Patients with melanoma often have cytotoxic T cells that are specific for melanocytic differentiation antigens in their tumors. One therapy against the disease is to expand (in number) and activate these melanoma-infiltrating cytotoxic T cells ex vivo before putting them back into the patient in a process known as adoptive cell transfer therapy (ACT). Although initially successful in killing tumor cells and leading to remission, ACT often eventually fails because the tumors develop resistance (see commentary by Ribas and Tumeh). One possible cause of remission is the outgrowth of a subpopulation of mutated melanoma cells that lack the targeted antigens. Landsberg et al. generated a mouse model of melanoma combined with an ACT protocol (with T cells specific for the tumor antigen gp100) that mimicked the phases of regression, remission, and relapse observed in human patients. Compared with primary melanomas, relapsed melanomas showed loss of gp100 and enhanced abundance of mRNAs encoding proinflammatory cytokines, including tumor necrosis factor–α (TNF-α). When transferred to new mice, relapsed melanoma cells regained gp100 and were susceptible to a new round of ACT before becoming resistant once again. Whole-transcriptome analysis of the various stages of melanomas showed the decreased expression of many melanocyte-specific genes, and immunohistochemistry showed the increased abundance in relapsed melanomas of the nerve growth factor receptor (NGFR, also known as p75 neurotrophin receptor), indicating the dedifferentiation of the tumor cells. Conditioned medium from tumor-infiltrating lymphocytes stimulated the appearance of NGFR and the loss of melanocyte markers (including gp100) on tumor cells, which was prevented by a blocking antibody against TNF-α. Inhibitor studies showed that the appearance of dedifferentiated melanoma cells did not require the proliferation or apoptosis of any subpopulation of cells. Treatment of human melanoma cell lines with TNF-α also increased the abundance of NGFR and reduced the abundance of gp100 in vitro. Together, these data suggest that resistance to ACT is not a result of the escape of genetically altered tumor cells that do not express specific antigens but rather is due to the reversible loss of targeted antigens as a result of proinflammatory cytokines in the tumor microenvironment. The authors suggest that modified ACTs that target both melanocytic and nonmelanocytic antigens may be more effective in targeting cells.
J. Landsberg, J. Kohlmeyer, M. Renn, T. Bald, M. Rogava, M. Cron, M. Fatho, V. Lennerz, T. Wölfel, M. Hölzel, T. Tüting, Melanomas resist T-cell therapy through inflammation-induced reversible dedifferentiation. Nature 490, 412–416 (2012). [PubMed]
A. Ribas, P. C. Tumeh, Tumours switch to resist. Nature 490, 347–348 (2012). [PubMed]