The Wnt/β-catenin pathway is the most frequently deregulated pathway in hepatocellular carcinoma (HCC). Inactivating mutations of the gene encoding AXIN1, a known negative regulator of the Wnt/β-catenin signaling pathway, are observed in about 10% of HCCs. Whole-genome studies usually place HCC with AXIN1 mutations and CTNNB1 mutations in the group of tumors with Wnt/β-catenin activated program. However, it has been shown that HCCs with activating CTNNB1 mutations form a group of HCCs, with a different histology, prognosis and genomic signature to those with inactivating biallelic AXIN1 mutations. We aimed to elucidate the relationship between CTNNB1 mutations, AXIN1 mutations and the activation level of the Wnt/β-catenin program.


We evaluated two independent human HCC datasets for the expression of a 23-β-catenin target genes program. We modeled Axin1 loss of function tumorigenesis in two engineered mouse models and performed gene expression profiling.


Based on gene expression, we defined three levels of β-catenin program activation: strong, weak or no activation. While more than 80% CTNNB1-mutated tumors were found in the strong or in the weak activation program, most of the AXIN1-mutated tumors (>70%) were found in the subgroup with no activation. We validated this result by demonstrating that mice with a hepatocyte specific AXIN1 deletion developed HCC in the absence of β-catenin induction. We defined a 329-gene signature common in human and mouse AXIN1 mutated HCC that is highly enriched in Notch and YAP oncogenic signatures.


AXIN1-mutated HCCs occur independently of the Wnt/β-catenin pathway and involve Notch and YAP pathways. These pathways constitute potentially interesting targets for the treatment of HCC caused by AXIN1 mutations.


Liver cancer has a poor prognosis. Defining the molecular pathways involved is important for developing new therapeutic approaches. The Wnt/β-catenin pathway is the most frequently deregulated pathway in hepatocellular carcinoma (HCC). Mutations of AXIN1, a member of this pathway, represent about 10% of HCC mutations. Using both human HCC collections and engineered mouse models of liver cancers with AXIN1 mutation or deletion, we defined a common signature of liver tumors mutated for AXIN1 and demonstrate that these tumors occur independently of the activation of the Wnt/β-catenin pathway.