Current Status and Future Prospects of Genome-Scale Metabolic Modeling to Optimize the Use of Mesenchymal Stem Cells in Regenerative Medicine

Abstract

Mesenchymal stem cells are a promising source for externally grown tissue replacements and patient-specific immunomodulatory treatments. This promise has not yet been fulfilled in part due to production scaling issues and the need to maintain the correct phenotype after re-implantation. One aspect of extracorporeal growth that may be manipulated to optimize cell growth and differentiation is metabolism. The metabolism of MSCs changes during and in response to differentiation and immunomodulatory changes. MSC metabolism may be linked to functional differences but how this occurs and influences MSC function remains unclear. Understanding how MSC metabolism relates to cell function is however important as metabolite availability and environmental circumstances in the body may affect the success of implantation. Genome-scale constraint based metabolic modeling can be used as a tool to fill gaps in knowledge of MSC metabolism, acting as a framework to integrate and understand various data types (e.g., genomic, transcriptomic and metabolomic). These approaches have long been used to optimize the growth and productivity of bacterial production systems and are being increasingly used to provide insights into human health research. Production of tissue for implantation using MSCs requires both optimized production of cell mass and the understanding of the patient and phenotype specific metabolic situation. This review considers the current knowledge of MSC metabolism and how it may be optimized along with the current and future uses of genome scale constraint based metabolic modeling to further this aim.

Keywords: MSCs; metabolic modeling; metabolism; metabolomics; personalized/precision medicine; tissue engineering.

FIGURE 1

Tri-lineage encompasses differentiation of MSCs. Mesenchymal stem cells are identified by their ability to differentiate into chondrocytes, adipocytes, and osteoblasts that in turn develop into cartilage, fat tissue and bone. PPARγ is the master regulator of adipogenesis, Runx2 for osteogenesis and Sox9 for chondrogenesis. Various expression markers are used as indicators of successful differentiation.

FIGURE 2

The inverse relationship of the main metabolic pathways in OD and AD. Differentiation toward one lineage can inhibit differentiation toward the other.

FIGURE 3

The immunoregulatory effects of hMSCs on immune cells.

FIGURE 4

The immunoregulatory secretome of hMSCs. Known soluble factors secreted by MSCs upon activation by T cells and other antigen presenting cells are shown.

FIGURE 5

Examples of therapeutic applications for hMSCs, including organs where cell transplants have been used for engraftment, some autoimmune disorders where the immunoregulatory and differentiation abilities of hMSCs have been utilized, and some of the paracrine factor-mediated influences that can be used in a therapeutic setting.

FIGURE 6

Possible ways to achieve metabolic manipulation of hMSCs. The resulting effects may vary depending on the original cell source.

FIGURE 7

The process of reconstructing a genome-scale metabolic model (GEM). Using a base human metabolic reconstruction along with experimental data and biochemical databases (e.g., KEGG or BiGG), a constraint-based model can be created. From this model, a GEM is derived, from which cellular functions can be studied. Newly acquired knowledge from the GEM can be used to create cell experiments to further validate or improve the model. The validated GEMs can then be used to further study cells for use in regenerative medicine.

FIGURE 8

Going from a large base model to a cell specific GEM. (A) Genome scale metabolic models build on the gene-protein reaction association and represent a set of reactions as a matrix with linked genes. Synthesis of metabolites is designated by a positive number and breakdown with a negative number. (B) Going from a large base model to a cell/condition specific model. (1) A relevant base model is chosen. This model is a summary of known metabolic reactions and forms a species specific metabolic reconstruction. This base model usually has the highest count of genes, reactions and metabolites. (2) A process aiming at reducing the size of the model starts. By considering biochemical and biophysical constraints, e.g., thermodynamic feasibility, some reactions stop being reversable and the reconstruction becomes a model. Other constraints applied at this stage relate to stoiciometricity and enzyme capacity. The number of genes, reactions and metabolites is reduced. (3) Data about transcripts and proteins present in a cell type and the availability of nutrients in medium lead to the removal of yet more reactions. A cell specific model is created. (4) Metabolomic data specific to the condition the cell is supposed to represent allows the magnitude of reactions to be predicted. Some will have higher flux rates than others. More reactions may be removed at this stage. The model now becomes condition specific.