Development of pancreatic and colon patient-derived xenograft (PDX) tumor microarrays (TMAs) from the NCI patient derived models repository

The National Cancer Institute's Patient Derived Models Repository (NCI PDMR; https://pdmr.cancer.gov) has developed a national repository of Patient-Derived Models (PDMs) currently comprised of over 1000 patient-derived xenograft (PDX), 450 organoid (PDOrg), 500 tumor cell culture (PDC), and 425 cancer associated fibroblast (CAF) models. Over 450 PDXs have matched PDOrg and/or PDCs allowing for complimentary/parallel in vivo/in vitro studies. These PDMs are clinically annotated with molecular information available in a public database for the extramural community with additional molecular features including OncoKB annotated mutations, microsatellite instability (MSI), human leukocyte antigen (HLA) typing, and clinically relevant fusions. Researchers can use these clinical and molecular features or perform their own independent analyses using the public data to aid in their selection of preclinical models. Due to the large number of NCI PDMR models within histologies and research community interest, we have developed histology-based PDX TMAs to further facilitate the selection of models for cancer research. Each TMA panel includes up to 60 unique PDX models, with two 1.5mm cores/model plus murine control tissue. Quality control (QC) assessment of the TMA cores is performed by a pathologist. Each core is reviewed with an initial pass/fail threshold set to ≥10% human tumor/core area with ≥500 tumor cells. TMA slides pass QC if they meet these criteria and ≥75% of the models have at least one passing core. TMA blocks are QC'd at regularly set intervals to ensure all distributable slides meet these requirements. The first PDX TMA panel available for distribution this year (PANC I) contains 60 pancreatic cancer PDXs (predominantly pancreatic adenocarcinoma [PAAD]) derived from primary and metastatic lesions from treatment naïve through heavily pretreated patients. KRAS mutated models include 28 G12D, 15 G12V, 10 G12R, 3 G12C, 1 Q61H, and 3 KRAS wildtype. Other genes frequently mutated in pancreatic cancer are also found in this cohort including TP53 , SMAD4 , and CDKN2A . Also in development are four colorectal cancer TMAs panels: (1) a general set of colon adenocarcinomas (COAD) with features including early onset, non-European ancestry, and MSI-High; (2) KRAS mutated COAD; (3) Treatment naïve COAD and wildtype APC COAD; and (4) Rectal adenocarcinoma models. These TMAs can be used to stratify models by therapeutic target, develop predictive markers or classify differential signaling in disease subtypes, integrative analysis of genomic and protein expression, and discover or validate biomarkers of disease in an efficient and cost-effective way. Targeted model selection is of high importance to better understand the biology of these cancers and improve preclinical drug testing and screening design to translate novel therapeutics from bench to clinic.