Introduction {#s1} ============ A major problem with all cancers is the spread of tumors from its site of origin. Cells that have acquired the ability to move around do not necessarily lose their former properties and can still proliferate locally. These cells are usually called metastatic cells. In contrast, cells that have gained the ability to spread through the body are often called tumor stem cells (TSCs). The fact that TSCs are also found in tumors in the liver ([@b5]), skin ([@b18]), pancreas ([@b9]) and breast ([@b19]) suggests that these cells play a general role in cancer development. Several features define the TSCs, such as the capacity to resist conventional therapies ([@b15]), the ability to form tumors when inoculated in immunodeficient animals ([@b22]), the ability to metastasize and differentiate ([@b7]), and its capacity to re-populate tumors upon transplantation ([@b23]). It is, therefore, not clear whether TSCs can be induced to differentiate and whether this can be exploited to reduce metastasis and cancer development. Another feature of the TSCs that must be overcome to cure the disease is their capacity to acquire and maintain a drug resistance profile. The ability of certain cells to differentiate and their general properties are both highly variable and influenced by cell culture conditions. Many features of TSCs can be changed under specific culture conditions, such as the cell's genetic background and culture media. Since the same types of cell are grown in different laboratories, it is important to find standardized conditions that will allow the most reliable assessment of the capacity of TSCs to differentiate or remain in a stem-like state. The most straightforward test to differentiate TSCs is to try to form tumors *in vivo*. However, this *in vivo* test cannot be used with cells that may be transformed by cell culture. In these cases, all *in vivo* experiments must be compared with controls grown *in vitro*. It is also likely that some cells will not differentiate when transplanted into the animal but would continue to grow *in vitro*. It is important to understand the mechanism behind such behavior, as it may indicate the ability of tumor cells to acquire the capacity to differentiate. Cellular senescence is characterized by irreversible growth arrest. Cells enter senescence as a result of DNA damage that occurs as a result of replication stress ([@b28]). It is accepted that after the original insult, the cell repairs DNA damage and can undergo senescence, which prevents them from proliferation. Several senescence inducing agents have been identified, such as reactive oxygen species (ROS) ([@b21]), gamma irradiation ([@b1]), interferon beta treatment ([@b11]), hydrogen peroxide ([@b25]) and UV light ([@b27]). An elevated concentration of ROS is found in many cancer cells, leading to the general assumption that accumulation of ROS causes cancer ([@b3]). UV light is known to induce DNA damage in skin cancer cells ([@b13]; [@b26]) and there is also evidence that ROS generated by UV light may influence the proliferation of skin cancer cells ([@b2]). Here, we show that the human melanoma cell line Mel-RM is prone to transformation upon serial passage in culture. When injected into immunodeficient mice, the cells form large tumors, which can be serially passaged upon transplantation. The first passage of these tumors into culture induces them to differentiate. This differentiation is associated with a reduced rate of proliferation and an increased rate of senescence. We show that the differentiation program induced by the loss of tumorigenicity includes a reprogramming of cells toward a less aggressive, dedifferentiated phenotype. However, we do not detect any significant changes in senescence or cell cycle upon acquisition of a TSC phenotype. Materials and Methods {#s2} ===================== Cell lines and culture conditions {#s2a} --------------------------------- Mel-RM is a spontaneously immortalized human melanoma cell line derived from a metastasis of the human skin melanoma ([@b23]). It was immortalized by transfection with the Simian Virus 40 Large T Antigen and selected for a high rate of spontaneous transformation ([@b16]). The cells were cultured in DMEM (GIBCO) supplemented with 10% FBS (GIBCO). Mel-RM/MelJ1 (MelJ1) were cultured in HFF1 medium (Invitrogen) in the presence of 0.2% NaHCO~3~ and 10% FBS, with 25 µg/ml of ascorbic acid added to the medium. Cells were tested and shown to be free of mycoplasma contamination and to contain less than 5% human