An official website of the United States government
Abstract Therapeutic antibody development requires selection and engineering of molecules with high affinity and other drug-like biophysical properties. Co-optimization of multiple antibody properties remains a difficult and time-consuming process that impedes drug development. Here we evaluate the use of machine learning to simplify antibody co-optimization for a clinical-stage antibody (emibetuzumab) that displays high levels of both on-target (antigen) and off-target (non-specific) binding. We mutate sites in the antibody complementarity-determining regions, sort the antibody libraries for high and low levels of affinity and non-specific binding, and deep sequence the enriched libraries. Interestingly, machine learning models trained on datasets with binary labels enable predictions of continuous metrics that are strongly correlated with antibody affinity and non-specific binding. These models illustrate strong tradeoffs between these two properties, as increases in affinity along the co-optimal (Pareto) frontier require progressive reductions in specificity. Notably, models trained with deep learning features enable prediction of novel antibody mutations that co-optimize affinity and specificity beyond what is possible for the original antibody library. These findings demonstrate the power of machine learning models to greatly expand the exploration of novel antibody sequence space and accelerate the development of highly potent, drug-like antibodies.
more »
« less
ANTIBODY DESIGN WITH CONSTRAINED BAYESIAN OPTIMIZATION
In therapeutic antibody design, achieving a balance between optimizing binding affinity subject to multiple constraints, and sequence diversity within a batch for experimental validation presents an important challenge. Contemporary methods often fall short in simultaneously optimizing these attributes, leading to ineffi- ciencies in experimental exploration and validation. In this work, we tackle this problem using the latest developments in constrained latent space Bayesian op- timization. Our methodology leverages a deep generative model to navigate the discrete space of potential antibody sequences, facilitating the selection of diverse, high-potential candidates for synthesis. We also propose a novel way of training VAEs that leads to a lower dimensional latent space and achieves excellent per- formance under the data-constrained setting. We validate our approach in vitro by synthesizing optimized antibodies, demonstrating consistently high binding affini- ties and preserved thermal stability.
more »
« less
- Award ID(s):
- 2145644
- PAR ID:
- 10544965
- Publisher / Repository:
- GEM workshop, ICLR 2024
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)In the global health emergency caused by coronavirus disease 2019 (COVID-19), efficient and specific therapies are urgently needed. Compared with traditional small-molecular drugs, antibody therapies are relatively easy to develop; they are as specific as vaccines in targeting severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); and they have thus attracted much attention in the past few months. This article reviews seven existing antibodies for neutralizing SARS-CoV-2 with 3D structures deposited in the Protein Data Bank (PDB). Five 3D antibody structures associated with the SARS-CoV spike (S) protein are also evaluated for their potential in neutralizing SARS-CoV-2. The interactions of these antibodies with the S protein receptor-binding domain (RBD) are compared with those between angiotensin-converting enzyme 2 and RBD complexes. Due to the orders of magnitude in the discrepancies of experimental binding affinities, we introduce topological data analysis, a variety of network models, and deep learning to analyze the binding strength and therapeutic potential of the 14 antibody–antigen complexes. The current COVID-19 antibody clinical trials, which are not limited to the S protein target, are also reviewed.more » « less
-
Abstract Pemphigus vulgaris (PV) is a blistering autoimmune disease that affects the skin and mucous membranes. The mechanisms by which PV antibodies induce loss of cohesion in keratinocytes are not fully understood. It is accepted that the process starts with antibody binding to desmosomal targets, which leads to its disassembly and subsequent structural changes to cell–cell adhesions. In vitro imaging of desmosome molecules has been used to characterize this initial phase. However, there remains an untapped potential of image analysis in providing us with more in-depth knowledge regarding biophysical changes after antibody binding. Currently, there is no quantitative framework from immunofluorescence images in PV pathology. Here, we seek to establish a correlation of biophysical changes with antibody pathogenicity by examining the effects of PV antibodies on adhesion molecules and the cytoskeletal network. Specifically, we introduced a data-driven approach to quantitatively evaluate perturbations in adhesion molecules following antibody treatment. We identify distinct imaging signatures that mark the impact of antibody binding on the remodeling of adhesion molecules and introduce a pathogenicity score to compare the relative effects of different antibodies. From this analysis, we showed that the biophysical response of keratinocytes to distinct PV antibodies is highly specific, allowing for accurate prediction of their pathogenicity. For instance, the high pathogenicity scores of the PVIgG and AK23 antibodies show strong agreement with their reported PV pathology. Our data-driven approach offers a detailed framework for the action of antibodies in pemphigus and paves the way for the development of effective diagnostic and therapeutic strategies.more » « less
-
One of the most important attributes of anti‐amyloid antibodies is their selective binding to oligomeric and amyloid aggregates. However, current methods of examining the binding specificities of anti‐amyloid β (Aβ) antibodies have limited ability to differentiate between complexes that form between antibodies and monomeric or oligomeric Aβ species during the dynamic Aβ aggregation process. Here, we present a high‐resolution native ion‐mobility mass spectrometry (nIM‐MS) method to investigate complexes formed between a variety of Aβ oligomers and three Aβ‐specific IgGs, namely two antibodies with relatively high conformational specificity (aducanumab and A34) and one antibody with low conformational specificity (crenezumab). We found that crenezumab primarily binds Aβ monomers, while aducanumab preferentially binds Aβ monomers and dimers and A34 preferentially binds Aβ dimers, trimers, and tetrameters. Through collision induced unfolding (CIU) analysis, our data indicate that antibody stability is increased upon Aβ binding and, surprisingly, this stabilization involves the Fc region. Together, we conclude that nIM‐MS and CIU enable the identification of Aβ antibody binding stoichiometries and provide important details regarding antibody binding mechanisms.more » « less
-
The emergence of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has triggered a global COVID-19 pandemic, challenging healthcare systems worldwide. Effective therapeutic strategies against this novel coronavirus remain limited, underscoring the urgent need for innovative approaches. The present research investigates the potential of cannabis compounds as therapeutic agents against SARS-CoV-2 through their interaction with the virus’s papain-like protease (PLpro) protein, a crucial element in viral replication and immune evasion. Computational methods, including molecular docking and molecular dynamics (MD) simulations, were employed to screen cannabis compounds against PLpro and analyze their binding mechanisms and interaction patterns. The results showed cannabinoids with binding affinities ranging from −6.1 kcal/mol to −4.6 kcal/mol, forming interactions with PLpro. Notably, Cannabigerolic and Cannabidiolic acids exhibited strong binding contacts with critical residues in PLpro’s active region, indicating their potential as viral replication inhibitors. MD simulations revealed the dynamic behavior of cannabinoid–PLpro complexes, highlighting stable binding conformations and conformational changes over time. These findings shed light on the mechanisms underlying cannabis interaction with SARS-CoV-2 PLpro, aiding in the rational design of antiviral therapies. Future research will focus on experimental validation, optimizing binding affinity and selectivity, and preclinical assessments to develop effective treatments against COVID-19.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
