Human antibody recognizing a quaternary epitope in the Puumala virus glycoprotein provides broad protection against orthohantaviruses

Abstract

The rodent-borne hantavirus Puumala virus (PUUV) and related agents cause hemorrhagic fever with renal syndrome (HFRS) in humans. Other hantaviruses, including Andes virus (ANDV) and Sin Nombre virus, cause a distinct zoonotic disease, hantavirus cardiopulmonary syndrome (HCPS). Although these infections are severe and have substantial case fatality rates, no FDA-approved hantavirus countermeasures are available. Recent work suggests that monoclonal antibodies may have therapeutic utility. We describe here the isolation of human neutralizing antibodies (nAbs) against tetrameric Gn/Gc glycoprotein spikes from PUUV-experienced donors. We define a dominant class of nAbs recognizing the "capping loop" of Gn that masks the hydrophobic fusion loops in Gc. A subset of nAbs in this class, including ADI-42898, bound Gn/Gc complexes but not Gn alone, strongly suggesting that they recognize a quaternary epitope encompassing both Gn and Gc. ADI-42898 blocked the cell entry of seven HCPS- and HFRS-associated hantaviruses, and single doses of this nAb could protect Syrian hamsters and bank voles challenged with the highly virulent HCPS-causing ANDV and HFRS-causing PUUV, respectively. ADI-42898 is a promising candidate for clinical development as a countermeasure for both HCPS and HFRS, and its mode of Gn/Gc recognition informs the development of broadly protective hantavirus vaccines.

Fig. 1:. Isolation and characterization of anti-Gn/Gc mAbs from PUUV-experienced donors.

(A) Phylogenetic tree of the M segment-encoded GPC sequences from Old World and New World hantaviruses (OWH/NWH) based on amino acid (aa) sequences shown in Table S1. ANDV, Andes virus; BAYV, Bayou virus; BCCV, Black Creek Canal virus; CHOV, Choclo virus; DOBV, Dobrava-Belgrade virus; HTNV, Hantaan virus; LGNV, Laguna Negra virus; MPRLV, Maporal virus; PHV, Prospect Hill virus; PUUV, Puumala virus; SEOV, Seoul virus; SNV, Sin Nombre virus; TULV, Tula virus. Scale bar represents 0.07 aa substitutions per site. (B) Frequency of hantavirus Gn/Gc-reactive B cells isolated with rVSV-mNG particles bearing PUUV or ANDV Gn/Gc from PUUV-naïve and PUUV-experienced donors. Fluorescence-activated cell sorting plots are gated on CD19⁺/CD20⁺/IgD⁻/IgM⁻/virus⁺ B cells. (C) Somatic hypermutation (SHM) load as determined by the number of V_H and V_L nucleotide substitutions away from the predicted V_H and V_L germline. Bar indicates median. (D) Binding reactivity of 180 isolated mAbs (25 nM) to rVSV-PUUV-Gn/Gc, as determined by ELISA. A₄₅₀, absorbance at 450 nm. The red dashed line indicates the threshold for designating binders (A₄₅₀=0.3). Bar indicates median. Averages for PUUV Gn/Gc binders (n=4) from two experiments. (E) Binding cross-reactivity of 135 isolated mAbs (25 nM) to rVSVs bearing Gn/Gc from the indicated viruses, as determined by ELISA. The red dashed line indicates the threshold for designating binders (A₄₅₀=0.3). ANDV, SNV: averages (n=2) from one experiment. Other viruses, averages (n=4) from two experiments.

Fig. 2:. Neutralization activity of PUUV Gn/Gc-reactive mAbs.

