A Bispecific Antibody That Simultaneously Recognizes the V2- and V3-Glycan Epitopes of the HIV-1 Envelope Glycoprotein Is Broader and More Potent than Its Parental Antibodies

Abstract

Broadly neutralizing antibodies (bNAbs) can prevent and control an HIV-1 infection, but their breadth is invariably too limited for use as monotherapy. To address this problem, bi- and trispecific antibody-like constructs have been developed. These engineered antibodies typically have greater breadth than the native bNAbs from which they were derived, but they are not more potent because they do not, in most cases, simultaneously engage more than a single epitope of the HIV-1 envelope glycoprotein (Env). Here, we describe a new class of bispecific antibodies targeting the V2-glycan (apex) and V3-glycan regions of the HIV-1 envelope glycoprotein (Env). Specifically, bispecific antibodies with a single-chain (scFv) form of the CAP256.VRC26.25 V2-glycan (apex) antibody on one antibody arm and a full V3-glycan Fab on the other arm neutralizes more HIV-1 isolates than the bNAbs from which they were derived. Moreover, these bispecific antibodies are markedly more potent than their parental bNAbs, likely because they simultaneously engage both the apex and V3-glycan epitopes of Env. Our data show that simultaneous engagement of two critical epitopes of a single Env trimer can markedly increase the potency of a bispecific antibody.IMPORTANCE Broadly neutralizing antibodies (bNAbs) can prevent a new HIV-1 infection and can at least temporarily suppress an established infection. However, antibody-resistant viruses rapidly emerge in infected persons treated with any single bNAb. Several bispecific antibodies have been developed to increase the breadth of these antibodies, but typically only one arm of these bispecific constructs binds the HIV-1 envelope glycoprotein trimer (Env). Here, we develop and characterize bispecific constructs based on well-characterized V2-glycan and V3-glycan bNAbs and show that at least one member of this class is more potent than its parental antibodies, indicating that they can simultaneously bind both of these epitopes of a single Env trimer. These data show that bispecific antibody-like proteins can achieve greater neutralization potency than the bNAbs from which they were derived.

Keywords: antibody neutralization; bispecific antibodies; broadly neutralizing antibodies; human immunodeficiency virus; human immunodeficiency virus 1.

FIG 1

A model of V2-glycan and V3-glycan bNAbs bound to Env. The BG505 Env trimer bound to PGT145 (PDB accession no. 5V8L) was aligned with the BG505 SOSIP trimer complex with 10-1074 (PDB accession no. 5T3Z) Fab. BG505 Env is shown in white (gp120) and gray (gp41), PGT145 is shown in light (light chain) and dark (heavy chain) blue. 10-1074 is shown in light (light chain) and dark (heavy chain) red. The green dotted line connects the C terminus of the PGT145 light chain to the C terminus of the 10-1074 heavy chain, a 79-Å distance that is bridged by the linker and hinge regions of the Fc domain in the BISC variants described below.

FIG 2

Bispecific constructs combining CAP256.VRC26.25 and 10-1074 are more potent than their parental antibodies. (A) Schematic representation of constructs used in this study. From left to right, wild-type antibody, a homodimeric scFv-Fc construct (e.g., CAP256.VRC26.25 scFv-Fc), a bispecific construct containing one scFv arm and one full antibody arm (i.e., CTM-1A-L10), a bispecific construct containing one scFv arm and one antibody arm with knob-in-hole (KIH) mutations (e.g., BISC-1A), a CrossMab construct of two full antibody arms with KIH mutations (i.e., BICR-1A). (B) TZM-bl neutralization curves of serial dilutions of wild-type CAP256.VRC26.25 (dark blue) or scFv-Fc constructs with various linker lengths against pseudotyped clade C CE1176 (left) and clade A BG505 (right) viruses. Infection is represented as the percentage of luciferase activity in the absence of inhibitor. Concentrations are shown as micrograms per milliliter. (C) TZM-bl neutralization curves as in panel B of cotransfected combinations (CTC) combining 10-1074 and CAP256.VRC26.25 scFv-Fc with various linker lengths. (D) TZM-bl neutralization curves as in panel B of bispecific constructs containing KIH mutations to stabilize heterodimerization. Values represent means ± standard errors of the means (SEM) (error bars) (n = 3). Data are representative of at least two independent experiments.

