Role of Alpha/Beta Interferons in the Attenuation and Immunogenicity of Recombinant Bovine Respiratory Syncytial Viruses Lacking NS Proteins

Viruses and cells.

Stocks of the Snook (53), 391-2 (31) and 127 (45) strains of BRSV were prepared in fetal calf kidney(FCK) cells as described previously (45). Wild-type (WT) recombinant BRSV (rBRSV) and viruses lacking either the NS1 gene (ΔNS1), the NS2 gene (ΔNS2), or both (ΔNS1/2) were derived from full-length cDNA of BRSV strain A51908 (36), variant Atue51908 (GenBank accession no. AF092942), as reported previously (8, 42). Stocks of rBRSV were prepared in Vero cell monolayers that were infected at a multiplicity of infection (MOI) between 0.1 and 0.5 in Dulbecco's modified Eagle's medium (Gibco-BRL, Paisley, United Kingdom) containing 2% heated fetal calf serum (HFCS), as described previously. All recombinant virus stocks were checked by reverse transcriptase PCR (RT-PCR) for the presence of mRNA for the NS proteins (data not shown), and all virus stocks were free from contamination with bovine viral diarrhea virus.

Virulent BRSV that produces clinical signs of respiratory disease in young gnotobiotic calves consisted of bronchoalveolar lavage (BAL) prepared from a gnotobiotic calf inoculated 6 days previously with the Snook strain of BRSV, which had been passaged on two previous occasions in gnotobiotic calves. The BAL was free from other viruses, mycoplasmas, and bacteria as assessed by inoculation of tissue culture or mycoplasmal or bacterial media.

Virus titers were determined by plaque assay on FCK or Vero cell monolayers in 35-mm-diameter petri dishes (53) or by immunostaining of triplicate virus samples on Vero cell monolayers in 96-well plates with anti-F protein monoclonal antibody (MAb) 19 (48), 48 h postinfection (p.i.).

Bovine nasal fibroblasts (NT cells) were obtained from six healthy calves during necropsy. Nasal turbinates were washed extensively with phosphate-buffered saline (PBS) containing 40 μg of gentamicin (Sigma-Aldrich, Gillingham, United Kingdom)/ml and 25 μg of amphotericin B (Sigma-Aldrich)/ml. The epithelium was removed from the bone and was incubated overnight at 4°C in modified minimal essential medium (MEM) with penicillin (10 IU/ml) (Gibco-BRL), streptomycin (10 μg/ml) (Gibco-BRL), gentamicin (40 μg/ml), 2.5 μg of amphotericin B/ml, 0.01 Kunitz units of DNase (DN-25) (Sigma-Aldrich)/ml, and 1 mg of dl-dithiothreitol (Sigma-Aldrich)/ml to eliminate mucus produced by the tissue. The epithelium was cut into 5-mm squares and was cultured in petri dishes with fibroblast medium (MEM, penicillin [100 IU/ml], streptomycin [100 μg/ml], nystatin [25 U/ml] [Sigma-Aldrich], glutamine [2 mM] [Gibco-BRL], 5% tryptose phosphate broth, and 25 mM HEPES [Gibco-BRL]), containing 10% HFCS. Following migration of fibroblasts from the explanted tissue, the pieces of tissue were removed and the cells were trypsinized, seeded in vented flasks with fibroblast medium containing 5% HFCS, and incubated at 37°C in 5% CO₂ in air.

Bovine bronchoalveolar macrophages (BAM) were obtained at necropsy from seven healthy cattle, aged between 6 weeks and 7 years, as described previously (54). Briefly, the lung was irrigated with 1,200 ml of PBS, and the BAL was filtered through sterile gauze. BAL cells were centrifuged at 1,200 × g for 15 min at 4°C, washed once with cold PBS, and resuspended in 20 ml of BAM medium (RPMI-1640-Glutamax I [Gibco-BRL] containing 10% of HFCS and 40 μg of gentamicin/ml). More than 98% of BAM cells were macrophages, as determined by light microscopy.

All cultures of NT cells and BAM were free from contamination with bovine viral diarrhea virus and mycoplasmas.

Replication of rBRSV in vitro.

To compare the replication of the NS deletion mutants with WT BRSV in vitro, NT or Vero cells were infected with each virus at an MOI of 0.01 and were incubated at 37°C for 6 days. The amount of virus produced each day from infected cells was determined by scraping the cells into the supernatant to obtain a cell suspension, which was then frozen at −70°C until assayed as described above.

