A natively paired antibody library yields drug leads with higher sensitivity and specificity than a randomly paired antibody library

Abstract

Deep sequencing and single-chain variable fragment (scFv) yeast display methods are becoming more popular for discovery of therapeutic antibody candidates in mouse B cell repertoires. In this study, we compare a deep sequencing and scFv display method that retains native heavy and light chain pairing with a related method that randomly pairs heavy and light chain. We performed the studies in a humanized mouse, using interleukin 21 receptor (IL-21R) as a test immunogen. We identified 44 high-affinity binder scFv with the native pairing method and 100 high-affinity binder scFv with the random pairing method. 30% of the natively paired scFv binders were also discovered with the randomly paired method, and 13% of the randomly paired binders were also discovered with the natively paired method. Additionally, 33% of the scFv binders discovered only in the randomly paired library were initially present in the natively paired pre-sort library. Thus, a significant proportion of “randomly paired” scFv were actually natively paired. We synthesized and produced 46 of the candidates as full-length antibodies and subjected them to a panel of binding assays to characterize their therapeutic potential. 87% of the antibodies were verified as binding IL-21R by at least one assay. We found that antibodies with native light chains were more likely to bind IL-21R than antibodies with non-native light chains, suggesting a higher false positive rate for antibodies from the randomly paired library. Additionally, the randomly paired method failed to identify nearly half of the true natively paired binders, suggesting a higher false negative rate. We conclude that natively paired libraries have critical advantages in sensitivity and specificity for antibody discovery programs.

Antibody repertoire analysis of mouse immunization protocols using microfluidics and molecular genomics

“…Despite the visual differences in the V-J usage patterns, V-J-gene abundance significantly correlated across all libraries generated using different methods, antigens, pre-and post-sort (Pearson correlation coefficients between 0.70 and 0.99, p-values <0.001). We further compared the V-J-gene usage with our previous study that used a different humanized mouse (Medarex HuMAb-Mouse) immunized with interleukin-21 receptor, 19 and we observed no correlation in V-J-gene usage between this current study and the previous study (Supplementary Figure S3). Together, this suggests that V-J-gene usage was minimally affected by FACS-sorting antigens or immunization methods.…”

“…Using our previously described high-throughput microfluidic antibody discovery platform, 12,18,19 B cells isolated from each tissue (>2 million total cells per immunization method; Supplementary Table S1) were encapsulated into droplets such that each droplet contained a single B cell ( Figure 1b). The heavy and light (kappa only) chain sequences from each cell were amplified within droplets to generate an scFv sequence that retains the proper heavy:light Ig pairing of the original B cell.…”

“…Clonal cluster analysis 18,19 reveals clonal lineages of similar antibody sequences, generated by processes such as affinity maturation and selection in vivo. Though our libraries were derived from different mice, and it is therefore impossible to distinguish clonal clusters from similar clones arising in different mice, clonal cluster analysis is still useful for visualizing similar sequences across different immunization methods and tissues.…”

Antibody discovery and engineering by enhanced CRISPR-Cas9 integration of variable gene cassette libraries in mammalian cells

“…Furthermore, antibody genes can be isolated from plasma cells, which may offer a unique compartment of antibody diversity, and, due to their high polarization toward antigen-specific variable genes, it may be possible to recover natural pairings at least for the most frequent clones by combinatorial screening. 42 While the limited number of plasma cells used to generate the library results in a relatively low diversity library, the high enrichment of antigen-specific clones makes it compatible for screening and discovery. In the future, deep sequencing of the input plasmid library, initial cell library as well as sequential enrichment rounds will provide more characterization and quantitation of library diversity.…”

Designing Feature-Controlled Humanoid Antibody Discovery Libraries Using Generative Adversarial Networks

“…A recent synthetic library approach implements random mutagenesis in which specific residues are allowed to vary in type following statistical rules for frequency of appearance by location in the antibody (commonly known as positional frequency analysis, PFA). 16,17,18 PFA and other related methods 19,20,21 do not take into account any interactions between residues except to the extent that such interactions limit the expressibility of the protein. While this widely explores the sequence space, it ignores how residue types interact to form stabilizing features such as hydrogen or ionic bonds.…”