An approach that has shown promise for answering these

An approach that has shown promise for answering these types of questions in other stem cell systems, and is becoming increasingly feasible with advanced computing power and improved tracking algorithms, is long-term in vitro imaging of individual stem and progenitor cells (Schroeder, 2011). Long-term imaging can enable the construction of cell lineage maps, tracking the progression from parent cell to daughter cells and revealing the emergence of properties of interest. For example, long-term imaging has provided data to detail the complex commitment hierarchy of embryonic (Ravin et al., 2008) and adult (Costa et al., 2011) neural stem cells, successfully captured instances of hemogenic endothelial cells giving rise to blood cells (Eilken et al., 2009), and allowed the comparison of rotenone time, migration speed, and growth kinetics between hematopoietic stem cell siblings (Scherf et al., 2012). Here, we utilize long-term in vitro imaging to address some of the outstanding questions regarding cell-size heterogeneity in commercially purified, culture-expanded BMSCs.
Materials and methods
Results
Discussion Through timelapse imaging and analysis of BMSC division and morphology, we observed both known and unanticipated changes in cell size over time. It is well established that cell size increases as a eukaryotic cell replicates genetic material while traversing through the normal cell replication cycle (Cooper, 2000), and that this cyclic increase in cell volume also occurs for putative BMSCs in vitro (Lee et al., 2011). However, we observed directly that an appreciable fraction of BMSCs did not replicate over ~6days in vitro (55.2%), and that most of the resulting progeny could be traced back to only 9% of the cells attached at day 1. The heterogeneity in both cell proliferation and lineage tree types (Fig. 3B) is not wholly unexpected, but is nonetheless striking. Note that one consequence of this low fraction of faster-replicators within a larger population of slower- (or non-) replicators means that a reported population doubling time poorly represents the number of times that individual, replicating BMSCs have divided. This distinction may be consequential for future studies of single cell profiling and therapeutic applications that require culture expansion of BMSCs and identification of bone marrow-derived mesenchymal stem cells with ostensibly low population doublings. Further, through these growth rate-tagged lineages, we concluded that smaller cells within the BMSC population were not a distinct subpopulation that existed at day 1, proliferating with a distinct doubling time and maintaining a distinct and smaller cell size. Likewise, the larger cells within that population were not a distinct and persistent lineage. Instead, we found that smaller cells became larger cells as those cells exited the cell cycle. In fact, larger cells not only grew and failed to divide within the observed 2days typical of 93% of the actively dividing population, but larger cells also showed a biochemical marker of senescence. We observed no indication over this limited timescale, or over the five passages summarized in Figs. 1 and 6B, that the smaller cells replicated to such an extent that this subpopulation remained small but reached a Hayflick limit (Hayflick and Moorhead, 1961). We note that the larger cells were those that had divided the least, but that we cannot report the full history of the larger cells that were seeded on day 1. We further note, as highlighted in our introduction, that initial heterogeneity of cell size and/or behavior within an in vitro culture of BMSCs can be attributed to several factors not considered explicitly here. These include ex vivo reprogramming (Zipori, 2010), genetic mutations (Wagner et al., 2010), differences among cells in commitment maturity (Ratajczak et al., 2008) toward a particular differentiation endpoint that can manifest as a wide distribution of metabolic expression profiles (McMurray et al., 2011), differences in cell cycle stages (Lee et al., 2011), and variation in culture conditions among many cells in a single tissue-culture dish (Bruder et al., 1997). Thus, understanding of how and when morphologically distinct cells within the in vitro BMSC population relate via cell division lineages, particularly under conditions of culture expansion that are intended to replicate that population for in vitro and in vivo applications, is of practical and scientific interest.