NT cells, at passes 3 to 14, were seeded at 6 × 10⁴ cells/cm² 24 h before infection with BRSV. After infection, NT cells were grown in fibroblast medium containing 2% of HFCS.

A suspension of BAM was infected at an MOI of 0.1 for 2 h at 37°C, washed twice with cold PBS, and resuspended in BAM medium to give a concentration of 10⁶ cells/ml. BAM were dispensed in 24-well plates or in 7-ml Teflon pots (Camlab, Over, United Kingdom) in a volume of 1 or 5 ml, respectively, and were incubated at 37°C in 5% CO₂ in air. At intervals after infection, BAM, grown in 24-well plates, were scraped into the supernatant and virus titers were determined after freezing and thawing as described above.

Induction of IFN-α/β in vitro.

In experiments to analyze induction of IFN-α/β in vitro, NT cells and BAM were infected as described above at an MOI of 0.1. Culture supernatants were harvested from cells at intervals, centrifuged at 700 × g for 10 min, and stored at −70°C until assayed for IFN-α/β bioactivity. To quantify IFN-α/β mRNA production, cells were resuspended in 350 μl of RNeasy lysis buffer (Qiagen, Crawley, United Kingdom)/10⁶ cells and were stored at −70°C until RNA extraction was performed. The following controls were included in all experiments: cells were cultured in medium only, inoculated with noninfected Vero cell lysate, inoculated with virus inactivated by exposure to UV light for 20 min, or inoculated with virus neutralized with a monoclonal antibody (MAb B4) directed against the F protein of RSV (27).

Analysis of IFN-α/β bioactivity.

IFN-α/β bioactivity was determined in an MxA-chloramphenicol acetyltransferase (MxA-CAT) reporter gene assay as described previously (20). Briefly, 10⁶ MDBK-t2 cells were seeded into 35-mm-diameter six-well plates and were cultured in 2 ml of medium (Eagle’s MEM [Gibco-BRL], 10 μg of blasticidin/ml, 100 IU of penicillin/ml, and 100 μg of streptomycin/ml) containing 10% of HFCS. After overnight incubation, the medium was replaced with 1 ml of test sample diluted 1:5 to 1:20 in medium containing 2% HFCS. Virus in test samples was inactivated by heating at 56°C for 30 min, prior to the assay. The heated samples were run in parallel with standard amounts (0.125 to 250 IU/ml) of recombinant bovine IFN-α1 (Novartis, Basel, Switzerland). All samples were set up in duplicate and were cultured for an additional 24 h at 37°C. CAT expression in cell lysates was determined by using a commercial enzyme-linked immunosorbent assay (ELISA) kit (Roche Molecular Biochemicals, Mannheim, Germany), and the level of IFN-α/β in the test samples was determined by reference to the IFN-α1 standard.

RT-PCR in real time for analysis of IFN-α and IFN-β mRNA.

Cytokine mRNA produced by NT cells or BAM following BRSV infection was quantified by using a method based on that of Kaiser et al. (26). Total RNA was extracted from cells lysed with RNeasy lysis buffer by using the RNeasy mini kit (Qiagen) with a DNase I digestion step (Qiagen) following the manufacturer's instructions. Purified RNA was eluted in 40 μl of RNase-free water and was stored at −70°C. Total RNA was quantified by using the Ribogreen Kit (Molecular Probes Europe BV, Leiden, The Netherlands) according to the manufacturer's instructions.

For both IFN mRNA- and 28S rRNA-specific amplification, primers and probes were designed by using the Primer Express software program (PE Applied Biosystems). Details of the probes and primers are given in Table 1. Primers and probes were designed from the consensus obtained by multiple-sequence alignments of the bovine IFN-α and IFN-β sequences available in GenBank (for IFN-α, A00145, A00146, A00147, A00148, X93087, X93088, X93089, E00133, E00134, E00135, M11001, and Z46508; for IFN-β, M15477, M15478, and M15479) by using the Wisconsin CGC 9.1 package (55). The primers and probes were designed to detect all of the nine known bovine IFN-α gene products or all of the three known bovine IFN-β gene products. The primers and probes corresponded to areas containing at least seven mutations compared with the consensus sequences of the other subfamilies of IFN, and their specificity was checked by using plasmids containing different segments of bovine IFN-α, -β, -ω, or -τ (data not shown). IFN and 28S probes were labeled with fluorescent reporter dye 5-carboxyfluorescein at the 5′ end and with the quencher N,N,N,N′-tetramethyl-6-carboxyrhodamine at the 3′ end. Quantification was based on the increased fluorescence detected by ABI PRISM 7700 Sequence Detection system due to hydrolysis of the target-specific probes by the 5′ nuclease activity of the rTth DNA polymerase during PCR amplification.

