Promoteur : Prof. Alain Vande Wouwer

Résumé :

Mammalian cell cultures are a representative source of a number of biopharmaceutical
products, including monoclonal antibodies (Niklas and Heinzle,
2012; Sidoli et al, 2004), viral vaccines (Vester et al, 2010), and hormones
(Nottorf et al, 2007). In the last decades, monoclonal antibodies (mAbs) are
increasingly used for medical research, diagnosis and therapy (Saleem and
Mustafa, 2008). The major percentage of the mAbs approved and under
trials is produced by mammalian cells because of their capacity for proper
protein folding, assembly and post-translational modification that result in
fully-active product. Thus, the quality and ecacy of a protein can be superior
when expressed in mammalian cells than in other hosts such as bacteria,
plants and yeast (Wurm, 2004). Monoclonal antibodies are mainly expressed
in chinese hamster ovary (CHO) cells (Chu and Robinson, 2001; Li et al, 2010;
Reichert, 2012), being other cell lines also employed, such as, hybridoma cells,
baby hamster kidney (BHK), human embryo kidney (HEK-293) and human
retinal cells (Wurm, 2004).
Knowledge on intracellular fluxes and on their distribution among the
network reactions is of critical importance in the process of investigating and
understanding cell metabolism. However, the experimental determination
of metabolic fluxes in mammalian cells is a very complex task due to the
high number of reactions involved, their highly bifurcated structure, and
the almost undetectable concentrations of most of the intermediates. All
these problems introduce the need for a tool to determine the metabolic
fluxes in the cell based on available and measurable data, giving rise to
metabolic flux analysis (MFA) method. Measurable data are usually obtained
from extracellular measurements, such as, cell density, substrate and product
concentrations.
Metabolic flux analysis has been a subject of intense research for three
decades and has been widely applied to investigate metabolic steady-state of
cells. However, it is understandable that cells change constantly to adapt to
the environment over a culture, and so, dynamics should be consider to better
understand cellular metabolism. In recent years, dynamic metabolic flux
analysis (DMFA) has been developed in order to evaluate the dynamic evolution
of the metabolic fluxes. Most of the proposed approaches are dedicated to
exactly determined or overdetermined systems. When an underdetermined
system is considered, the literature suggests the use of dynamic flux balance
analysis (DFBA). The main challenge of this approach is to determine an ap2
propriate objective function, which remains valid over the whole culture. In
this thesis, we propose an alternative dynamic metabolic flux analysis based
on convex analysis, DMFCA, which allows the determination of bounded intervals
for the fluxes using the available knowledge of the metabolic network
and information provided by the time evolution of extracellular component
concentrations. Smoothing splines and mass balance dierential equations
are used to estimate the time evolution of the uptake and excretion rates from
this experimental data. The main advantage of the proposed procedure is that
it does not require additional constraints or objective functions, and provides
relatively narrow intervals for the intracellular metabolic fluxes. DMFCA
is applied to experimental data from hybridoma HB58 cell perfusion cultures,
in order to investigate the influence of the operating mode (batch and
perfusion) on the metabolic flux distribution.
The metabolic network models, while very useful for explanatory and
descriptive purposes of particular cell’s behavior aspects, have also an important
structural complexity to be handled for the development of optimization
and control purposes. To deal with the complexity, the macroscopic
modeling approach (Bastin and Dochain, 1991) is often used due to the simpler
structure as they directly connect extracellular substrates and products
through the definition of a set of chemical ”macroreactions” representing
the cell metabolism as a whole and not taking into account the intracellular
metabolism.
The second main objective of this thesis is to derive dynamic models of
cell cultures based upon the concept of elementary flux modes to translate
the metabolic network into macroscopic bioreactions linking extracellular
substrates to products. The main steps in the development of such models
are (1) the determination of the minimum number of macroreactions (2) the
selection of adequate elementary flux modes, among the large number of
candidates, to represent these macroreactions and (3) the identification of
kinetic laws to fully explain the observed data. In this study, these questions
are addressed based on both data- and knowledge-driven approaches. The
methodology is illustrated with an experimental case-study, including four
dierent hybridoma cell perfusion cultures. The resulting model reproduces
well the experimental data, and could be used for process monitoring or
control.