1
|
Yu X, Vangjeli K, Prakash A, Chhaya M, Stanley SJ, Cohen N, Huang L. Protein language models enable prediction of polyreactivity of monospecific, bispecific, and heavy-chain-only antibodies. Antib Ther 2024; 7:199-208. [PMID: 39036071 PMCID: PMC11259759 DOI: 10.1093/abt/tbae012] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/09/2024] [Revised: 05/10/2024] [Accepted: 05/20/2024] [Indexed: 07/23/2024] Open
Abstract
Background Early assessment of antibody off-target binding is essential for mitigating developability risks such as fast clearance, reduced efficacy, toxicity, and immunogenicity. The baculovirus particle (BVP) binding assay has been widely utilized to evaluate polyreactivity of antibodies. As a complementary approach, computational prediction of polyreactivity is desirable for counter-screening antibodies from in silico discovery campaigns. However, there is a lack of such models. Methods Herein, we present the development of an ensemble of three deep learning models based on two pan-protein foundational protein language models (ESM2 and ProtT5) and an antibody-specific protein language model (PLM) (Antiberty). These models were trained in a transfer learning network to predict the outcomes in the BVP assay and the bovine serum albumin binding assay, which was developed as a complement to the BVP assay. The training was conducted on a large dataset of antibody sequences augmented with experimental conditions, which were collected through a highly efficient application system. Results The resulting models demonstrated robust performance on canonical mAbs (monospecific with heavy and light chain), bispecific Abs, and single-domain Fc (VHH-Fc). PLMs outperformed a model built using molecular descriptors calculated from AlphaFold 2 predicted structures. Embeddings from the antibody-specific and foundational PLMs resulted in similar performance. Conclusion To our knowledge, this represents the first application of PLMs to predict assay data on bispecifics and VHH-Fcs.
Collapse
Affiliation(s)
- Xin Yu
- Corresponding authors. Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States. E-mail: ; E-mail:
| | - Kostika Vangjeli
- Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States
| | - Anusha Prakash
- Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States
| | - Meha Chhaya
- Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States
| | - Samantha J Stanley
- Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States
| | - Noah Cohen
- Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States
| | - Lili Huang
- Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States
| |
Collapse
|
2
|
Minot M, Reddy ST. Meta learning addresses noisy and under-labeled data in machine learning-guided antibody engineering. Cell Syst 2024; 15:4-18.e4. [PMID: 38194961 DOI: 10.1016/j.cels.2023.12.003] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/30/2023] [Revised: 07/21/2023] [Accepted: 12/07/2023] [Indexed: 01/11/2024]
Abstract
Machine learning-guided protein engineering is rapidly progressing; however, collecting high-quality, large datasets remains a bottleneck. Directed evolution and protein engineering studies often require extensive experimental processes to eliminate noise and label protein sequence-function data. Meta learning has proven effective in other fields in learning from noisy data via bi-level optimization given the availability of a small dataset with trusted labels. Here, we leverage meta learning approaches to overcome noisy and under-labeled data and expedite workflows in antibody engineering. We generate yeast display antibody mutagenesis libraries and screen them for target antigen binding followed by deep sequencing. We then create representative learning tasks, including learning from noisy training data, positive and unlabeled learning, and learning out of distribution properties. We demonstrate that meta learning has the potential to reduce experimental screening time and improve the robustness of machine learning models by training with noisy and under-labeled training data.
Collapse
Affiliation(s)
- Mason Minot
- ETH Zurich, Department of Biosystems Science and Engineering, Basel 4056, Switzerland
| | - Sai T Reddy
- ETH Zurich, Department of Biosystems Science and Engineering, Basel 4056, Switzerland.
| |
Collapse
|
3
|
Irvine EB, Reddy ST. Advancing Antibody Engineering through Synthetic Evolution and Machine Learning. JOURNAL OF IMMUNOLOGY (BALTIMORE, MD. : 1950) 2024; 212:235-243. [PMID: 38166249 DOI: 10.4049/jimmunol.2300492] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Received: 07/19/2023] [Accepted: 10/20/2023] [Indexed: 01/04/2024]
Abstract
Abs are versatile molecules with the potential to achieve exceptional binding to target Ags, while also possessing biophysical properties suitable for therapeutic drug development. Protein display and directed evolution systems have transformed synthetic Ab discovery, engineering, and optimization, vastly expanding the number of Ab clones able to be experimentally screened for binding. Moreover, the burgeoning integration of high-throughput screening, deep sequencing, and machine learning has further augmented in vitro Ab optimization, promising to accelerate the design process and massively expand the Ab sequence space interrogated. In this Brief Review, we discuss the experimental and computational tools employed in synthetic Ab engineering and optimization. We also explore the therapeutic challenges posed by developing Abs for infectious diseases, and the prospects for leveraging machine learning-guided protein engineering to prospectively design Abs resistant to viral escape.
