Translating Lung Cancer Molecular Variability into Personalized Signaling Signatures Towards Efficient Design of Individualized Treatments

Introduction: Lung cancer is one of the leading causes of death. Lung cancer patients often develop resistance to the FDA-approved targeted monotherapies and chemotherapies due to inter-patient heterogeneity.

Results: Using protein expression data obtained from hundreds of cancer tissues and cell lines, including HCC827 and H2286 lung tumors, we computed protein-protein network reorganization in every tumor using an information-theoretic approach. We show that patient-specific protein-protein networks, named patient-specific signaling signatures (PaSSS) can differ significantly between lung cancer patients. These PaSSSs can include several distinct altered signaling subnetworks each. Based on the resolved PaSSSs personalized targeted drug combinations can be designed.

We show that simultaneous targeting of central proteins from each altered subnetwork in every PaSSS is essential to selectively inhibit growth of lung cancer cells in-vitro and in-vivo. We show that the predicted and efficient targeted drug combination for one lung cancer can be less efficient for another and vice versa. Furthermore, we demonstrate that the PaSSS-based drug combinations lead to induced expression of T cell markers in immune cells co-cultured with lung cancer cells, pointing to higher efficiency of the immune response in the presence of individualized drug combinations.

Conclusion: Our preliminary results suggest that the PaSSS-based analysis allows the design of efficient and patient-specific targeted drug therapies. Moreover, these therapies may induce the immune response in a patient-specific manner.