Background
Antibodies blocking immune checkpoint molecules such as CTLA-4 have been shown to be effective in several cancer types, with some patients displaying durable complete regression. However, many patients do not respond to treatment. It is not known what molecular events control the response nor which co-treatments are likely to combine effectively with checkpoint blockade. Current strategies involve empirically testing different combinations of checkpoint blocking antibodies with other immunotherapeutic strategies or conventional anti-cancer drugs. We provide an alternative approach.
Methods
Through performing network analysis of gene expression data from responding versus non-responding AB1-HA mesothelioma tumours from mice treated with anti-CTLA-4, we identified genetic modules and hub genes within these modules that were associated with responsiveness. We subsequently identified synergistic anti-CTLA-4/drug combinations using two different approaches: first, by pinpointing drugs that modulated hub genes within these response-associated modules, and second, by interrogating overlaps in the modular response patterns and drug-perturbation signatures in drug repurposing databases. The approaches were validated by testing the identified drugs in vivo, in combination with anti-CTLA-4 in murine cancer models.
Results
We identified and validated several drugs that increased the response rate to anti-CTLA-4 in a highly synergistic manner. We identified four drug classes with the capacity to increase the cure rate from 10% for anti-CTLA-4 alone to 60-80% as combination therapy. These repurposed drugs are normally used in completely unrelated conditions such as cardiovascular or neurologic diseases.
Conclusions
Together, our results show that using network analysis of gene expression data from immunotherapy-responsive tumours generates testable hypotheses for the identification of novel synergistic drug combinations.