RT-PCR was performed by using Reverse Transcriptase qPCR Master Mix kit (Eurogenetec, Seraing, Belgium). The RT-PCR mixture consisted of 1× Master Mix Buffer (containing deoxynucleoside triphosphates, deoxynucleoside triphosphate buffer, and Hot Goldstar DNA polymerase), 0.25 U of Moloney murine leukemia RT/μl, 0.1 U of RNase inhibitor/μl, primers (concentrations, 300 nM for 28S, 400 nM for IFN-α, and 200 nM for IFN-β), a 100 nM concentration of the specific probe, 15 ng of RNA, and RNase-free water so that the volume added up to 25 μl. To quantify the RNA, the following cycle profile was used: 1 cycle of 48°C for 30 min and 96°C for 10 min and 40 cycles of 95°C for 15 s and 60°C for 1 min.

To generate standard curves for IFN and 28S RNA, plasmids containing the specific amplified segment were used and were serially diluted in RNase-free water to obtain between 10¹ and 10¹⁰ copies in 5 μl. Each RT-PCR experiment contained Neo template controls, test samples, and serial dilutions of plasmids, and each was run in triplicate. Regression analyses of the mean values of replicate RT-PCRs for serial dilutions of the specific plasmids were used to generate standard curves. Results are expressed as the number of copies of IFN mRNA/10⁶ copies of 28S rRNA.

Calves and experimental design.

Gnotobiotic, BRSV-seronegative calves were derived, reared, and maintained individually in plastic isolators as described previously (12). To evaluate the virulence of different strains of WT BRSV and rBRSV, groups of two to five gnotobiotic calves were infected at 2 to 3 weeks of age with approximately 10⁶ PFU of virus in a volume of 20 ml, 10 ml administered intranasally (i.n.) and 10 ml administered intratracheally (i.t.). Nasopharyngeal swabs and 5 ml of blood were obtained daily to measure IFN-α/β and to monitor virus excretion from the nasopharynx. As controls, calves were inoculated i.n. and i.t. with 20 ml of control tissue culture fluid.

A clinical examination was performed twice a day following virus infection. Calves were killed, 6 days after infection, by intravenous injection of sodium pentobarbital (Euthatal; Merial Animal Health Ltd., Harlow, United Kingdom). At postmortem examination, macroscopic lung lesions were recorded on a standard lung diagram, and the extent of pneumonic consolidation was expressed as the percentage of pneumonia. BAL was collected by irrigating the lungs from each calf with 400 ml of PBS, as described previously (49). Cells, prepared from 100 ml of the BAL by centrifugation at 1,200 × g for 15 min at 4°C, were resuspended in 5 ml of lung buffer (50). A sample of the BAL supernatant was stored at −20°C for antibody detection. Three pieces of pneumonic lung taken from different lobes were homogenized in lung buffer to give a 20% (wt/vol) suspension. Samples from nasopharyngeal swabs, BAL cells, and lung homogenates from calves inoculated with rBRSV or different strains of BRSV were inoculated onto Vero or FCK cells, respectively, to determine virus titers.

In experiments, to evaluate the immunogenicity of rBRSV, four groups of two or three gnotobiotic calves were inoculated i.n. and i.t. with WT rBRSV, rBRSVΔNS1, rBRSVΔNS2, or a suspension of noninfected Vero cell lysate as described above. Ten days after infection with rBRSV, calves were removed from the plastic isolators, mixed, and reared in a high-security, barrier-maintained building. Six weeks after immunization, calves were challenged i.n. and i.t. with 10³ PFU of virulent BRSV Snook in BAL. Serum samples and heparinized blood were obtained at 3-week intervals to detect antibodies and to perform T-cell assays. Following challenge, nasopharyngeal swabs were obtained daily and calves were killed 6 days after challenge to determine the extent of gross pneumonic consolidation and the extent of virus infection in the lower respiratory tract.

All experiments were performed under the regulations of the Home Office Scientific Procedures Act (1986) of the United Kingdom.

