Biomedical Engineering Reference
In-Depth Information
Figure 2.2. Biotechnology products have a large number of structural attributes that may
impact product performance. (a) Some structural attributes of monoclonal antibodies are
indicated, suchas pyroglutamine, oxidation, glycation, deamidation, glycosylation, and clipping
of C-terminal lysines. An oxidation at one site is circled in blue. (b) A decision tree for the impact
of that oxidation on product activity is shown. (See the insert for color representation of this
figure.)
these decision trees. For example, immunogenicity is difficult to predict on the basis of
attributes, but the impact of immunogenicity can be evaluated in terms of clinical risk
[23-25]. Risk assessments, whether for activity or safety, are challenging and require the
use of many sources of information such as prior knowledge, related product or platform
data, in vitro and in vivo biological characterization of product variants, and clinical data.
In many cases, no one source of information would allow a meaningful risk assessment.
However, the integration of multiple sources of information may facilitate a useful
assessment of risk. A matrix approach can be informative [26], evaluating product lots
generated at varying points throughout the development for biological impact across a
variety of studies. This information can give confidence to proposed mechanisms of
action and structure-function relationships. These data may be integrated around
mechanism of action models and potentially use Bayesian statistical approaches.
For products that share significant sequence homology to related products, a platform
approach to attribute impact may be very useful. Monoclonal antibodies may present
opportunities [26] for this, but the nature of the targets, patient population, disease state(s),
and the role of effector functions need to be considered in any extrapolations. Studies
demonstrating that a product attribute is rapidly modified in vivo (e.g., deamidation or
oxidation of a specific residue) may allow for increasing levels of that attribute over
product shelf life.Comparisons to relatedendogenousmoleculesmayalsobe informative.
Clinical studies on product variants would provide the strongest linkage between
product attributes and clinical performance. In this topic [27], an example of using
structural characterization of variants in timed samples for variant pharmacokinetics is
given. Such studies do not require purification of variants and additional clinical studies.
These studies may allow the evaluation of variant effects on pharmacokinetics and, in
some cases, variant effects on pharmacodynamics. In addition, during standard clinical
Search WWH ::




Custom Search