cells based on HLA typing (Dynal). Passaging is performed once a week. *In vivo* tumor formation assays {#s2b} -------------------------------- Stable luciferase-expressing Mel-RM and Mel-RM/MelJ1 cells were generated as follows. Mel-RM and Mel-RM/MelJ1 cells were transfected with pGL2 (Promega), which contains the luciferase gene, using the Amaxa Basic Nucleofector Kit for primary cells (Amaxa). The cells were placed in standard medium for one day and then placed in a medium containing 10% FCS for two days to remove toxic components. The cells were then selected in DMEM containing 1 µg/ml puromycin for 7 days, after which the cell population was 100% luciferase positive. Cell differentiation {#s2c} -------------------- Cells were cultured in the presence of 4 µM of Retinoic Acid (RA) (Sigma Aldrich) for 2 weeks. RA induces cell cycle arrest ([@b14]), a phenotype that is expected to result in senescence and is consistent with the finding that TSCs are highly resistant to senescence ([@b6]). It was also shown that the loss of tumorigenicity occurs in cultures with longer growth periods, during which they may have acquired additional genetic changes ([@b24]). To generate the control cells, the RA treated cells were subsequently grown for 2 weeks in the absence of RA. Senescence assays {#s2d} ----------------- β-Galactosidase activity was detected using a kit according to the manufacturer's protocol (Cell Signalling Technology). Briefly, cells were fixed with 0.2% glutaraldehyde/2% formaldehyde, stained with β-Galactosidase substrate and incubated overnight at 37°C. Senescent cells were detected by staining cells with the Cell Proliferation Dye eFluor 670 according to the manufacturer's protocol (eBioscience). Cells were harvested by trypsinization, washed with PBS and resuspended in PBS with 4% FBS. A total of 100,000 cells were stained for 30 minutes at 4°C. Following the staining, cells were washed once and analyzed by flow cytometry. Cell cycle analysis {#s2e} ------------------- Cells were cultured in a 1∶1 mix of medium and ethanol for 4 days before collection. Cells were then harvested, washed and fixed with 70% ethanol and stored at −20°C for 16 hours. After washing, cells were incubated in PBS containing 10 µg/ml RNase A (Roche), 1 mM DTT (Sigma), 25 µg/ml of PI (Sigma) and incubated at 37°C for 30 minutes before analysis. Chromosome preparation and karyotype analysis {#s2f} --------------------------------------------- Cells were treated with 1 mg/ml of colcemid for 6 hours before harvesting. Cells were swollen in PBS and incubated with 0.075 M KCl for 20 minutes at room temperature. Cells were centrifuged and washed once with PBS and the pellet was resuspended in a drop of Carnoy's solution (methanol/glacial acetic acid, 3∶1). The cells were dropped onto a glass slide, dried overnight and stained with Giemsa stain for 5 minutes. Metaphase spreads were captured using Zeiss Imaging System with Axiovision software. For karyotyping, metaphases were analyzed with a spectral imaging system as described in ([@b10]). Images were prepared using the Zeiss Imaging System with Adobe Photoshop software. The analysis was performed in five independent experiments. For analysis of aneuploidy in karyotypes, a total of 22 chromosomes (2p, 2q, 3, 5, 7, 8, 9, 10, 11, 12, 15, 17, 18, 19, 20, 21, X) were investigated. Chromosome numbers in G1, S and G2/M phase cells were determined by cell sorting. The following parameters were used for the analysis of the flow cytometry results: debris exclusion (FSC-A versus FSC-H), and cell cycle distribution (G0/G1 phase versus S-phase versus G2/M). Tumorigenicity assays {#s2g} --------------------- To generate tumors in immunodeficient mice, cells were resuspended in PBS and injected into the right flank. After formation of tumors, they were passaged into nude mice. Tumor growth was monitored by bioluminescence imaging (BLI) using the Xenogen system, and tumor tissue was examined for metastasis. To pass the cells through the animal, they were trypsinized, harvested in PBS with 5% FCS and pelleted by centrifugation. After cells were pelleted, the cell pellet was resuspended in PBS with 5% FCS and kept at 37°C for 2 hours before injection into animals. Flow cytometry {#s2h} -------------- Flow cytometric analysis was performed on an LSR II flow cytometer (Becton Dickinson). Cells were incubated with antibodies for 30 minutes at 4°C. The following monoclonal antibodies were used: APC conjugated anti-CD54, FITC conjugated anti-CD324 (E-Cadherin), anti-CD324 (N-Cadherin), anti-