As therapeutics, antibodies stand out for their specificities, affinities, biological activities, expression levels, stability and pharmacokinetics (Cao et al., 2022; Cunningham et al., 2021). These properties, combined with the modular nature of antibodies, providing engineering flexibility, have led to a consistent increase in the number of antibodies and their derivatives entering clinical trials and being approved (Cao et al., 2022). However, notwithstanding potent activities, antibodies can display characteristics, independently of their biological activities, that cause trial failure (Sun and Benet, 2020). This has resulted in a focus on properties predicting favorable therapeutic outcomes, which, falling under the rubric of “antibody developability”, include desirable features such as high solubility, covalent integrity, stability, high expression titers, low immunogenicity and low polyreactivity/polyspecificity.

Polyreactivity and polyspecificity involve the unintended binding of antibodies to either off-target/unrelated biomolecules, or formulation excipients (Cunningham et al., 2021; Mieczkowski et al., 2023). While often conflated, polyreactivity refers to a monoclonal antibody's ability to bind many unrelated self and foreign antigens, usually with lower affinity, and is attributed to a flexible antigen-binding pocket, excessive charge or hydrophobicity (Cunningham et al., 2021; Mieczkowski et al., 2023). Within the context of the immune system, this may be advantageous as polyreactive antibodies play a crucial role in immune defense against pathogens, allowing them to cover a broader antigenic space (Cunningham et al., 2021; Mieczkowski et al., 2023), particularly during early phases of the humoral immune response. Generally, the low affinity of early response antibodies is compensated for by their expression as polyvalent, high avidity IgMs, which mature into high-affinity, high specificity IgG molecules lacking self-reactive specificities following T cell engagement (Cunningham et al., 2021; Mieczkowski et al., 2023). An antibody with polyspecificity, on the other hand, recognizes one or a limited number of targets which may or may not be related, by sequence or structure, to the expected target. This is usually unexpected and can result in surprising toxicities (Cunningham et al., 2021; Finlay et al., 2019; Lickliter et al., 2020; Li et al., 2019; Norden et al., 2024). The anti-PD1 antibody, camrelizumab, for example, is unique among clinical PD1 antibodies in causing capillary hemangiomas, found to be due to the unexpected recognition and potent activation of vascular endothelial growth factor receptor 2 (VEGFR2), and eliminated by CDR optimization (Cunningham et al., 2021; Finlay et al., 2019). Whereas polyreactivity can be identified by testing against a relatively small panel of targets, polyspecificity requires testing against the full proteome, or at least the secreted and membrane bound subset, usually by array-, display- or flow cytometry-based methods, in order to identify unexpected interactions (Cunningham et al., 2021; Norden et al., 2024; Freeth and Soden, 2020; Credle et al., 2022; Tucker et al., 2018).

Some clinical antibodies in development exhibit unwanted low affinity polyreactivity to unrelated molecules (Cunningham et al., 2021). This can lead to antibodies binding healthy tissues or unintended targets, causing unwanted off-target effects, including toxicity, reduced efficacy or severe adverse reactions. Polyreactivity can also reduce effective antibody concentrations by sequestering antibodies in low affinity off-target binding to abundant molecules. Polyreactive antibodies can also trigger immune responses resulting in anti-drug antibodies, further rendering treatment ineffective and potentially causing immune-related adverse events, leading to inflammation and tissue damage. Eliminating polyreactivity is consequently crucial to ensuring the safety and effectiveness of therapeutic antibodies (Sigounas et al., 1994; Avery et al., 2018; Harvey et al., 2022; Fernandez-Quintero et al., 2023; Dyson et al., 2020; Ridker et al., 2017).

The classic method of evaluating polyreactivity of lead clinical antibodies is ELISA (Enzyme-Linked Immunosorbent Assay) (Mouquet et al., 2010). However, confounding factors can occur when detecting non-specific interactions by ELISA, in part due to signal amplification mediated by secondary antibodies. This is a well-known property of indirect ELISAs (Hayrapetyan et al., 2023), which may exacerbate and overestimate low binding signals in polyreactivity assessment. Here we compared results obtained by Surface Plasmon Resonance (SPR), which lacks signal amplification, to ELISA. Since SPR is mass-based, we selected a panel of defined biotinylated probes with known molecular weights, avoiding non-defined probes such as lipopolysaccharide. Our probes allowed us to transition from an ELISA to an SPR assay for the evaluation of antibody polyreactivity. Both assays recapitulated findings from previous studies assessing polyreactivity of clinical antibodies (Jain et al., 2017) with varying degrees of sensitivity. However, SPR may not be sensitive enough to flag antibodies with intermediate polyreactive binding activity that may fail further downstream in the pre-clinical or clinical validation process.