We here show that metabolomics can discern the effects of genetic differences of cell lines by highlighting metabolic similarities and differences of pluripotent cell types compared to somatic cells. We employed multiple mass spectrometry-based techniques to overcome analytical challenges, such as efficient throughput and metabolite coverage as all analytical methods possess coverage limitations 
. Direct infusion nanoESI-mass spectrometry enabled a rapid analysis of different lipid structures and confirmed that the pluripotent cells are strikingly similar in terms of lipid metabolite signatures relative to fibroblasts indicating near complete metabolic reprogramming of iPSCs. However, it also revealed some potentially important differences between iPSCs and the parental fibroblasts and more subtle distinctions between mESCs and iPSCs.
While the nanoESI-MS method is an effective preliminary approach to observe variation between sample groups due to speed and MS/MS coverage, it retains some limitations. The lack of chromatographic separation and low mass spectrometric resolution causes isobaric and isomeric structures to form single peaks in the mass spectra, complicating correlation of ion intensities to a single structure. Direct infusion analysis is also particularly susceptible to ion suppression, potentially compromising appropriate reflection of compound concentration 
. Application of (lower throughput) GC-TOF MS and LC-QTOF MS chromatography-based analytical methods minimized the potential impact of these limitations and expanded metabolite coverage beyond the capabilities of either individual system. The GC-TOF MS method provided the greatest number of identified compounds, which is a reflection of a highly developed infrastructure designed specifically for the applied system and method 
. While MS/MS libraries are available for use with the LC-QTOF MS system, identification was greatly hampered by both library coverage and acquisition of MS/MS spectra. The LC-QTOF MS uses a quadrupole to isolate a particular ion for generation of MS/MS data, and if a particular ion does not achieve the relative intensity necessary to trigger isolation in a data-dependent MS/MS acquisition mode, the MS/MS data necessary for annotation will not be available. Subsequently, while the current analysis covers many structures, there are still many other metabolites that were not annotated and may provide additional insight into the metabolic state of iPSC, mESC, and m15 fibroblast cell types.
Several reports manifest large metabolic differences between fibroblast or other somatic cells and pluripotent cells, for example with respect to energy metabolism 
. Metabolomic research groups have stated a range of other significant differences after nuclear reprogramming, most notably on the level of nucleotides and purines 
and membrane lipids 
, but neither group identified metabolic differences in lactate or glucose metabolism. In accordance with these reports, we have observed very large differences in metabolic phenotypes between mouse embryonic fibroblasts and embryonic stem cells ( and Figure S1
) which mostly conferred metabolic reprogramming in amino acid and lipid metabolism, but not in compounds such as taurine, reported as being different in extracellular footprints 5–7 days after nuclear reprogramming by Folmes et al. 
. We detected intracellular taurine levels with two independent analytical techniques, LC-QTOF MS and GC-TOF MS. Neither of these methods found quantitative metabolic differences for taurine or energy-related metabolism. Authors have noted the similarity between energy metabolism in stem cells and cancer cells with respect to the use of anaerobic glycolysis for generating ATP rather than by mitochondrial oxidation 
, similar to the well-known Warburg effect in cancer cells 
. It has been pointed out to be of critical importance to control for available oxygen levels in cell cultures if hypotheses about hypoxic metabolism are to be tested, for example in cancer cells 
. To our knowledge, however, none of the reports in stem cell metabolism actually monitored and controlled the level of oxygen available to cells, including results reported here.
This work detailed many metabolic similarities and differences between iPSCs, mESCs, and embryonic fibroblasts beyond those metabolites covered recently 
by encompassing additional structures including lipids, polyamines and amino acids. While parts of the metabolome coverage were overlapping, quantitative results revealed conflicting trends to this previous publication 
. For example, in the study presented here we did not find any increase in free fatty acids, unlike a seven-fold difference reported before between different pluripotent cell lines 
, and we also could not confirm broad changes in hydroxyl acid levels when comparing pluripotent cell lines to fibroblasts. While we did observe statistically significant changes in 5′-methylthioadenosine levels in iPSCs relative to ESCs, the effect was less than two-fold and in the opposite direction from the previously published analysis 
It is unlikely that conflicting results between our data and the recently published study 
are due to the analytical methodologies used in our research, because we have utilized the broadest line of technologies that were yet applied to stem cell metabolism. NanoESI-MS, GC-TOF MS, and HILIC- and RP liquid chromatography-QTOF MS based technologies facilitated expansion of metabolite profiling to encompass a chemically diverse array of cellular metabolites and enabled an in-depth investigation and comparison of iPSC, m15 fibroblast, and mESC metabolite profiles. Application of this global approach delineated some of the features which, despite their similarity, distinguish iPSCs from mESCs and identified candidate mechanisms that may be subject to more targeted, focused methods to attain a more clear understanding of potential cause and biological impact. Additional studies of metabolic phenotypes and genomic differences of iPSCs, ESCs, and embryonic fibroblasts are critical to further advances in our understanding of pluripotent cell metabolism.
