Hepatocellular carcinoma-associated AXIN1 mutations drive low levels of Wnt/β-catenin pathway activity that allow for niche-independent growth and YAP/TAZ signaling

In healthy cells, AXIN1 organizes assembly of a large destruction complex that mediates proteolysis of the transcriptional co-activator {beta}-catenin to prevent inappropriate Wnt/{beta}-catenin pathway activation. In hepatocellular carcinoma (HCC), AXIN1 mutations (11%) associate with a poor-prognosis subtype that is molecularly distinct from {beta}-catenin-mutant HCC (28-40%). How AXIN1 deficiency drives HCC formation has remained highly debated. Here, we address this issue by introducing HCC-associated AXIN1 and CTNNB1 mutations in human liver cancer cells and liver-derived organoids. We show that different mutant AXIN1 classes activate varying degrees of Wnt signaling, although at lower overall levels than CTNNB1 mutations. Strikingly, premature stop codons in 5 coding regions do not classify as knock-out mutations but drive alternative translation of an N-terminally truncated AXIN1 variant with partially retained suppressor activity. All AXIN1 variants endow liver progenitor organoids with the capacity to grow in the absence of R-spondin and Wnt, indicative of aberrant Wnt/{beta}-catenin pathway activation. Additionally, induced Wnt/{beta}-catenin pathway activation inversely correlates with YAP/TAZ-mediated signaling, thus leaving higher residual YAP/TAZ activity in AXIN1-mutant versus CTNNB1-mutant cells. We conclude that AXIN1 mutations drive physiologically relevant Wnt/{beta}-catenin signaling in HCC, while providing a permissive environment for YAP/YAZ signaling, thereby distinguishing AXIN1 mutations from those in CTNNB1.