Deep Mining of Complex Antibody Phage Pools Generated by Cell Panning Enables Discovery of Rare Antibodies Binding New Targets and Epitopes

Abstract

Phage display technology is a common approach for discovery of therapeutic antibodies. Drug candidates are typically isolated in two steps: First, a pool of antibodies is enriched through consecutive rounds of selection on a target antigen, and then individual clones are characterized in a screening procedure. When whole cells are used as targets, as in phenotypic discovery, the output phage pool typically contains thousands of antibodies, binding, in theory, hundreds of different cell surface receptors. Clonal expansion throughout the phage display enrichment process is affected by multiple factors resulting in extremely complex output phage pools where a few antibodies are highly abundant and the majority is very rare. This is a huge challenge in the screening where only a fraction of the antibodies can be tested using a conventional binding analysis, identifying mainly the most abundant clones typically binding only one or a few targets. As the expected number of antibodies and specificities in the pool is much higher, complementing methods, to reach deeper into the pool, are required, called deep mining methods. In this study, four deep mining methods were evaluated: 1) isolation of rare sub-pools of specific antibodies through selection on recombinant proteins predicted to be expressed on the target cells, 2) isolation of a sub-pool enriched for antibodies of unknown specificities through depletion of the primary phage pool on recombinant proteins corresponding to receptors known to generate many binders, 3) isolation of a sub-pool enriched for antibodies through selection on cells blocked with antibodies dominating the primary phage pool, and 4) next-generation sequencing-based analysis of isolated antibody pools followed by antibody gene synthesis and production of rare but enriched clones. We demonstrate that antibodies binding new targets and epitopes, not discovered through screening alone, can be discovered using described deep mining methods. Overall, we demonstrate the complexity of phage pools generated through selection on cells and show that a combination of conventional screening and deep mining methods are needed to fully utilize such pools. Deep mining will be important in future phenotypic antibody drug discovery efforts to increase the diversity of identified antibodies and targets.

Rapid Evaluation of Antibody Fragment Endocytosis for Antibody Fragment–Drug Conjugates

“…Various approaches have been employed in an attempt to improve the selection of internalizing antibodies. Biopanning on tumor cells for the selection of phenotypic antibodies and internalizing antibodies have been reported [21,41]. However, with the complexity of the phage antibody pool from the cell panning due to the diversity of the cell surface proteins exemplified by Ljungars et al [41], the subsequent screening of antibodies is often hampered due to difficulties in obtaining the membrane protein in sufficient quantity and quality, as well as the need to generate antibodies from the phage antibodies.…”

A bispecific IgG format containing four independent antigen binding sites

“…CD40-, OX40-, and 4-1BB-specific scFv displayed on phages were obtained from the n-CoDeR scFv library 18 , using a combination of recombinant proteins (in-house produced) and target-expressing cells (in-house constructed) for selection. After 3 consecutive selection rounds, genes encoding the scFv were used for conversion to soluble scFv as described previously 29 .…”

Public Baseline and shared response structures support the theory of antibody repertoire functional commonality

“…Clonotyping can also be used to detect antigen-specific immunoglobulins, through the identification of expanded clones after vaccination, or those present in unusually high proportions in individuals immune to certain diseases. Explorations of expanded lineages have yielded high-affinity antibodies and T cells against numerous pharmacologically interesting antigens, such as HIV proteins [6], cluster of differentiation proteins [18], botulinum neurotoxin serotype A [19], proteins implicated in type-1 diabetes [20], and many more.…”