Arvind Sivasubramanian, Patricia Estep, Heather Lynaugh, Yao Yu, Adam Miles, Josh Eckman, Kevin Schutz, Crystal Piffath, Nadthakarn Boland, Rebecca Hurley Niles, Stéphanie Durand, Todd Boland, Maximiliano Vásquez, Yingda Xu, Yasmina Abdiche

November 21, 2016

The successful discovery of therapeutic antibodies depends on identifying binders that exhibit both high affinity and broad epitope diversity. Campaigns that achieve wide “epitope coverage” across a target antigen increase the probability of uncovering antibodies with distinct biological functions, mechanisms of action, or intellectual property potential.

Epitope specificity is an intrinsic property of the antibody–antigen interface that cannot be rationally optimized through engineering; therefore, it must be selected appropriately at the earliest stages of discovery. Yet traditional immunization approaches often reveal epitope diversity only late in discovery, after considerable time and cost. High-throughput epitope binning, which sorts antibodies into families based on their pairwise blocking behavior, offers rapid, empirical means to assess diversity upfront to enable informed selection of functionally differentiated leads.

In a collaborative study, researchers from Adimab, Wasatch Microfluidics, and Rinat-Pfizer characterized the epitope coverage of a diverse panel of antigen-specific antibodies derived from Adimab’s yeast-based human synthetic IgG library. The study directly compared the epitope diversity of an in vitro antibody repertoire to that of antibodies obtained from traditional animal immunizations.

Experimental design and analytical framework

To assess epitope diversity, the study used hen egg white lysozyme (HEL) as a model antigen. HEL was chosen because it has been structurally characterized in depth, with 20 published antibody–HEL co-crystal structures providing a rich reference framework for comparison.

The analysis combined three main components:

1. Diverse antibody library: A panel of 350 anti-HEL antibodies was selected from Adimab’s full-length human IgG yeast display library. The clones spanned a broad affinity range (with apparent K_D values ranging from 82 pM to 69 nM) and represented a wide variety of VH:Vk germline pairings, illustrating the underlying diversity of the synthetic repertoire.
2. High-throughput epitope mapping: Antibodies were evaluated using surface plasmon resonance (SPR) imaging in a 384-ligand array format, which allowed hundreds of pairwise cross-blocking interactions to be tested simultaneously. This epitope binning approach grouped antibodies according to overlapping or distinct binding footprints. Key findings were confirmed on a biolayer interferometry (BLI) platform to ensure consistency across biosensor technologies.
3. Structural benchmark comparison: The experimental data were compared with predictions derived from the structural benchmark of known HEL–antibody complexes. This enabled a direct correlation between observed blocking patterns and established epitope locations, revealing how extensively the library covered the antigen’s accessible surface.

Across the combined dataset, over 144,000 analyte–ligand interactions were analyzed on label-free biosensors. This represents one of the most comprehensive high-throughput epitope binning studies reported to date and provides a detailed view of the epitope diversity achievable from a human in vitro antibody library.

Key findings

Broad and comprehensive epitope coverage

Antibodies from Adimab’s in vitro yeast library overlapped with, and extended beyond, known structural epitopes, demonstrating that nearly the entire solvent-accessible surface of HEL is antigenic. Quantitatively, antibodies derived from the library covered >75% of the HEL accessible surface area, confirming the breadth of antigen recognition achievable with synthetic repertoires. 

Expanded diversity and fine specificity

The binning analysis identified 17 distinct “blocking profiles” compared to the 11 profiles predicted from the literature (the structural benchmark), and all except one of these 11 profiles were represented in the 17 profiles obtained from the yeast library clones, indicating broader and finer-grained epitope coverage.

Two findings exemplified this diversity:

• Profile 0 (“universal sandwichers”): Approximately 10 % of clones did not cross-block any controls, revealing binding to previously uncharacterized epitopes on the back view of HEL.
• Profile 5* clones (e.g., 1E4): Antibodies that distinguished between two highly similar epitopes bound by control antibodies F10.6.6 and HyHEL-5, demonstrating fine epitope discrimination.

Comparison to immune-derived antibodies

When compared directly, Adimab’s in vitro yeast library produced a greater number of unique blocking profiles and comparable overall coverage to the benchmark set, which itself represented decades of immune-derived antibody structures. These data demonstrate that synthetic antibody libraries can deliver epitope breadth equivalent to that of animal immunization, while providing the advantages of speed, reproducibility, and defined sequence diversity.

Implications for therapeutic antibody discovery

1. Informed lead selection: Comprehensive epitope mapping enables early selection of clones that collectively represent functional and mechanistic diversity. This reduces the likelihood of redundancy and supports rational portfolio decisions before resource-intensive optimization begins.
2. Maximizing options for development: Broad epitope coverage offers multiple pathways for lead progression based on target biology, cross-reactivity, or developability criteria. It also broadens opportunities for novel intellectual property positions.
3. Validation of synthetic library platforms: The study offered direct evidence that Adimab’s yeast-based in vitro library recapitulates the diversity of immune repertoires. Its performance has translated into practical outcomes across multiple programs, including:

• MM-151, an oligoclonal anti-EGFR cocktail of three antibodies targeting epitopes distinct from cetuximab and panitumumab.
• HT-19, an anti-HER2 antibody-drug conjugate binding an epitope different from those of trastuzumab and pertuzumab.
• A broadly neutralizing anti-Staphylococcus aureus antibody targeting a conserved patch across α-hemolysin and leukocidins, enabling cross-reactivity despite low sequence homology among toxins. This bnAb demonstrates superior effectiveness in combating S. aureus infections compared to antibodies targeting α -hemolysin alone.

4. Establishing a benchmark for high-throughput analytics: This work also highlighted the power of array-based biosensor assays and advanced data-integration tools to interrogate antibody repertoires at scale. Such methods have since become integral to early-stage screening workflows across the industry.

Conclusion

This study established a new benchmark for evaluating antibody diversity and guiding the discovery of therapeutic antibodies representing a diverse set of functional epitopes. By combining yeast-based synthetic library design, high-throughput SPR imaging, and structure-guided analysis, the authors demonstrated that in vitro antibody platforms can achieve, and in some respects exceed, the epitope diversity of animal immunizations.

These findings underscore the critical role of high-throughput epitope binning in modern discovery pipelines. Comprehensive mapping of the antibody–antigen landscape enables more strategic lead selection, facilitates differentiation, and ultimately accelerates the path to clinical development.

For more details, access the full article in mAbs.

Post-study note: Examples of Adimab antibodies targeting specific or unique epitopes include: 

1. NPX887: A bispecific T cell engager, B7-H7 x CD3. B7-H7 has an epitope associated with a desired activity. NextPoint Therapeutics generated multiple leads and determined that NPX887 provided superior functional activity due to its epitope (Vekaria et al. J Immunother Cancer, 2024;12. DOI: 10.1136/jitc-2024-SITC2024.1318
2. TTX-080: HLA-G antibody that blocks ILT2 and ILT4 interactions and is specific to HLA-G. (Paul Widboom, PEGS Summit 2025. Discovery of TTX-080: Developing a Highly Specific Clinical HLA-G Antibody in Collaboration with Tizona Therapeutics). 
3. SRK-181: Binds proTGFB1 regardless of the binding partner it is complexed with. This requires a specific epitope (Martin et al. Sci Transl Med . 2020 Mar 25;12(536). DOI: 10.1126/scitranslmed.aay8456.