Researchers have developed a high-throughput, in vivo technology that they hope could significantly speed the timeline for discovering potential new antibody-based drugs. The technology, developed through an eight-year collaboration between the University of Leeds and AstraZeneca, allows fragments of antibodies to be screened in bacteria for their likely susceptibility to aggregation during manufacture or storage, much earlier in the drug discovery process.

“There is tremendous excitement about this approach,” said Sheena Radford, FRS, director of the Astbury Centre for Structural Molecular Biology, at the University of Leeds. “We are letting evolutionary selection change the sequence of the proteins for us and that might make some of them more useful as drug therapies. Importantly for industry, nature does the hard work—obviating the need for so-called rational engineering which is time- and resource-intensive. Radford and colleagues reported on the technology in Nature Communications, in a paper titled, “An in vivo platform to select and evolve aggregation-resistant proteins.”

Antibody-based medicines now comprise over half of first-time approvals, and seven of the ten highest-grossing pharmaceuticals in 2018 were based on antibodies or antibody-like scaffolds, the authors noted. Developing protein-based drugs is not without challenges, however, particularly when produced at a large scale. Stresses encountered during manufacturing can disrupt the structure of the protein molecule, leading to aggregation and loss of activity, and potentially failure of the candidate at a late stage of development. “As proteins are subjected to various stresses during manufacturing that increase the risk of protein misfolding and aggregation, overcoming aggregation (which may be associated with low protein stability and/or low solubility) is a major hurdle in the development of biopharmaceuticals,” the scientists noted. Unfortunately, the prediction of what the investigators term “poorly-behaved” biologics is difficult, and searching for sequences with desired properties is labor-intensive and time-consuming.”

Research lead David Brockwell, PhD, associate professor in the Astbury Centre for Structural Molecular Biology further noted, “Antibody therapeutics have revolutionized medicine. They can be designed to bind to almost any target and are highly specific. But a significant problem has been the failure rate of candidates upon manufacturing at industrial scale. This often only emerges at a very late stage in the development process—these drugs are failing at the last hurdle. But our research is turning that problem on its head.”

Scientists developing antibody-based drugs against a particular target are not always restricted to identifying a single protein sequence. There is often an array of similar antibodies with the same target-binding ability. Researchers may have a range of proteins to screen to determine which are more likely to progress through the development process.

The University of Leeds and AstraZeneca team has developed a tripartite β-lactamase enzyme assay (TPBLA) to evaluate which sequences have a propensity to aggregate. The candidate sequence is cloned into the center of an enzyme that breaks down antibiotics, in E. coli. This provides a way of directly linking antibiotic resistance of the bacteria to how aggregation-prone the antibody fragment is. “ … the test protein is fused inframe between the two domains of the E. coli periplasmic enzyme β-lactamase (βLa),” the authors explained. “This assay thus directly links the aggregation-propensity of the test protein to the susceptibility of the bacterium to β-lactam antibiotics.”

A simple readout—bacterial growth on an agar plate containing antibiotic—then gives an indication of whether the protein will survive the manufacturing process. If the antibody proteins are susceptible to stress, they will unfold or clump together, become inactive, and the antibiotic will kill the bacteria. But if the protein chain is more stable, the bacteria will exhibit antimicrobial resistance and continue to grow in the presence of the antibiotic.

The surviving bacteria can then be harvested and the cloned protein sequence identified. The genes of protein sequences hosted in the bacterial cells that have shown resistance to aggregation are then sequenced and scored, to select the best performing sequences. After verifying that the new antibody would still retain its binding affinity to the original disease-causing target, it can be taken forward for further development. The whole cycle takes about a month.

In their Nature Communications paper, the researchers described use of TPBLA with both therapeutically relevant proteins and proteins involved in aggregation diseases, to demonstrate use of the assay to assess the aggregation propensity of different protein structural scaffolds, including monoclonal antibody fragments that are only slightly different in amino acid sequence, but which have very different aggregation properties. “Importantly, the approach does not require any structural knowledge or prior biophysical information about the protein of interest,” they commented. “Most importantly, no perturbant such as increased temperature, pH, or chemical denaturant is used to accelerate aggregation, allowing identification of sequence characteristics that trigger innate (unaccelerated) aggregation pathways.”

The basic approach can also be taken a step further, and the assay used as a directed evolution screen, by which the evolutionary pressure is applied by the antibiotic, to select for protein bacteria that produce protein variants which do not aggregate. The team used the TPBLA as a screen for directed evolution experiments to select for sequences that are aggregation-resistant. “This powerful tool can thus screen and evolve ‘manufacturable’ biopharmaceuticals early in industrial development,” the authors stated. “By comparing the mutational profiles of three different immunoglobulin scaffolds, we show the applicability of this method to investigate protein aggregation mechanisms important to both industrial manufacture and amyloid disease.”

Radford commented, “The collaboration that has existed between the team of scientists within the University of Leeds and AstraZeneca demonstrates the power of industry and academia working together to tackle what has been one of the major roadblocks preventing the efficient and rapid development of these powerful therapeutic molecules.” The sequence information generated through the work will be added to a database, which could be paired with artificial intelligence and machine learning tools to identify patterns in protein sequences that would indicate whether a particular protein can be scaled up for pharmaceutical production, without the need for experimentation. “That is our next challenge and one we are tackling right now,” Radford said.

As the authors stated, “This will allow a greater understanding of the relationship between sequence, solubility, and aggregation, the developability of promising biologic candidates, and the prediction of mutations that may cause protein aggregation disease.”

David Lowe, PhD, senior director, R&D, and who led the work at AstraZeneca, added, “The screening system that we have developed here is a great example of industry and academia working together to solve important challenges in the development of potential new medicines. By combining AstraZeneca’s antibody discovery and screening expertise, together with the Astbury Centre’s world-leading knowledge of protein structure and aggregation, we have produced a high throughput method for rapidly evolving proteins with better biophysical properties that has the potential for wide scientific applicability.”

The researchers suggested that TPBLA will have utility as a research tool to help better understand the mechanisms of aggregation, which still aren’t understood more than 30 years since the introduction of the first IgG antibodies into the clinic. “We have shown here that the TPBLA is a powerful method by which to identify (using the TPBLA alone), or re-engineer (using the TPBLA as a directed evolution screen) inherently manufacturable proteins.”