Serology.

The presence of antibodies to BRSV in sera and BAL was determined by ELISAs with a lysate prepared from FCK cells infected with the Snook strain of BRSV and mock-infected cells as control antigen, as described previously (50). Neutralizing antibodies to BRSV were determined by a plaque reduction assay on FCK cells by using heat-inactivated serum as described previously (27).

ELISPOT assays.

CD4⁺ T cells were isolated from peripheral blood mononuclear cells (PBMCs) after centrifugation on a density gradient (50), followed by staining with MAb CC8 (mouse immunoglobulin G2a [IgG2a], specific for bovine CD4⁺ T cells) (24) and anti-mouse IgG2a supermagnetic particles (Miltenyi-Biotech) as described previously (62). The purity of the cells as evaluated by flow cytometry was > 98%. ELISPOT assays were conducted in triplicate wells of Multiscreen-HA plates (Millipore). The wells were coated overnight at room temperature with anti-bovine IFN-γ MAb CC330 (0.22 μg/well) (38). Unbound antibody was removed by washing with PBS containing 0.05% Tween 20 (PBST), and the coated wells were blocked with PBS containing 5% bovine serum albumin for 2 h at room temperature. The plates were then washed with PBST and were incubated at room temperature with RPMI medium-10% HFCS until cells were added. To detect IFN-γ production by BRSV-specific CD4⁺ T cells, PBMCs were incubated for 90 min at 37°C with the Snook strain of BRSV at an MOI of 0.01 or with control uninfected FCK cell lysate, washed, and irradiated with 20 Gy from a ¹³⁷Cs source. These irradiated antigen-pulsed PBMCs (10⁵ cells/well) and freshly prepared CD4⁺ T cells (1 × 10⁵ to 4 × 10⁵/well) were added to each well in a 100-μl volume. Controls consisted of irradiated antigen-pulsed PBMCs alone or CD4⁺ T cells alone. After overnight incubation at 37°C with 5% CO₂, plates were washed extensively with PBST and with water. Biotinylated mouse anti-bovine IFN-γ MAb CC302 (38) (0.56 μg/well) diluted in 1% bovine serum albumin in PBS was added to each well, and the plate contents were incubated overnight at 4°C. Excess antibody was removed by washing with PBST. Vectastain ABC peroxidase PK-4000 kit (Vector Laboratories, Burlingame, Calif.) was mixed and added to each well in a 100-μl volume. After incubation for 1.5 h at room temperature, plates were washed with PBST and the spots were developed with 3-amino-9-ethyl carbazole (Sigma-Aldrich) according to the manufacturer's instructions. Plates were dried overnight and read by using an ELISPOT reader and AID 2.9 software (AutoImmun Diagnostika, Strassberg, Germany). The number of BRSV-specific IFN-γ-secreting CD4⁺ T cells was determined by subtracting the mean number of spots/10⁶ CD4⁺ T cells in control antigen wells from the mean number of spots in BRSV-stimulated wells.

Histology.

Lung tissue for histology was fixed in 10% neutral buffered formalin and paraffin wax embedded, and sections were stained with hematoxylin and eosin.

NS deletion mutants of BRSV are attenuated in IFN-inducing cells.

Previous studies have demonstrated that, whereas rBRSVs lacking one or both of the NS genes are only slightly attenuated in BSR T7/5 cells compared with the parental full-length virus, their growth is severely impeded in MDBK cells (42). We have extended these observations by comparing the growth of NS deletion mutants with that of WT virus in Vero cells, characterized by the lack of IFN-α/β genes (14, 59), and in low-passage bovine NT cells, which are competent for the production of IFN-α/β and which are used for the isolation of BRSV from clinical material (35). There were no significant differences in the growth of the NS deletion mutants and the parental full-length virus or the Snook strain of BRSV in Vero cells (Fig. 1). All viruses reached peak titers of between 8 × 10⁵ and 5 × 10⁶ PFU/ml after 3 or 4 days. For all of the viruses, giant cells were observed 2 days after infection and an increasing cytopathic effect was observed from day 3 to day 6. In contrast, the growth of all three deletion mutants was impeded in NT cells compared with that in Vero cells (Fig. 1). Furthermore, there were differences in the level of attenuation of the NS deletion mutants in NT cells. Thus, the growth of the ΔNS2 and ΔNS1/2 mutants was impeded to a greater extent than that of the NS1 deletion mutant. Whereas ΔNS1 mutant reached a peak titer of 4 × 10³ PFU/ml, 2 days after infection, the titers of ΔNS2 and ΔNS1/2 viruses decreased until day 4 after infection, when virus could no longer be recovered. The growth of BRSV Snook in NT and Vero cells was similar. However, WT rBRSV grew to lower titers in NT cells than in Vero cells. Cytopathic effects were observed in NT cells infected with WT BRSVs 3 to 6 days after infection but not in NT cells infected with the NS deletion mutants.

