dc.description.abstract | "Metabolomics" can be defined as the comprehensive analysis of metabolites in cells, tissue
and organism, at specific times and conditions. This technology is frequently used for disease
biomarker discovery, pharmacokinetic studies, and metabolic disorder studies within human
biology. Mammalian cell line metabolomics is a strongly emerging field in human biology
that is used to study cell metabolism under altered metabolism as a result of gene deletion
or over expression of certain genes, and to study cell responses to various drug treatments.
The ultimate requirement in metabolomics, regardless of domain, is providing quantitative
data of the metabolite pool. This can only be obtained by comprehensive quantitative
analysis of targeted metabolite classes. The overall aim of this PhD project has been to study
the effect of combinatorial treatment with cisplatin and ATX-101 (a new, potential anticancer
drug) on two cancer cell lines, DU145 (prostate) and Jjn3 (multiple myeloma), as well
as a non-cancer cell line, Hek293 (embryonic kidney) at the metabolite pool level, with an
emphasis on the quantitative analysis of the metabolites within the central carbon
metabolism. As the methodology to perform a metabolomics study of mammalian cell lines
was not in place at the initiation of the PhD project, the major focus has been to develop a
human cell line mass spectrometry (MS)-based metabolomics workflow. This has resulted in
the development of three comprehensive quantitative targeted MS-based methods – gas
chromatography (GC) and capillary ion chromatography (capIC) coupled to tandem mass
spectrometry (MS/MS) – for the analysis of more than 100 primary metabolites in glycolysis,
the pentose phosphate pathway, tricarboxylic acid (TCA) cycle, the amino acid biosynthesis
pathways, and the nucleotide pool. In addition, strategies were developed to reduce
analytical drift in MS analysis which involves derivatization. Furthermore, mammalian cell
sampling protocols that yield a high energy charge and which are compatible with capICMS/
MS separation have been developed; thus, this work is expected to make a significant
contribution to the quantitative targeted metabolomics field.
Metabolome data can be considered as high-variability data as variation is being
introduced during varying growth conditions, aberrations in sampling, and analytical drift.
Particular emphasis has therefore been placed on developing strategies to reduce variation
at certain stages of the metabolomics workflow.
First, two different GC-MS/MS methods were developed. Both employed a standard
mixture derivatized with isotope coded derivatization (ICD) reagents spiked into real
samples, allowing absolute quantification. The targeted quantitative methods aimed at
analyzing sugars, amino and non-amino organic acids. The first method employed methyl
chloroformate (MCF) and methanol (MeOH) as derivatization reagents, aiming strictly at
amino acids and non-amino organic acids. The second GC-MS/MS method utilized the
trimethylsilylating reagent N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA), which
derivatizes sugars in addition to amino acids and non-amino organic acids. As both
derivatization methods also included the use of ICD reagents, d3MCF/d4MeOH and d9MSTFA,
the analytical precision of the methods was consequently improved. The ICD strategy can
therefore be a complementary approach for normalization to the more well-established
isotope dilution strategy. The use of MS/MS also significantly shortened the analysis time for
the MCF method compared to the original MCF EI GC-MS scan method.
Second, a capIC-MS/MS method was developed for the analysis of primary
metabolites belonging to the phosphometabolome (i.e., nucleotides, sugar phosphates and
phospho-carboxylic acids) and a few organic acids. The precision proved to be excellent for a
capillary flow system, and for most metabolites, the precision was high even at low
concentrations. The novel capIC-MS/MS method can possibly serve as a successor to other
more frequently used LC-MS methods (e.g., ion-pair reverse phase LC-MS and hilic LC-MS)
for these specific metabolites, as capIC-MS/MS proved to be a very sensitive, robust and
reproducible separation technique.
Third, an excessive amount of salt in samples is known to result in ion suppression,
which can affect the separation in analyses that strongly depend on the ionic strength of the
mobile phase. The latter was an issue for the capIC system, which retains inorganic salt on
the column. Therefore, it was necessary to develop capIC-compatible sampling methods
without causing metabolite leakage during sampling. As few sampling methods exist for
mammalian cell lines, and because there is a lack of consensus in sampling strategies; this
work also focused on developing methods using optimal quenching methods, salt removal
techniques and using the optimal extraction solvents for both adherent and suspension
grown cells. The suspension cells were more thoroughly tested because they are more
susceptible to leakage, because they are subjected to a vacuum pressure force before complete quenching. The mammalian cell culture sampling method developed during this
PhD study uses fast filtration under controlled vacuum pressure, saline for washing, a rapid
de-ionized (DI) water rinse, and 50% acetonitrile (ACN) in DI-water for complete extraction
of cells' phosphometabolome. Chromatographic performance was dramatically improved,
and it was verified that a quick DI-water rinse was tolerated by the cells, in addition to the
fast filtration. Also, the number of cells used per filter had a larger impact on the sampled
metabolome than the volume of DI-water used for rinsing. The maximum number of cells for
the given filter was determined to be 2 million cells, and the energy charge (EC; a measure of
the cells' energy state and cell integrity) obtained was determined to be 0.94.
Fourth, the cellular stress response was studied at the metabolite pool level on two
human cancer cell lines DU145 (prostate) and Jjn3 (multiple myeloma), and the noncancerous
cell line Hek293 (embryonic kidney) exposed to cisplatin and two APIM-peptides
(the functionally active ATX-101 and non-functional ATX-A). Data from these cell lines were
generated using the MCF GC-MS/MS and capIC-MS/MS methods, as well as a non-targeted
negative electrospray ionization (ESI) LC quadrupole time-of-flight (Q-TOF) MS analysis. The
main focus of this particular study was to monitor changes in the central carbon metabolism,
caused by ATX-peptide stress, using the target approach (MCF GC-MS/MS and capIC-MS/MS
methods, respectively). Interestingly, only minor changes were observed with the targeted
methods in the central carbon metabolism metabolite pool of DU145 and Hek293. Further,
the non-targeted method utilized both reverse phase (RP)-LC and hilic/amide LC for
separation. The non-targeted approach was used as a rough estimator of variation between
and in-between biological replicas, and to check whether there was clustering resulting from
the stress agent. The non-targeted data show a time-dependent clustering for Hek293 (i.e.,
the metabolism was not affected by APIM-peptide treatment), and that APIM-peptide had
an effect on the metabolism in DU145 cells; thus, suggesting alteration in other metabolic
pathways in DU145 (e.g., the fatty acid metabolism). The Jjn3 data produced suspiciously
low and varying metabolite abundances, determined to be caused by a non-optimal
sampling protocol. The Jjn3 general growth data, however, shows that the glucose uptake is
affected by APIM-peptide treatment. The generated Jjn3 MS data from the case study,
together with capIC separation issues related to excess inorganic salt in the biological
extracts, initiated the development of the suspension cell sampling protocol.
This PhD study contributes to the field of metabolomics by improving the GC-MS
methodology for the analysis of amino acids, non-amino organic acids, and sugars by
increasing speed, sensitivity and precision. A novel capIC-MS/MS method was established for
the robust and sensitive analysis of the analytically challenging phosphometabolome, and
sampling protocols for mammalian cells that are tailored for capIC-MS/MS analysis were
developed. Altogether, these methods represent a workflow, which was used to evaluate
potential metabolism alterations in three different mammalian cell lines stressed with a
potential drug lead. | nb_NO |