John Langenfeld, MD, associate professor, Division of Thoracic Surgery, UMDNJ-Robert Wood Johnson Medical School
R
|
ecent studies have linked the growth of cancer to a specific population of cancer cells that reside within a tumor. These specialized cells can form tumors on repeated injections into nude mice while the remaining cancer cells do not form tumors. The behavior of these specialized tumor cells resembles that of stem cells, so they have been named “tumor stem cells.” This data provides intriguing new insight into how tumors grow and will likely lead to novel treatment modalities. Present day chemotherapy is designed to treat the whole tumor. Future tumor therapy may be designed to treat only the “tumor stem cells.”
Many fundamental questions need to be answered before progress can be made for therapies targeted against “tumor stem cells.” It is clear that some tumor cells can initiate tumor growth while others do not. It is not clear if the tumor-initiating cell population represents one pure population or whether more than one cell population is present. We do not know whether cancers are composed of distinct lineages with different biological activity It is also not clear whether more than one population of cells can initiate tumor growth. Another important question is whether “tumor stem cells” differentiate and, if so, what cell types do they differentiate into. There is evidence that in brain carcinomas, neural precursors differentiate. However, in other solid tumors, differentiation needs to be further elucidated.
Figure 1. By FACS, GFP expressing cells from the Oct4/GFP cell line were sorted and plated at a single cell density on glass cover slips. We then examined whether the Oct4/GFP cells differentiated into a neural lineage. (A) Colonies were immunostained for the neural marker NeuN, The NeuN expressing cells are red and GFP cells are green. Note: some cells lost GFP expression and became NeuN positive.
In my laboratory, we have been studying the “tumor stem cell” population in non-small cell lung carcinomas. We found that surface markers used to identify tumor stem cells in other solid tumors were not helpful in isolating a homogenous cell population. We have developed, in lung cancer cell lines, a new method to isolate and study “tumor stem cells.” Our model utilizes the activity of the embryonic transcription factor Oct4 that is only present in pluripotent stems cells. Oct4 is required to maintain normal embryonic stem cells in an undifferentiated state. The Oct4 promoter was placed upstream of green fluorescent protein (GFP). We developed cells lines that express GFP, which indicates an active Oct4 promoter in those cells. The Oct4 cells are isolated by FACS and placed into cell culture. Some of the cells lose GFP expression, which suggests differentiation. We then isolate the Oct4/GFP (+) and GFP (-) population. We found that the Oct4/GFP (+) cells, but not the GFP (-) cells, form tumors in nude mice.
We have also identified 2 other distinct cell populations within the lung cancer cell lines. One subpopulation of cancer cells expresses the mesoderm marker of differentiation—smooth muscle actin (SMA). The other subpopulation expresses the neural precursor, nestin. Using promoters specific for nestin or SMA, we isolated each individual subpopulation. We found that the nestin subpopulation formed tumors in nude mice while the SMA population did not. These data show that more than one subpopulation of cells is able to initiate tumor growth. This challenges the present hypothesis that only the “tumor stem cell” subpopulation can initiate tumor growth.
In solid tumors, the ability of a tumor stem cell population to differentiate needs to be better elucidated. We show from a single cell that the Oct4 tumor stem cell population is able to differentiate into both mesenchymal and neural lineages. The Oct4 cells differentiate into neural progenitor cells expressing nestin. Therefore, the Oct4 cells can differentiate into cancer cells that initiate tumor growth or cells that do not initiate tumor growth in nude mice.
Our studies show that more than one population of cancer cells within a lung tumor is able to initiate tumor growth. The Oct4 cells differentiate into tumor initiating and non-initiating cells. Future therapy for lung and other cancers will likely target specific subpopulations of cancer cells. Therapy may block the ability of tumor initiating cells to grow or even induce differentiation into benign cell types. Further studies are needed to better delineate the mechanism regulating specific subpopulations of cells found in cancers.
John Langenfeld earned his medical degree from Rush Medical School in Chicago. He completed a general surgery residency at UMDNJ, and thoracic surgery training at West Virginia University and Memorial Sloan Kettering Cancer Institute. He performed a research fellowship at UMDNJ and Memorial Sloan Kettering Cancer Institute. He has been on the faculty of UMDNJ-Robert Wood Johnson Medical School since 1999, specializing in thoracic oncology surgery. His research is on lung cancer.
