P1-01-05: Conditioning by the Tumor Environment Turns Invariant Natural Killer T Cells into Negative Regulators of Anti-Tumor Immunity Elicited by Treatment.

Abstract

Background: Local suppression in the tumor microenvironment remains an important obstacle to success of immune-based therapy. We have previously shown in the 4T1 model of metastatic breast cancer that combination radiotherapy (RT) to the primary tumor and CTLA-4 blockade can elicit a robust anti-tumor effector response that can inhibit metastases and prolong survival. However, complete cure and long-term memory was seen in only a small fraction of mice, prompting us to investigate new ways to improve the response elicited by treatment. iNKT cells are known to be powerful regulators of tumor immunity as a result of their ability to rapidly secrete a broad range of cytokines, but the mechanisms regulating their switch from a stimulatory to a regulatory role remain poorly defined. We have previously shown that iNKT cells negatively regulate the response to treatment with RT and CTLA-4 blockade in 4T1 tumor-bearing mice (Clin Cancer Res 15:597, 2009). Here, we test the novel hypothesis that conditioning of iNKT cells to perform this regulatory role occurs within the tumor microenvironment. Specifically, we address whether 4T1 tumor cells, directly or indirectly, promote iNKT cell regulatory phenotype. Methods: For in vitro experiments, 4T1 tumor or LPS-matured bone marrow-derived dendritic cells (BMDC) were loaded with α-galactosylceramide and used to stimulate the NKT DN32.D3 hybridoma cells. In some wells, BMDCs were supplemented with 4T1 conditioned media or loaded with tumor lysates from 4T1 cells. For in vivo experiments, wild-type and iNKT-deficient mice were inoculated s.c with 4T1 cells for enumeration of DCs on immunostained sections. Phenotypic profile of DCs within tumor and draining lymph nodes were assessed by flow cytometry. Results: 4T1 cells are positive for CD1d by real-time RT-PCR; surface CD1d were barely detectable by flow cytometry but sufficient to induce activation of DN32.D3 cells when irradiated 4T1 cells were loaded with α-GalCer. BMDC supplemented with conditioned media from untreated or irradiated 4T1 cells, or loaded with tumor lysates from irradiated 4T1 cells were also able to activate DN32.D3 cells, and this effect was partially inhibited by blocking CD1d mAb (1B1) in a dose-dependent manner. In vivo, we found that iNKT-deficient mice showed significantly (p=0.04) higher densities of tumor infiltrating DCs than wild type mice. Moreover, DC in tumor and tumor-draining lymph nodes expressed markers of maturation in the absence but not presence of iNKT cells. Conclusion: Results indicate that a 4T1 derived factor, possibly a lipid antigen that can be presented directly by tumor cells themselves or indirectly by BMDCs induces activation of iNKT cells in vitro. In vivo, data suggests that iNKT cells downregulate the response to treatment by controlling the population of DC present in the tumor and draining lymph nodes. Since DC are essential for cross-presentation of the tumor antigens released by RT-induced cell death, reduced DC numbers may impair the activation of anti-tumor T cells. Supported by DOD Postdoctoral Award W81XWH-10-1- Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P1-01-05.