ABSTRACT
Palivizumab, a humanized murine monoclonal antibody that recognizes antigenic site II on both the prefusion (pre-F) and postfusion (post-F) conformations of the respiratory syncytial virus (RSV) F glycoprotein, is the only prophylactic agent approved for use for the treatment of RSV infection. However, its relatively low neutralizing potency and high cost have limited its use to a restricted population of infants at high risk of severe disease. Previously, we isolated a high-potency neutralizing antibody, 5C4, that specifically recognizes antigenic site Ø at the apex of the pre-F protein trimer. We compared in vitro and in vivo the potency and protective efficacy of 5C4 and the murine precursor of palivizumab, antibody 1129. Both antibodies were synthesized on identical murine backbones as either an IgG1 or IgG2a subclass and evaluated for binding to multiple F protein conformations, in vitro inhibition of RSV infection and propagation, and protective efficacy in mice. Although 1129 and 5C4 had similar pre-F protein binding affinities, the 5C4 neutralizing activity was nearly 50-fold greater than that of 1129 in vitro. In BALB/c mice, 5C4 reduced the peak titers of RSV 1,000-fold more than 1129 did in both the upper and lower respiratory tracts. These data indicate that antibodies specific for antigenic site Ø are more efficacious at preventing RSV infection than antibodies specific for antigenic site II. Our data also suggest that site Ø-specific antibodies may be useful for the prevention or treatment of RSV infection and support the use of the pre-F protein as a vaccine antigen.
IMPORTANCE There is no vaccine yet available to prevent RSV infection. The use of the licensed antibody palivizumab, which recognizes site II on both the pre-F and post-F proteins, is restricted to prophylaxis in neonates at high risk of severe RSV disease. Recommendations for using passive immunization in the general population or for therapy in immunocompromised persons with persistent infection is limited because of cost, determined from the high doses needed to compensate for its relatively low neutralizing potency. Prior efforts to improve the in vitro potency of site II-specific antibodies did not translate to significant in vivo dose sparing. We isolated a pre-F protein-specific, high-potency neutralizing antibody (5C4) that recognizes antigenic site Ø and compared its efficacy to that of the murine precursor of palivizumab (antibody 1129) matched for isotype and pre-F protein binding affinities. Our findings demonstrate that epitope specificity is an important determinant of antibody neutralizing potency, and defining the mechanisms of neutralization has the potential to identify improved products for the prevention and treatment of RSV infection.
INTRODUCTION
Palivizumab is the only licensed monoclonal antibody (MAb) for prophylaxis of infants at risk for severe respiratory syncytial virus (RSV) disease. Palivizumab is given monthly by intramuscular injection at a monthly dose of 15 mg/kg of body weight (1) and is expected to maintain a trough level of about 40 μg/ml (2). This regimen prevents about 50% of hospitalizations among premature infants infected with RSV (1). It is a humanized antibody on a human IgG1 framework and was derived from the murine antibody 1129 (3). This MAb was originally isolated from a mouse immunized with RSV A2 and vaccinia virus recombinants expressing the RSV F glycoprotein (4).
The RSV F glycoprotein is a class I fusion protein that initially folds into an 11-nm-tall prefusion (pre-F) conformation comprised of a trimer of F1-F2 heterodimers. RSV F is required for viral entry into susceptible cells and undergoes a massive unidirectional rearrangement, resulting in the formation of a stable antiparallel 6-helix bundle structure that pulls the membranes anchored by the N-terminal F1 fusion peptide and the C-terminal F1 transmembrane domain together to mediate membrane fusion. Because the pre-F molecule is metastable, this rearrangement can also occur spontaneously, resulting in a 16-nm-tall stable postfusion (post-F) molecule that can be found on the viral envelope (5). Therefore, RSV virions have both pre-F and post-F molecules on their surface. The pre-F molecule has at least 6 discrete regions on its surface targeted by neutralizing antibodies, and 3 of those are also present on surfaces shared with the post-F molecule.
Motavizumab (also known as MEDI-524; developed by MedImmune, LLC) is an MAb engineered by altering 13 amino acids of palivizumab. These alterations resulted in an affinity much higher than that of palivizumab, an off rate lower than that of palivizumab, and a neutralizing potency 18-fold higher than that of palivizumab (6, 7). Even though motavizumab has more potent neutralizing activity in vitro, its ability to protect cotton rats in vivo is not higher than that of palivizumab at a given dose (7). Therefore, the dose used in clinical studies was the same as that of palivizumab. In this study, we asked whether using an antibody with a different epitope specificity and a greater neutralization potency would translate into better in vivo protection and allow dose sparing of passively administered MAb.
Site II—the target of palivizumab, motavizumab, and 1129—is a helix-turn-helix motif (8) present on both the pre-F and post-F conformations (Fig. 1A). Approximately 50% of the pre-F and post-F protein surfaces overlap one another, including antigenic sites II and IV (9). Antigenic site Ø is present exclusively on the pre-F conformation and is located on the membrane-distal apex of the trimer. It consists of a helix from the F1 subunit and an unstructured region from the F2 subunit and is considered a supersite because multiple potent neutralizing MAbs can bind from different angles and rotational orientations (10). MAb 5C4 was identified as a site Ø-specific murine MAb (10) and has nearly 50-fold higher neutralizing activity than palivizumab or its murine parent, antibody 1129. To compare the in vitro and in vivo potencies of antibodies with distinct specificities, 5C4 (site Ø specific) and 1129 (site II specific) were synthesized on identical murine backbones as either an IgG1 or IgG2a subclass to match the antibody and origin with the species being tested and eliminate any differences in Fc-mediated antibody function. The intent is to evaluate the MAbs' ability to reduce viral loads or prevent RSV disease in mice on the basis of the epitope being targeted. We show that the pre-F conformation-specific MAb targeting the apical site Ø epitope provides greater protection in vivo than an MAb directed to antigenic site II, even though the MAbs showed similar affinities to the F protein.