Induction of IFN-α/β following infection of bovine NT cells with NS deletion mutants of BRSV.

Since the growth of the NS deletion mutants in IFN-competent NT cells was impeded compared with that in Vero cells, we investigated the ability of the mutant viruses to induce IFN-α/β in NT cells. NT cells, at a passage level of less than 14, were infected with WT or mutant viruses at an MOI of 0.1, and IFN production was analyzed by using an MxA-CAT bioassay. IFN-α/β could not be detected in the supernatant of NT cells infected with WT BRSV or rBRSVΔNS1 (Fig. 2). In contrast IFN-α/β was detected in the supernatant from NT cells infected with the ΔNS2 or ΔNS1/2 mutant. As seen previously (Fig. 1), the growth of the ΔNS2 and ΔNS1/2 mutants was impeded in NT cells (Fig. 2B). Nevertheless, infectious virus was essential for the induction of IFN-α/β, as UV-inactivated viruses (results not shown) or antibody-neutralized viruses failed to induce IFN (Fig. 2C and D). The concentrations of IFN-α/β produced by NT cells treated with UV-inactivated or antibody-neutralized NS deletion mutants were less than 0.9 IU/ml, which was similar to that induced by NT cells alone or by cells exposed to uninfected Vero cell lysate (results not shown).

IFN-α/β was produced between 6 and 12 h p.i. and reached a plateau 12 h after infection with rBRSVΔNS2 or 24 h after infection with rBRSVΔNS1/2 (Fig. 2A). Although the ΔNS1 mutant did not induce detectable IFN-α/β in NT cells, the ΔNS1/2 mutant induced between 4 and 15 times more IFN-α/β than did rBRSVΔNS2. IFN-α/β induction by NT cells infected with rBRSVΔNS1/2 was influenced by the MOI, with increasing levels of IFN induced at higher MOIs (results not shown). However, even at an MOI of 1.0, IFN-α/β production was not detected in the supernatant of NT cells infected with ΔNS1 virus or WT rBRSV (results not shown).

Since infection of NT cells with WT rBRSV did not induce IFN-α/β, the ability of other WT BRSVs to induce IFN was investigated. The Snook, 391-2, and 127 strains of BRSV grew to similar titers in NT cells (Fig. 2F) and induced little or no IFN-α/β at an MOI of 0.1 (Fig. 2E). Thus, IFN-α/β production was not detected in NT cells infected with the Snook strain of BRSV and the highest levels of IFN-α/β induced by BRSV strains 391-2 and 127 were 3.75 and 2.15 IU/ml, respectively.

Induction of IFN-α/β by BAM infected with NS deletion mutants.

Infection of BAM with NS deletion mutants resulted in induction of IFN-α/β in a pattern similar to that seen in NT cells (Fig. 3A). However, whereas infection of NT cells with rBRSVΔNS1 induced little or no IFN-α/β, this virus induced low levels of IFN in BAM. The ΔNS2 mutant induced four to seven times more IFN than the ΔNS1 mutant and the ΔNS1/2 mutant induced 1.5-fold more IFN-α/β than did the ΔNS2 mutant. Although the NS deletion mutants appeared to be rapidly inactivated in BAM, as infectious virus could not be recovered 48 h after infection (Fig. 3B), infectious virus was essential for induction of IFN-α/β. Thus, UV-inactivated or antibody-neutralized viruses did not induce IFN-α/β in BAM (results not shown).

As seen in NT cells, WT rBRSV induced little or no IFN-α/β in BAM (Fig. 3A). However, this virus did not survive in BAM as well as other WT strains of BRSV did (Fig. 3B and C). Although two of the WT strains, 391-2 and 127, induced low levels (<10 IU/ml) of IFN-α/β in BAM (Fig. 3A), this did not appear to influence their survival in these cells. This is to be expected, since previous studies demonstrated that the NS proteins mediate resistance against IFN-α/β (3, 42).

