Bio-orthogonal Metabolic Labeling Reveals Platinum Chemotherapy Dysregulates Spatiotemporal ECM Dynamics in Ovarian Cancer

Introduction: : One of the deadliest forms of gynecological cancers in women is ovarian carcinoma with high grade serous ovarian carcinoma (HGSC) being the most common subtype. Before surgery, 70% of patients with HGSC are treated with platinum chemotherapy and over 80% of these patients develop resistance to treatment within 1 year. Chemoresistance in HGSC has been associated with the dysregulation of the tumor extracellular matrix (ECM) composition and structure which critically contributes to a cancer cell’s ability to resist treatment, ultimately leading to over 90% of deaths from cancer. The ECM is a non-cellular 3D network secreted by cells essential to structural and biochemical support of the surrounding cells, but the tumor ECM is biochemically and biomechanically distinct compared to normal tissue ECM. ECM remodeling signatures are associated with an increased risk of treatment failure and patient mortality in HGSC. The effects of ECM composition and structure have been well explored and linked to reduced therapeutic efficacy; however, the effects of ECM changes associated with chemotherapy are still unknown. Therefore, technologies to determine the spatiotemporal dynamics of newly synthesized ECM in response to chemotherapy in HGSC are desperately needed. Our objective is to assist in the development and functional characterization of cancer biomimetic tissue-engineered modeling necessary for performing studies profiling spatiotemporal ECM dynamics in response to chemotherapy utilizing metabolic labeling technologies. In the absence of these approaches, the development of effective anticancer therapies that mitigate chemotherapy resistance in HGSC will remain difficult.

Materials and

Methods: : The 3D extracellular matrix (3D-ECM) is a biomimetic 3D hybrid hydrogel technology that was developed through the chemical crosslinking of collagen I and plasma-derived fibrinogen ensuring viscoelastic and fibrillar microstructure which are critical for 3D mechanosignaling. HGSC cell lines (OVCAR8, CAOV3, OVCAR3) or platinum sensitive/resistant patient derived xenografts lines (OV81 and OV81-CP40, respectively) were co-cultured with cancer-associated fibroblasts (CAFs) for 7 days with or without cisplatin treatment (5 uM). Newly synthesized proteins secreted in response to chemotherapy were identified through bio-orthogonal metabolic labeling. To do this, 3D-ECM cultures were incubated with a non-canonical amino acid analog (L-azidohomoalanine, AHA). Newly synthesized proteins containing AHA were fluorescently labeled by copper-free cycloaddition with dibenzocycloctyne-amine (DBCO-647) which were further imaged with super resolution confocal microscopy and analyzed by machine learning software (Fig A). Flow cytometry was performed to investigate Epithelial to Mesenchymal Transition (EMT) in platinum sensitive/resistant pairs co-cultured with CAFs and exposed to PBS control or platinum treatment to investigate changes in E-Cadherin and N-Cadherin expression. Proteomic analysis of enriched AHA-labeled proteins utilized a cleavable biotin-alkyne linker and NeutrAvidin beads to selectively enrich AHA-labeled nascent proteins. Immunofluorescence was further performed to validate the newly synthesized proteins identified by proteomic analysis and correlate them to known ECM proteins.

Results, Conclusions, and Discussions:: The 3D-ECM model was engineered to recapitulate HGSC response to treatment parameters such as cell-cell interactions between the cancer and stroma cell types and high collagen I content with associated increased stiffness. Through bio-orthogonal metabolic labeling, newly synthesized AHA proteins (magenta) were significantly increased in response to cisplatin treatment compared to PBS controls in multiple HGSC cells (Fig B). Newly synthesized AHA proteins were significantly higher in cisplatin resistant OV81-CP40 compared to OV81 in the absence of cisplatin treatment, highlighting a reprogramming of resistant cells to secrete dysregulated ECM dynamics even in the absence of further treatment (Fig C). In both cell types, addition of cisplatin treatment significantly enhanced AHA protein secretion. Importantly, using machine learning software we tracked the ancestry of AHA proteins to identify which cell type is most likely the secretor and identified that cisplatin treatment significantly increases the secretor profile of HGSC cells and in co-culture of resistant cells with CAFs, both cells significantly increase their secretion profile, indicating paracrine signaling between both cell types. Furthermore, it was identified that while HGSC secreted more AHA proteins under cisplatin treatment, cells underwent EMT, with reduction of epithelial markers (E-cadherin) and increase of mesenchymal markers (N-Cadherin) (Fig D). Importantly, cisplatin-resistant OV81-CP40 had a significantly higher mesenchymal phenotype than OV81 in response to cisplatin, aligning with the previous AHA protein secretion and ancestry results. Proteomics analysis of enriched AHA proteins highlighted that this technique enables us to identify not only high abundance ECM proteins but also transient and low abundant ECM proteins, and results were validated via immunofluorescence co-staining (Fig E). Ultimately, these results suggest that platinum chemotherapy itself dysregulates spatiotemporal ECM dynamics and such changes could be linked to an increase in chemoresistance. In conclusion, our biomimetic tissue-engineered 3D-ECM model paired with metabolic labeling enables tracking of spatiotemporal ECM dynamics in response to treatment. These results are intended to have an important positive impact by providing strong evidence-based proof of principle for further development of therapeutics with specificity towards chemotherapy-induced ECM changes and ultimately improving chemotherapy effectiveness and mitigating chemoresistance.

Acknowledgements and/or References (Optional):: This project was supported by 1P30GM145398-02