Binding of 5C4 and 1129 to F protein. (A) The epitopes for 1129 and 5C4 are shown on a surface representation of the prefusion (pre-F) and postfusion (post-F) RSV F structures. MAb 5C4 binds to site Ø on the pre-F protein on the basis of the crystal structure of the 5C4–pre-F protein complex. Antibody 1129 binds to site II on the pre-F and post-F proteins on the basis of the crystal structure of the motavizumab-peptide and the motavizumab–pre-F protein complex. (B and C) Antibody binding to purified recombinant post-F (B) and pre-F (C) proteins was determined by an indirect ELISA. The reactivities of 5C4 and 1129 were expressed as EC50s (in micrograms per milliliter). (D and E) Human RSV-infected HEp-2 cells were labeled with 5C4 or 1129 to detect F protein on the membranes of cells at 24 h after infection. GAM-tetramethyl rhodamine isocyanate was added as a secondary antibody to MAb-labeled cells and detected by superresolution fluorescence microscopy (D), or GAM-FITC was added as a secondary antibody to MAb-labeled cells and detected by flow cytometry (E). (F) Antibody binding to F protein on the surface of cryopreserved RSV A2 virions was detected by a sandwich ELISA. RSV was captured by anti-F MAb 3G3 coated on 96-well plates. MAb 5C4-HRP or 1129-HRP was added to the 3G3-bound virus for antigenicity analysis. The flow cytometry data are representative of those from three independent experiments. OD, optical density.
RESULTS
MAbs 5C4 and 1129 binding and neutralizing potencies.Isotype-matched 5C4 and 1129 were produced in transiently transfected HEK293 cells. The binding affinities of recombinant 5C4 and 1129 to the pre-F and post-F proteins were then determined. MAb 5C4 exclusively bound the pre-F proteins (Fig. 1B), whereas 1129 bound both the pre-F and post-F proteins with similar affinities (Fig. 1C). We also determined, by confocal microscopy (Fig. 1D) and flow cytometry (Fig. 1E), that both 5C4 and 1129 could bind to wild-type F on the surface of RSV-infected HEp-2 cells. However, results from a sandwich enzyme-linked immunosorbent assay (ELISA) showed that the binding of 5C4 to virions was significantly lower than that of 1129 (Fig. 1F), indicating that the F on virions is unstable and spontaneously flips from the pre-F into the post-F conformation, which is consistent with prior reports (5, 11).
The RSV-neutralizing (NT) and fusion-inhibiting potencies of recombinant 5C4 and 1129 were measured using recombinant green fluorescent protein (GFP) reporter viruses on HEp-2 cells (12). MAb 5C4 demonstrated about 50-fold greater NT activity than 1129 (Fig. 2A) and an approximately 4 times greater inhibition of cell-to-cell spread (Fig. 2B).
MAbs 5C4 and 1129 inhibit the infection and spread of human RSV (hRSV) in vitro. (A) Neutralization of human RSV infection in HEp-2 cells by antibodies, as detected by flow cytometry, reported as the 50% inhibitory concentration (IC50; in micrograms per milliliter). (B) Antibodies inhibit the cell-to-cell spread of human RSV in HEp-2 cells after initial infection of cells. At 6 h postinfection, cell monolayers were washed to remove unattached virus. Serial dilutions of antibodies were then added to the medium. Inhibition was detected by flow cytometry and recorded as the IS50 (in micrograms per milliliter).
Antibody prophylaxis of RSV infection in the BALB/c mouse model.We first found that 5C4 and 1129 exhibited similar neutralizing half-lives in BALB/c mice (Fig. 3). To determine the potency of 5C4 and 1129 for prophylaxis, BALB/c mice were intranasally inoculated with 107 PFU of RSV A2 24 h after the administration of 5C4 or isotype-matched 1129 at a dose of 1.5 mg/kg or 15 mg/kg.
Kinetics of antibody clearance from serum in a mouse model. MAbs were administered via i.p. injection at a dose of 15 mg/kg/mouse. Five mice were tested for each antibody. Serum samples were collected from the tail vein every 2 days from 1 day after i.p. injection to day 13 after i.p. injection. All of the serum specimens were tested for the neutralization of RSV A2-mKate. (A) Kinetics of 5C4-IgG2a and 1129-IgG2a clearance from serum; (B) kinetics of 5C4-IgG1 and 1129-IgG1 clearance from serum.
Nasal turbinates and lungs were harvested from each group and homogenized to investigate the infectious viral loads by plaque assay at day 5 after RSV challenge (Fig. 4E and F). MAb 5C4 at both the high and low doses significantly decreased the RSV titers in the nasal turbinates compared to those in the group treated with phosphate-buffered saline (PBS) (P < 0.001) (Fig. 4E). MAb 5C4 at 15 mg/kg consistently achieved a mean 4-log10 reduction in the RSV copy number in nasal turbinates compared to that for the PBS-treated control group (P < 0.001) and was 1,000-fold more effective than 1129 at the same dose (P < 0.01). MAb 5C4 at the lower dose resulted in a nearly 10-fold reduction, similar to that achieved by 1129 at the higher dose (P > 0.05). These data suggest that 5C4 has significantly more potent antiviral activity than the equivalent dose of 1129.
