Structural basis for potent cross-neutralizing human monoclonal antibody protection against lethal human and zoonotic severe acute respiratory syndrome coronavirus challenge

Abstract

Severe acute respiratory syndrome coronavirus (SARS-CoV) emerged in 2002, and detailed phylogenetic and epidemiological analyses have suggested that it originated from animals. The spike (S) glycoprotein has been identified as a major component of protective immunity, and 23 different amino acid changes were noted during the expanding epidemic. Using a panel of SARS-CoV recombinants bearing the S glycoproteins from isolates representing the zoonotic and human early, middle, and late phases of the epidemic, we identified 23 monoclonal antibodies (MAbs) with neutralizing activity against one or multiple SARS-CoV spike variants and determined the presence of at least six distinct neutralizing profiles in the SARS-CoV S glycoprotein. Four of these MAbs showed cross-neutralizing activity against all human and zoonotic S variants in vitro, and at least three of these were mapped in distinct epitopes using escape mutants, structure analyses, and competition assays. These three MAbs (S109.8, S227.14, and S230.15) were tested for use in passive vaccination studies using lethal SARS-CoV challenge models for young and senescent mice with four different homologous and heterologous SARS-CoV S variants. Both S227.14 and S230.15 completely protected young and old mice from weight loss and virus replication in the lungs for all viruses tested, while S109.8 completely protected mice from weight loss and clinical signs in the presence of viral titers. We conclude that a single human MAb can confer broad protection against lethal challenge with multiple zoonotic and human SARS-CoV isolates, and we identify a robust cocktail formulation that targets distinct epitopes and minimizes the likely generation of escape mutants.

FIG. 1.

Mapping of neutralizing epitopes on the SARS-CoV S glycoprotein recognized by human MAbs through sequence analysis and cross-competition studies. (A) Sequence analysis of the amino acid differences in the SARS-CoV S glycoproteins of zoonotic and human epidemic isolates and their associations with binding to neutralizing human MAbs. The graphic representation of the SARS-CoV S glycoprotein shows the locations of the variant amino acids in the RBD and heptad repeat 2 (HR2). Three neutralizing domains have previously been identified by using murine MAbs and antisera targeting the S glycoprotein. Domain I is localized in the N terminus of the S1 domain between amino acid residues 130 and 150; however, the mechanism by which these antibodies neutralize infectivity remains unknown (10). Domain II includes the RBD (residues 318 to 510), where antibodies likely block binding to the SARS-CoV receptor ACE2, based on studies describing the crystal structures of two different neutralizing MAbs in complex with the RBD (13, 15, 32). Finally, domain III includes HR2 (residues 1143 to 1157) and is likely neutralized by disturbing the interaction between HR2 and HR1, thereby abolishing fusion activity (20). Amino acids associated with recognition by MAbs are color coded as follows: blue, group I MAbs; red, group II MAbs; green, group III MAbs; yellow, group IV MAbs. MAb groups are shown in Table 2 and defined in the text under “Identification of cross-neutralizing MAbs.” (B) Cross-competition of MAbs binding to the SARS-CoV S glycoprotein. Shown are the percentages of inhibition of the binding of 3 biotinylated MAbs (S109.8, S227.14, and S230.15; concentration, 0.1 μg/ml) to the recombinant SARS-CoV S glycoprotein by a panel of 23 unlabeled competing MAbs (at a saturating concentration [5 μg/ml]) belonging to groups I through VI. Error bars, standard deviations from triplicates. ctr, control.

FIG. 2.

Locations of neutralization escape variant mutations and effects on the structure of the SARS-CoV RBD. (A and B) The S109.8 escape variant mutations T332I and K333N (A) and the S230.15 escape variant mutation L443R (B) were mapped onto the structure of the SARS-CoV RBD. The changed amino acid residues are shown in red. (C) The locations of all the important amino acid residues associated with the cross neutralizing MAbs are highlighted in the SARS-CoV RBD. Yellow, amino acid residues associated with S109.8; red, amino acid residues associated with S227.14; purple, amino acid residues associated with S230.15.

FIG. 3.

