Genomic and transcriptomic analysis of a diffuse pleural mesothelioma patient-derived xenograft library

Background Diffuse pleural mesothelioma (DPM) is an aggressive malignancy that, despite recent treatment advances, has unacceptably poor outcomes. Therapeutic research in DPM is inhibited by a paucity of preclinical models that faithfully recapitulate the human disease. Methods We established 22 patient-derived xenografts (PDX) from 22 patients with DPM and performed multi-omic analyses to deconvolute the mutational landscapes, global expression profiles, and molecular subtypes of these PDX models and compared features to those of the matched primary patient tumors. Targeted next-generation sequencing (NGS; MSK-IMPACT), immunohistochemistry, and histologic subtyping were performed on all available samples. RNA sequencing was performed on all available PDX samples. Clinical outcomes and treatment history were annotated for all patients. Platinum-doublet progression-free survival (PFS) was determined from the start of chemotherapy until radiographic/clinical progression and grouped into < or >= 6 months. Results PDX models were established from both treatment naive and previously treated samples and were noted to closely resemble the histology, genomic landscape, and proteomic profiles of the parent tumor. After establishing the validity of the models, transcriptomic analyses demonstrated overexpression in WNT/beta-catenin, hedgehog, and TGF-beta signaling and a consistent suppression of immune-related signaling in PDXs derived from patients with worse clinical outcomes. Conclusions These data demonstrate that DPM PDX models closely resemble the genotype and phenotype of parental tumors, and identify pathways altered in DPM for future exploration in preclinical studies.