TP53 status aligns with morphology and metaplastic programs in colorectal cancer organoids, and MDM2 P2 promoter usage reflects p53 activity

Abstract

Background: In colorectal cancer (CRC), TP53 mutation accompanies chromosomal instability. How TP53 status relates to tumor morphology and lineage programs, and whether a p53 activity readout could refine MDM2-inhibitor use, are unclear. MDM2 is transcribed from a constitutive (P1) and a p53 inducible (P2) promoter. Methods: Sixty-three organoid samples from 21 CRC patients were classified by deep learning morphology (VGG16), array based MDM2 Splicing Index and TP53 targeted resequencing. In organoids, p53 activity was scored from 12 direct p53 targets and promoter usage as the array P2 index. TCGA-COAD/READ and cell lines used RNA seq transcript ratios (P2_index; cell-line P2 index). Validation used same-patient arrays (GSE128213), TCGA-COAD/READ (n = 624), GSE39582 and an independent biobank (GSE294511). Results: TP53 mutations were almost confined to P1 dominant samples (40 of 42 mutants). Morphology followed TP53 status (median Type1 fraction 0.854 (mutant) versus 0.422, specimen level; P = 0.013 adjusted for chromosome 20q and aneuploidy). With only five samples discordant, TP53 status and promoter usage could not be separated (P = 0.97). The morphology gene score followed TP53 status in 88 GSE294511 organoids and TCGA-COAD/READ, identifying mucinous carcinoma; gastric-metaplasia genes followed TP53 status, not P2, in both cohorts but not discovery organoids. The p53 target module of the P2 signature followed P2 beyond TP53 status in both organoid cohorts, whereas the independent 12-gene p53 activity score did so in GSE294511 (P = 0.012) but not discovery organoids (P = 0.094). P2 dominant samples, including two TP53-mutant ones, were p53-active, and nutlin induced P2 only in p53 proficient cells. In TCGA, P2_index was higher in wild-type tumors but did not follow p53 activity beyond total MDM2. In wild-type cell lines P2 followed MDM2 dependency beyond total MDM2, and in wild type GSE294511 organoids idasanutlin sensitivity, though not beyond total MDM2. Conclusions: In CRC organoids, morphology followed TP53 status, with no detectable contribution of MDM2 promoter usage, as did gastric-metaplastic programs in an independent biobank and tumors. In organoids, P2 usage reflected p53 activity, beyond genotype mainly in the biobank and two TP53-mutant P2-dominant specimens; whether it improves on MDM2 expression for MDM2-inhibitor selection requires testing.