Prophylactic efficacy of 5C4-IgG2a and 1129-IgG2a against human RSV. BALB/c mice (n = 5) were injected with different doses of 5C4-IgG2a, 1129-IgG2a, or PBS as an untreated control at 24 h before i.n. challenge with 107 PFU of human RSV A2. Antibodies were delivered at doses of 15 mg/kg and 1.5 mg/kg via i.p. injection. (A) Percent weight change (as a percentage of the weight at day 0) was monitored until 12 days postinfection. (B to D) Comparison of weight loss between the different groups at days 6, 7, and 8, respectively. The y axis indicates the percentage of the body weight at the time of challenge. (E and F) The viral load was quantified as the number of PFU per gram of nasal turbinates (Nose) (E) and per gram of lung (F) by plaque assay at day 5 after challenge. Two-tailed P values were calculated by the unpaired Student's t test. ***, P < 0.001; **, 0.001 ≤ P ≤ 0.01; *, 0.01 < P ≤ 0.05. Each treated group was first compared to the untreated group, and then each treated group was compared to the other treated groups. The red asterisks show the comparison between 1129 and 5C4 groups.
The virus titers in the lungs were quantified by plaque assay (Fig. 4F). Both antibodies at the 15-mg/kg dose significantly reduced the viral loads in the lungs by up to 6 log10 compared to those in the lungs of the PBS-treated control group (P < 0.001) (Fig. 4F). MAb 5C4 at 1.5 mg/kg reduced the viral titers in the lungs by 3 log10 (P < 0.01). These data were consistent with the viral loads measured as the genome copy numbers by quantitative PCR (data not shown). Although both 5C4 and 1129 were effective in reducing viral replication in the lung, 5C4 had much greater antiviral potency. At the 15-mg/kg dose, 5C4 was able to completely eliminate the virus from the lower and upper airways, such that no virus could be isolated from nasal turbinates.
The weights of 5 mice in each group were monitored from day 0 to day 12 after infection. As shown in Fig. 4A, the average weight of mice prophylactically treated with 1129 at a dose of 1.5 mg/kg declined about 10% by day 6. The mice also had significant signs of illness, with ruffled fur and hunched posture, similar to the mice in the unprotected PBS-treated control group. However, in mice prophylactically treated with 5C4 at both the high and low doses and in mice treated with 1129 at the high dose, there was no significant weight loss (Fig. 4B to D). Antibody mice treated with 5C4 had significantly less weight loss than mice treated with 1129 at the same dose (P < 0.05). In addition, low-dose 5C4 was more effective than high-dose 1129 at preventing weight loss (P < 0.05) (Fig. 4B). Weight loss was inversely correlated with the magnitude of virus recovered from the lungs and nasal turbinates.
Evaluation of lung histopathology suggested that 5C4 prevented early the monocytic infiltration into the lung interstitium and alveolar spaces that can be seen during acute infection (Fig. 5). In addition, the concentrations of 9 cytokines and chemokines in homogenized lung tissues were investigated by ELISA at day 5 after RSV challenge as a measure of pathogenicity (Fig. 6). Gamma interferon (IFN-γ)-induced protein 10 (IP-10) and macrophage inflammatory protein 1α (MIP-1α) have previously been correlated with disease severity in mice following RSV infection (13). The concentrations of IP-10 (Fig. 6H) in mice treated with 5C4 were significantly lower than those in mice in the PBS-treated control group (P < 0.05 for the lower dose, P < 0.01 for the higher dose), whereas no significant differences in the concentrations were observed between mice treated with 1129 and mice in the PBS-treated control group. Although the concentrations of MIP-1α (Fig. 6I) in all mice treated with 5C4 or 1129 at a dose of 1.5 mg/kg were significantly lower than those in mice in the PBS-treated control group, the 5C4-treated animals had lower levels than the 1129-treated animals (P < 0.05). These data demonstrate that the cytopathology and immune inflammatory responses were lower in 5C4-protected mice than in 1129-treated mice. Even at the low dose of 1.5 mg/kg, 5C4 efficiently reduced the level of virus-related inflammation in vivo.
Qualitative analysis of lung histopathology in RSV-infected mice treated prophylactically with MAbs. At day 5 after inoculation, formalin-fixed lung sections were stained with hematoxylin and eosin. (A) Lung sections from the PBS-treated control group (the RSV-infected, untreated group) demonstrated perivascular and interstitial infiltrates. Scattered areas of peribronchial mononuclear cell inflammation were also seen. In many areas, the air spaces contained numerous monocytes and neutrophils. (B) Sections of lungs from RSV-infected mice treated with 1129 at the lower dose showed histopathological features similar to those seen in the PBS-treated control group. (C) Sections of lungs from RSV-infected mice treated with the high dose of 1129 demonstrated mild interstitial infiltrates. (D and E) Sections of lungs from RSV-infected mice treated with 5C4 at the lower dose (D) and the higher dose (E) showed minimal or no perivascular or interstitial infiltrates.
