Engineered Bispecific Antibodies with Exquisite HIV-1-Neutralizing Activity

Abstract

While the search for an efficacious HIV-1 vaccine remains elusive, emergence of a new generation of virus-neutralizing monoclonal antibodies (mAbs) has re-ignited the field of passive immunization for HIV-1 prevention. However, the plasticity of HIV-1 demands additional improvements to these mAbs to better ensure their clinical utility. Here, we report engineered bispecific antibodies that are the most potent and broad HIV-neutralizing antibodies to date. One bispecific antibody, 10E8V2.0/iMab, neutralized 118 HIV-1 pseudotyped viruses tested with a mean 50% inhibitory concentration (IC50) of 0.002 μg/mL. 10E8V2.0/iMab also potently neutralized 99% of viruses in a second panel of 200 HIV-1 isolates belonging to clade C, the dominant subtype accounting for ∼50% of new infections worldwide. Importantly, 10E8V2.0/iMab reduced virus load substantially in HIV-1-infected humanized mice and also provided complete protection when administered prior to virus challenge. These bispecific antibodies hold promise as novel prophylactic and/or therapeutic agents in the fight against HIV-1.

Figure 1. HIV CrossMAbs possess potent and broad antiviral activity against a diverse panel of 118 Tier-2 HIV-1 Env pseudoviruses

(A) Schematic of an HIV CrossMAb and list of examples of parental antibodies from which each CrossMAb was derived. (B) IC₅₀ (top panel) and maximum percent inhibition (MPI, bottom panel) comparison of select HIV CrossMAbs and their parental mAbs. Asterisks refer to data obtained from other sources (3BNC117 MPI data from personal communication with Michel Nussenzweig and PGT151 IC₅₀ and MPI data from Blattner et al., 2014). Error bars indicate median ± interquartile range. (C) Percent of viruses neutralized (based on IC₅₀ values) by 10E8/P140 and 10E8/iMab, and their parental mAbs. Neutralization by penta-mix is included as a reference (Klein et al., 2012).

Figure 2. Neutralization studies to elucidate the mechanism of action of a potent HIV CrossMAb

(A) Neutralization of the representative HIV-1 pseudovirus 246F C1G by 10E8/P140, parental mAbs individually, or parental mAbs in combination. (B) Neutralization of the representative pseudovirus TF7 by bispecific antibodies comprised of a 10E8 antibody moiety and one of several host cell receptor targeting antibody moieties. (C) Neutralization of the representative HIV-1 pseudovirus TF4 by MPER-binding mAbs or CCR5-anchored MPER-binding bispecific antibodies.

Figure 3. 10E8 mAb and 10E8-containing CrossMAbs exhibit physicochemical heterogeneity

(A) SEC analysis of 10E8/P140, 3BNC117/iMab, 4E10/P140 and 10E8/iMab. (B) SEC analysis of parental mAbs iMab, 10E8 and P140. (C) SEC analysis of 10E8_V1.0/iMab and 10E8_V1.0/P140 and mAb variant 10E8_V1.0. (D) Percent of viruses of a 118 Tier-2 HIV-1 Env pseudovirus panel neutralized (based on IC₅₀ values) by 10E8/iMab and 10E8/P140 and engineered variants 10E8_V1.0/iMab and 10E8_V1.0/P140.

Figure 4. Engineering HIV CrossMAb variants with improved developability, activity and manufacturability potential

(A) SEC analysis and (B) percent of a 118 Tier-2 HIV-1 Env pseudovirus panel neutralized by the originally identified 10E8/iMab and 10E8/P140 and engineered variants 10E8_V2.0/iMab and 10E8_V1.1/P140. (C) Percent of a panel of 200 clade C Env pseudoviruses neutralized by 10E8_V2.0/iMab and 10E8_V1.1/P140. The neutralization profiles of these two candidates against the 118 virus panel from Figure 1D are overlaid for ease of comparison. (D) Serum concentration of the indicated HIV CrossMAb after 100 µg intraperitoneal administration to mice. Error bars represent standard error of mean.

Figure 5. HIV CrossMAb 10E8_V2.0/iMab exhibits therapeutic efficacy in vivo

(A) Changes in plasma viral RNA from baseline at week 0. Grey lines represent data from each mouse, whereas the red lines represent the mean for each group. NSG humanized mice were infected with JR-CSF at week −4 and the antibody treatment consisting of weekly 500 µg antibody injections (arrows) began at week 0. (B) Comparison of the therapeutic efficacy of 10E8_V2.0/iMab with the comparator groups. Columns represent changes in viral load. * p < 0.05, ** p < 0.01, *** p < 0.001 as determined by the Mann-Whitney test. (C) Mutations in HIV-1 Env associated with resistance after viral rebound. Colored amino acids and dashes indicate the positions of mutations and deletions, respectively. All mutations are aligned to the JR-CSF sequence and numbered according to the HXB2 sequence. (D) IC₅₀ concentrations (µg/mL) of the antibodies listed in the top row against wild-type HIV-1_JR-CSF or mutants containing the indicated mutations in the HIV-1_JR-CSF envelope shown in the left column. Red indicates IC₅₀ < 0.2 µg/mL, orange indicates IC₅₀ between 0.2 and 2 µg/mL, yellow indicates IC₅₀ between 2 and 20 µg/mL and white indicates IC₅₀ > 20 µg/mL.

Figure 6. HIV CrossMAb 10E8_V2.0/iMab protects humanized mice against repeated systemic HIV-1 challenges

(A) Kaplan-Meier plot depicting the percentage of aviremic mice in 10E8_V2.0/iMab-treated versus PBS-treated challenged mice. Mice received 200 µg of 10E8_V2.0/iMab every week from week 0 to week 8 (black arrows) and were challenged at day 1 and weeks 4 and 6 (blue arrows). PBS control mice were challenged once at day 1 (red arrow). Statistics were calculated by log-rank test. (B) Quantification of 10E8_V2.0/iMab plasma concentration by ELISA. The dashed line represents the limit of detection at 0.11 µg/mL. Values below the limit of detection are arbitrarily plotted at 0.09 µg/mL. Error bars represent the standard deviation.

Figure 7. Percent of large panels of multi-clade HIV-1 Env pseudoviruses neutralized by antibodies currently in development for HIV-1 prevention

Antiviral coverage of 10E8_V2.0/iMab and 10E8_V1.1/P140 are reported in this manuscript; antiviral coverage of all other molecules presented are from previously published data (Gardner et al., 2015; Kong et al., 2015; Mouquet et al., 2012a; Rudicell et al., 2014; Scheid et al., 2011; Sok et al., 2014; Wu et al., 2010).