Aurora kinase A inhibition as a synthetic lethality strategy in ARID1A-mutated gastroenteropancreatic neuroendocrine carcinoma

Chemotherapy with cisplatin (CDDP) and etoposide (ETO) is the standard treatment for gastroenteropancreatic neuroendocrine carcinoma (GEP-NEC). Despite high initial response rates, treatment-related toxicity and acquired resistance remain significant clinical challenges. No molecularly targeted therapies are currently established for this aggressive cancer. Here, we evaluated Aurora kinase inhibition with alisertib (ALI) as a synthetic lethality strategy in ARID1A-deficient GEP-NEC. Subcutaneous and orthotopic xenograft models were established in immunodeficient mice using the GEP-NEC-derived NT-38 cell line, and the therapeutic efficacy of ALI was compared with CDDP + ETO. Tumor responses were assessed by immunohistochemistry, Western blotting, and RNA sequencing. To confirm ARID1A dependency, transient knockdown was induced in the pancreatic neuroendocrine NT-18LM cell line, demonstrating that ALI sensitivity is significantly enhanced in ARID1A-deficient cells. ALI achieved antitumor efficacy comparable to chemotherapy and was well tolerated, with minimal weight loss relative to CDDP + ETO. Distant metastases, an early feature of GEP-NEC, developed in five animals over six weeks, two of which were treated with ALI and three controls. Transcriptomic profiling revealed convergence of both treatments on signal transduction, focal adhesion, receptor tyrosine kinase, and VEGFA-VEGFR2 signaling, while ALI uniquely enriched pathways related to pancreatic secretion, lipid metabolism, protein processing in the endoplasmic reticulum, and extracellular matrix organization. These findings establish Aurora kinase inhibition as mechanistically distinct and selectively effective in ARID1A-deficient GEP-NEC. Given its efficacy, favorable tolerability, and ARID1A-dependent specificity, ALI may represent a promising alternative to platinum-based chemotherapy, offering a strong rationale for further development of mechanism-driven combination strategies for GEP-NEC.