SACF and GILA assays on AML12 cells show limited predictive value for mouse liver genotoxicity

Abstract

Hepatocellular carcinoma (HCC) has been observed in neonatal mice following the integration of recombinant Adeno-Associated Viruses (rAAV) into the Rian locus. rAAV-related oncogenic risk for patients remains unclear, and the lack of relevant in vitro methods hinders its proper assessment. The soft agar colony-forming (SACF) assay and the growth in low attachment assay (GILA) monitor anchorage-independent growth, a hallmark of transformed adherent cells, and have been previously proposed to assess the tumorigenicity of CRISPR/Cas9-edited human MCF10A cells. Here, we introduce murine versions of SACF and GILA as surrogate in vitro systems to evaluate the risk of HCC development following genome editing or rAAV induced insertional mutagenesis. Selected tumor suppressors linked to HCC onset in vivo were edited through CRISPR/Cas9 in the hepatic murine cell line AML12. The knockout of neurofibromin (Nf2) and the dual inactivation of tumor protein p53 (Tp53) and phosphatase and tensin homolog (Pten) induced anchorage-independence, while the editing of Axin1, Ctnnb1 (coding for β-catenin), and tuberous sclerosis complex 1 (Tsc1) did not promote growth in anchorage-free conditions. Additionally, we generated stable AML12 and MCF10A clones with the rAAV genome respectively integrated into Rian and MEG8, the human homolog of Rian; however, these clones did not show anchorage independence when seeded in SACF and GILA. Overall, the murine SACF and GILA exhibit low predictive value for HCC development, failing to detect rAAV- and tumor-suppressors-associated oncogenicity. While further optimization may improve assays performance, these results highlight the need for more appropriate in vitro methodologies to accurately evaluate rAAV genotoxicity.