1 protein plays at least two roles during reovirus infection of Peyer's patches: σ1 mediates adherence to M cells (34), which promotes delivery of virus across the epithelial barrier (3, 74), and then σ1 serves as the attachment protein required for infection of target cells in the mucosa (7). It is tempting to think that the anti-σ1 MAbs blocked infection by preventing M-cell adherence in the passive feeding assay, but we could not rule out the possibility that the MAbs might have entered the mucosa in association with viral particles and neutralized virus within the Peyer's patch. In addition, the fact that passively fed IgG as well as IgA MAbs were equally protective in this assay does not imply that both isotypes would be protective in normal, actively immunized mice, because significant amounts of IgG are not present in the normal mouse intestine (31). Indeed, mice that were naturally immunized by an oral dose of reovirus T1L and protected against subsequent oral challenge had anti-σ1 IgG (but not IgA) in serum and anti-σ1 IgA (but not IgG) in intestinal secretions (60).

The backpack tumor method provided clear evidence for the distinct roles of IgA and IgG in mucosal protection. We recognize that this method has drawbacks; hybridoma tumors result in abnormally high levels of IgA or IgG in blood and high levels of IgA in bile and intestinal contents (32, 43). For example, the levels of antireovirus IgG in the sera of the backpack tumor mice in this study were about 40-fold higher than those previously observed in mice orally immunized with T1L, and fecal antireovirus IgA levels were about 80-fold higher than in immunized mice (60). On the other hand, a major advantage of the backpack method is that it provides a distribution of antibodies that resembles that of naturally immunized animals in that dimeric or polymeric IgA MAbs are secreted into the intestinal lumen in association with secretory component, whereas IgG MAbs are confined to the blood and presumably the interstitial tissues of the intestine (1, 32). In this study, mice with anti-σ1 IgA hybridoma tumors were completely protected against entry of virus into Peyer's patches, whereas the Peyer's patches of mice with anti-σ1 IgG hybridoma tumors became infected despite extremely high levels of specific IgG in serum. Thus, the presence of specific IgA in secretions was crucial for protection.

This requirement for S-IgA for protection is consistent with previous observations of reovirus infectivity in suckling mice. Systemic injection of IgG MAb 5C6 into suckling mice prior to oral challenge with reovirus T1L prevented viral infection of the central nervous system but failed to prevent infection of Peyer's patches (63, 64). Suckling mice nursed on mucosally immunized dams (whose milk contained antireovirus IgA) were protected against oral challenge with T3D, whereas those nursed on systemically immunized dams (whose milk presumably lacked antireovirus IgA) were not protected (24). This difference occurred in spite of the fact that both groups of pups ingested milk containing antireovirus IgG that was transferred into the circulation (24).

It must be emphasized that in normal, actively immunized animals, S-IgA would not act alone. We observed that mice with backpack tumors producing anti-σ1 IgG, although not completely protected against oral challenge, showed significantly reduced levels of virus in Peyer's patches compared to control mice with no tumors. This presumably reflects the capacity of serum-derived, antireovirus IgG to inhibit target cell infection and cell-to-cell spread of the virus in the Peyer's patch mucosa (72). The significance of the partial protection provided by anti-σ1 IgG may be tempered by the fact that serum (and presumably tissue) IgG levels were abnormally high in mice bearing IgG backpack tumors, but it should be noted that partial protection attributable to serum IgG was previously observed in orally immunized, IgA-deficient mice (60). It is also important to note that mice secreting irrelevant (anti-CTB) S-IgA from backpack tumors showed lower levels of virus in Peyer's patches after oral challenge than did mice lacking tumors. This is consistent with our previous finding that the Peyer's patches of nonimmunized, IgA-deficient mice showed higher levels of viral infection after oral reovirus challenge than did normal mice (60). Taken together, these observations suggest that specific anti-σ1 IgG in serum and nonspecific S-IgA in secretions can contribute to intestinal defense against reovirus but that neither is sufficient to prevent infection completely. We can conclude, nonetheless, that neutralizing anti-σ1 S-IgA in intestinal secretions is indeed sufficient for complete protection of the adult mouse Peyer's patch mucosa from infection by reovirus T1L.