Constructing Metabolic Models for Pluripotent Stem Cells

Objectives 

● Explain the general overview of what building metabolic models consists of 

● Describe how genome-scale metabolic models have been applied to pluripotent stem cells

● Communicate the implications and applications of metabolic modeling in stem cell research

● Identify the future direction of metabolic modeling for human embryonic stem cells (hESCs) 

Summary  

Throughout the semester in the course BIT 495 (001) Special Topics in Biotechnology: Metabolic  Modeling, different metabolic models have been created to outline the metabolic pathways induced and  metabolites utilized by different microbial communities whether that includes a single microbe or  multiple microbes that are interacting in a mixed bag community. A majority of the genome-scale  metabolic models constructed in this course have been done using the Department of Energy’s (DOE’s) KBase which is one of the newer computational programs that allows for metabolic models to be created  for a certain organism after entering a genomic sequence (could be annotated and/or downloaded from a  database such as NCBI) and then using KBase’s provided application to ‘Build a Metabolic Model (MM)’  to understand which pathway in a metabolic network is fulfilled; ‘Gapfill a MM’ with Media or certain  reactions to optimize biomass generated by an organism in case the current biomass is at 0 or around 0;  and ‘Run Flux Based Analysis (FBA)’ to see which metabolic reactions occur, are blocked, or may have  increased the flux after going through a gapfill. There are also other applications within KBase that check  for other abnormalities or provide more metabolic information about the microbial community in research  such as if there are reversible reactions, if the reactions performed by the organism(s) are followed  through and/or balanced, or if there are any auxotrophies in the predicted model.  

Since much of the focus in this course was on microbes, I was wondering if metabolic modeling  can also be applied to other types of cells including human cells initially with a focus on healthcare  applications, but now on regenerative cells. Due to the focus on regenerative cells, published scientific  articles focused on embryonic stem cells or just pluripotent stem cells that tend to stem from embryonic  cells. Prior to providing more information about metabolic analysis in pluripotent cells, there are still  ethical concerns regarding stem cell research especially due to the nature of harvesting stem cells from  embryos and also having to navigate conducting human trials. Despite this issue, certain doctors have  been able to make a breakthrough by reprogramming mouse and human fibroblasts to induce pluripotency  by introducing 4 genes that encode Yamanaka factors (Oct4, Sox2, KIf4, and c-Myc) using retroviral  vectors (Hsu et al., 2016). In general, pluripotent cells, especially since they can give rise to cell growth  and cell differentiation based on manipulated characteristics, are seen as potential for scientific research  because understanding how they are stimulated or even inhibited in growth allows for understanding of  general research pertaining to development, aging, and cancer treatment (He et al., 2018).  

Just like microbes, genomic metabolic models have been applied to stem cell research, but it is  much newer and hence have less published articles on general results, conclusions, and discussions on the  construction of genome-scale metabolic network for general human embryonic stem cells (hESCs).  Recent research has discovered through the simulation and analysis of GEM, global metabolic  characteristics such as essential metabolites used by the cells or other network motifs have been identified  (He et al., 2018). It has also been understood through this research that stem cell metabolism is an active  regulatory mechanism that is related to lineage commitment, specification, and self-renewal rather than  just being a byproduct of a mechanism, so GEMs can be constructed for these cells (He et al., 2018).  Even relating back to the course even though much of research relating to GEM construction,  reconstruction, and reprogramming of hESCs is done by other applications that is not KBase, the fact that  stem cell metabolism having an active regulatory mechanism possibly related to lineage commitments  also brings up the point of how KBase and other applications provides phenotypic analysis through  phylogenetic trees. Some general applications of generic and personalized GEMs have been used to  capture dynamic changes between the prime and native states within mouse embryonic stem cells that  have gone through rewiring, to describe the metabolic status from transcriptomics data and tissue-or cell specific models to help identify the underlying cellular mechanisms in complex diseases, and identify  selective anticancer drugs for cancer patients whose personalized GEM has been constantly reconstructed  (He et al., 2018).  

Specific studies have included ones performed by He et al., 2018 included using robust multi array average (RMA) method to normalize all gene expression, using Python script to process the microarray dataset to input in the tINIT algorithm, and using the Human Metabolic Reaction 2.0 (HMR  2.0) to perform metabolic modeling since now was the largest biochemical reaction database for human  metabolism. Just through reconstruction and numbers alone, there were six reconstructed hESC-GEMs  that accounted for 3722-5104 reactions that involved 3188-3892 metabolites associated with 2045-2404  genes. Also compared to cancer cells and tissue cell lines, hESC models were observed to contain more specific reactions, metabolites, and genes. Some general conclusions that were pulled from the modeling  the hESC cells included observed similarities in activity between ESC and cancer cells; for example,  subsystems with higher activity in both cell types can be tagged on with cell proliferation and may  include nucleotide metabolism and aminoacyl-tRNA biosynthesis which is an essential process to  maintain self-renewal of the cells.

Other pathways that were seen to have similar activity and may have  been closely related to cellular pluripotency include pentose phosphate pathway, cysteine and methionine  metabolism, and omega-6 fatty acid metabolism. Interestingly, there was a depletion in arginine and  proline for hESC pluripotency (specifically in H9) after 48 hours that was downregulated by NANOG and  OCT4. After testing germ layers that were affected and it being concluded that only the marker genes of  the ectoderm were upregulated and others left unaffected, it suggested that arginine and proline, even  though non-essential amino acids, may potentiate ectoderm differentiation. Lastly, though hESC tends to  be mainly glycolytic, more specifically anaerobic glycolysis regulated by bio-energetic signaling and fatty  acid metabolism, but the metabolic modeling showed 20 kinds of four-node motifs in the oxidative  phosphorylative subnetwork while only six were found in the glycolytic subnetwork (He et al., 2018).  

