A non-neutralizing glycoprotein B monoclonal antibody protects against herpes simplex virus disease in mice

Figure 1. Isolation and characterization of B cells specific for HSV-2. HSV-specific BCRs were isolated from GC B cells in inguinal lymph nodes, memory B cells in spleens, and plasmacytes in BM of mice after ΔgD-2 vaccinations. Single GC and memory B cells were introduced into Nojima cultures, while single plasmacytes were subjected to a single-cell RT-PCR method. (A) Representative flow diagrams for GC B cells (left), memory B cells (middle), and plasmacytes (right). GC and memory B cells were pre-gated on B220+CD138– cells and B220+ cells, respectively. (B) The reactivity of culture supernatant IgGs against HSV-2(G) infected (HSV+) and uninfected (HSV–) Vero cell lysates was assessed by ELISA. Representative ELISA screening for single GC B cell cultures is shown. Each dot represents a single B cell (n = 672). Gray dots on the HSV+ column represent samples that also bound HSV– lysates. The dotted line indicates a reactivity threshold determined as mean + 6SD of B cell–negative, mock-cultured culture supernatants. (C) Distributions of VH gene segment use by HSV-specific GC B cells (top, n = 61) and memory B cells (bottom, n = 15). (D) Distributions of VH mutation frequency for HSV-specific…

Figure 2. HSV-2(G) lysate binding by rAbs. rAbs were generated from selected samples of single plasmacytes (BMPC-23) and single B cell cultures of GC or memory B cells (others). Serially diluted rAbs were tested for binding to HSV2-G lysates (HSV+, filled circles) and uninfected Vero lysates (HSV–, open circles) by ELISA. Representative data (OD values) of at least 2 independent experiments and interpolated 50% maximal effective concentration (EC50) values are shown.

Figure 3. Functional characteristics of HSV-specific rAbs. (A) mAbs with a murine IgG1 (open symbols) or IgG2c (filled symbols) Fc (1 mg/mL) were tested for their ability to induce FcγRIV activation when incubated with HSV-2–infected Vero cells. Each antibody was tested at least twice in duplicate, and mean results are shown (P = 0.03 comparing the fold induction elicited by IgG2c vs. IgG1 for the subset of 18G4, 19G7, 22D10, 33B8, 35H7, and BMPC-23; Wilcoxon’s matched-pairs signed rank test). (B) A subset of antibodies was tested for their ability to neutralize HSV in the absence (left) or presence (right) of complement. Results are shown as percent neutralization relative to control and are the mean of duplicate wells at each concentration. (C) Female C57BL/6 mice received 750 μg of the IgG1 or the IgG2c version of indicated antibody 1 day before an LD90 challenge (5 × 104 PFU/mouse) with HSV-2(4674). Percentage survival is shown; n = 10 mice per group, 2 independent experiments. (D) Mice were treated i.p. with 250 or 500 μg of BMPC-23 or equivalent concentration of control (VD60 lysate–vaccinated) immune serum 24 hours after an LD90 skin challenge (n = 5 mice per group with 2…

Figure 4. BMPC-23 protects against vaginal HSV-2 infection. Medroxyprogesterone-pretreated female mice (n = 5 per group) were treated i.p. with 250 or 500 μg of BMPC-23, 250 μg of 5E7, or 500 μg of immune serum obtained from control VD60 lysate–vaccinated mice 24 hours before vaginal inoculation with HSV-2(4674). Mice were scored daily for signs of disease using the following scale: 1, erythema at inoculation site; 2, hair loss, erythema, edema, urinary retention; 3, severe edema, hair loss, lesion formation, constipation and urinary retention, hind-limb paresis; 4, severe ulceration, hind-limb paralysis; and 5, death. (A) Bar graphs show disease scores for mice at each day. (B) Survival is compared with that of the VD60 control–treated mice by Gehan-Breslow-Wilcoxon test, ***P < 0.001.

