Surprisingly, we showed that tumor derived exosome (TDE) -and not a tumor derived soluble Napabucasin in vivo factor- determines MDSCs Stat3-dependent suppressive activity. Moreover,

we could demonstrate that, in both mice and humans, membrane Hsp72 from TDE triggers Stat3 activation in MDSCs in a TLR2/MyD88 dependent manner through an autocrine production of IL-6. Accordingly, targeting exosome production in vivo using dimethylamiloride blunts the suppressive activity of MDSC and enhanced the efficacy of cyclophosphamide treatment in three different mouse tumor models. Finally, we also demonstrated that this mechanism supporting suppressive MDSCs activity is relevant in cancer patients. Collectively, our findings show for the first time in both mice and human settings that membrane TDE associated Hsp72 restrained tumor immune surveillance by supporting MDSCs suppressive functions. O175 Immune Cell Homing in Preinvasive HPV Disease Cornelia Trimble 1 , Christopher Thoburn1, Shiwen Peng1, Nicole Hanson3, Jodie Tassello3, Denise Frosina3, Ferdynand Kos1, Rachael Clark2, Achim Jungbluth3 1 Johns Hopkins University

School of Medicine, Baltimore, MD, USA, 2 Dermatology, PARP inhibitor Brigham and Women’s Hospital, Boston, MA, USA, 3 Ludwig Institute for Cancer Research, Memorial Sloan-Kettering Cancer Center, New York, NY, USA Globally, human papillomavirus (HPV) causes more human malignancies than any other virus. High grade cervical intraepithelial neoplasia (CIN2/3) occurs only in the setting of persistent mucosal infection with an oncogenic strain of HPV, and presents a compelling opportunity to test immunotherapies because expression of two viral proteins, E6 and E7,

are functionally required to initiate and maintain disease. We have a large prospective cohort of subjects with CIN2/3 who are followed for a brief, 15-week window prior to definitive excision of the cervical squamocolumnar junction (cervical conization or LEEP procedure). Not all dysplastic lesions progress to cancer; 25% of HPV16+ CIN2/3 undergo complete regression in this timefrqme. However, systemic HPV16 E6 and E7 T cell responses are marginally detectable, and do not correlate with lesion regression. However, CIN2/3 does recruit inflammatory infiltrates. Memory (-)-p-Bromotetramisole Oxalate T cells accumulate in dysplastic mucosa, and spectratyping provides strong evidence that these often contain clonally expanded populations. In our cohort, intraepithelial CD8+ infiltration at t0 was predictive of regression by twk15. In contrast, in lesions that failed to regress in the study window, inflammatory infiltrates were restricted to the cervical stroma, whilst intraepithelial CD8+ infiltrates were minimal. Detectable IFNγ immune responses to E6 and E7 measured in patient-matched peripheral blood obtained at the same visits did not correlate with lesional CD8+ infiltrates.