ABSTRACT
There are no therapeutic predictive biomarkers or representative preclinical models for high-grade gastroenteropancreatic neuroendocrine neoplasms (GEP-NEN), a highly aggressive, fatal, and heterogeneous malignancy. We established patient-derived (PD) tumoroids from biobanked tissue samples of advanced high-grade GEP-NEN patients and applied this model for targeted rapid ex vivo pharmacotyping, next-generation sequencing, and perturbational profiling. We used tissue-matched PD tumoroids to profile individual patients, compared ex vivo drug response to patients' clinical response to chemotherapy, and investigated treatment-induced adaptive stress responses.PD tumoroids recapitulated biological key features of high-grade GEP-NEN and mimicked clinical response to cisplatin and temozolomide ex vivo. When we investigated treatment-induced adaptive stress responses in PD tumoroids in silico, we discovered and functionally validated Lysine demethylase 5 A and interferon-beta, which act synergistically in combination with cisplatin. Since ex vivo drug response in PD tumoroids matched clinical patient responses to standard-of-care chemotherapeutics for GEP-NEN, our rapid and functional precision oncology approach could expand personalized therapeutic options for patients with advanced high-grade GEP-NEN.
ABSTRACT
Pancreatic neuroendocrine neoplasms (PanNENs) are the second most common malignancy of the pancreas. Surgery remains the only curative treatment for localized disease. For patients with inoperable advanced or metastatic disease, few targeted therapies are available, but their efficacy is unpredictable and variable. Exploiting prior knowledge on pathogenetic processes involved in PanNEN tumorigenesis, we tested buparlisib (PI3K inhibitor) and ribociclib (CDK4/6 inhibitor), as single agents or in combination, in different preclinical models. First, we used cell lines representative of well-differentiated (INS-1E, NT-3) and poorly differentiated (BON-1) PanNENs. The combination of buparlisib with ribociclib reduced the proliferation of 2D and 3D spheroid cultures more potently than the individual drugs. Buparlisib, but not ribociclib, induced apoptosis. The anti-proliferative activity of the drugs correlated with downstream target inhibition at mRNA and protein levels. We then tested the drugs on primary islet microtissues from a genetic PanNET animal model (Men1-defective mice) and from wild-type mice: the drug combination was effective against the former without altering islet cell physiology. Finally, we treated PanNET patient-derived islet-like 3D tumoroids: the combination of buparlisib with ribociclib was effective in three out of four samples. Combined targeting of PI3K and CDK4/6 is a promising strategy for PanNENs spanning various molecular and histo-pathological features.
ABSTRACT
Abdominal surgeries are lifesaving procedures but can be complicated by the formation of peritoneal adhesions, intra-abdominal scars that cause intestinal obstruction, pain, infertility, and significant health costs. Despite this burden, the mechanisms underlying adhesion formation remain unclear and no cure exists. Here, we show that contamination of gut microbes increases post-surgical adhesion formation. Using genetic lineage tracing we show that adhesion myofibroblasts arise from the mesothelium. This transformation is driven by epidermal growth factor receptor (EGFR) signaling. The EGFR ligands amphiregulin and heparin-binding epidermal growth factor, are sufficient to induce these changes. Correspondingly, EGFR inhibition leads to a significant reduction of adhesion formation in mice. Adhesions isolated from human patients are enriched in EGFR positive cells of mesothelial origin and human mesothelium shows an increase of mesothelial EGFR expression during bacterial peritonitis. In conclusion, bacterial contamination drives adhesion formation through mesothelial EGFR signaling. This mechanism may represent a therapeutic target for the prevention of adhesions after intra-abdominal surgery.
