The emerging need of new and improved cancer models has accelerated the optimization of cancer cell line reprogramming. So far, two different ways for reprogramming somatic cells to pluripotency have been described, by somatic cell nuclear transfer ((SCNT) implantation of a somatic nucleus cell into a enucleated oocyte)3,14 or by using ectopic expression of TFs to produce iPS cells (Fig. 1). Whereas just a subset of cancer cells have been successfully reprogrammed by SCNT, the second approach has been more successful since there is no need for the use of oocytes or blastocysts and also due to increased reprogramming efficiency15. Whether these differences are due to technical issues or cancer-tissue specificity is still not understood. Despite the challenges, the first report showing successful reprograming of human malignant cells appeared in 2010 when iPS cells were generated using retroviral delivery of OSKM into a chronic myeloid leukemia (CML) cell line, KBM716. Injection of the resulting CML-iPS cells subcutaneously into NOD-SCID mice produced teratomas that contained cells from all three germ layers. The CML-iPS cells were able to differentiate in vitro into cells that express CD43 (pan-T cell marker), CD45 (hematopoietic lineage marker), as well as the stem cell marker CD34, indicating a restoration of differentiation ability into hematopoietic lineages. The parental CML cell lines were dependent on the BCR–ABL pathway; however, the CML-iPS cell lines showed resistance to imatinib, an inhibitor of BCR–ABL signaling, despite expressing the BCR–ABL gene. The CML-iPS cells regained sensitivity to imatinib when differentiated in vitro to hematopoietic lineage cells, suggesting that oncogenic dependency of the CML-iPS cells depended on differentiation status of the cells17.
Thereafter, other reports verified generation of iPS cell lines from different cancer cells lines or primary tumor cells; melanoma18, gastric cancer19, glioblastoma20, and pancreatic ductal adenocarcinoma (PDAC)21, among others. Not all reports have described the use of iPS as a model for cancer but solely the success of using this technique to generate pluripotent cells. However, some studies have gone further and showed the use of cancer cell-derived iPS cells to model the disease (Fig. 2a). Since PDAC lacks an early disease progression model, Kim et al. generated iPS cells from PDAC tumor samples that upon differentiation underwent early stages of pancreatic cancer. When injected into immunodeficient mice, the iPS cells formed intra-epithelial neoplasias (PanINs) which progressed to more invasive stages. Furthermore, they generated and studied pancreatic organoids from the PanIN cells to identify biomarkers and pathways useful for early detection of the disease, since a major problem in diagnosing pancreatic cancer is the lack of symptoms until late-stage disease. The HNF4a network was discovered to be significantly activated in early to intermediate stages of PDAC development, indicating this method may be useful for identifying novel targets for diagnosis and treatment21.
However, it has been demonstrated that reprogramming cancer cells can be challenging and some cell lines need other factors such as NANOG, LIN28, BCL2, or KRAS in addition to the classical OSKM factors. The reprogramming problems might originate from the presence of genomic instability, including accumulation of mutations and DNA damage, and epigenetic modifications22.
Cancer is not solely driven by genetic defects, extensive research has identified epigenetic changes in cancer cells23. Considering that cell transformation, cellular reprograming, and pluripotency are related concepts strongly linked to epigenetic regulation, it has been suggested that reprogramming could epigenetically reset cancer cells. Zhang et al. could show that the sarcoma cell lines SAOS2, HOS, MG63, SW872, and SKNEP could be reprogrammed into iPS cells, which were capable of generating connective tissue and hematopoietic cells24. When sarcoma-derived iPS were differentiated into osteogenic and adipogenic lineage and followed by in vivo subcutaneous transplantation into mice, no tumor growth was observed, suggesting reprogramming could overcome the tumorigenic capacity of the parental sarcoma cells24. In the same manner, non-small cell lung cancer cell lines were reprogrammed and reversion of their epigenome and transcriptome resulted in reduced tumorigenicity25. Therefore, because of the extensive epigenetic remodeling that oncogenes and tumor suppressors undergo via reprogramming, this method has been proposed as a therapeutic strategy. However, it has also been shown that the same approach can lead to the opposite phenotype, resulting in a more invasive and aggressive disease and even resistance to currently available inhibitors, such as seen in PDAC, juvenile myelomonocytic leukemia (JMML), and CML19,21,26,27. Even though ectopic expression of pluripotency genes such as reprogramming factors can be achieved using non-integrating techniques28, we have to take in consideration that the OSKM factors are also implicated as potent cancer drivers. In addition, it has been suggested that incomplete reprogramming could lead to cancer development. Using a mouse model with inducible OSKM factors, Ohnishi et al. 29 showed that long-term (4 weeks) in vivo activation of OSKM factors resulted in reprogramming and teratoma formation in vivo, however, a shorter late activation in vivo for 3–9 days resulted in tumor development in various somatic tissues consisting of undifferentiated dysplastic cells and global changes in DNA methylation patterns. The tumor cells could be reprogrammed into iPS cells by OSKM activation for 2 weeks in vitro. When the iPS cells were injected into blastocysts, they gave rise to non-neoplastic normal kidney cells in the chimeric mice29. These findings suggest that epigenetic processes associated with iPS cell reprogramming may also drive cancer development and that this genetic transformation is reversible. Consequently, taking this approach as a potential treatment has to be carefully revised. Due to tumor heterogeneity we might be selecting for clonal subpopulations explaining the discrepancy of reports whether reprogramming cancer cells may alleviate the tumorigenic potential or not.
Source : https://www.nature.com/articles/s41420-017-0009-2