Background

The distribution of carbon fluxes determines the metabolic functionality of cells. Our project aims at the analysis and the mathematical description of mammalian cell culture metabolism. A thorough understanding of the intracellular reaction network and its regulation is the basis for rational metabolic engineering strategies for process optimization.

Cells, especially eucaryotes are highly complex systems. In contrast to the traditional reductionist approach, systems biology tries to investigate the interaction of the system components. This approach perceives, that even if we know all single parts of the cell, we will be unable to describe its phenotype without knowing their interactions and the dynamics of its parts. However, already the simple procaryotes, like e.g. E. coli, have about 4.400 genes. That accounts for at least 4.400 transcripts and 4.400 proteins, interacting with about 1.000 metabolites. The huge number of possible interactions will not be manageable without a formalism. The only way to quantitatively handle this complexity is by the use of mathematical formalism, i.e. building models.

Approach

Systems biology only works with a close connection between experimental observations and mathematical modelling (Fig. 1). The modelling process begins with hypothesis on the cellular interactions, i.e. its metabolic network, enzyme kinetics, inhibitions, transcriptional regulation, etc. These hypothesis form the inner part of the model and mostly include parameters that cannot be observed experimentally. In Fig. 1 this part is reflected in the function f. This model has now to be connected with the observable surrounding, i.e. the available measurements y and possible experimental input variables a^inp.

Fig. 1. The model building process is based on hypotheses (forming the function f), the available measurement (y) and possible experimental conditions (a^inp).

With this model in hand hypotheses of cellular metabolism can be testet in silico and then be verified or falsified by an experiment (Fig. 2).

Fig. 2. All in silico predictions have to be verified / falsified by the experiment.

The determination of metabolic fluxes is a challenge – Cells dispose a large variety of enzymes forming different pathways. Some of these pathways are bidirectional, e.g glycolysis and gluconegenesis. From only external flux measurements it will not be possible to determine a possible simultaneous flux. The same holds true for parallel routes that have a common substrate and product, e.g. glycolysis and the Entner-Doudoroff pathway (both from glucose to the end product pyruvate). To distinguish between these pathways, 13C tracer experiments will be performed. Using a specifically labeled substrate and tracing the labeling through the intermediates enables the quantification of metabolic fluxes. These experiments have to be perfomed at a metabolic stationary state, i.e. the intracellular concentrations and fluxes do not change over time. Then, after some time, the labeling patterns will also reach a stationary state, that is specific for a certain flux distribution.

The labeling will be measured using GC- and LC-MS(/MS) methods. The inverse problem, finding a flux distribution that explains the measured labelings is perfomed using mathematical modelling (i.e. isotopomer balancing) and parameter optimization. The sofware 13C FLUX is used.

Essential when working with biological systems is the question of errors. All available measurements, the extracellular fluxes (determined from concentration measurements) as well as the measured labeling is error prone data. Consequently the estimated flux distribution will also contain errors. To see whether a flux can be identified from the available measurements, all calculations are accompanied by error propagation. The 13C FLUX tool therefor calculates all sensitivities. This can also be exploited for experimental design, i.e. determine the most information containing labeling pattern for the input substrates. Fig. 3 gives an overview of the experimental and mathematical interactions and calculations.

Fig. 3. Having a model in hand, various calculations can be performed – first of all parameter estimation (flux calculation), error propagation and experimental design.

System

Mammalian cells (MDCK and susceptible human cell lines) and their interactions with influenza virus are used as model systems. Such systems offer a multitude of advantages and challenging aspects:

1. Influenza is a serious cause of morbidity and mortality. While much is known about the molecular biology of influenza, and its genetic shift and drift responsible for epidemics, far less is known about the interaction of the virus and the host cells.

2. MDCK cells are ideal hosts for influenza and yield high quantities of the virus. The cell line is adherent and large-scale culture systems were established for culturing. Such systems include fixed bed reactor, fluidized bed reactor and microcarrier culture. This way, defined experimental conditions for large-scale and reproducible analysis are ensured, a prerequisite for systems biology.

3. In 1995, WHO proposed that the possibility of establishing an acceptable mammalian cell line should be investigated and encouraged both the development of optimum cell growth conditions and the examination of the specific requirements for influenza virus culture in these cells.

Experimental Techniques

The experimental aspect of the project focuses on metabolite, enyme activity and flux analysis. Metabolite analysis will be carried out applying GC-MS(/MS) as well as HPLC. Various groups work on the quantification of intracellular compounds. The database of the Max Planck Institute in Golm contains the retention times and GC-MS spectra of 360 identified metabolites.

Flux analysis will be carried out using 13C-labelled substrates and analyzing the mass isotopomer distributions in central carbon metabolites and free amino acids. Enzyme activities will be measured by enzyme assays in cell extracts, protein concentration changes will be determined with 2D gels.

Mathematical Modelling

In metabolite and flux analysis, between 200 and 1000 mass isotopomers are measured in each sample. Another 10 measurements will result from enzyme activities, from protein comparison another 100-200 measurements will be available. To cope with that amount and include it into a comprehensive analysis, a multi-scale model including mechanistic enzyme kinetics, protein concentration (activity) will be used.

Related Literature

Wiechert, W. C-13 metabolic flux analysis Metab Eng, 2001, 3, 195-206

Wiechert, W.; Möllney, M.; Petersen, S. & de Graaf, A.A. A universal framework for C-13 metabolic flux analysis Metab Eng, 2001, 3, 265-283

Sheikh, K.; Forster, J. & Nielsen, L.K. Modeling hybridoma cell metabolism using a generic genome-scale metabolic model of Mus musculus Biotechnol Prog, 2005, 21, 112-121

Nöh, K.; Wahl, A. & Wiechert, W. Computational tools for isotopically instationary (13)C labeling experiments under metabolic steady state conditions. Metab Eng, 2006, 8, 554-577

Kopka, J.; Schauer, N.; Krueger, S.; Birkemeyer, C.; Usadel, B.; Bergmüller, E.; Dörmann, P.; Weckwerth, W.; Gibon, Y.; Stitt, M.; Willmitzer, L.; Fernie, A.R. & Steinhauser, D. GMD@CSB.DB: the Golm Metabolome Database. Bioinformatics, 2005, 21, 1635-1638

Erban, A.; Schauer, N.; Fernie, A.R. & Kopka, J. Nonsupervised Construction and Application of Mass Spectral and Retention Time Index Libraries From Time-of-Flight Gas Chromatography-Mass Spectrometry Metabolite Profiles. Methods Mol Biol, 2006, 358, 19-38

Interconnections with other projects of Bioprocess Engineering

Hierarchical structures:

Influenza Virus Replication in MDCK cells


Coupled Processes:

Influenza Vaccine Production in Microcarrier Systems
Quantitative Analysis of Energy Metabolism of Animal Cells





back to overview on bioprocess engineering