Collapse
Affiliation(s)
- Edward B Irvine
- Department of Biosystems Science and Engineering, ETH Zürich, Basel, Switzerland
| | - Sai T Reddy
- Department of Biosystems Science and Engineering, ETH Zürich, Basel, Switzerland
| |
Collapse
|
4
|
Matsunaga R, Ujiie K, Inagaki M, Fernández Pérez J, Yasuda Y, Mimasu S, Soga S, Tsumoto K. High-throughput analysis system of interaction kinetics for data-driven antibody design. Sci Rep 2023; 13:19417. [PMID: 37990030 PMCID: PMC10663500 DOI: 10.1038/s41598-023-46756-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/01/2023] [Accepted: 11/04/2023] [Indexed: 11/23/2023] Open
Abstract
Surface plasmon resonance (SPR) is widely used for antigen-antibody interaction kinetics analysis. However, it has not been used in the screening phase because of the low throughput of measurement and analysis. Herein, we proposed a high-throughput SPR analysis system named "BreviA" using the Brevibacillus expression system. Brevibacillus was transformed using a plasmid library containing various antibody sequences, and single colonies were cultured in 96-well plates. Sequence analysis was performed using bacterial cells, and recombinant antibodies secreted in the supernatant were immobilized on a sensor chip to analyze their interactions with antigens using high-throughput SPR. Using this system, the process from the transformation to 384 interaction analyses can be performed within a week. This system utility was tested using an interspecies specificity design of an anti-human programmed cell death protein 1 (PD-1) antibody. A plasmid library containing alanine and tyrosine mutants of all complementarity-determining region residues was generated. A high-throughput SPR analysis was performed against human and mouse PD-1, showing that the mutation in the specific region enhanced the affinity for mouse PD-1. Furthermore, deep mutational scanning of the region revealed two mutants with > 100-fold increased affinity for mouse PD-1, demonstrating the potential efficacy of antibody design using data-driven approach.
Collapse
Affiliation(s)
- Ryo Matsunaga
- Department of Bioengineering, School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
- Department of Chemistry and Biotechnology, School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
| | - Kan Ujiie
- Department of Bioengineering, School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
| | - Mayuko Inagaki
- Department of Bioengineering, School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
| | - Jorge Fernández Pérez
- Department of Bioengineering, School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
| | - Yoshiki Yasuda
- Department of Chemistry and Biotechnology, School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
| | - Shinya Mimasu
- Biologics Engineering, Discovery Intelligence, Astellas Pharma Inc., 21, Miyukigaoka, Tsukuba-shi, Ibaraki, 305-8585, Japan
| | - Shinji Soga
- Biologics Engineering, Discovery Intelligence, Astellas Pharma Inc., 21, Miyukigaoka, Tsukuba-shi, Ibaraki, 305-8585, Japan
| | - Kouhei Tsumoto
- Department of Bioengineering, School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan.
- Department of Chemistry and Biotechnology, School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan.
- The Institute of Medical Science, The University of Tokyo, Tokyo, 108-8639, Japan.
| |
Collapse
|
5
|
Svilenov HL, Arosio P, Menzen T, Tessier P, Sormanni P. Approaches to expand the conventional toolbox for discovery and selection of antibodies with drug-like physicochemical properties. MAbs 2023; 15:2164459. [PMID: 36629855 PMCID: PMC9839375 DOI: 10.1080/19420862.2022.2164459] [Citation(s) in RCA: 6] [Impact Index Per Article: 6.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/15/2022] [Revised: 12/22/2022] [Accepted: 12/29/2022] [Indexed: 01/12/2023] Open
Abstract
Antibody drugs should exhibit not only high-binding affinity for their target antigens but also favorable physicochemical drug-like properties. Such drug-like biophysical properties are essential for the successful development of antibody drug products. The traditional approaches used in antibody drug development require significant experimentation to produce, optimize, and characterize many candidates. Therefore, it is attractive to integrate new methods that can optimize the process of selecting antibodies with both desired target-binding and drug-like biophysical properties. Here, we summarize a selection of techniques that can complement the conventional toolbox used to de-risk antibody drug development. These techniques can be integrated at different stages of the antibody development process to reduce the frequency of physicochemical liabilities in antibody libraries during initial discovery and to co-optimize multiple antibody features during early-stage antibody engineering and affinity maturation. Moreover, we highlight biophysical and computational approaches that can be used to predict physical degradation pathways relevant for long-term storage and in-use stability to reduce the need for extensive experimentation.
