Targeting RSV-neutralizing B cell receptors with anti-idiotypic antibodies

Abstract

Respiratory syncytial virus (RSV) causes lower respiratory tract infections with significant morbidity and mortality at the extremes of age. Vaccines based on the viral fusion protein are approved for adults over 60, but infant protection relies on passive immunity via antibody transfer or maternal vaccination. An infant vaccine that rapidly elicits protective antibodies would fulfill a critical unmet need. Antibodies arising from the VH3-21/VL1-40 gene pairing can neutralize RSV without the need for affinity maturation, making them attractive to target through vaccination. Here, we develop an anti-idiotypic monoclonal antibody (ai-mAb) immunogen that is specific for unmutated VH3-21/VL1-40 B cell receptors (BCRs). The ai-mAb efficiently engages B cells with bona fide target BCRs and does not activate off-target non-neutralizing B cells, unlike recombinant pre-fusion (preF) protein used in current RSV vaccines. These results establish proof of concept for using an ai-mAb-derived vaccine to target B cells hardwired to produce RSV-neutralizing antibodies.

Keywords: B cell sorting; CP: Immunology; anti-idiotype; bispecific antibody; germline targeting; neutralizing antibodies; respiratory syncytial virus; vaccines.

Graphical abstract

Figure 1

Characterization of anti-idiotypic antibodies specific for antibodies encoded by VH3-21 and VL1-40 genes (A) IgGs purified from hybridoma supernatants 2C1, 2C2, 2F1, and 1D3 were evaluated for their ability to bind to a panel of recombinant mAbs by ELISA. The scale indicates the A₄₅₀ signal as measured by ELISA. Data shown are average of duplicate wells from one biological replicate. (B) ai-mAbs were fluorescently labeled with allophycocyanin (APC) conjugated to a unique oligonucleotide barcode and used to stain naive B cells from human PBMCs. The frequency of APC⁺ B cells is shown. The gating strategy is shown in Figure S2. (C–E) APC⁺ naive B cells were sorted, and the paired BCR transcripts were obtained through next-generation sequencing. Shown are (C) the percentage of VH3-21/VL1-40 pairs among all BCR sequences sorted by each ai-mAb bait (blue), (D) the percentage of BCRs expressing an HC derived from the VH3-21 gene sorted by each ai-mAb (green), and (E) the percentage of BCRs expressing an LC derived from the VL1-40 gene sorted by each ai-mAb bait (teal). The number of B cells analyzed is shown at the bottom of (C)–(E). The frequency of naive B cells expressing these genetic features identified by high-throughput unbiased sorting is included for comparison in (C)–(E). Data are representative of one biological replicate for each ai-mAb in (B)–(E).

Figure 2

Crystal structures of ai-mAbs in complex with ADI-19425 (A and C) Surface representation of ADI-19425 Fab bound to (A) 1D3 Fab and (C) 2C1 Fab. Views are looking down on the ADI-19425 paratope. Residues within 5 Å of (A) 1D3 or (C) 2C1 are outlined in white. (B and D) Buried surface area (BSA) plots shown as stacked bar graphs of ADI-19425 HC (top) and ADI-19425 LC (bottom) bound by (B) 1D3 Fab or (D) 2C1 Fab. Interactions with the HC of either 1D3 or 2C1 are shown as solid-colored bars, while interactions with the LC of either 1D3 or 2C1 are shown as empty bars. Residues participating in hydrogen-bond interactions are labeled with an “H.” ADI-19425 HC and LC sequences are shown underneath the graph. VH3-21 and VL1-40 sequences are boxed. See also Figure S3 and Table S1.

Figure 3

Engineering a VH3-21/VL1-40 BCR-targeting bispecific ai-mAb (A) Schematic of the bispecific platform used, created with BioRender. (B) Anion exchange (AEX) chromatogram of bispecific ai-mAb production. Representative data from one biological replicate are shown. (C–F) Binding of ADI-19425 Fab and the indicated chimeric Fabs to recombinant 2C1-IgG (C), 1D3-IgG (D), IgG from peak 1 (E), or IgG from peak 2 (F) was measured by BLI. Representative data are from one biological replicate.

