Synergistic Neutralization of Pertussis Toxin by a Bispecific Antibody In Vitro and In Vivo

Abstract

Bispecific antibodies are a rapidly growing class of therapeutic molecules, originally developed for the treatment of cancer but recently explored for the treatment of autoimmune and infectious diseases. Bordetella pertussis is a reemerging pathogen, and several of the key symptoms of infection are caused by the pertussis toxin (PTx). Two humanized antibodies, hu1B7 and hu11E6, bind distinct epitopes on PTx and, when coadministered, mitigate disease severity in murine and baboon models of infection. Here we describe the generation of a bispecific human IgG1 molecule combining the hu1B7 and hu11E6 binding sites via a knobs-in-holes design. The bispecific antibody showed binding activity equivalent to that of the antibody mixture in a competition enzyme-linked immunosorbent assay (ELISA). A CHO cell neutralization assay provided preliminary evidence for synergy between the two antibodies, while a murine model of PTx-induced leukocytosis definitively showed synergistic neutralization. Notably, the bispecific antibody retained the synergy observed for the antibody mixture, supporting the conclusion that synergy is due to simultaneous blockade of both the catalytic and receptor binding activities of pertussis toxin. These data suggest that a hu1B7/hu11E6 bispecific antibody is a viable alternative to an antibody mixture for pertussis treatment. Despite vaccination, pertussis infection continues to cause ϳ195,000 deaths worldwide, primarily of infants (1). Of the estimated 16 million cases of pertussis each year, ϳ95% occur in the developing world. Even in developed countries, the disease incidence has increased dramatically over the last decade, reaching prevaccination levels in some countries (2, 3). This rise has been attributed to shortcomings of the current acellular vaccine (4) as well as pathogen adaptation (5). In both cases, high levels of circulating disease place young infants at risk, as this population is the most susceptible to severe disease. An antibody therapeutic could be used to treat seriously ill infants in the developing world and to prevent disease in high-risk areas.Pertussis toxin (PTx) is one of several virulence factors secreted by the Gram-negative bacterium Bordetella pertussis. PTx is directly responsible for suppression of the innate immune system (6) and for systemic leukocytosis, which is the key clinical indicator of severe disease and appears to be directly responsible for pulmonary hypertension and organ failure (7). In addition, low titers of PTx-neutralizing antibodies correlate with susceptibility to clinical infection (8). We previously developed a binary mixture of two humanized anti-pertussis toxin antibodies which was able to mitigate whooping cough in mouse and baboon models of infection (9). The antibody hu11E6 blocks binding of the toxin to host cells, while the antibody hu1B7 interferes with the catalytic pathway. At a dose selected to demonstrate efficacy but not synergy, the individual antibodies and the mixture were able to completely suppress leukocytosis in a murine ...

Humanised monoclonal antibodies neutralise pertussis toxin by receptor blockade and reduced retrograde trafficking

“…Prior experiments indicated the PTx-induced CHO cell clustering morphology is completely neutralised by a 1,000-fold molar excess of hu11E6 or >5,000-fold molar excess of hu1B7 in the presence of PTx Wagner et al, 2016). Prior experiments indicated the PTx-induced CHO cell clustering morphology is completely neutralised by a 1,000-fold molar excess of hu11E6 or >5,000-fold molar excess of hu1B7 in the presence of PTx Wagner et al, 2016).…”

“…Indeed, both appear to act by altering PTx interactions with target cells: hu11E6 by blocking the initial toxin-receptor interaction and hu1B7 by altering toxin internalisation and retrograde trafficking in more subtle ways(Figure 3a).Although effector functions do not appear to be involved in hu1B7 or hu11E6 neutralisation of PTx toxicity, the interaction of the Fc domain with Fc receptors may increase the protection conferred by antitoxin antibodies through clearance of immune complexes(Ganesan et al, 2012;Montero-Julian et al, 1995). CHO cells are highly susceptible to PTx intoxication and exhibit altered morphology in the presence of very low PTx concentrations, as low as 4 pM(Wagner et al, 2016). We anticipate that larger complexes may also be formed through 2:2 binding stoichiometry, because each PTx molecule can bind two antibodies (Figure 2a)and it is reasonable that a single antibody can bind two PTx molecules, based on binding of a bispecific antibody composed of the hu1B7 and hu11E6 binding sites to two PTx molecules(Wagner et al, 2016).…”

“…which they neutralise PTx will aid in their development and contribute to our understanding of humoral protection conferred by PTx-containing vaccines. We anticipate that larger complexes may also be formed through 2:2 binding stoichiometry, because each PTx molecule can bind two antibodies (Figure 2a)and it is reasonable that a single antibody can bind two PTx molecules, based on binding of a bispecific antibody composed of the hu1B7 and hu11E6 binding sites to two PTx molecules(Wagner et al, 2016). Indeed, both appear to act by altering PTx interactions with target cells: hu11E6 by blocking the initial toxin-receptor interaction and hu1B7 by altering toxin internalisation and retrograde trafficking in more subtle ways(Figure 3a).Although effector functions do not appear to be involved in hu1B7 or hu11E6 neutralisation of PTx toxicity, the interaction of the Fc domain with Fc receptors may increase the protection conferred by antitoxin antibodies through clearance of immune complexes(Ganesan et al, 2012;Montero-Julian et al, 1995).…”

“…The neutralising properties of both antibodies rely solely on their abilities to bind secreted PTx, as they do not directly bind B. pertussis bacteria or confer complement-mediated bactericidal activity(Figure 1). Although the antibody concentrations in an animal experiment are easy to determine (e.g., 20-μg antibody delivered to the~2-ml mouse blood volume results in an initial serum concentration of~10 μg ml −1 ;Wagner et al, 2016), the relevant serum PTx concentration during infection is less apparent, which makes the impact of immune complex formation in vivo harder to assess. Each antibody, either by itself or in combination, formed at least small immune complexes with two antibodies and one PTx molecule required to give an ELISA signal(Figure 2a).…”

“…Humanised versions of these antibodies, hu1B7 and hu11E6, retained low nanomolar affinities for PTx, prevented leukocytosis when delivered individually in a mouse infection model (Nguyen et al, 2015), and exhibited synergy when coadministered to mice in a toxin challenge model (Wagner, Wang, Bui, & Maynard, 2016). The combination of both humanised antibodies dramatically halted leukocytosis when administered to adolescent baboons 3 days after experimental infection with B. pertussis (Nguyen et al, 2015).…”