Collapse
Affiliation(s)
- Hristo L. Svilenov
- Laboratory of General Biochemistry and Physical Pharmacy, Faculty of Pharmaceutical Sciences, Ghent University, Gent, Belgium
| | - Paolo Arosio
- Department of Chemistry and Applied Biosciences, ETH Zürich, Zürich, Switzerland
| | - Tim Menzen
- Coriolis Pharma Research GmbH, Martinsried, 82152, Germany
| | - Peter Tessier
- Department of Pharmaceutical Sciences, University of Michigan, Ann Arbor, MI, USA
- Biointerfaces Institute, University of Michigan, Ann Arbor, MI, USA
- Department of Chemical Engineering, University of Michigan, Ann Arbor, MI, USA
- Department of Biomedical Engineering, University of Michigan, Ann Arbor, MI, USA
| | - Pietro Sormanni
- Centre for Misfolding Diseases, Yusuf Hamied Department of Chemistry, University of Cambridge, Cambridge, UK
| |
Collapse
|
6
|
Harvey EP, Shin JE, Skiba MA, Nemeth GR, Hurley JD, Wellner A, Shaw AY, Miranda VG, Min JK, Liu CC, Marks DS, Kruse AC. An in silico method to assess antibody fragment polyreactivity. Nat Commun 2022; 13:7554. [PMID: 36477674 PMCID: PMC9729196 DOI: 10.1038/s41467-022-35276-4] [Citation(s) in RCA: 11] [Impact Index Per Article: 5.5] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/27/2022] [Accepted: 11/22/2022] [Indexed: 12/12/2022] Open
Abstract
Antibodies are essential biological research tools and important therapeutic agents, but some exhibit non-specific binding to off-target proteins and other biomolecules. Such polyreactive antibodies compromise screening pipelines, lead to incorrect and irreproducible experimental results, and are generally intractable for clinical development. Here, we design a set of experiments using a diverse naïve synthetic camelid antibody fragment (nanobody) library to enable machine learning models to accurately assess polyreactivity from protein sequence (AUC > 0.8). Moreover, our models provide quantitative scoring metrics that predict the effect of amino acid substitutions on polyreactivity. We experimentally test our models' performance on three independent nanobody scaffolds, where over 90% of predicted substitutions successfully reduced polyreactivity. Importantly, the models allow us to diminish the polyreactivity of an angiotensin II type I receptor antagonist nanobody, without compromising its functional properties. We provide a companion web-server that offers a straightforward means of predicting polyreactivity and polyreactivity-reducing mutations for any given nanobody sequence.
Collapse
Affiliation(s)
- Edward P Harvey
- Department of Biological Chemistry and Molecular Pharmacology, Blavatnik Institute, Harvard Medical School, Boston, MA, 02115, USA
| | - Jung-Eun Shin
- Department of Systems Biology, Harvard Medical School, Boston, MA, 02115, USA
| | - Meredith A Skiba
- Department of Biological Chemistry and Molecular Pharmacology, Blavatnik Institute, Harvard Medical School, Boston, MA, 02115, USA
| | - Genevieve R Nemeth
- Department of Biological Chemistry and Molecular Pharmacology, Blavatnik Institute, Harvard Medical School, Boston, MA, 02115, USA
| | - Joseph D Hurley
- Department of Biological Chemistry and Molecular Pharmacology, Blavatnik Institute, Harvard Medical School, Boston, MA, 02115, USA
| | - Alon Wellner
- Department of Chemistry, University of California, Irvine, CA, 92697, USA
- Department of Molecular Biology & Biochemistry, University of California, Irvine, CA, 92697, USA
- Department of Biomedical Engineering, University of California, Irvine, CA, 92692, USA
| | - Ada Y Shaw
- Department of Systems Biology, Harvard Medical School, Boston, MA, 02115, USA
| | - Victor G Miranda
- Department of Biological Chemistry and Molecular Pharmacology, Blavatnik Institute, Harvard Medical School, Boston, MA, 02115, USA
| | - Joseph K Min
- Department of Systems Biology, Harvard Medical School, Boston, MA, 02115, USA
| | - Chang C Liu
- Department of Chemistry, University of California, Irvine, CA, 92697, USA
- Department of Molecular Biology & Biochemistry, University of California, Irvine, CA, 92697, USA
- Department of Biomedical Engineering, University of California, Irvine, CA, 92692, USA
| | - Debora S Marks
- Department of Systems Biology, Harvard Medical School, Boston, MA, 02115, USA.
- Broad Institute of Harvard and MIT, Cambridge, MA, 02142, USA.
| | - Andrew C Kruse
- Department of Biological Chemistry and Molecular Pharmacology, Blavatnik Institute, Harvard Medical School, Boston, MA, 02115, USA.
| |
Collapse
|
7
|
Gao B, Han J, Reddy ST. Learning what not to select for in antibody drug discovery. CELL REPORTS METHODS 2022; 2:100258. [PMID: 35880020 PMCID: PMC9308151 DOI: 10.1016/j.crmeth.2022.100258] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Indexed: 11/25/2022]
Abstract
Identifying antibodies with high affinity and target specificity is crucial for drug discovery and development; however, filtering out antibody candidates with nonspecific or polyspecific binding profiles is also important. In this issue of Cell Reports Methods, Saksena et al. report a computational counterselection method combining deep sequencing and machine learning for identifying nonspecific antibody candidates and demonstrate that it has advantages over more established molecular counterselection methods.
Collapse
Affiliation(s)
- Beichen Gao
- Department of Biosystems Science and Engineering, ETH Zurich, Basel 4058, Switzerland
| | - Jiami Han
- Department of Biosystems Science and Engineering, ETH Zurich, Basel 4058, Switzerland
| | - Sai T. Reddy
- Department of Biosystems Science and Engineering, ETH Zurich, Basel 4058, Switzerland
| |
Collapse
|