Further studies on the interaction of rBRSV with BAM showed that there were large variations in the production of IFN-α/β by cells from different animals (Fig. 4). Nevertheless, the ΔNS2 mutant consistently induced significantly higher levels of IFN than did WT rBRSV. These variations may be related to the age and/or immune status of the cattle.

Induction of IFN-α/β mRNAs in NT and BAM cells.

Production of IFN-α/β was confirmed by analysis of IFN-α and IFN-β mRNA in virus-infected NT cells with Taq-Man PCR. Following infection with the ΔNS2 or ΔNS1/2 mutant, NT cells produced both IFN-α and IFN-β mRNA (Fig. 5A), with the levels of IFN-β mRNA being greater than those for IFN-α. Although transcripts for IFN-α or -β were very low in NT cells infected with WT rBRSV (<1.2 copies/10⁶ copies of 28S RNA) (Fig. 5A), IFN-β mRNA (6.16 copies/10⁶ copies of 28S RNA) could be detected in NT cells infected with the ΔNS1 mutant. Low levels (0.78 to 3.7 copies/10⁶ copies of 28S RNA) of IFN-β mRNA could also be detected in NT cells exposed to antibody-neutralized NS deletion mutants. These low levels of IFN-α or -β transcripts may be due to residual DNA contamination following treatment with DNase or incomplete neutralization of the mutant viruses. These results demonstrate that IFN-β mRNA is strongly synthesized and translated in NT cells infected with rBRSV lacking the NS2 gene but not in cells infected with WT virus, suggesting that the NS2 protein interferes with transcriptional activation of the IFN-β promoter.

Production of IFN-α/β by BRSV-infected BAM was confirmed by analysis of IFN-α and IFN-β mRNA. Whereas there was little or no IFN-α or -β mRNA in BAM infected with WT rBRSV or with the Snook strain of BRSV, transcripts for IFN-α were detected in BAM infected with the NS deletion mutants 24 h p.i. and decreased dramatically between 24 and 48 h p.i. (Fig. 5B). Levels of IFN-α mRNA in BAM infected with the ΔNS1, ΔNS2, and ΔNS1/2 mutants were 226-, 1,345-, and 1,831-fold greater, respectively, than that induced by WT rBRSV.

Differences in the types of IFNs produced by NT cells and BAM in this study were consistent with the known differences in the IFN response of fibroblasts and leukocytes. Thus, IFN-β is synthesized mainly by fibroblasts, which appear to produce IFN-α only following IFN receptor activation by IFN-β (15). In contrast, IFN-α is synthesized predominantly by leukocytes and is independent of IFN-β synthesis in these cells (15).

WT rBRSV and ΔNS1 rBRSV inhibit production of IFN by rBRSVΔNS1/2.

In order to determine if expression of the NS proteins by rBRSV inhibits the rBRSVΔNS1/2-mediated production of IFN-α/β, IFN levels were compared in rBRSVΔNS1/2-infected NT cells with or without superinfection by BRSV expressing the NS1 and/or the NS2 protein. Bovine NT cells were infected with the ΔNS1/2 mutant at an MOI of 0.05 for 2 h; the inoculum was removed and replaced by fresh medium containing 2% HFCS. Four hours p.i., cells were superinfected with WT rBRSV, rBRSVΔNS1, or rBRSVΔNS2 at an MOI of 0.5 or with fresh medium. Virus was adsorbed for 2 h at 37°C, the inoculum was removed, and the cells were incubated with fresh medium containing 2% HFCS. IFN-α/β production was determined at 24 and 48 h by using the MxA-CAT bioassay. Superinfection of rBRSVΔNS1/2-infected NT cells with either WT virus or the NS1 deletion mutant inhibited IFN-α/β production by 50 to 60% (Fig. 6). In contrast, superinfection with the NS2 deletion mutant increased IFN production by approximately 40%.

NS deletion mutants of BRSV are attenuated in calves.

