Drug resistance, a major challenge in cancer therapy, is typically attributed to mutations and genetic heterogeneity. Emerging evidence suggests that dynamic cellular interactions and group behavior also contribute to drug resistance. However, the underlying mechanisms remain poorly understood. Here, we present a new mathematical approach with game theoretical underpinnings that we developed to model real-time growth data of non-small cell lung cancer (NSCLC) cells and discern patterns in response to treatment with cisplatin. We show that the cisplatin-sensitive and cisplatin-tolerant NSCLC cells, when co-cultured in the absence or presence of the drug, display dynamic group behavior strategies. Tolerant cells exhibit a ‘persister-like’ behavior and are attenuated by sensitive cells; they also appear to ‘educate’ sensitive cells to evade chemotherapy. Further, tolerant cells can switch phenotypes to become sensitive, especially at low cisplatin concentrations. Finally, switching treatment from continuous to an intermittent regimen can attenuate the emergence of tolerant cells, suggesting that intermittent chemotherapy may improve outcomes in lung cancer.

Mitochondria are dynamic organelles that constantly fuse and divide, forming dynamic tubular networks. Abnormalities in mitochondrial dynamics and morphology are linked to diverse pathological states, including cancer. Thus, alterations in mitochondrial parameters could indicate early events of disease manifestation or progression. However, finding reliable and quantitative tools for monitoring mitochondria and determining the network parameters, particularly in live cells, has proven challenging. Here, we present a 2D confocal imaging-based approach that combines automatic mitochondrial morphology and dynamics analysis with fractal analysis in live small cell lung cancer (SCLC) cells. We chose SCLC cells as a test case since they typically have very little cytoplasm, but an abundance of smaller mitochondria compared to many of the commonly used cell types. The 2D confocal images provide a robust approach to quantitatively measure mitochondrial dynamics and morphology in live cells. Furthermore, we performed 3D reconstruction of electron microscopic images and show that the 3D reconstruction of the electron microscopic images complements this approach to yield better resolution. The data also suggest that the parameters of mitochondrial dynamics and fractal dimensions are sensitive indicators of cellular response to subtle perturbations, and hence, may serve as potential markers of drug response in lung cancer.

Summary: Tumor heterogeneity and cisplatin resistance are a major cause of tumor relapse and poor survival. Here we show that in lung adenocarcinoma (LUAD), paxillin (PXN) and integrin beta 4 (ITGB4) are associated with tumor progression, and cisplatin resistance. Silencing PXN and ITGB4 render cisplatin tolerant cells sensitive, and immunologically neutralizing ITGB4 improves sensitivity. The N-terminal half of PXN is intrinsically disordered and interacts with ITGB4 to regulate expression of USP1 and VDAC1 which are required for maintaining genomic stability and mitochondrial function in LUAD. By virtual screening an FDA-approved compound library, we identified compounds that interact with PXN in silico and attenuate cisplatin resistance in LUAD cells. RNAseq analysis identified a double negative feedback loop between ITGB4 and microRNA miR-1-3p, suggesting that bistability could lead to stochastic switching between cisplatin-sensitive and resistant states in these cells. The data highlight an alternate, non-genetic, mechanism underlying chemoresistance in lung cancer.

Squamous cell carcinomas (SCC) of the lung (LSCC) and head and neck (HNSCC) are very prevalent with poor prognosis and limited treatment options. In both cancer types, Ephrin receptor A2 (EPHA2) is known to be overexpressed and exhibit opposing effects via two distinct signaling mechanisms. While it can inhibit cancer cell survival and migration by ligand-dependent signaling through tyrosine kinase phosphorylation at Y588 and Y772, it can promote tumor progression and cell migration in a ligand-independent manner via phosphorylation at S897. Variable ABnormal morphology (VAB-1) is the C. elegans ortholog of the human ephrin receptor (EPHR) that interacts genetically and biochemically in a dose-dependent manner with the axon guidance receptor, SAX3, the worm ortholog of ROBO. Double mutants of vab-1 (EPHR)/ sαx-3 (ROBO) are synthetic lethal, underscoring the interaction between the two signaling pathways which prompted us to investigate their role in SCC. Using biochemical and biophysical techniques, we show that EPHA2 and ROBO1 reside in the same complex and interact physically to form a functional heterodimer in LSCC and HNSCC. Furthermore, we show that treating squamous cells with the SLIT2, ligand of ROBO1, hinders phosphorylation of EPHA2 at S897, and thereby, attenuates cell proliferation. Interestingly, SLIT2 can interact with EPHA2 and attenuate the proliferation of cells that have low ROBO1 expression. Additionally, SLIT2 can act synergistically with the EPHA2 inhibitor, Ensartinib to attenuate cell growth in LSCC and HNSCC cells. Taken together, the data suggest that SLIT2 may serve as a novel therapeutic for LSCC and HNSCC. Here, we propose to stratify patients for treatment with SLIT2 and/or Ensartinib, based on their EPHA2 and ROBO1 expression levels in the diseased tissue. Thus 85% of LSCC cases can be treated with combination of SLIT2+Ensartinib and 55% of HNSCC cases can be treated with either SLIT2 or Ensartinib. Furthermore, EPHA2 and ROBO1 may represent novel theranostics in these two diseases. One sentence summary: Heterodimerization of EPHA2 and ROBO1 receptors attenuates growth of squamous cell carcinomas of the lung and head and neck.