Impact of passive IgG2a antibody prophylaxis on cytokine and chemokine production in lung. Lung homogenates from samples tested in the assay whose results are presented in Fig. 5 were evaluated by ELISA for the cytokines and chemokines IFN-γ, tumor necrosis factor alpha (TNF-α), interleukin-4 (IL-4), IL-5, IL-6, IL-10, IL-13, IP-10, and MIP-1α. The detection limits of these cytokines and chemokines were 0.8, 2.2, 0.6, 0.5, 1.0, 1.5, 2.0, 0.8, and 12.1 pg/ml, respectively. Only high-dose 5C4 prophylaxis fully suppressed the production of inflammatory cytokines. Two-tailed P values were calculated by the unpaired Student's t test. ***, P < 0.001; **, 0.001 ≤ P ≤ 0.01; *, 0.01 < P ≤ 0.05. The concentrations of the cytokines and chemokines in each treated group were first compared to those in the untreated group. Then the concentrations in each treated group were compared to those in each of the other treated groups. The red asterisks show the comparison between 1129 and 5C4 groups.
MAbs 5C4 and 1129 built on the murine IgG1 background were also evaluated. As expected, the in vitro binding affinities of the IgG1 and IgG2a versions were similar to one another (data not shown). The efficacy of 5C4 and 1129 delivered as IgG1 was evaluated in mice by passive transfer prior to infection (Fig. 7 and 8). The data are similar to those shown for the IgG2a versions. MAb 5C4-IgG1 was more potent than 1129-IgG1 in protecting mice from RSV infection at equivalent doses, and the low 1.5-mg/kg dose of 5C4-IgG1 as well as the 15-mg/kg dose of 1129-IgG1 protected BALB/c mice from RSV infection.
Prophylactic efficacy of 5C4-IgG1 and 1129-IgG1 against human RSV. BALB/c mice (n = 5) were injected with different doses of 5C4-IgG1, 1129-IgG1, or PBS as an untreated control 24 h before intranasal (i.n.) infection with 107 PFU of human RSV A2. Antibodies were injected at a dose of 15 mg/kg or 1.5 mg/kg via i.p. injection. (A) Percent weight change (as a percentage of the weight at day 0) was monitored until 12 days postinfection. (B to D) Comparison of the weight loss between the different groups at days 6, 7, and 8, respectively. The y axis indicates the percentage of the body weight at the time of challenge. (E and F) The viral load was quantified as the number of PFU per gram of nasal turbinates (Nose) (E) and per gram of lung (F) by plaque assay at day 5 after challenge. Two-tailed P values were calculated by the unpaired Student's t test. ***, P ≤ 0.001; **, 0.001 ≤ P ≤ 0.01; *, 0.01 < P ≤ 0.05. Each treated group was first compared to the untreated group, and then each treated group was compared to each of the other treated groups. The red asterisks show the comparison between 1129 and 5C4 groups.
Impact of passive IgG1 antibody prophylaxis on cytokine and chemokine production in lung. Lung homogenates were derived for the lungs of the mice used in the assay whose results are presented in Fig. 7. The levels of production of IFN-γ, tumor necrosis factor alpha (TNF-α), interleukin-4 (IL-4), IL-5, IL-6, IL-10, IL-13, IP-10, and MIP-1α were tested. The detection limit of these cytokines and chemokines were 0.8, 2.2, 0.6, 0.5, 1.0, 1.5, 2.0, 0.8, and 12.1 pg/ml, respectively. Two-tailed P values were calculated by the unpaired Student's t test. ***, P < 0.001; **, 0.001 ≤ P ≤ 0.01; *, 0.01 < P ≤ 0.05. The concentrations of cytokines and chemokines in each treated group were first compared to those in the untreated group. The concentrations in each treated group were then compared to those in each of the other treated groups. The red asterisks show the comparison between 1129 and 5C4 groups.
These data demonstrate that the greater in vitro neutralizing potency of the site Ø-specific 5C4 MAb over the site II-specific 1129 MAb translates to greater in vivo potency when the IgG isotype is controlled for.
Inhibition of RSV infection and propagation by 5C4 and 1129 in vitro.We next evaluated the ability of 5C4 and 1129 to inhibit RSV infection and propagation in vitro by continuous monitoring. Serially diluted 5C4 and 1129 were added at either 1 h before infection (to inhibit infection) or 6 or 18 h after infection (to inhibit propagation). The culture was continuously monitored with a high-content analysis (HCA) instrument. To enable a visual comparison, we chose to display data obtained at 24 h, 36 h, and 47 h postinfection (Fig. 9A to C).
MAbs 5C4 and 1129 inhibit infection and propagation of human RSV in vitro. Serial dilutions of 5C4 or 1129 were added either 1 h before RSV infection (−1 h) (A) or 6 h (B) or 18 h (C) after RSV infection, and the fluorescence intensity of each well was detected by use of a real-time high-content analysis (HCA) system. From left to right, the graphs show the results obtained at 24 h, 36 h, and 47 h postinfection. (D) The 50% inhibitory concentration (IC50) for each group at each time point. These results are representative of those from three total experiments.
In this assay, it was determined that 5C4 was more potent than 1129 in inhibiting RSV infection. A 0.16-μg/ml dose of 5C4 efficiently inhibited RSV infection, whereas 1129 required at least 8 μg/ml to achieve a similar level of inhibition (Fig. 9A and D, −1-h group). At the same concentration, 5C4 exhibited a more than 50-fold increase in neutralizing activity compared to that of 1129 at between 18 h and 47 h after infection.
When the antibodies were added at 6 h postinfection, the inhibition activity of 5C4 was about 10-fold greater than the activity of 1129 at 36 h after infection (Fig. 9B and D, 6-h group). MAbs 5C4 and 1129 could not fully inhibit RSV propagation when the antibodies were added at 18 h after infection at any concentration (Fig. 9C and D, 18-h group), with no difference in the inhibitory activities of the MAbs being shown. These data demonstrate that the efficacy of 5C4 and 1129 was dependent on the intervention time point. The later that the MAbs were added after RSV inoculation, the less that they inhibited infection and/or propagation. The differences between 5C4 and 1129 also decreased as the time since infection increased.