Since the NS deletion mutants appeared to have an attenuated phenotype in both NT cells and BAM, the replication of the mutants in the respiratory tract of 2-week-old, BRSV-seronegative, gnotobiotic calves was evaluated. Groups of three to six calves were inoculated simultaneously by the i.n. and i.t. routes with WT rBRSV or with the NS deletion mutants, and three calves in each group were killed 6 days after infection. The remaining calves were kept in isolation until challenged with virulent virus, 6 weeks after immunization. For comparison, the ability of other WT strains of BRSV to infect gnotobiotic calves was also investigated. Following infection, clinical signs of respiratory disease were not observed in any of the calves apart from those infected with the calf-passaged Snook strain of BRSV (Snook BAL), who developed increased respiratory rates 6 days p.i. These calves had extensive pneumonic consolidation at postmortem, 7 days after infection. WT rBRSV was not as virulent as either calf-passaged BRSV Snook or tissue culture-grown BRSV Snook but was not as attenuated as two other WT strains of BRSV studied (Table 2). Thus, the replication of WT rBRSV in the upper respiratory tract was similar to that in calves inoculated with a 10-fold-lower dose of tissue culture-grown BRSV Snook. Furthermore, virus was isolated in lower titers and less frequently from the lower respiratory tract of calves infected with WT rBRSV than that found in calves infected with a 10-fold-lower dose of tissue culture-grown BRSV Snook (Table 2). However, WT rBRSV appeared to replicate to higher titers in the bovine respiratory tract than the 391-2 or 127 strain of BRSV, although the numbers of calves infected with the latter viruses were small (Table 2).

The replication of the single-deletion NS mutants in the bovine respiratory tract was significantly reduced compared with that of WT rBRSV, and virus could not be isolated from either the nasopharynx or the lungs of calves infected with the double-deletion mutant (Table 2). Peak virus titers in the nasopharynx of calves infected with the single deletion mutants were 300- to 600-fold less than that in calves infected with WT rBRSV, and there were no significant differences in the replication of the ΔNS1 and ΔNS2 mutants. Virus was not detected in BAL cells from calves inoculated with either the NS1 or the NS2 deletion mutants but was isolated in low titers from a single sample of lung tissue from one calf in each group.

Whereas the lungs from all three calves inoculated with WT rBRSV showed some gross pneumonic consolidation, there was little or no gross pulmonary pathology in calves inoculated with any of the NS deletion mutants. Microscopic lesions in lungs from calves infected with WT rBRSV were similar to those reported previously in gnotobiotic calves experimentally infected with BRSV Snook (53) Thomas et al., 1984) and were characterized by a proliferative and exudative bronchiolitis with some accompanying alveolar collapse and peribronchiolar infiltration by mononuclear cells. Microscopic lesions were not observed in the lungs of any of the calves inoculated with the NS deletion mutants, and the pulmonary histology of these calves was indistinguishable from that in calves inoculated i.n. and i.t. with lysates from mock-infected Vero cells.

Mucosal immunization with rBRSV lacking either NS1 or NS2 protects against challenge with virulent BRSV.

Groups of two or three gnotobiotic calves that had been inoculated i.n. and i.t. with WT, ΔNS1, ΔNS2 rBRSV, or noninfected Vero cell lysate 6 weeks previously were challenged with BRSV Snook in BAL. Despite the highly restricted replication of the ΔNS1 and ΔNS2 mutants in the respiratory tract of calves, infection with these viruses induced a serum antibody response (Table 3; Fig. 7A and B) and primed CD4⁺ T cells for production of IFN-γ (Fig. 7D). Although there were no detectable differences in the level of attenuation of the ΔNS1 and ΔNS2 mutants, the ΔNS2 mutant appeared to be more immunogenic than the ΔNS1 mutant. Thus, the number of BRSV-specific, IFN-γ-producing CD4⁺ T cells was greater in two of the calves immunized 6 weeks previously with the ΔNS2 mutant than in the three calves immunized with the ΔNS1 mutant (Fig. 7D). Furthermore, the levels of BRSV-specific serum antibodies detected by ELISA in calves immunized 6 weeks previously with the ΔNS2 mutant were similar to those induced by WT rBRSV and were eightfold greater than those induced by the ΔNS1 mutant (Table 3). Serum-neutralizing antibodies in calves immunized with the ΔNS2 mutant were only sixfold less than those induced by WT rBRSV, whereas the titer of neutralizing antibody induced by the ΔNS1 mutant was 80-fold less than that induced by WT virus (Table 3). The IgG1 serum antibody response detected by ELISA appeared to be delayed in calves immunized with the NS deletion mutants compared with that in calves immunized with WT rBRSV (Fig. 7A). However, BRSV-specific IgG2 serum antibodies were detected earliest in calves immunized with the ΔNS2 mutant and titers were always highest in these animals, although the differences were not statistically significant (Fig. 7B).