Prophylactic treatment of lethal SARS-CoV infection in 12-month-old BALB/c mice with 25 μg of cross-neutralizing MAbs. (A to C) Body weights of mice infected with icUrbani (A), icGZ02 (B), or icHC/SZ/61/03 (C) were measured daily after passive transfer of 25 μg of MAb S109.8 (+), S227.14 (○), S230.15 (×), or D2.2 (□). (D and E) Lung tissues were harvested from infected mice on day 2 (D) and day 5 (E) postinfection and were assayed for infectious virus. Error bars, standard deviations (n = 3).

FIG. 4.

Prophylactic treatment of lethal SARS-CoV infections in 12-month-old BALB/c mice with 250 μg of cross-neutralizing MAbs. (A to C) Body weights of mice infected with icUrbani (A), icGZ02 (B), or icHC/SZ/61/03 (C) were measured daily after passive transfer of MAbs S109.8 (+), S227.14 (○), S230.15 (×), and D2.2 (□), all at 250 μg/mouse, given alone or as a 1:1:1 cocktail of the three neutralizing MAbs (▵). (D and E) Lung tissues were harvested from infected mice on day 2 (D) and day 5 (E) postinfection and were assayed for infectious virus. Error bars, standard deviations (n = 3).

FIG. 5.

Prophylactic treatment of lethal SARS-CoV infections in 10-week-old BALB/c mice with 25 μg of cross-neutralizing MAbs. (A) Body weights of mice infected with MA15 were measured daily after passive transfer of 25 μg of MAb S109.8 (+), S227.14 (○), S230.15 (×), or D2.2 (□). (B) Lung tissues of mice infected with MA15 or icHC/SZ/61/03 were harvested on day 2 postinfection and assayed for infectious virus. (C) Lung tissues of mice infected with MA15 were harvested on day 4 postinfection and assayed for infectious virus. Error bars, standard deviations (n = 3). *, only one animal out of three had detectable virus titers.

FIG. 6.

Postinfection treatment of 12-month-old BALB/c mice infected with SARS-CoV. (A) Body weights of mice infected with GZ02 were measured daily after passive transfer of 250 μg of MAb S230.15 at day −1 (+), day 0 (○), or day 1 (×), 2 (□), or 3(▵) postinfection. (B) Lung tissues of mice infected with GZ02 were harvested on days 2 and 4 postinfection and were assayed for infectious virus. Error bars, standard deviations (n = 5). *, only one animal out of five had detectable virus titers.

FIG. 7.

Light photographs of preterminal bronchioles (PB) in the lungs of 12-month-old BALB/c mice that had received 250 μg of a human MAb prior to SARS-CoV infection and were sacrificed 5 days postinoculation. Virus-induced peribronchiolar inflammation (solid arrows) is evident in mice treated with the control MAb D2.2 and infected with icUrbani (A), icGZ02 (C), or icHC/SZ/61/03 (E). Numerous hyaline membranes (dashed arrows) are present in the alveolar airspaces of mice treated with the control MAb. No inflammation or hyaline membrane formation can be observed in mice treated with 250 μg of S230.15 and subsequently infected with icUrbani (B), icGZ02 (D), or icHC/SZ/61/03 (F). AL, alveoli; AD, alveolar ducts; BV, blood vessels. Tissues were stained with hematoxylin and eosin. Magnification, ×100.

FIG. 8.

Light photographs of preterminal bronchioles (PB) and terminal bronchioles (TB) in the lungs of 12-month-old BALB/c mice that received 250 μg of a human MAb after infection with SARS-CoV and were sacrificed 5 days postinoculation. (A) No inflammation or hyaline membrane formation can be observed in mice treated with 250 μg of S230.15 on the day of infection with icGZ02. (B through D) Increasing-virus induced peribronchiolar inflammation (solid arrows) is evident in mice treated with 250 μg of MAb S230.15 at day 1 (B), 2 (C), or 3 (D) postinfection. AL, alveoli; AD, alveolar ducts; BV, blood vessels. Tissues were stained with hematoxylin and eosin. Magnification, ×100.