(A) Binding activity of mAbs (25 nM) to rVSV-PUUV-Gn/Gc, as determined by ELISA (see Fig. 1D), plotted against IC₅₀ values calculated from rVSV-PUUV-Gn/Gc neutralization curves displayed as averages (n≥4) from at least two experiments (also see Fig. S3). Data points are colored according to neutralization activity with data points in shades of blue indicating IC₅₀ > 1 nM and points in shade of green indicating IC₅₀ < 1 nM. Antibodies with IC₅₀ values >100 nM are designated as non-neutralizers (N.N.). Dashed red lines indicate the threshold for PUUV Gn/Gc binders (A₄₅₀=0.3; x-axis) and potent neutralizers (IC₅₀=1 nM; y-axis). (B) Frequency of mAbs grouped according to their neutralization activity. Number in the center of the pie chart denotes the total number of mAbs analyzed. (C) Frequency of prioritized mAbs grouped according to their neutralization activity towards rVSVs bearing divergent OWH and NWH Gn/Gc. Number in the center of the pie chart indicates the total number of mAbs analyzed. Each group is represented as a segment proportional to the group size. (D) Neutralizing activity of 50 prioritized mAbs (IC₅₀ values) against the indicated rVSVs. mAbs with IC₅₀>10 nM are designated non-neutralizers (N.N.), and mAbs leaving an un-neutralized virus fraction are depicted as striped squares. An un-neutralized virus fraction is defined as a residual normalized virus infection of >5% at the highest mAb concentration tested. Averages of IC₅₀ values (n≥6) from at least two experiments.

Fig. 3:. Epitope binning of nAbs by competition analysis.

(A) Heatmap of competitive pairwise mAb binding studies to rVSV-PUUV-Gn/Gc. Percent mAb competition by the second mAb (columns) was normalized to the binding ELISA value for un-competed binding by the first mAb (rows). Antibodies were clustered according to their Pearson correlation-generated relatedness scores using Morpheus software (Broad Institute, Cambridge, USA). Averages (n=2-4) from one to two experiments. (B) Proportion of mAbs competing with ADI-42898 or ADI-42098 was determined in a high-throughput yeast competition assay by binding of IgG-displaying yeast cells to soluble PUUV Gn^H/Gc pre-complexed with F(ab)s of ADI-42898 or ADI-42098. Number in the center of the pie chart denotes the total number of mAbs analyzed. Each group (I-IV) is represented as a segment proportional to the group size. (C) Heatmap showing the capacity of selected mAbs to compete with ADI-42898 or ADI-42098 for binding to PUUV Gn^H/Gc from the experiments in panel B. mAbs marked in bold were tested in both competition-binding studies.

Fig. 4:. Epitope mapping of nAbs by escape mutant analysis.

(A) Schematic representation of the organization of lattices of Gn/Gc tetramers in viral particles (top, en face view; bottom, side view). Gn, green; Gn capping loop, purple; Gc domain I, red; Gc domain II, yellow; Gc domain III, blue; Gc fusion loops, orange. The approximate position of the viral membrane is indicated with a black line. (B) Ribbon representation of the ANDV Gn^H/Gc pre-fusion dimer (modeled from its monomer, PDB: 6Y5W; (14)) with the two-fold molecular axis of the dimeric spike indicated (“2”). Residues exchanged in the neutralization-escape mutants are shown as blue (Gn) or green (Gc) spheres. (C–D) The Gn^H:Gc and Gc:Gc interfaces in panel B (dotted rectangles) were enlarged for clarity and rotated as indicated to show neutralization-escape mutants in the Gn capping loop (C) and Gc (D). In panel C, the cd fusion loop in Gc associated with the capping loop in Gn is shown in orange. (E) The PUUV Gc post-fusion trimer with the three-fold molecular axis indicated (“3”) (PDB: 5J81; (20)). Labeling of Gc domains and escape mutations as in panel B. (F) Heatmap showing the capacity of mAbs to neutralize rVSV-PUUV-Gn/Gc bearing neutralization-escape Gn/Gc mutants (also see Fig. S5). mAbs were grouped according to competition Groups I-IV (see Fig 3). Group I = ADI-42898 competitors; Group II = ADI-42098 competitors; Group III = ADI-42898/ -42098 non-competitors; Group IV = undefined/ Gn^H/Gc non-binders. Neutralization averages (n=6) from two experiments.

Fig. 5:. Negative-stain electron microscopy (nsEM) of scFv:PUUV Gn^H/Gc complexes.

