(The) study of gastric cancer genome heterogeneity using Patient-Derived Xenograft (PDX) and PDX-derived organoids

Abstract

Gastric cancer is a heterogeneous disease accounting for a high proportion of cancer-related death worldwide. Patient-derived xenograft (PDX) and organoids are effective cancer models for investigating the heterogeneity of gastric cancer, but PDX suffers from low throughput, and organoids suffer from normal tissue contamination. PDX-derived organoid (PDXO) is an alternative solution to cancer organoids in that no normal cell depletion is needed. In addition, given that additional mutations can accumulate in cancer models, whether the genome of PDX and PDXO accurately represent the original tumor is an important issue to confirm. In this study, I established 30 PDXOs, and longitudinally analyzed 29 sets of matched normal, original tumor, PDX, and PDXO in a whole genome scale. PDXOs were free from normal tissue contamination, but the final "take rate" of the establishment was 51% due to mouse cell contamination. WGS analysis revealed that PDX and PDXO accurately represented the original tumor in that both shared much of the major driver mutations together as a stable clone. Two molecular subtypes of gastric cancer, chromosome instability (CIN) and microsatellite instability (MSI), were identified from analyzing stable clones of the models. Besides the major clone, PDX and PDXO also harbored numerous subclonal mutations in a whole genome scale. MSI PDXO accumulated more short insertion-deletion mutations during the passage, and CIN PDX and PDXO accumulated more structural variations compared to the MSI subtype. Signature analysis of single nucleotide variants revealed environment specific signatures, or new MSI-driver gene associated signatures exclusively active in PDX and PDXO. Our study provides a valuable resource for gastric cancer research, validate the similarity of PDX and PDXO to the original tumor, and highlight dynamic subclonal mutational processes active in PDX and PDXO.