Therapeutic efficacy of 5C4 and 1129 in RSV-infected BALB/c mice.In the postinfection treatment studies, 5C4 and 1129 were administered 1 day after inoculation with 107 PFU of RSV A2 per mouse, using doses of 5 mg/kg or 15 mg/kg. The body weight changes of the mice in each group and the RSV titers of the mice in each group were detected by the same methods used to assess prophylaxis.
As shown in Fig. 10A, the weight curves showed the earlier recovery of the mice treated with 5C4 or 1129 at both doses than the mice receiving the PBS control. MAb 5C4 was more effective than 1129 at either dose in reducing the percentage of weight lost (P < 0.01) (Fig. 10B to D). This finding was consistent with the impact of these antibodies on virus replication in both the lungs and nasal turbinates. Treatment with 5C4 after infection resulted in the absence of detectable virus in the lungs or nasal turbinates at day 5 and a significant reduction in the number of genome copies in the lung (Fig. 10E to G). The differences in postinfection efficacy between 5C4 and 1129 were particularly apparent in preventing virus replication in the nasal turbinates. Virus could not be recovered from the nasal turbinates of mice treated with 5C4 at either dose. However, virus was uniformly recovered from the nasal turbinates of 1129-treated mice at a level more than 2 log10 above the limit of detection (Fig. 10E).
Therapeutic efficacy of 5C4-IgG2a and 1129-IgG2a against human RSV. Mice (n = 10) received 5C4-IgG2a or 1129-IgG2a at two dose levels (15 mg/kg and 5 mg/kg) by i.p. injection at day 1 after infection. Untreated groups were injected with PBS. (A) The weight change of 5 mice in each group was monitored until 12 days after infection. (B to D) Comparison of the weight change between the different groups at days 6, 7, and 8, respectively. The y axis indicates the ratio of the weight at day 6, 7, or 8 to that at day 0. Five mice in each group were euthanized on day 5, and then the nasal turbinates and lungs of the mice were harvested. (E and F) RSV replication in nasal turbinates (Nose) (E) or lung homogenates (F) was detected by plaque assay. Two-tailed P values were calculated by the unpaired Student's t test. ***, P < 0.001; **, 0.001 ≤ P ≤ 0.01; *, 0.01 < P ≤ 0.05. Each treated group was first compared to the untreated group, and then each treated group was compared to each of the other treated groups. The red asterisks show the comparison between 1129 and 5C4 groups.
The postinfection therapeutic efficacy of the IgG1 versions of 5C4 and 1129 was also evaluated, and the results for weight loss and virus isolation obtained for the IgG1 MAbs were similar to those obtained for the IgG2a MAbs. The IgG1 MAbs were slightly less effective in clearing genomes detected by PCR than the IgG2a MAbs, suggesting a potential role for Fc-mediated antibody function (Fig. 11). Overall, the data showed that 5C4 substantially reduced the pulmonary and nasal turbinate viral loads compared to those found in the control group, prevented weight loss, and was significantly more effective at clearing infectious virus from the nasal turbinate than 1129. These data suggest that 5C4 and other site Ø-specific MAbs have the potential to be used as therapeutics and/or prophylactics that could reduce viral shedding and, consequently, interrupt transmission.
Therapeutic efficacy of 5C4-IgG1 and 1129-IgG1 against human RSV. Mice (n = 10) received 5C4-IgG1 or 1129-IgG1 at two dose levels (15 mg/kg and 5 mg/kg) by i.p. injection at day 1 after infection. Untreated groups were injected with PBS. (A) The weight loss of 5 mice in each group was monitored until 12 days after infection. (B to D) The weight loss between the different groups was compared at days 6, 7, and 8, respectively. The y axis indicates the percentage of the body weight at the baseline, prior to challenge. Five mice in each group were euthanized on day 5, and the nasal turbinates and lungs of the mice were harvested. (E and F) RSV titers in nasal turbinates (Nose) (E) and lungs (F) were detected by plaque assay. Two-tailed P values were calculated by the unpaired Student's t test. ***, P ≤ 0.001; **, 0.001 ≤ P ≤ 0.01; *, 0.01 < P ≤ 0.05. Each treated group was first compared to the untreated group, and then each treated group was compared to each of the other treated groups. The red asterisks show the comparison between 1129 and 5C4 groups.
DISCUSSION
In this study, we compared the in vivo potencies of RSV F-directed MAbs 5C4 and 1129 in a murine model of RSV infection. MAb 5C4 exclusively recognizes the pre-F conformation at antigenic site Ø and was originally isolated from mice immunized with a recombinant adenovirus vector expressing a wild-type version of F, including the transmembrane domain and cytoplasmic tail (10). MAb 5C4 competes for binding with other recognized site Ø MAbs (D25 and AM22) that are of human origin (10). Antibody 1129 is the murine precursor of palivizumab (3, 4), the licensed MAb used to prevent severe RSV disease in high-risk infants. Antibody 1129 targets antigenic site II, which is present on both the pre-F and the post-F conformations of the RSV F glycoprotein (10). To focus the comparison of the functions of the MAbs on the specificity of antibody binding and avoid the influence of constant regions, particularly the Fc domain, both MAbs were built on matching IgG1 or IgG2a murine backbones and produced from transfected HEK293 cells. The affinities of recombinant 5C4 and 1129 for binding stabilized pre-F protein in vitro were similar, as both antibodies neutralized by fusion inhibition at a postattachment stage. This provided the opportunity to compare the role of epitope specificity on immunity, as measured by neutralizing antibody titers. Improving the understanding of in vivo neutralizing potency is important for defining the basic mechanisms of virus neutralization and has practical implications for vaccine antigen design and the choices of therapeutic MAbs for clinical use.
