Cancer Stem Cells (CSCs), or Tumor Initiating Cells (TICs) are typically defined functionally. That is, they are usually defined as the subset of tumor cells that can be serially xenotransplanted through mice and give rise to the full heterogeneity of the original tumor (or in this case cell line).
There are many viable strategies for identifying such a subpopulation, for example, one may enrich for CSCs/TICs by treating with escalating doses of chemotherapy or one may look for subpopulations of cells that express high levels of developmental transcription factors. If you are interested in identifying a potential CSC/TIC compartment within MKN-49 cells, you should design experiments that you think could enrich for the TIC population and then test those sub-populations using serial xenograft studies and/or 3D cell culture assays.
However, if you are only asking how one would isolate a previously identified TIC/CSC compartment from MKN-49 cells, the answer is much simpler. You should first identify cell surface markers for the population of interest, for example CD44 and CD24 (see a review like Singh, Cancer Letters, 2013, http://www.sciencedirect.com/science/article/pii/S0304383513002784). Next, perform FACS collecting the population that you are interested in (for a generalizable method see Vidal, et al., JoVE, 2014, http://www.jove.com/video/51332/isolation-of-cancer-stem-cells-from-human-prostate-cancer-samples).
CD44(high)/CD24(low)/Lin(-) population should be isolated by FACS. Further, CD44v8-10 expressing cells have the higher tumorigenic and invasive phenotype, which reflects the "stemness."
Hi, First of all, you need to identify specific cell surface markers for example CD44 or CD133 in this gastric cancer cell line, then you can sort these population by FACS or with low accuracy MACS.
I suggest that you be careful about drawing conclusions about CSCs in cell lines. Serum differentiates stem cells. The differentiated cells in cell lines have typically acquired self-renewal. Because of these things, your stem cell population may not even be present anymore.