Differences in the degree and range of detected metabolic changes in stem cell reprogramming may not only be founded in differences in cell culture conditions, but may also relate to the differences in cell lines investigated. For example, line-to-line variability in induced pluripotent stem cells has also been demonstrated repeatedly; for example, different ways to derive iPS-cells directly influence x-inactivation and epigenetic variability 
. In analogy to the apparently conflicting results on stem cell metabolism, proteomics analyses also showed substantial differences between each iPS cell line studied, in addition to differences between individual embryonic stem cell lines 
. Overall, all reports so far concur that metabolic reprogramming is a key feature in reprogramming cells into a pluripotent state which may involve a wide range of metabolic pathways. The extent of which pathways are mostly reprogrammed appear to be dependent on the actual culture conditions, time points and cell types being studied. Indeed, taking together these different reports, metabolomics may be a tool to determine the phenotypic proximity of different iPSC lines to embryonic stem cells, in addition to informing about metabolic reprogramming events from fibroblast cells.
An area of intense interest in the stem cell field has been defining the properties of iPSCs relative to ESCs. Just how similar are iPSCs and ESCs? Based on previous work at the cell biological, gene expression, and epigenetic levels, the two pluripotent stem cell types are remarkably similar, although not identical. We detected higher levels of phosphatidylcholines comprised of three or more double bonds in mESCs relative to m15 fibroblasts, confirming recently reported results 
. Such increase in unsaturation levels of polyunsaturated PCs may alter membrane fluidity and hence can contribute to physiological changes that are relevant for stem-cell phenotypes. However, the more detailed investigation presented here also reveals that this up-regulation of membrane lipids did not extent to phosphatidylethanolamines in either pluripotent cell type. Indeed, sometimes just the opposite regulation was found as for the 18:1/18:2 PC and 18:1/18:2 PE lipids, especially when comparing mESCs to fibroblasts. iPSCs displayed the same tendency, supporting the concept that the iPSC model reflects ESCs in terms of regulating membrane composition. Such differential changes may also be due to different activity levels of methylating enzymes, because PCs may originate from PEs by methylation steps catalysed by the enzyme phosphatidylethanolamine N-methyltransferase, a ~20 kDa transmembrane-spanning enzyme located mainly in the endoplasmic reticulum. PCs may otherwise originate de novo
from CDP-choline and diacylglycerols, but the exact origin of PCs in these cell types has not been determined yet. DNA- and Histone-methyltransferase activities are well known to be involved in stem cell developmental phenotypes 
. It is interesting that the magnitude of PE-dependent methyltransferase activity was less pronounced in iPSCs than in mESCs, suggesting the genetic reprogramming of the iPSC line was not entirely adequate to yield identical mESC lipid profiles.
Metabolic differences between the iPSC and mESC lines were also evident in primary metabolism, specifically for amino acids. Both cell types yielded lower amounts of free amino acids compared to the m15 fibroblasts, but several amino acids were present at even lower levels in iPSCs relative to mESCs. While this dissimilarity might point to differences in protein synthesis (i.e. differences in amino acid consumption), differences in amino acid transporter activities would be another potential explanation given that both essential and non-essential polar uncharged amino acids displayed the most substantial differences. Statistically significant changes in both putrescine and 5-methylthioadenosine indicate that the polyamine biosynthesis pathway is another deviation between mESCs and iPSCs. This effect is notable due to polyamine biosynthesis involvement in cell proliferation and differentiation 
The body of literature in stem cell metabolism is growing. However, at this point mechanistic interpretations have not established which parts of metabolic reprogramming are a necessary condition for pluripotency and which metabolic differences are consequence rather than causal factor in cellular de-differentiation. Detailed investigations on mechanistic aspects of stem cell metabolism, including flux studies and studies under hypoxia versus normoxic conditions, may be needed to reveal the underlying metabolic prerequisites for pluripotency.