The results of the comparison of 5C4 and 1129 were similar whether the comparisons were performed on the IgG2a or IgG1 background, supporting the concept that neutralizing activity is the primary activity that correlates with virological control. It also suggests that complement binding and other Fc-mediated functions have a relatively minor impact in this model. The affinity of binding of 5C4 to the pre-F protein is similar to that of 1129 (Fig. 1C), and both antibodies are equally able to bind virus-infected cells whether they are adherent or nonadherent (Fig. 1D and E). Overall, 1129 bound better to thawed stocks of the cryopreserved virus representative of RSV used for in vitro assays and in vivo challenge (Fig. 1F). Nevertheless, in vitro 5C4 had neutralizing potency about 50-fold greater than that of 1129 and a potency for inhibiting cell-to-cell spread about 10-fold greater than that of 1129 (Fig. 2). Therefore, binding affinity does not correlate with functional activity or virological control and does not explain the greater in vivo neutralizing potency of 5C4. We speculate that the improved functional activity of the pre-F protein-specific antibody recognizing site Ø is based on access. The site Ø epitope is at the apex of the pre-F protein trimer, whereas site II is on the side. On a crowded virion surface, the location of the epitope and the angle of antibody binding may significantly impact the effectiveness of neutralization.
The finding that improved in vitro neutralizing potency directly translated into in vivo neutralizing potency and protection was not necessarily expected. Published studies comparing motavizumab to palivizumab did not show a corresponding improvement in in vivo potency relative to in vitro potency (6, 7). Motavizumab has a 70-fold improvement in F binding and a 20-fold improvement in neutralizing activity in vitro compared to palivizumab. Wu et al. reported on a comparison of palivizumab and motavizumab performed in vivo (7). Although motavizumab showed greater potency in vivo than palivizumab, it was estimated to be only about 3-fold more effective in vivo, and subsequent clinical trials of motavizumab used the same dose of 15 mg/kg used for palivizumab treatment. Although the affinity of 5C4 binding to the pre-F protein is lower than that of motavizumab binding (14), the neutralizing activity of 5C4 is 10-fold greater than that of motavizumab in vitro (10), and 5C4 at the dose of 1.5 mg/kg used in the current study reduced the virus titers in the lungs nearly 1,000-fold more than the equivalent dose of 1129. As shown for motavizumab at a dose of 20 mg/kg, 5C4 at a dose of 15 mg/kg completely prevented viral replication in the upper respiratory tract. Considering that motavizumab prophylaxis in healthy term Native American infants resulted in an 87% reduction in RSV-related hospitalization (15), the use of site Ø-specific MAbs should be considered for the prevention of RSV infection in infants.
Because of the extreme potency of 5C4, it was also evaluated for its potential as a therapeutic agent given postinfection. First, the impact of 5C4 on virus propagation in vitro was greater than that of 1129 (Fig. 10). In vivo, treatment 1 day after challenge not only resulted in a reduction in the day 5 virus load in the lungs and nasal turbinates but also reduced the amount of weight lost. Therefore, it is reasonable to consider the use of site Ø-specific MAbs for the treatment of severe RSV disease during primary infection or RSV infections that occur in immunocompromised patients. It is possible that even if the immune-mediated component of RSV-mediated disease is not attenuated, it could reduce the duration of virus shedding and possibly reduce nosocomial transmission risks.
Conclusion.This study has determined that antibodies against antigenic site Ø have greater neutralizing activity than the licensed (palivizumab) and optimized (motavizumab) site II-specific MAbs in vitro and in vivo. This suggests that site Ø-directed antibodies may provide greater efficacy than currently available antibodies for RSV prophylaxis, allowing the indication for use to extend to broader target populations, possibly including therapeutic uses. Importantly, these studies support the concept that vaccine antigens designed to elicit antibody responses to antigenic site Ø and other prefusion F-specific surfaces will provide a greater level of protective immunity than antigens that primarily elicit responses restricted to antigenic site II.
MATERIALS AND METHODS
Mice and cells.BALB/c female mice between 8 and 10 weeks old were purchased from The Jackson Laboratory (Bar Harbor, ME) and cared for at the NIAID Vaccine Research Center Animal Care Program. Experimental groups were age matched. HEp-2 cell lines were purchased from ATCC, and HEK293 FreeStyle cells were purchased from Invitrogen (Carlsbad, CA).
Ethics statement.The NIH Animal Care and Use Committee approved all animal protocols used in the present study. Animal use and all procedures (VRC-13-443 and VRC-14-482) were approved by the Vaccine Research Center Animal Care and Use Committee (ACUC) in accordance with NIH policy and Animal Research Advisory Committee (ARAC) guidelines. The Vaccine Research Center is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) and is in compliance with the Animal Welfare Act and Public Health Service Policy on the humane care and use of laboratory animals. The National Institutes of Health Intramural Research Program Office of Laboratory Animal Welfare (OLAW) assurance number is D16-00602 (A4149-01).