(A) Exemplary nsEM 2D-classes of PUUV Gn^H/Gc bound to scFvs of the indicated mAbs. scFvs, turquoise; Gn^H, green with the capping loop in purple; Gc domain I, red; domain II, yellow; domain III, blue; fusion loops, orange. (B) 3D reconstructions of scFv:Gn^H/Gc complexes are shown in transparent surface representations (light gray) with the structure of ANDV Gn^H/Gc docked into the density and pseudo-colored as described in panel A. Corresponding neutralization-escape mutations are shown in the closeups. (C) Modeled interactions of the indicated scFvs with tetrameric PUUV Gn^H/Gc complexes are shown in surface-shaded representation and colored as in panel A. En face and side views are shown.

Fig. 6:. Neutralization potency of mAbs against authentic Old World hantaviruses (OWH).

Potency of mAbs to block OWH PUUV (A), HTNV (B), and DOBV (C) infection of VeroE6 cells (mAb concentrations of 100, 4, and 0.04 nM). Data points are colored according to the mAb’s competition group assignments defined in Fig. 3: Group I, green; Group II, blue; Group III, purple; Group IV, pink. Averages ± standard deviation (s.d.), n=2-10 from one to five experiments (PUUV); n=4 from two experiments (HTNV); n=4-6 from two to three experiments (DOBV). (D) Heatmap of IC₅₀ values isolated from PUUV, HTNV, and DOBV dose-response neutralization curves (A-C) derived by non-linear regression analysis. Data points are colored according to the mAb’s neutralization potency. mAbs with IC₅₀ values >5 nM are designated non-neutralizing mAbs (N.N.); mAbs whose neutralization curves did not adhere to a sigmoidal dose-response fit and for which an IC₅₀ could not be computed are designated as N.C. mAbs leaving an un-neutralized virus fraction are shown as striped data points on the map. An un-neutralized virus fraction is defined as a residual normalized virus infection at the highest mAb concentration tested of >5%. N.D., not determined.

Fig. 7:. In vivo protective efficacy of human nAbs in a PUUV bank vole challenge model.

(A) Bank voles were administered a single 25 mg/kg dose (i.p.) of the indicated nAbs (n=12 from two experiments) or PBS vehicle (n=9 from two experiments) followed by a challenge with PUUV/Suo (500 FFU, s.c.) at 4 h post nAb administration. Animals were sacrificed at 3 days post challenge and serum and organ viral RNA loads were determined by RT-qPCR detecting the PUUV S segment. Bars indicate median. (B) Bank voles were challenged with PUUV/Suo (500 FFU, s.c.) followed by treatment with a single 25 mg/kg dose (i.p.) of the indicated nAbs (n=12 from two experiments) or PBS vehicle (n=11 from two experiments) at 24 h post challenge. Animals were sacrificed at 3 days post challenge, and viral RNA loads were determined as above. Bars indicate median. Untreated versus nAb-treated animals, un-paired Kruskal-Wallis test: **, P=0.0021; ***, P=0.0002; ****, P<0.0001.

Fig. 8:. Broad pan-hantavirus neutralization and protection by ADI-42898.

(A) Potency of ADI-42898 to block infection of Vero cells by rVSVs bearing Gn/Gc from the indicated NWH. Viruses were exposed to a 3-fold mAb dilution series starting at 100 nM. Averages ± s.d, n=14 from five experiments (ANDV); n=9 from three experiments (SNV); n=12 from four experiments (CHOV). (B) Potency of ADI-42898 to block ANDV (strain Chile-9717869) and SNV (strain CC107) infection of HUVECs with a 3-fold mAb dilution series starting at 300 nM. Averages ± s.d, n=4 from two experiments. (C) Syrian golden hamsters were challenged with ANDV (200 PFU, i.m.; strain Chile-9717869) followed by treatment with a single dose of ADI-42898 (~25 mg/kg or ~6 mg/kg, i.p.) at 3 days post-virus exposure. Mortality of hamsters was monitored for 35 days. Averages from two experiments, n=5 (~25 mg/kg dose); n=8 (~6 mg/kg dose), n=10 (vehicle). Untreated versus mAb-treated animals, Mantel-Cox test: ***, P=0.0003; ****, P<0.0001. (D) Serum virus titers of Syrian golden hamsters on day 8 post ANDV challenge were determined by plaque assay. Untreated versus ADI-42898-treated animals, un-paired Kruskal-Wallis test: *, P=0.03; **, P=0.002. L.O.D., limit of detection.