Broad and potently neutralizing monoclonal antibodies isolated from human survivors of New World hantavirus infection

Abstract

New World hantaviruses (NWHs) are endemic in North and South America and cause hantavirus cardiopulmonary syndrome (HCPS), with a case fatality rate of up to 40%. Knowledge of the natural humoral immune response to NWH infection is limited. Here, we describe human monoclonal antibodies (mAbs) isolated from individuals previously infected with Sin Nombre virus (SNV) or Andes virus (ANDV). Most SNV-reactive antibodies show broad recognition and cross-neutralization of both New and Old World hantaviruses, while many ANDV-reactive antibodies show activity for ANDV only. mAbs ANDV-44 and SNV-53 compete for binding to a distinct site on the ANDV surface glycoprotein and show potently neutralizing activity to New and Old World hantaviruses. Four mAbs show therapeutic efficacy at clinically relevant doses in hamsters. These studies reveal a convergent and potently neutralizing human antibody response to NWHs and suggest therapeutic potential for human mAbs against HCPS.

Keywords: Andes virus; Bunyavirus; Sin Nombre virus; antibody; cross-reactivity; hantavirus; infection.

Figure 1.. NWH-reactive mAbs isolated from individuals previously infected with SNV or ANDV exhibit diverse patterns of neutralization potency, cross-reactivity, and mechanisms of neutralization

(A and B) Neutralization potency of SNV-reactive mAbs (A) or ANDV-reactive mAbs (B) to pseudotyped VSV particles (pVSV/SNV or pVSV/ANDV), SNV strain SN77734 (SNV), or ANDV strain Chile-9717869 (ANDV). Antibodies in a dilution series of decreasing concentrations were incubated with pVSVs or authentic virus, the suspension was used to inoculate cells, and then GFP⁺ cells or plaques were counted to determine relative infectivity. IC₅₀ or IC₉₀ values were obtained using non-linear fit analysis, with the top of the curve constrained to 100 and the bottom of the curve constrained to 0, using Prism software version 7 (GraphPad Software). The colors indicate the relative potency of the antibody. The data shown are average values from 2–3 independent experiments. Binding to proteins for each hantavirus species was determined by flow cytometric analysis. Gn and Gc were displayed on the surface of mammalian cells and incubated with decreasing concentrations of mAb. EC₅₀ binding values were obtained using non-linear fit analysis, with the bottom of the curve constrained to 0, using Prism software. The value for %PE⁺ cells was determined by gating on cells stained only with secondary antibodies. The colors indicate the relative potency of the antibody. > indicates that the neutralization or reactivity was not detected at the highest concentration tested, 20 μg/mL. NT indicates that the mAb was not tested. The data shown are average values from 3 independent experiments. PCDH-1 blocking (%) was determined through a flow cytometric assay, in which mAbs were added at saturating concentration before the addition of the soluble PCDH-1 domain, sEC-1 labeled with Alexa Fluor 647 dye. PCDH-1 blocking was defined by the reduction of the maximal binding score to <50% of un-competed binding (green boxes). The data shown are average values from 3 independent experiments. The fusion index (%) was determined by adding mAbs to Vero cells transfected with cDNAs encoding SNV or ANDV Gn/Gc, and then inducing fusion through exposure to medium with low pH and counting the percentage of multinucleated cells by fluorescent microscopy. The percentage of multinucleated cells in mAb-treated samples then was divided by the percentage of multinucleated cells in a non-mAb-treated sample (representing maximal fusion index). Fusion-inhibiting mAbs were defined by the reduction of the maximal fusion index to <50% of non-mAb-treated cells (yellow boxes). The data shown are the average values from 3 independent experiments. (C) Representative binding curves of neutralizing antibodies mediating complete (top) or incomplete (bottom) neutralization for authentic SNV (left) or ANDV (right). The dotted lines indicate 10% or 50% relative infectivity. The data shown are average values for technical replicates ± SDs. The experiment was performed 2–3 times independently with similar results; one experiment is shown. See also Figures S1–S3 and Tables S1 and S2.

Figure 2.. NWH-reactive mAbs show binding and neutralizing activity against OWH species

(A and B) Binding potency of mAbs isolated from SNV-immune (A) or ANDV-immune (B) human individuals for four OWH species: Puumala (PUUV), Dobrava-Belgrade (DOBV), Hantaan (HTNV), and Seoul (SEOV). EC₅₀ binding values were obtained using non-linear fit analysis, with the bottom of curve constrained to 0, using Prism software. The data are shown as average values from 3 independent experiments. The colors indicate the relative potency of the antibody. > indicates no detectable reactivity at concentrations >20 μg/mL. (C) Neutralization potency of broadly reactive mAbs to HTNV strain 76–118, DOBV strain Dobrava, PUUV strain K27, or SEOV strain SR-11. Neutralization potency was determined with a plaque reduction neutralization assay. The neutralizing mAb concentration indicates the concentration at which there was a ≥50% reduction in plaque count by each mAb. The data shown are average values for technical replicates ± SDs. The experiments were performed 2–3 times independently with similar results. See also Figures S1 and S2.

Figure 3.. NWH-reactive mAbs bind to at least eight major antigenic sites on Gn or Gc based on competition-binding analysis

(A) Unlabeled antibodies were incubated with Expi293F cells transfected with cDNA encoding ANDV Gn/Gc at saturating concentrations and then competed with a second antibody labeled with Alexa Fluor 647. Percent competition was analyzed and quantified using flow cytometry as compared to the un-competed binding of the second mAb. Competing antibodies were defined as those with <33% of the maximal un-competed binding in the presence of an unlabeled first antibody (black). Non-competing antibodies were defined as those with >66% of the maximal un-competed binding (white). Intermediate competing antibodies were defined as those with 33%–66% of the maximal un-competed binding (gray). Antibodies are colored based on neutralization potency to ANDV and SNV. Antibodies were clustered based on the Pearson correlation generated relatedness score based on normalized competition values. The values shown are averages from 3 independent experiments. (B) Binding to Gn displayed on the surface of Expi293F cells was determined using a flow cytometry-based binding assay, as described previously. Binding to recombinant Gc was determined using an ELISA. EC₅₀ binding values were obtained using non-linear fit analysis with the bottom of the curve constrained to 0, using Prism software. The colors indicate the relative potency of the antibody. > indicates that neutralization or reactivity was not detected at the highest concentration tested, 20 μg/mL. The data shown are the average values from 3 independent experiments.

Figure 4.. Four NWH mAbs protect Syrian hamsters in ANDV challenge

(A) 8-week-old Syrian hamsters (n = 6 per treatment group) were inoculated with 200 PFU of ANDV intramuscularly (i.m.), and 5 mg/kg of indicated mAb was administered intraperitoneally (i.p.) at 3 and 8 dpi. Animals were treated with a dengue-specific mAb (rDENV 2D22) to serve as an isotype control. The statistical analysis was done using a log-rank (Mantel-Cox) test comparing each group to the control (rDENV 2D22); *p < 0.01; **p < 0.01; ***p < 0.001; ns, non-significant. (B) Body weight measurements averaged for each treatment group. The data shown are mean ± S.E.M. for each treatment group. (C) Lungs and livers were collected upon euthanasia and used to determine viral titer in tissue. Dots with black borders indicate animals that were found dead or euthanized according to Institutional Animal Care and Use Committee (IACUC) protocol before the termination of the study. The dotted line indicates the limit of detection. See also Figure S4.