Oncolytic adenovirus decreases the proportion of TIM-3 subset of tumor-infiltrating CD8 T cells with correlation to improved survival in patients with cancer.

Background: Oncolytic viruses are a potent form of active immunotherapy, capable of invoking antitumor T-cell responses. Meanwhile, less is known about their effects on immune checkpoints, the main targets for passive immunotherapy of cancer. T-cell immunoglobulin and mucin domain-3 (TIM-3) is a coinhibitory checkpoint driving T-cell exhaustion in cancer. Here we investigated the effects of oncolytic adenovirus on the TIM-3 checkpoint on tumor-infiltrating immune cells and clinical impact in patients with cancer receiving oncolytic immunotherapy.

Methods: Modulation of TIM-3 expression on tumor-infiltrating immune cells was studied preclinically in B16 melanoma following intratumoral treatment with Ad5/3∆24-granulocyte-macrophage colony-stimulating factor oncolytic adenovirus. We conducted a retrospective longitudinal analysis of 15 patients with advanced-stage cancer with tumor-site biopsies before and after oncolytic immunotherapy, treated in the Advanced Therapy Access Program (ISRCTN10141600, April 5, 2011). Following patient stratification with regard to TIM-3 (increase vs decrease in tumors), overall survival and imaging/marker responses were evaluated by log-rank and Fisher's test, while coinhibitory receptors/ligands, transcriptomic changes and tumor-reactive and tumor-infltrating immune cells in biopsies and blood samples were studied by microarray rank-based statistics and immunoassays.

Results: Preclinically, TIM-3 tumor-infiltrating lymphocytes (TILs) in B16 melanoma showed an exhausted phenotype, whereas oncolytic adenovirus treatment significantly reduced the proportion of TIM-3 TIL subset through recruitment of less-exhausted CD8 TIL. Decrease of TIM-3 was observed in 60% of patients, which was associated with improved overall survival over TIM-3 increase patients (p=0.004), together with evidence of clinical benefit by imaging and blood analyses. Coinhibitory T-cell receptors and ligands were consistently associated with TIM-3 changes in gene expression data, while core transcriptional exhaustion programs and T-cell dysfunction were enriched in patients with TIM-3 increase, thus identifying patients potentially benefiting from checkpoint blockade. In striking contrast, patients with TIM-3 decrease displayed an acute inflammatory signature, redistribution of tumor-reactive CD8 lymphocytes and higher influx of CD8 TIL into tumors, which were associated with the longest overall survival, suggesting benefit from active immunotherapy.

Conclusions: Our results indicate a key role for the TIM-3 immune checkpoint in oncolytic adenoviral immunotherapy. Moreover, our results identify TIM-3 as a potential biomarker for oncolytic adenoviruses and create rationale for combination with passive immunotherapy for a subset of patients.