Preview

PDF, English Download (4MB) | Terms of use

Abstract

The hypothesis of the cancer stem cell (CSC) suggests that neoplastic clones are maintained by a rare fraction of tumor cells with stem cell properties. CSCs could represent disseminated dormant tumor cells without clinical signs of tumor progression. We used a ret transgenic mouse spontaneous melanoma model, in which 25% of transgenic mice develop skin tumors with metastases in lymph nodes (mLN), liver, lungs and bone marrow (BM). Mice older than 20 weeks without macroscopic tumors contain in the BM tyrosinase related protein (TRP)-2-specific effector memory CD8+ T cells and show no further melanoma progression. This suggests a potential role of dormant tumor cells in the maintenance of memory CD8+ T cells. We found that TRP-2+CD133+ melanoma cells represent less than 1.5% of all cells in primary skin tumors and mLN. The majority of these cells were Ki67— suggesting thereby that these cells could remain in a dormant state. We found an increased expression of the major regulator for cell survival, self-renewal, and tumor growth, HIF-1a in TRP-2+CD133+ melanoma cells in large tumors in comparison with those in smaller tumors. To investigate whether TRP-2+CD133+ melanoma cells are disseminated in the BM of ret transgenic mice, we performed a triple immunofluorescence staining. We found that only 40% of mice without macroscopic tumors contained TRP-2+CD133+ melanoma cells in the BM. In contrast, all tumor bearing mice contained TRP-2+CD133+ melanoma cells. TRP-2+CD133+ melanoma cells were detected in 2 of 712 (0.238%) and 4 of 1285 (0.311%) disseminated melanoma cells in the BM of mice without and with macroscopic tumors, respectively. We confirmed the dormant state of TRP-2+CD133+ melanoma cells based on the negative expression of Ki67 and PCNA. Proteins p16 and p27, which are typically located in the nuclei of dormant cells, were found in the cytoplasmic compartment of TRP-2+CD133+ melanoma cells indicating their highly malignant phenotype. Investigating the interaction between memory CD8+ T cells with disseminated melanoma cells in the BM, we found that TRP-2+Ki67— melanoma cells were co-localized with memory CD8+ T cells both in mice without and with macroscopic tumors. The proportion of memory CD8+ T cells interacting with TRP-2+Ki67— melanoma cells was lower (less than 15%) in the BM of these mice. Quantitative analyses revealed that although certain IFN-g-producing CD8+ T cells interacted either with single TRP-2+ melanoma cells or the smallest cluster of melanoma cells (2-5 TRP-2+ cells), none of these T cells produced perforin. Only two TRP-2-specific CD8+ T cells produced perforin, but none of them were co-localized either with TRP-2+ melanoma cells or TRP-2+CD133+ melanoma cells. Furthermore, memory CD8+ T cells located within the large cluster of 50 TRP-2+ melanoma cells were unable to produce both perforin and IFN-g. These findings suggest that tumor microenvironment might neutralize CD8+ T cell reactivity. In conclusion, our data demonstrate the existence of a subpopulation of CD133+ melanoma cells in ret transgenic mice. Dormant TRP-2+ melanoma cells are able to interact with CD8+ T cells in the BM of tumor-bearing mice.