Monoclonal antibodies (mAbs) are the dominant class of biotherapeutics with increasing clinical use, recently celebrating the 100th mAb approval. Behind the therapeutic success of these molecules lies a time-consuming and costly process.
The average cost of developing therapeutics based on monoclonal antibodies is $650-$750 million, and it takes from 8 to 9 years.
Keep on reading to get a (brief) introduction to antibody therapeutics, different techniques for antibody engineering, and context on how computational approaches are now steadily transforming this rapidly expanding field.
Table of contents:
Before we get to the engineering part, let’s answer this essential question first: What are antibodies?
These naturally occurring proteins were evolved to recognize a wide variety of molecular surfaces - that’s how they defend organisms against diseases.
What’s so special about antibodies? As such they are key to the functioning of the adaptive immune systems of jawed vertebrates. The primary purpose of antibodies is to identify non-self molecular surfaces to neutralize noxious material (like viruses).
This quality makes antibodies incredibly valuable for clinical diagnostic and therapeutic domains.
The versatile binding nature of antibodies opens new opportunities to develop molecules to bind and modify the function of virtually any molecule - which is the premise of diagnostics and drug development.
Antibody engineering aims to discover and modify monoclonal antibody (mAb) sequences or structures that bind a therapeutic target, modulating its pathological function.
The idea here is to produce highly specific antibodies characterized by optimal processing, stability, and tolerance (with some of the functions not necessarily being associated with ‘natural antibodies’ and having to be engineered-in).
Antibody engineering is revolutionizing the diagnosis and immunotherapy for developing the treatment of various diseases, especially cancer therapy. The main challenge in the production of therapeutic antibodies and antibody-derived drugs is achieving the highest objective response rate in patients and the lowest toxicity.
The first antibodies engineered by scientists to fight against a specific disease were approved by the US Food and Drug Administration (FDA) already back in 1986.
Since then, industrial and academic laboratories have carried out many different programs for developing therapeutic antibodies. Antibody engineering is today a central approach to fighting some of the most serious diseases, including cancer.
By using the latest, most sophisticated experimental techniques, scientists can produce fully human monoclonal antibodies that come with a much lower immunogenicity level than the early mAbs engineered on mice.
In recent years, therapeutic antibodies have become the primary class of new drug development programs - and led to the production of some of the bestselling drugs in the pharmaceutical market. Seven of the top ten best-selling drugs in 2018 were mAbs.
The global therapeutic monoclonal antibody market was valued at $115.2 billion in 2018 and is expected to generate revenue of $300 billion by 2025. Recently, 100th monoclonal antibody was approved by the FDA, with hundreds more in various stages of clinical trials.
The market for therapeutic antibody drugs has grown dramatically as new therapeutics became approved for treating many human diseases - including many types of cancer, autoimmune, metabolic, and infectious diseases.
The following techniques play the primary role in antibody engineering today:
The antibody phage display method is based on integrating a gene sequence coding for a particular antibody into the DNA sequence of a filamentous bacteriophage, allowing its expression on the surface of the bacteriophage capsid.
The immobilized antigens are washed with phages, and those with specificity towards the antigen bind, and can be washed away. Such binding phages form the basis for further iterations where diversity is introduced to the binding site.
The basis for this method is the creation of a ‘suitable’ phage library. These are created on the basis of current knowledge of antibody diversity.
Naïve libraries are largely non-specific and these can be created by sampling a representative set of antibodies reflecting diversity in an organism (typically human).
Target-specific libraries are developed on basis of antibodies with known specificities, introducing diversity in the binding site with the goal of developing an even better binder than the original antibody.
This technique is based on challenging the immune system of an animal with an antigen (typically mouse), inducing them to produce antibodies against it. The antigen-specific B cells are selected in this case.
Antibodies originating from such animal immunizations might induce an immune response in humans, as a result of its non-human origin. Such sequences, therefore, undergo ‘humanization’ where the original animal antibody is modified to increase its similarity to human antibodies and thus reduce its immunogenicity.
Bioinformatics offers an alternative to experimental time- and resource-consuming antibody engineering processes. The holy grail of antibody drug design is to produce antibodies purely in silico, given a target antigen.
Clearly, such a feat of engineering would reduce the years of development to only a handful of confirmatory experiments.
Currently, computational methods are not replacing the experimental pipelines. But they are already an indispensable part of many therapeutic pipelines. The prediction of biophysical properties, structures, and other multiple parameters, facilitates the research decision process while developing such therapeutics.
Scientists can use established bioinformatics methods such as homology modeling, protein-protein docking, or protein interface prediction for rational antibody design.
Pharmaceutically-focused computational approaches help to assess the immunogenicity and developability properties of antibodies.
Moreover, the increasing volume of structural, sequence and experimental data on antibodies deposited in the public domain provides the foundation for continuous improvement and benchmarking of such data-driven approaches.
A good example of that is the currently generated volume of next-generation sequencing (NGS) of B-cell receptor (antibody) repositories. NGS provides a snapshot of millions of antibody sequences sampled from the theoretically possible 1012–1018 antibody sequences in a human repertoire.
Such samples provide an unprecedented view of the natural antibody diversity in healthy individuals as well as in response to disease and vaccine challenges.
Computational antibody design methods are now becoming a standard in pharmaceutical discovery processes and provide time- and cost-efficient methods of guiding experimental approaches.
As the value of antibodies as therapeutics increases, faster and more accurate methods are badly needed to keep delivering novel therapies to the clinic. The more attrition there is in the current drug development pipelines, the less affordable and timely will the future treatments be.
To reduce attrition, current experimental approaches would have to radically increase throughputs and reduce costs. Therefore, the most realistic and promising avenue of rethinking the drug discovery process currently is the continuous improvement and adoption of bioinformatics solutions for antibody drug discovery pipelines.
Even at the stage when the computational methods don’t replace the experiments entirely, they provide a rapid and cost-efficient way to accelerate drug discovery. If bioinformatics can provide even seemingly modest acceleration of 1%, on a typical decade-long development scale this still means delivering a drug months earlier to the benefit to many patients.
But the direction in which computational antibody tools are developing indicates that we should soon be able to talk of improvements on the scale of years rather than months.
Keep an eye on our Resources section for more educational materials on computational approaches to antibody engineering.