Overall this specific study stated that this was the first hESC GEM to be constructed and hence  there is room for improvement, but just being able to construct genome-scale metabolic models for  pluripotent stem cells provides us more insight to metabolic analysis in human cells without being too  invasive beforehand and also apply it to other proliferating cells such as benign through malignant cancer  cells and their metabolism which can help with better understanding of personalized drug and/or cell  therapy against cancer. There are even current studies that are trying to apply E-flux into studying human  metabolism or even construction models through HapMap and NCI-60 to have a better understanding of  high expression levels in cells (Yizhak et al., 2014). Building cell-specific models can be incorporated  using PRIME (Personalized Reconstruction of Metabolic Models) by utilizing both molecular and  phenotypic data for tailoring to cell-specific GSMMs (Yizhak et al., 2014). Through this method, over  280 GSMMs of cancer and normal proliferating cells which are then evaluated for their ability to predict  any metabolic phenotypes such as proliferation rate, drug response, and biomarkers on individual levels  (Yizhak et al., 2014). There are articles focused on metabolic modeling for cancer cells such as Yizhak et  al., 2014 and Nilsson et al., 2017 that go beyond the scope of this summary, but through metabolic  modeling we can have a better understanding of metabolic activity in pluripotent human cells in vitro. 

Questions 

1. What are the current applications of metabolic modeling of pluripotent stem cells?

Currently GEnome-scale Metabolic modeling (GEM) is being constructed for pluripotent stem  cells, especially human embryonic stem cells to predict and understand the metabolic pathways  that are induced, and which metabolites are utilized during cell proliferation and even self renewal. For example, the study conducted by He et al. highlighted how at least 20 different  motifs were visualized in the oxidative phosphorylative subnetwork while only 6 were found in  that of the glycolytic subnetwork even though it has always been understood that hESC tend to  mainly follow glycolytic metabolic pathway. Also, in addition to applications such as KBase that  can be used for metabolic modeling, other platforms can be used such as the Human Metabolic  Reaction 2.0, the tINIT algorithm, and Personalized Reconstruction of Metabolic Models.  

2. What are the current ethical concerns surrounding human stem cell research and how have  scientists gone about combating these ethical issues if available? 

Current ethical concerns surround using human embryonic stem cells for research since first of,  research would require human trials which is already hard to navigate and second of, harvesting  stem cells can also cause a bit of a problem due to certain morals against manipulating cells  pertaining to the embryo. Currently certain scientists have looked into other possibilities of reprogramming mouse or human fibroblasts to induce pluripotency and in turn, be studied. It is  also helpful that metabolic modeling to allow for pathway predictions before carrying outgrowth experiments that will manipulate cell metabolism in vitro because it will allow for more  assurance.  

3. What are the current implications of metabolic modeling for pluripotent stem cell research?

Successful or even working progress observed from constructing, programming, and  reconstructing GEMs for pluripotent stem cells allow for scientists to expand to applying  metabolic modeling in cancer cell research since both types of cell lines share certain similarities.  Additionally, being able to compare cancer and normal proliferating cells allows for testing cells  for their metabolic phenotypes such as proliferation rate, drug response, and biomarkers.  Knowing this information will allow scientists to have a better understanding of cell metabolism  and know what metabolic activity to assess in different cancer patients and possibly have a better  understanding of drug therapy targets in these patients since people can respond differently to the  same treatment.  

4. Why may it be important to construct, program, and reprogram a genome-scale metabolic  network for pluripotent stem cells? 

Again, it is important to perform GEMs and constantly update them for pluripotent stem cells  because it is a growing field and the more research done, the more data present in metabolic  modeling platforms to perform these analyses on several types of pluripotent stem cells before  conducting in vitro experiments that manipulate human cells or incorporate general human patient  trials. Additionally, general research will allow for one to understand general cell potential for  pluripotency and self-renewal that is associated with specific metabolic activity. In long-term  research, it will allow for potential research pertaining to ethical development, aging, and cancer  treatment. 

References 

He, Yangzhige, et al. “Revealing the Metabolic Characteristics of Human Embryonic Stem Cells by  Genome‐Scale Metabolic Modeling.” FEBS Letters, vol. 592, no. 22, 2018, pp. 3670–3682.,  https://doi.org/10.1002/1873-3468.13255.  

Hsu, Yi-Chao, et al. “Mitochondrial Resetting and Metabolic Reprogramming in Induced Pluripotent  Stem Cells and Mitochondrial Disease Modeling.” Biochimica Et Biophysica Acta (BBA) - General Subjects, vol. 1860, no. 4, 2016, pp. 686–693.,  

https://doi.org/10.1016/j.bbagen.2016.01.009.  

Lengauer, T., and C. Hartmann. “Bioinformatics.” Comprehensive Medicinal Chemistry II, 2007, pp.  315–347., https://doi.org/10.1016/b0-08-045044-x/00088-2.  

Nilsson, Avlant, and Jens Nielsen. “Genome Scale Metabolic Modeling of Cancer.” Metabolic  Engineering, vol. 43, 2017, pp. 103–112., https://doi.org/10.1016/j.ymben.2016.10.022.  Yizhak, Keren, et al. “Phenotype-Based Cell-Specific Metabolic Modeling Reveals Metabolic Liabilities  of Cancer.” ELife, vol. 3, 2014, https://doi.org/10.7554/elife.03641.