FIG 3

Bispecific constructs combining CAP256.VRC26.25 with V3 glycan antibodies are generally more potent than their parental antibodies. (A) TZM-bl neutralization curves as in Fig. 1. Curves are shown for viruses pseudotyped with CE1176 and BG505 for combinations of CAP256.VRC26.25 with 10-1074, PGT121, and PGT128. Values represent means ± SEM (n = 3). Data are representative of at least two independent experiments. (B to D) IC₅₀ (B), IC₈₀ (C), and IC₉₅ (D) values are plotted for each isolate tested. Median neutralization efficiency values are indicated by black bars. Values represent means from two independent experiments of triplicates. (E to G) Comparison of bispecific construct potency to parental antibody potency. Viruses were ranked by sensitivity to the bispecofic constructs and IC₈₀ values for parental antibodies and each BISC-1A (E), BISC-1B (F), and BISC-1C (G) were plotted against this ranking. Data points above the bispecific line indicate lower potency of the parental components compared to the bispecific construct.

FIG 4

Characterization of bispecific constructs combining PGT145 with V3 glycan antibodies. (A) TZM-bl neutralization curves as in Fig. 1. Curves are shown for viruses pseudotyped with CE1176 and BG505 using combinations of PGT145 with 10-1074, PGT121, or PGT128. Values represent means ± SEM (n = 3). Data are representative of at least two independent experiments. (B to D) IC₅₀ (B), IC₈₀ (C), and IC₉₅ (D) values are plotted for each isolate tested. Median neutralization efficiency values are indicated by black bars. Values represent means from two independent experiments of triplicates. (E to G) Comparison of bispecific construct potency to parental antibody potency. Viruses were ranked by potency to the parental antibodies and IC₈₀ values for each parental component and BISC-2A (E), BISC-2B (F), and BISC-2C (G) were plotted against this ranking. Data points above the bispecific line indicate lower potency of the parental components compared to the bispecific construct.

FIG 5

A bispecific construct BISC-1A binds BG505 SOSIP trimers and cell-expressed Env more efficiently than its components. (A and B) Env binding ELISAs. Plates were coated with BG505 gp120 (A) or BG505 SOSIP trimer (B) and then incubated with serial dilutions of the indicated antibodies or BISC-1A. Binding was detected with an HRP-conjugated secondary antibody. The absorbance at 450 nm is shown on the y axes in panels A and B. Data are representative of two independent experiments, and error bars represent range. (C) HEK293T cells were transfected to express BG505 Env lacking its cytoplasmic tail (BG505 ΔCT). Cells were harvested and stained with CAP256.VRC26.25 scFv, 10-1074, or BISC-1A. Binding was determined by flow cytometry with a FITC-conjugated secondary antibody. MFI, mean fluorescence intensity. (D and E) BG505 SOSIP trimer-coated ELISA plates were preincubated with 1 μg/ml of the constructs indicated in the legend. Binding of serial dilutions CAP256.VRC26.25-scFv-mFc (D) or 10-1074-mFc (E) was then measured. Baseline BG505 SOSIP binding, in the absence of a competing human antibody, is shown in gray (no antibody [No Ab]). Data are representative of two independent experiments, and error bars represent range. (F and G) HEK293T cells were transfected to express BG505 ΔCT. Cells were harvested and preincubated with serial dilutions of human Fc antibodies or BISC-1A before being stained with CAP256.VRC26.25-scFv-mFc (F) or 10-1074-mFc (G). Binding was determined by flow cytometry with an APC-conjugated secondary antibody. Data are representative of two independent experiments, and error bars represent range.

FIG 6

Theoretical breadth of BISC constructs from individual neutralization data available on CATNAP. (A) Using the CATNAP database, we analyzed the number of isolates neutralized by each component of BISC-1A (CAP256.VRC26.25 and 10-1074), BISC-1B (CAP256.VRC26.25 and PGT121), and BISC-1C (CAP256.VRC26.25 and PGT128) using an IC₈₀ of <20 μg/ml as the cutoff. These results are represented as pie charts showing the percentage of isolates neutralized by each antibody individually, as well as those neutralized by both antibodies. Below each chart is the total number of isolates included in the analysis and the theoretical coverage percentage of viruses neutralized by at least one of the component antibodies. (B) Potency-breadth analysis of experimentally determined IC₈₀ neutralization values provided in Table S2 in the supplemental material for the indicated constructs as analyzed with CombiNAber online tools.