A targeted screen to identify relevant oncogenes in RAS activated liver cancer

Background:

Activation of RAS signaling is a common event in human cholangiocellular carcinoma (ICC) and hepatocellular carcinoma (HCC). However, gene mutations within the RAS signaling pathway are less common as they occur with a frequency of 15 – 20% in human tumor samples. Of note, activation of RAS signaling is associated with a poor prognosis in liver cancer patients. Do date, it remains unclear if patients within this “RAS activation” subgroup benefit from specific therapies such as MEK inhibitor treatment or other targeted therapies. To mimics RAS activated hepatocarcinogenesis, we generated the RPK mouse model that combines liver-specific expression of oncogenic Kras with genetic inactivation of the tumor suppressors Rb and p53. Upon Cre-mediated gene recombination, adult mice developed liver tumors resembling human hepatocellular carcinomas, intrahepatic cholangiocarcinomas, as well as tumors of mixed HCC/ICC differentiation. Do identify potential target genes in the context of activated RAS signaling we employed a targeted screening approach and validated the functional relevance of candidate tumor suppressors in vitro and in vivo.

Method:

Liver tumors from Rblox/lox;p53lox/lox;KrasLSL-G12D/+ (RPK) mice and liver tissue of age matched control mice were microdissected and analyzed for RNA and protein expression. A curated subset of top upregulated genes in mRNA microarray analysis of RPK liver tumors were chosen for further investigation. Primary cells lines from RPK mice were transfected with lentiviral shRNA constructs and analyzed for proliferation and clonogenic assay. For in vivo analysis mice were inject with a transposon construct harbouring a CreER and an inducible shRNA by hydrodynamic tail vein injection.

Results:

Microarray analysis showed several highly upregulated genes in RPK tumors in comparison to controls. shRNA constructs targeted against top upregulated genes showed a significant decrease in cell proliferation and clonogenic capacity with constructs targeted against Dmbt1. Knock-down of Dmbt1 mRNA and protein was verified by qPCR and western blot. In vivo know-down using shDmbt1 transposon constructs showed a significantly longer survival in comparison to controls.

Conclusions:

Altered RAS signaling is found in human HCC and ICC with poor prognosis. In our screen, we identified Dmbt1 as a potent oncogene. Interestingly, Dmbt1 has been described as a tumor suppressor in other cancers by a mechanism that involves interaction with the immune system. We were able to identify, a novel cell-autonomous oncogenic role of Dmbt1 in the liver. Dmbt1 is upregulated in a subset of human HCC and many ICC, where it might serve as an oncogene. The mechanism of Dmbt1 function in hepatocarcinogenesis and its potential role as a therapeutic target are under further investigation.