Virus.Viral stocks were prepared and maintained as previously described (16). The titer of RSV A2 used for prophylaxis and treatment in the mouse model was 1.02 × 108 PFU/ml. RSV A2 expressing the red fluorescence Katushka protein (mKate), RSV A2-mKate, was constructed on the basis of RSV L19-mKate, which was kindly provided by Martin Moore at Emory University and previously described by Hotard et al. (17). The titer of RSV A2-mKate used for flow cytometry-based neutralization was 3.8 × 107 PFU/ml. RSV expressing the green fluorescence protein (GFP), based on RSV A2 (RSV A2-GFP), was constructed by Mark Peeples and Peter Collins (18), and the titer of RSV A2-GFP was 8.9 ×107 PFU/ml.
Construction of antibody expression plasmids.Monoclonal antibody 5C4 was isolated as previously described (10). The antibody 5C4 and 1129 heavy and light variable regions were subcloned into pVRC8400 expression plasmids containing in-frame mouse constant domains.
Expression and purification of antibodies.Plasmids expressing the heavy and light chains for 5C4 and 1129 in the context of the selected IgG isotype constant regions were cotransfected into HEK293 FreeStyle cells with the TrueFect transfection reagent (United BioSystems, MA) and maintained in suspension at 37°C for 6 to 7 days. Cell supernatants were harvested and passed over protein A agarose (Thermo Fisher Scientific Inc., MA). Bound antibodies were washed with PBS and eluted with IgG elution buffer into 1/10 volume of 1 M Tris-HCl, pH 8.0.
Expression and purification of pre-F (DS-Cav1) and post-F proteins.The expression and purification of the pre-F and post-F proteins were performed as previously reported (10, 14, 19). Briefly, plasmids expressing pre-F or post-F constructs were transfected into HEK293 FreeStyle cells in suspension. After 4 to 5 days, cell supernatants were harvested, centrifuged, filtered, and concentrated. The initial purification was performed using a Ni2+-nitrilotriacetic acid (NTA) resin (Qiagen, Valencia, CA) and an elution buffer consisting of 20 mM Tris-HCl, pH 7.5, 200 mM NaCl, and 250 mM imidazole, pH 8.0. The elution sample was then concentrated and further purified over StrepTactin resin according to the manufacturer's instructions (Novagen, Darmstadt, Germany). The protein was further purified on a Superdex 200 gel filtration column (GE Healthcare, United Kingdom) with a running buffer of 2 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.02% NaN3. The trimeric fraction was concentrated and stored at −80°C.
Immunofluorescence analysis.HEp-2 cells cultured on coverslips were incubated with or without RSV A2 at 37°C for 24 h. The cells were then fixed with 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO) and permeabilized with 0.1% Triton X-100 (Amresco LLC, OH) in PBS. The samples were blocked with 10% goat serum in PBS for 15 min, incubated with 5C4 or 1129 (dilution, 1:500) for 30 min, and then labeled with fluorescein isothiocyanate (FITC) secondary antibodies (goat anti mouse IgG-FITC [GAM-FITC]; Sigma-Aldrich, MO) for 30 min. After being washed three times, the cells were stained with 0.5% 4′,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific Inc., MA). The fluorescence signals were detected with a superresolution laser scanning confocal microscope (SP8 STED 3X; Leica Biosystems, Germany).
Pre-F and post-F protein binding assay.A kinetic ELISA was used to test the binding of the monoclonal antibodies to the RSV pre-F and post-F proteins as described previously (10, 14). Briefly, 96-well Ni2+-NTA-coated plates (Thermo Fisher Scientific Inc., MA) were coated with 100 μl RSV post-F or pre-F protein (1 μg/ml) for 1 h at room temperature. One hundred microliters of serially diluted antibody was added to each well, and the plates were incubated for 1 h at room temperature. Bound antibodies were detected by incubating the plates with 100 μl horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) or HRP-conjugated anti-human IgG (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for 1 h at room temperature. Then, 100 μl of Super AquaBlue ELISA substrate (eBioscience, CA) was added to each well and the plates were immediately read at 405 nm using a Dynex Technologies microplate reader (Chantilly, VA). Between steps, the plates were washed with PBS-Tween 20.
Sandwich ELISA.The plates were coated with 100 μl MAb 3G3 (0.1 μg/ml; 3G3 is an MAb isolated from a mouse and has been optimized for virus capture in combination with PCR analysis [unpublished data]) in 20 mM phosphate buffer, pH 7.4 (16.2 mM Na2HPO4 and 3.8 mM NaH2PO4), at 37°C for 2 h and were washed with PBS-Tween 20 one time. They were then blocked with 350 μl PBS containing 2% bovine serum albumin for 2 h at 37°C and washed. Two hundred microliters of 1 × 106 PFU the RSV A2 strain was added to the antibody-coated plates, and the plates were incubated at 37°C for 1 h. The plates were washed five times. Then, 100 μl of 5C4-HRP or 1129-HRP was added to the plates and the plates were incubated at 37°C for 30 min. The plates were washed as described above, and 100 μl of tetramethylbenzidine substrate solution was added to the wells. The reaction was stopped by adding 50 μl of 2 M H2SO4 after incubation at 37°C for 15 min, and the absorbance at 450 nm was measured by use of a reference wavelength of 620 nm.
