of these stresses in the physiological response of the organism, although there was a
marked increase in the expression of certain genes, upon genomic analysis [62]. In a
recent report, E. coli cells exposed to CO 2 levels above 70 mbar CO 2 partial pressure
in the inlet gas led to reduced biomass yields and rapid accumulation of acetate, even
under non-overflow and fully aerobic conditions [63].
Interaction Between Multiple Gradients Finally, the results of Limberg and
colleagues show that when pH gradients are coupled to oxygen limitation,
C. glutamicum loses its robustness against dissolved oxygen fluctuations [64],
leading to yield losses of up to 40%. This implies that the study of concentration
gradients in fermentation should be conducted in a multi-faceted manner, to consider
all possible gradients and the necessary combinations among them to arrive at a more
holistic conclusion for each strain. There is also a close correlation between pCO 2
levels, pH, base addition, and osmolality in large-scale CHO cell cultures which
affect the metabolic lactate shift (transition from lactate production to lactate consumption) [65, 66].
4 Framework for Bioprocess Scale-Down Studies
4.1 Characterization of the Large Scale
A good characterization of the large-scale bioprocess is important to conclude proper
scale-down experiments which really imitate the large scale (see Fig. 2). Since the
scale-down data is only as good as the environment it mimicked, it is absolutely
necessary to characterize both the cellular state and specific heterogeneity (gradient
profiles) in the larger scale. Standard analytical methods of the medium and gas
composition and the derivation of cell specific rates need to be complemented by
direct monitoring of the physiological state of the cells. A proper scale-down
methodology should be based on the similarity of cellular responses, all at the
level of metabolism, protein expression, and population heterogeneity between
laboratory and industrial scale. In order to avoid false conclusions and to reduce
the risk of scale-up, robustness analyses must be used to assess the final batch-tobatch variability. This complex problem can only be solved if digital approaches
(digital twin) can be coupled with a large number of experiments.
In the past, there were a large number of approaches to simulate these gradients
occurring in the industrial bioprocess in scale-down systems, see reviews by
Neubauer and Junne [3], Lara et al. [46], Delvigne and Noorman [67]. All these
systems achieve oscillating conditions regarding the specifically investigated parameters, i.e. the specific parameters which were the focus of the investigators. Different
priorities were set depending on the specific approach. In multi-compartment reactors, the dominant parameter is the residence time distribution in different compartments where cells are located within a defined period of time. In more-compartment
stirred tank systems, the zones are characterized by a previously defined state, e.g. in
10
P. Neubauer et al.
marked increase in the expression of certain genes, upon genomic analysis [62]. In a
recent report, E. coli cells exposed to CO 2 levels above 70 mbar CO 2 partial pressure
in the inlet gas led to reduced biomass yields and rapid accumulation of acetate, even
under non-overflow and fully aerobic conditions [63].
Interaction Between Multiple Gradients Finally, the results of Limberg and
colleagues show that when pH gradients are coupled to oxygen limitation,
C. glutamicum loses its robustness against dissolved oxygen fluctuations [64],
leading to yield losses of up to 40%. This implies that the study of concentration
gradients in fermentation should be conducted in a multi-faceted manner, to consider
all possible gradients and the necessary combinations among them to arrive at a more
holistic conclusion for each strain. There is also a close correlation between pCO 2
levels, pH, base addition, and osmolality in large-scale CHO cell cultures which
affect the metabolic lactate shift (transition from lactate production to lactate consumption) [65, 66].
4 Framework for Bioprocess Scale-Down Studies
4.1 Characterization of the Large Scale
A good characterization of the large-scale bioprocess is important to conclude proper
scale-down experiments which really imitate the large scale (see Fig. 2). Since the
scale-down data is only as good as the environment it mimicked, it is absolutely
necessary to characterize both the cellular state and specific heterogeneity (gradient
profiles) in the larger scale. Standard analytical methods of the medium and gas
composition and the derivation of cell specific rates need to be complemented by
direct monitoring of the physiological state of the cells. A proper scale-down
methodology should be based on the similarity of cellular responses, all at the
level of metabolism, protein expression, and population heterogeneity between
laboratory and industrial scale. In order to avoid false conclusions and to reduce
the risk of scale-up, robustness analyses must be used to assess the final batch-tobatch variability. This complex problem can only be solved if digital approaches
(digital twin) can be coupled with a large number of experiments.
In the past, there were a large number of approaches to simulate these gradients
occurring in the industrial bioprocess in scale-down systems, see reviews by
Neubauer and Junne [3], Lara et al. [46], Delvigne and Noorman [67]. All these
systems achieve oscillating conditions regarding the specifically investigated parameters, i.e. the specific parameters which were the focus of the investigators. Different
priorities were set depending on the specific approach. In multi-compartment reactors, the dominant parameter is the residence time distribution in different compartments where cells are located within a defined period of time. In more-compartment
stirred tank systems, the zones are characterized by a previously defined state, e.g. in
10
P. Neubauer et al.