Sponsored content brought to you by
In 1986, the FDA approved the first therapeutic monoclonal antibody.1 Now, nearly 40 years later, “antibodies have become the fastest-growing class of biological drugs approved for the treatment of a wide range of diseases, from cancer to autoimmune conditions,” according to a team of experts from the Pennsylvania-based Geisinger Commonwealth School of Medicine.2
Although therapeutic antibodies offer a wide range of benefits, from high affinity and specificity to multiple mechanisms of attack on diseases, scientists need tools that “address the challenge of screening millions of antibody-producing cells to find the best candidates,” says Bob Chen, PhD, senior director, discovery systems at OmniAb.
Traditionally, scientists produced monoclonal antibodies by applying hybridoma technology to mice, subsequently testing the therapeutic potential of the antibodies, one by one. Such methods, however, “cause the loss of valuable immune repertoire information and time,” Chen says. “Instead, we have the ability to utilize AI-empowered single-cell screening to analyze millions of B-cells and efficiently identify high-affinity clones directly from our animals, including transgenic mice and rats, as well as cows and chickens. Then, we use AI to further expand the insights gained from a few hundred well-validated sequences, broadening our understanding over a larger space.”
To analyze the potential therapeutic antibodies, OmniAb developed xPloration®, which incorporates AI in a platform that can collect thousands of antibody variants from millions of single B cells. This platform includes about 1.5 million microcapillaries that are just 40 micrometers in diameter and one millimeter long. “This enables robust spatial separation of
single cells for various assays,” Chen explains. “Our most common use case is a selective binding assay for antibodies.”
Then, automated imaging and AI-based machine vision algorithms, including classification and segmentation, identify the B cells that are creating antibodies with the desired features, such as binding a particular target. “These AI models are trained on expert-labeled data,” Chen notes. “So, we can use AI to democratize the skill set of expert users to all users of the platform.” B cells identified with the desired phenotype are collected with a proprietary laser recovery technology.
Selecting the sequences
OmniAb turns to Biological Intelligence™, the interplay between rational genetic design and immunization, to generate the initial large, high-quality antibody repertoire. “To get the most out our diverse repertoires, we need high-throughput tools,” according to Chen. For the large-scale mining of the immune repertoires, OmniAb utilizes xPloration, in which a single run can screen millions of B cells and identify thousands of potential hits.
Then OmniAb can use its suite of in silico tools, OmniDeep™, which integrates structural modeling, multi-species antibody databases, molecular dynamics simulations, AI, and machine and deep learning sequence models, to aid in narrowing down to hundreds or thousands of antibodies that possess the desired affinity, specificity, and other features that enhance the odds of efficiently developing a novel therapy. For example, one of OmniAb’s collections of data produced about 10 million antibody sequences, generating 5,000 hits which were sorted with xPloration and characterized with hundreds of affinity measurements, resulting in hundreds of high affinity and well behaved clones.
OmniAb is spearheading the use of AI with immunized animals, using deep learning to strengthen the core ability of the company. “After AI learns the space, it can suggest a large number of clones to test,” Chen says. “To improve the iteration-cycle time, we are using expression libraries to evaluate AI-selected sequences on the order of thousands of selections.”
Moreover, OmniAb’s approach features two fundamental capabilities. “We analyze antibodies secreted by B cells against membrane-based targets, such as GPCRs and ion channels, in their native format on the surface of cells,” Chen continues. “Plus, we have the throughput to find rare binding events—if they exist.”
The vast antibody space offers millions of opportunities to develop therapies that precisely target diseases. Developing the best therapeutic antibodies over the next 40 years, though, depends on efficiently finding the best ones.
Learn more www.omniab.com
References
- Lu, R-M., Hwang, Y-C., Liu, I-J., et al. Development of therapeutic antibodies for the treatment of diseases. Journal of Biomedical Science 27: 1. (2020).
- Sharma, P., Joshi, R.V., Pritchard, R., et al. Therapeutic antibodies in medicine. Molecules 28(18): 6438. (2023).