Fluorescence plate reader neutralization assay.The neutralization activity of the monoclonal antibodies or serum samples was measured using HEp-2 cells infected with a subtype A Katushka-expressing RSV (RSV A2-mKate) reporter. RSV A2-mKate was diluted to 1:6.5 and then added to serial 4-fold dilutions (beginning with a dilution of 1:10) of antibody or serum in 96-well plates, and the plates were incubated at 37°C for 1 h. The antibody concentrations ranged from 100 to 0.000381 ng/μl, and the serum dilutions ranged from 1:10 to 1: 2,621,440. After 1 h, 50 μl of the virus-antibody or virus-serum mixture was added to each well of a 384-well plate which had been seeded with 1.5 × 104 HEp-2 cells in 100 μl per well. The fluorescence intensity of the infection was monitored at 22 h in a microplate reader at an excitation wavelength of 588 nm and an emission wavelength of 635 nm (SpectraMax Paradigm; Molecular Devices, Sunnyvale, CA). Virus neutralization was measured as the reduction of the fluorescence of the samples compared to the fluorescence for the virus control. Data were analyzed and the 50% effective concentration (EC50) for neutralization was calculated by curve fitting with GraphPad Prism software (GraphPad Software Inc., San Diego, CA).
Inhibition of RSV spreading.The inhibition of RSV spreading activity by the monoclonal antibodies was measured as a function of RSV A2-GFP infection in HEp-2 cells. RSV A2-GFP was diluted 1:16 and then added to each well of a 96-well plate containing HEp-2 cells seeded at 5 × 104/100 μl per well. Then, they were incubated at 37°C for 6 h. The plate was washed once with minimal essential medium (MEM). Serial 4-fold dilutions (beginning with a dilution of 1:10) of antibody were added. Antibody concentrations ranging from 100 to 0.000381 ng/μl were evaluated, and serum dilutions ranged from 1:10 to 1: 2,621,440. After 36 h, cells were detected by flow cytometry. Prior to assessment by flow cytometry, cells were treated with trypsin to achieve a single-cell suspension, and the data were analyzed by curve fitting and nonlinear regression (GraphPad Prism software; GraphPad Software Inc., San Diego, CA) to determine the infection ratio at a given antibody concentration. Results were reported as the concentration resulting in a 50% inhibition of spreading (IS50).
Assay of antibody clearance from serum in a mouse model.To evaluate in vivo potency, 1129 or 5C4 was administered via intraperitoneal (i.p.) injection at a dose of 15 mg/kg/mouse. Five mice were tested for each antibody. Serum samples were collected from the tail vein every 2 days from the 1st day after i.p. injection to day 13 after i.p. injection. All of the serum samples were tested for the neutralization of RSV A2-mKate with a fluorescence plate reader. The data were analyzed, and the EC50 for neutralization was calculated by curve fitting with GraphPad Prism software (GraphPad Software Inc., San Diego. CA).
Prophylaxis experimental design.Evaluation of prophylactic immunization was performed as previously described (20, 21). After anesthesia with isoflurane, BALB/c mice were intranasally (i.n.) inoculated with 100 μl of 108 PFU/ml human RSV A2 on day 0. At 24 h before RSV i.n. infection, animals received 1129 or 5C4 via i.p. injection at a dosage of 1.5 mg/kg or 15 mg/kg, while the control group received PBS. At 5 days after RSV challenge, the animals were sacrificed and their lungs and nasal turbinates were harvested. The lungs were homogenized by use of a gentleMAC dissociator (Miltenyi Biotec Company, MA) and were prepared in 2 ml MEM containing 10% fetal bovine serum. The lung suspensions were used to examine lung viral titers by plaque assay by titration on HEp-2 cells and genome content by detection of the N gene by real-time PCR. Cytokines in lung homogenates were detected by use of a Luminex bead array (Assaygate, Inc., MA). Nasal turbinates were ground with a mortar and pestle in 2 ml MEM, and the virus titer of the nasal turbinate supernatant was examined by plaque assay as described above. The weight of each mouse was measured daily through day 12 postchallenge.
Treatment experimental design and examination.After anesthesia with isoflurane, BALB/c mice were i.n. inoculated on day 0 with 100 μl of 108 PFU/ml human RSV A2. At 24 h after RSV challenge, mice were administered 5 mg/kg or 15 mg/kg of 1129 or 5C4 i.p., while PBS was injected as a control. At 5 days after RSV challenge, the animals were sacrificed, and their lungs and nasal turbinates were harvested and samples were processed as described above.
Histopathology.At 5 days after RSV challenge, the animals were anesthetized and sacrificed, and their lungs were harvested. Then, the lung tissue was inflated and fixed in 10% formalin and transverse sections (thickness, 10 μm) were stained with hematoxylin and eosin.
Plaque assays.Lung homogenates were diluted 10-fold (concentration range, 1 to 1:1,000), and then 50 μl was added to each well in 12-well plates which had been seeded with HEp-2 cells at 2 × 105/100 μl per well 1 day before and the plates were incubated at room temperature for 1 h. One milliliter of 0.75% methylcellulose was then added to each well, and the plates were incubated at 37°C for 4 days, at which time 1 ml of 10% formalin was added to each well for 1 h. After the cells were rinsed with tap water, the cell monolayer was stained with hematoxylin for 20 min and then eosin for 2 to 5 min. After washing three times, the plates were air dried for 24 h and plaques were counted under a dissecting microscope.
ACKNOWLEDGMENTS
Support for this work was provided by the National Natural Science Foundation of China (81361120408 and 81401668) and the U.S. National Institute of Allergy and Infectious Diseases Intramural Research Program.
M.Z., Z.-Z.Z., M.C., J.S.M., B.S.G., and N.-S.X. are named inventors on a Chinese patent pending for MAb 5C4.
FOOTNOTES
- Received 30 January 2017.
- Accepted 9 May 2017.
- Accepted manuscript posted online 24 May 2017.
- Copyright © 2017 American Society for Microbiology.
