Conventional therapies of cancer rely upon radiotherapy and chemotherapy. Such treatments supposedly mediate their effects via the direct elimination of tumor cells. Here we show that anthracyclin, X-Rays and platinum based-therapies mediate optimal antitumor efficacy against established mouse tumors in immunocompetent but not immunocomprised littermates. We addressed whether TLR4, a Toll Like Receptor for which loss of function alleles have been described in humans, control the immune response against dying tumor cells. In vitro, cross-presentation of tumor antigens derived from dying tumor cells to specific CD4+ and CD8+ T cells could be obtained provided that wild type (WT) dendritic cells (DC) but not TLR4-/- or MyD88-/- DC were added to the cocultures. Inoculation of irradiated (but not live) tumor cells into the footpad primed draining lymph node T cells for interferon γ production in WT but not TLR4-/- littermates. Moreover, local tumor irradiation induced a specific T cells activation in WT but not TLR4-/- littermates. Following chemotherapy or radiotherapy of established tumors (CT26 colon carcinoma, EL4 thymoma, TS/A breast carcinoma and GOS osteosarcoma), tumor growth delay depended upon host DC and was observed in WT or Trif-/- mice but not in TLR4-/- or MyD88-/-counterparts. TLR4 was mandatory for efficient processing and cross-presentation of tumor antigens from dying cells (but not for antigen uptake or DC maturation). We hypothetized that damage associated molecular patterns should be released from dying tumor cells to bind to TLR4 and mediate antigen presentation and tumor rejection. Using neutralizing Ab and specific siRNA, we demonstrated that secretion of high mobility group box 1 (HMGB1) alarmins by dying tumor cells could account for TLR4-dependent antitumor immune responses in vivo. We next addressed whether the TLR4 status could predict the therapeutic outcome in a series of 280 breast cancer females treated with systemic anthracyclins. Patients carrying a TLR4 loss-of-function allele (Asp299Gly) manifested a higher frequency of metastases five years after surgery (40% versus 26.5% in non mutated patients, p = 0.04). In addition, the Kaplan Meier analysis of metastasis free survival showed an overall significantly lower percentage of metastases-free patients in the mutated TLR4 group (Log rank test, p = 0.03). These results delineate a clinically relevant immunoadjuvant pathway triggered by tumor cell death.

98th AACR Annual Meeting-- Apr 14-18, 2007; Los Angeles, CA