Ask the Expert: Developability: Evaluating Specificity, Immunogenicity, Functionality, and Manufacturability for Lead Candidate Selection

View PDF

Campbell Bunce (chief scientific officer at Abzena) joined BPI for an Ask-the-Expert webinar on 21 May 2019 to discuss his company’s approach to developability assessment, an increasingly popular way to identify leading drug candidates. Bunce explained what developability assessment is and how Abzena uses it to identify candidate liabilities and manage both clinical and financial risk during key stages of drug development.

Bunce’s Presentation
The high attrition rate of drug candidates from discovery to clinical proof-of-concept is well known. The degree of failure makes it crucial for developers to predict risk factors that could present in later, more costly stages of manufacture and clinical evaluation. Thus, companies increasingly are turning to developability assessment, combining in silico analysis with in vivo and ex vivo experiments that characterize lead drug candidates with the highest chance of success.

Developability studies establish two primary concerns: whether a drug candidate can be manufactured in a stable form to required scales in the appropriate form and at a suitable cost, and whether the drug will work as designed while meeting safety requirements. The hope is that understanding drug candidates more completely early on will mitigate attrition and benefit stakeholders, including regulatory agencies, investors, and patients.

To that end, Abzena uses developability assessment to evaluate four key attributes tied to drug clinical success: specificity, functionality, immunogenicity, and manufacturability. Abzena evaluates those criteria holistically so that it does not produce, e.g., drugs that have strong functionality profiles but are chemically unstable or difficult to express.

Abzena begins with some combination of in silico modeling and high-throughput screening to identify risks while controling up-front costs. Front-loading the analytical work could seem burdensome, but it can identify problems that otherwise might not appear until at-scale manufacturing. In silico studies could highlight, for instance, deamidation or oxidation risks. Gathering such findings early on helps Abzena determine whether to pursue corrective designs or cancel a program that’s unlikely to succeed.

In silico bioinformatics analyses alongside ex vivo assays also support immunogenicity risk profiling. Early stage ex vivo assays allow ranking of candidates based on induction of CD4 T cell responses from a broad population of human cells. The results enable researchers to design out such liabilities, then clone and express multiple design alterations — again helping to determine whether a candidate merits further study.

Developability assessment effectively helps drug developers construct a comprehensive drug profile early in development. By anticipating future liabilities, developers can increase significantly the likelihood of drug success while ensuring feasibility and safety for a range of important stakeholders.

Questions and Answers
Is there a formal immunogenicity database for monoclonal antibody (MAb) therapeutics? I do not think so. There is much literature on antidrug antibody response that can be fed back to improve preclinical assays toward a more predictive position. They are used most robustly for ranking relative risks against multiple variants. Abzena has its own database of known T cell epitopes, and we use proprietary bioinformatics algorithms to identify and evaluate relative risk in silico.

What are the most important aspects of developability assessment? The two keys at the early stage are stability and functionality. If a molecule is nonfunctional, then there is no point in developing it; if it is not stable, then it is not likely to be functional for long. Other manufacturability and productivity routes can be managed somewhat in downstream processing. But it is good to know that a molecule is stable and does what we expect it to do.

To what extent might developability assessment affect development timelines? The answer depends on when and at what level of liability the risks are identified. Applying a rigorous developability approach early on takes three to five months. One good in silico process helps us design out liabilities before we start making material, so that time is insignificant in the context of the full lifespan of a program.

What innovations could take developability from a predominantly later stage to an earlier stage part of the process? One area for improvement is silico analysis. The more we learn from drugs in development, the more we can feed back into bioinformatics platforms. We are constantly doing so at Abzena.

The second strategy would be to perform many tests using little material. Miniaturizing assays and high-throughput screening will help us answer questions without spending money making a lot of sample material.

More Online 
The full presentation of this webcast can be found on the BioProcess International website at the link below.

Watch the full webcast now.