Figure 4

Bispecific ai-mAb-specific B cell sorting and characterization (A) A fluorescently-labeled 1D3/2C1 bispecific ai-mAb was used to stain and sort naive human B cells; a representative sorting plot is shown, with FMO representing a fluorescence minus one (no ai-mAb) staining control. See Figure S4 for the gating strategy. (B) From the sorted cells, the percentages of productive VH3-21 HCs (green) and VL1-40 LCs (teal) are shown. From successfully recovered pairs, the percentages of VH3-21/VL1-40 (blue) are shown. Each data point represents the frequency of gene usage from 3 independent sorting experiments. Bars represent the mean ± standard deviation. Naive BCRs obtained from unbiased high-throughput sequencing of 3 unique donors are shown for comparison (gray). Significant differences were determined using Student’s two-tailed t tests. ^∗∗p = 0.001–0.01, ^∗∗∗p = 0.0001–0.001, ^∗∗∗∗p < 0.0001. (C) VH3-21/VL1-40 BCRs from (B) were expressed as recombinant mAbs and tested for binding to RSV preF and postF by BLI. Data shown are average from 2 biological replicates. (D) mAbs from (C) were tested for their ability to compete for binding with ADI-14337 to preF by BLI. Each data point represents a biological replicate, and the bars represent mean. The red dotted line represents the average ADI-14337 competition against itself. D25 is an RSV preF site Ø mAb and is included as a negative control. (E) Side view of a representative nsEM 3D class of the DS-Cav1-MLR_24 Fab complex. (F) Superposition of the crystal structure of ADI-19425 Fab bound to preF (PDB: 6APD) into the representative 3D volume. The crystal structure is colored as follows: RSV preF (midnight blue), ADI-19425 HC (dark red), and ADI-19425 LC (salmon). Density was also seen for the T4 fibritin trimerization domain, which is modeled by the foldon region of the GlyProPro(10)-foldon crystal structure (PDB: 1NAY) in green. See also Figures S4–S6 and Table S2.

Figure 5

VH3-21/VL1-40 mAb RSV neutralization and affinity (A) Neutralization curves of selected VH3-21/VL1-40 mAbs against RSV-A (left) and RSV-B (right). Points shown are the mean of 2 technical replicates from a single biological replicate. (B) Mean IC₅₀ values of selected mAbs against RSV-A and RSV-B. Each dot represents a biological replicate (n = 3–4), and bars represent mean ± standard deviation. The dotted line represents maximum tested concentration. (C) Steady-state analysis of VH3-21/VL1-40 mAb binding response to RSV-A preF measured by BLI. Each data point represents the mean response from two biological replicates. (D) Steady-state kinetic analysis of neutralizing VH3-21/VL1-40 mAbs binding to RSV-A (solid lines) and RSV-B (dashed lines) preF measured by BLI. Each data point represents the mean response from two biological replicates. (E) Apparent affinity of neutralizing VH3-21/VL1-40 mAbs to RSV-A/RSV-B preF. See also Figure S7.

Figure 6

Activation of on- and off-target BCR-expressing cell lines by calcium flux assays (A–C) Staining of DG75 B cell lines transduced to express the indicated BCRs, shown as histograms, with signals normalized to mode. Shown is a representative plot of a single biological replicate. (A) Staining with the 1D3 ai-mAb. (B) Staining with the 2C1 ai-mAb. (C) Staining with the bispecific ai-mAb. (D–F) Calcium flux in response to the addition of the 1D3 ai-mAb D, 2C1 ai-mAb E, or the 1D3/2C1 bispecific ai-mAb F as immunogen. Signals shown are representative of 2 biological replicates. (G–I) Calcium flux in non-neutralizing RSV preF-specific BCR-expressing cell lines as indicated. Cell lines were stimulated with RSV-A preF (G), the bispecific ai-mAb (H), and, as a positive control, an α-IgG Fcγ F(ab′)₂ (I). Signals shown are representative of 2 biological replicates.

Figure 7

RSV neutralization competition assays MLR_24 was formulated at its IC₅₀ (green, open) or IC₈₀ (green, shaded) concentration alone or in combination with 5 non-neutralizing, pre-F-binding mAbs at matched concentrations “mAb Mix” (blue open, blue shaded) (Figures S7A and S7B) and tested for inhibition of RSV-A (A) and RSV-B (B) in a plaque reduction assay. Each dot represents a single biological replicate carried out in triplicate, and bars represent mean ± standard deviation. Significance was determined by Mann-Whitney tests comparing MLR_24 to mAb Mix at each concentration.