Applying Mechanistic Models in Bioprocess Development - Measurement, Monitoring, Modelling, and Control of Bioprocesses - page 139

parameters that lead to the right critical quality attributes'' [ 2 ]. The term ''critical''
should be interpreted as ''having a significant influence on final product quality.''
Changing the process within the design space is therefore not considered as a
change. As a consequence, no regulatory postapproval of the process is required
for a change within the design space. Almost naturally, this opens up the possi-
bility of increased use of optimization methods for pharmaceutical processes in the
future, methods that have been used for a long time in, for example, the chemical
industry [ 3 ].
Small-molecule (MW \ 1,000) drug substances (APIs, NCEs) are typically
produced via organic synthesis. In such a production system, the available process
knowledge is often relatively large. Process systems engineering (PSE) methods
and tools—especially those relying on mechanistic models to represent available
process knowledge—are therefore increasingly applied in the frame of pharma-
ceutical process development and innovation of small-molecule drugs [ 4 ], with the
aim of shortening time to market while yielding an efficient production process. In
essence, mechanistic models rely on deterministic principles to represent available
process knowledge on the basis of mass, energy, and momentum balances; given
initial conditions, future system behavior can be predicted.
It is, however, not the intention here to provide a detailed review on mecha-
nistic models for biobased production processes of pharmaceuticals. There are
excellent textbooks and review articles on the general principles of mechanistic
modeling of fermentation processes [ 5 - 8 ], biocatalysis [ 9 , 10 ], and mammalian
cell culture [ 11 ].
Biotechnology research has resulted in a new class of biomolecular drugs—
typically larger molecules, also called biologics or NBEs—which includes
monoclonal antibodies, cytokines, tissue growth factors, and therapeutic proteins.
The production of biomolecular drugs is usually complicated and extremely
expensive. The level of process understanding is therefore in many cases lower,
compared with small-molecule drug substances, and as a consequence, PSE
methods and tools relying on mechanistic models are usually not applied to the
same extent in production of biomolecular drugs, despite the fact that quite a
number of articles have been published throughout the years on the development
of mechanistic models for such processes.
This chapter focuses on the potential use of mechanistic models within bio-
based production of drug products, as well as the use of good modeling practice
(GMoP) when using such mechanistic models [ 12 ]. A case study with the yeast
model by Sonnleitner and Käppeli [ 1 ] is used to illustrate how a mechanistic
model can be formulated in a well-organized and easy-to-interpret matrix notation.
This model is then analyzed using uncertainty and sensitivity analysis, an analysis
that serves as a starting point for a discussion on the potential application of such
methods. Strategies for mechanistic model-building are highlighted in the final
discussion.