br Material and methods br Results

Material and methods
Results
Discussion Aging-associated phenotype change in hBMSCs is one of the most important considerations in furthering the advancement of cell-based therapy and is a subject of numerous studies and reviews (Baker et al., 2015; Bara et al., 2014). In regard to the in vivo aging, there is a general agreement between most studies in that the cell fitness declines with aging. For example, with age of the bone marrow the frequency of CFUs declines, susceptibility to senescence increase, and npy have reduced migration and adhesion capacity (Baker et al., 2015). The clinical performance of BMSCs from aged individuals has also been shown in most instances to be less than cells from younger counterparts (Baker et al., 2015). In vitro aging is also very critical in influencing therapy because manufacturing of hBMSCs in a quantity suitable for clinical use requires extensive in vitro expansion of cells. In vitro culturing exposes cells to numerous artificial stresses. For example, harvesting cells using enzymatic treatment (usually trypsin, collagenase, accutase, or mixture of them) exposes cells to a condition where the surface proteins could be partially cleaved and this may initiate a cascade of biological events (Huang et al., 2010). The effect of in vitro passaging on hBMSC\'s phenotype has been examined in numerous occasions (Bruder et al., 1997; Lo Surdo and Bauer, 2012; Lo Surdo et al., 2013; Wagner et al., 2010). We also observed, among others, that metabolic pathways are largely altered, where many proteins in the glycolytic, pentose phosphate, and TCA pathways were shown to be largely upregulated in late passages. An expression trend similar to ours for some of the proteins has already been reported by other authors. Using 2D gel techniques, Sun and coworkers showed the expression levels of pyruvate kinase and enolase increase with serial sub-culturing of hBMSCs (Sun et al., 2006). In a separate report, Madeira et al. used a comparable technique to report similar induction of GAPDH, aldolase, and aldo-keto reductase with in vitro aging of cells (Madeira et al., 2012). Proteins in the metabolic networks are traditionally considered to have housekeeping functions, although this notion has recently evolved as they are recognized to play vital roles in proliferation, self-renewal, differentiation, and aging (Shyh-Chang et al., 2013). For example, hiPSCs experience a high glycolytic flux during oxidative stress. During differentiation of hiPSCs, genes in the pentose phosphate and lipid biosynthesis pathways become upregulated (Zhang et al., 2012). Further, one of the changes during reprograming of somatic cells is the progressive increase in glycolytic genes. In fact, the metabolic network is claimed to be the major barrier during reprograming (Hansson et al., 2012; Zhang et al., 2012). The metabolic network has also been shown to play important roles in hBMSC biology. Glycolysis is the major source of energy during chondrogenic differentiation and hypoxia suppresses osteogenic differentiation of hBMSCs (Pattappa et al., 2011). In regards to self-renewal and expansion, some reports have already indicated that the underlying metabolic pathways have a significant effect on population doublings, frequency of primitive cells in a colony, colony size, and number of colonies in culture media (Estrada et al., 2012; Pattappa et al., 2011; Sanchez-Arago et al., 2013). BMSCs retain the flexibility to use both oxidative phosphorylation and glycolytic pathways (Pattappa et al., 2011; Schop et al., 2009). Under normoxic conditions, non-differentiating BMSCs depend more on glycolysis (~70%) and part of the reason may be that cells require not just ATP, but also other metabolic byproducts. Besides, glycolysis protects cells from excessive generation of ROS, the primary cause of cellular damage and senescence (Pattappa et al., 2011; Shyh-Chang et al., 2013). The fact that enzyme content in the glycolytic network increased with passage seems to indicate that glycolytic pathway may be preferred even more so as hBMSCs are subjected to stress caused by long-term passaging. Interestingly, gene expression assays in BMSCs also pointed to similar bioenergetic alteration trends with passaging (Will, 2010). Therefore, as we seek to gain clearer understanding of the relationship between metabolic networks and some key features of hBMSCs; the differentially regulated metabolic enzymes identified in this study may be key starting targets.