Following challenge with virulent BRSV, there was a rapid increase in serum antibody titers in all immunized calves. Although serum antibody titers, as detected by ELISA, were similar in all immunized calves 6 days after challenge serum-neutralizing titers in calves immunized with WT rBRSV were 10- or 60-fold greater than those in calves immunized with the ΔNS2 or ΔNS1 mutant, respectively (results not shown). Differences in IgA titers in BAL from immunized calves were also detected 6 days after challenge. Thus, IgA titers in BAL from calves immunized with the ΔNS2 mutant were similar to those in calves immunized with WT rBRSV and were significantly higher than those in BAL from calves immunized with the ΔNS1 mutant or the control calves (P < 0.05) (Fig. 7C). BRSV-specific IgG2 antibodies were not detected in BAL from any of the calves.

Following challenge, clinical signs of respiratory disease were not observed in any calves. However, calves were killed on day 6 after challenge, which is at the time that clinical signs of disease are usually first detected. Nevertheless, pneumonic lesions were present in all of the control calves, whereas there was little or no pneumonic consolidation in calves immunized with either WT rBRSV or with either the ΔNS1 or the ΔNS2 mutant (Fig. 8). Animals previously immunized with WT rBRSV were highly resistant to subsequent challenge with BRSV Snook (Table 3). Following challenge, virus was not isolated from the nasopharynx or the lung tissue of calves immunized with WT rBRSV but was isolated in low titers from BAL cells of one calf. In contrast, control calves excreted high titers of virus from the nasopharynx starting from day 3 after challenge and high titers of virus were isolated from BAL cells and from all lung samples studied at postmortem (Table 3). Calves previously immunized with either of the NS deletion mutants were completely protected against BRSV infection in the lower respiratory tract. However, the level of protection against upper respiratory tract infection was not as great as that seen in calves immunized with WT rBRSV. Although the differences were not statistically significant, the level of protection against upper respiratory tract infection appeared to be greater in calves immunized with the ΔNS2 mutant than in those immunized with the ΔNS1 mutant. Thus, following challenge, all of the calves previously immunized with the ΔNS1 mutant excreted virus from the nasopharynx, whereas only one of the three calves immunized with the ΔNS2 mutant excreted virus.

Induction of IFN-α/β in vivo.

Following infection of gnotobiotic calves with the different rBRSV, levels of IFN-α/β in serum and nasopharyngeal secretions obtained daily after infection and in BAL obtained at postmortem, 6 days after infection, were analyzed by using the MxA-CAT bioassay. Although viruses lacking the NS2 or NS1/2 gene induced high levels of IFN-α/β in NT cells or BAM in vitro, IFN-α/β could not be detected in sera, nasopharyngeal swabs (results not shown), or BAL (Table 2) from calves infected with these viruses or in calves infected with the ΔNS1 mutant. In contrast IFN-α/β could be detected in BAL, obtained 6 days after infection, from calves infected with WT rBRSV or BRSV Snook, which did not induce IFN-α/β in vitro (Table 2). The large variations in the concentrations of IFN-α/β seen in BAL from calves infected with WT rBRSV may be the result of variations in the volume of PBS that was recovered during the BAL procedure. In general, IFN-α/β was detected in BAL only from animals in which the virus titers were greater than 10³ PFU of BAL cells/ml. These results suggest that, in vivo, induction of IFN-α/β in the lung correlates with the level of virus replication.

Induction of IFN-α/β appeared quite late after infection with WT BRSVs. Thus, in two of the mock-immunized calves that were infected with calf-passaged BRSV Snook, IFN-α/β could not be detected in BAL on day 1 or 3 of infection (results not shown), but 15.4 and 58.3 IU of IFN/ml of BAL were detected in these animals on day 6. The correlation between the induction of IFN-α/β and viral replication, in vivo, was supported by the observation that IFN-α/β could be detected in both nasal secretions and BAL from immunized calves that were not protected against challenge with virulent virus. Thus, low levels of IFN-α/β were detected in nasal secretions on days 3 and 6 after challenge of calves that had been previously immunized with either the ΔNS1 mutant or with noninfected Vero cell lysate (Fig. 9A). Similarly, IFN-α/β was detected only in BAL from control calves 6 days after challenge with virulent BRSV Snook and not in those calves previously immunized with either the ΔNS1 or ΔNS2 mutants (Fig. 9B) or WT rBRSV (<0.7 IU/ml).