Hepatocellular carcinoma, the most common form of primary liver cancer, is associated with a poor prognosis, making the precise identification of the molecular pathways involved in its development essential. Among these, the Wnt/β-catenin pathway is the most frequently deregulated in this tumor. Inactivating mutations of the AXIN1 gene, which encodes a negative regulator of this pathway, are found in approximately 10% of hepatocellular carcinomas. Genome-wide studies typically classify these tumors, together with those harboring CTNNB1 mutations (the gene encoding β-catenin), within the group of cancers displaying an activated Wnt/β-catenin program. However, earlier observations have shown that these two categories differ in their histology, prognosis, and genomic signature. The authors therefore sought to clarify the relationship between CTNNB1 mutations, AXIN1 mutations, and the actual level of activation of the β-catenin program.
To this end, the expression of a 23-gene program of β-catenin target genes was assessed in two independent human datasets (the GSE62232 and TCGA-LIHC collections). In parallel, Axin1 loss of function was modeled in mice using two genetic models that inactivate the gene specifically in hepatocytes, either at the embryonic stage or in adulthood following tamoxifen induction, followed by gene expression profiling. Based on these profiles, three levels of β-catenin program activation were defined: strong, weak, or absent. Whereas more than 80% of CTNNB1-mutated tumors fell into the strong or weak activation categories, more than 70% of AXIN1-mutated tumors belonged to the subgroup lacking activation. The mouse models confirmed this finding: mice with hepatocyte-specific Axin1 inactivation develop hepatocellular carcinoma without β-catenin induction, and only tumors arising from Apc deletion showed significant induction of target genes.
This work establishes a 329-gene signature shared by human and murine AXIN1-mutated tumors. This signature, also enriched in genes associated with invasion, tumor stemness, and poor prognosis, proves to be strongly enriched in oncogenic signatures of the Notch and YAP pathways, as confirmed by the nuclear activation of Sox9 and YAP/TAZ as well as the overexpression of their target genes in the tumors. The authors conclude that AXIN1-mutated hepatocellular carcinomas develop independently of the Wnt/β-catenin pathway and involve the Notch and YAP pathways, which thus represent potentially attractive therapeutic targets for this category of tumors.