Figure 5. BMPC-23 binds glycoprotein. (A) HSV-protective rAbs BMPC-23 (orange), 22D10 (blue), and 33B8 (red) were tested for binding to gB in a Luminex binding assay. MFI, median fluorescence intensity. Representative mean data of at least 2 independent experiments are shown. (B) Western blotting was performed with uninfected Vero or HSV-2–infected Vero cell lysates as the antigen and probed with BMPC-23. Molecular weight markers are indicated on the left, and the arrow denotes the monomeric form of gB (~116 kDa); the gel is representative of 2 independent experiments. (C) The binding kinetics of Fab of BMPC-23 was evaluated by biolayer interferometry at each of the indicated concentrations.

Figure 6. Isolation and characterization of gB-specific BCRs from GC B cells. HSV gB-specific GC B cells were identified by screening of HSV-reactive Nojima culture samples by a Luminex assay. Inhibition of BMPC-23 by gB-specific rAbs was assessed in a multiplex binding assay as described in Methods. The gB-specific rAbs (filled symbols) were grouped by their inhibitory capacity: strong (HSV010-4, -7, and -34; left), weak (HSV010-6, -9, -13, -14, and -28; middle), and no inhibition (HSV010-20; right). Inhibition by self (BMPC-23, open circles) and an irrelevant, hapten-specific mAb (H33Lγ1, open squares) is coplotted in each panel. The y axis indicates MFI percentage of maximal binding, determined as the mean MFI in the presence of H33Lγ1. Dotted lines indicate MFI values corresponding to 50%, 90%, and 99% inhibition, respectively.

Figure 7. Mapping of binding sites for BMPC-23 and HSV010-13. (A) Structures of HSV-1 gB in the prefusion (6Z9M) and postfusion (2GUM) states were aligned on domain VI to illustrate the orientation of domain IV relative to the viral and host cell membranes. (B) Models built with cryo-EM densities for BMPC-23 and HSV010-13 were aligned, and a single Fab from each structure is shown. Densities were only of sufficient quality to build models of the Fv domains (Supplemental Figure 3). (C) Domain IV and Fv domains of BMPC-23 as well as a neutralizing mAb, 93k (6vni), which recognizes domain IV of the alphaherpesvirus varicella zoster virus, were extracted and aligned to domain IV of HSV-1 gB in the prefusion state. The side view is the same as in A, while the top view is rotated 90° with the top moving toward the reader. The prefusion gB is shown as a transparent surface, and the Fv-bound domain IV is shown as a cartoon. BMPC-23 and HSV010-13 are occluded from binding domain IV in the prefusion state, while the 93k epitope is accessible.

Figure 8. Human IgG1 BMPC-23 protects against HSV-2 in FcγR-humanized mice. FcγR-humanized mice received 750 μg of human IgG1 BMPC-23 or a comparable concentration of IgG isolated from pooled HSV-seropositive [hIgG(HSV+)] or HSV-seronegative [hIgG(HSV–)] human serum 24 hours before lethal challenge on the skin with HSV-2(4674) [n = 6 for BMPC-23 and n = 3 each for hIgG(HSV+) and hIgG(HSV–)]. (A) The percentage of mice that survived was compared by Gehan-Breslow-Wilcoxon test, *P < 0.05. (B) Bar graphs show the disease scores for each group of mice (percentage of total mice in the group) on days 5, 8, and 11.

Figure 9. Combination of BMPC-23 with anti-gD neutralizing immune serum provides comparable protection to BMPC-23 alone. Antibodies against gD were enriched by passing of pooled serum obtained from mice vaccinated with recombinant adjuvanted gD protein over Protein L and gD-lectin columns. (A and B) The indicated concentrations of murine IgG2c BMPC-23, 5E7, or gD serum antibody were assayed for neutralization activity by plaque assay in duplicate (A) and ability to activate the FcγRIV (B). Mice were treated i.p. with 250 μg of control serum (VD60 lysate–vaccinated mice), 250 μg of 5E7, 250 μg of gD serum antibody, 250 or 500 μg of BMPC-23, or a combination of 250 μg of gD serum antibody and BMPC-23 each 1 day before LD90 challenge (5 × 104 PFU/mouse) with HSV-2(4674). (C and D) Percentage survival (C) and disease scores for individual mice (D); n = 5 mice per group. Survival is compared with that of VD60 control–treated mice by Gehan-Breslow-Wilcoxon test, **P < 0.01.