Bispecific antibodies: a mechanistic review of the pipeline.

The term bispecific antibody (bsAb) is used to describe a large family of molecules designed to recognize two different epitopes or antigens. BsAbs come in many formats, ranging from relatively small proteins, merely consisting of two linked antigen-binding fragments, to large immunoglobulin G (IgG)-like molecules with additional domains attached. An attractive bsAb feature is their potential for novel functionalities - that is, activities that do not exist in mixtures of the parental or reference antibodies. In these so-called obligate bsAbs, the physical linkage of the two binding specificities creates a dependency that can be temporal, with binding events occurring sequentially, or spatial, with binding events occurring simultaneously, such as in linking an effector to a target cell. To date, more than 20 different commercialized technology platforms are available for bsAb creation and development, 2 bsAbs are marketed and over 85 are in clinical development. Here, we review the current bsAb landscape from a mechanistic perspective, including a comprehensive overview of the pipeline.

Improved breadth and potency of an HIV-1-neutralizing human single-chain antibody by random mutagenesis and sequential antigen panning.

Several human monoclonal antibodies can neutralize a range of human immunodeficiency virus type 1 (HIV-1) primary isolates but their potency and related ability to suppress generation of HIV-1 escape mutants is significantly lower than the activity of antiretroviral drugs currently in clinical use. Recently, a human Fab, X5, was identified and found to neutralize primary isolates from different clades. Further improvement of the potency and breadth of HIV-1 neutralization by this antibody could be critical for its potential use in the treatment of HIV-1-infected patients. However, increasing potency of an antibody by selection from libraries may lead to a decrease in the breadth of neutralization. In an attempt to solve this problem, we subjected a random mutagenesis library of the scFv X5 to sequential rounds of selection on non-homologous HIV-1 envelope glycoproteins (Envs) dubbed sequential antigen panning (SAP). By using SAP, we identified two scFv antibodies, m6 and m9, that were tested with a panel of 33 diverse primary HIV-1 infectious isolates in an assay based on a reporter cell-line expressing high levels of CD4, CCR5 and CXCR4. The IC(50) was less than 50 microg/ml for 21 (m6) and 19 (m9) out of 29 isolates from group M (subtypes A-C, F, G and CRF-01AE) and one isolate from group N; three isolates from group O were not significantly inhibited at 50 microg/ml. The average IC(50) values for the two antibodies were significantly (p<0.001, n=29) lower compared to scFv X5. Their inhibitory activity does not appear to be related to the HIV-1 subtype, coreceptor usage or the disease stage. m9 inhibited infection of peripheral blood mononuclear cells by the primary isolates JRCSF, 89.6 and BR020 with IC(90) of 4, 6 and 25 microg/ml, respectively; for a single-round infection by pseudovirus, the IC(90) for JRSCF, 89.6, YU2 and HXBc2 was 15, 5, 15 and 5 microg/ml, respectively. In these two assays the IC(90) for m9 was, on average, two- to threefold lower than for scFv X5. These results demonstrate that both the potency and the breadth of HIV-1 neutralization of one of the few known potent broadly cross-reactive human monoclonal antibodies, scFv X5, could be improved significantly. However, only experiments in animal models and clinical trials in humans will show whether these new scFvs and the approach for their identification have potential in the development of prophylactics and therapeutics for HIV-1 infections.