NEW YORK (GenomeWeb) – An international team led by researchers at Yale University has confirmed that suspected tumor suppressor genes previously found to be mutated in in vitro cells are most likely to cause liver cancer when they are mutated in animal cells.
While recurrent mutated genes (RMGs) are well known tumor suppressors or oncogenes, researchers have not previously performed experiments on the genes within the native tumor environment of a living organism.
In this new study, published yesterday in Science Advances, the team directly mapped functional cancer genome variants of tumor suppressors in the autochthonous mouse liver using parallel CRISPR-Cas9 genome editing. Because adeno-associated viruses (AAVs) can efficiently infect the liver after intravenous injection, the team chose liver hepatocellular carcinoma as a model to test in the study.
After excluding known oncogenes, the team first identified the top 50 RMGs and then designed a library of single guide RNAs (sgRNAs) that targeted the different genes.
Using AAVs to carry the library targeting the tumor suppressors that are significantly mutated in human cancers, the team intravenously injected the solution into the livers of fully immunocompetent LSL-Cas9 mice. All 27 mice that received the AAV library developed liver cancer and passed away within four to five months. Using a fluorescent dissecting scope, the researchers found liver tumors, abdominal tumors, sarcomas, and ear tumors in the dead mice.
The researchers then used molecular inversion probe (MIP) sequencing of the estimated 70-base pair regions surrounding the predicted cut site of each sgRNA to map out the mutational landscape of the tumors that killed the mice. They noted that MIP capture sequencing enables a direct quantitative analysis of mutations induced by the Cas9-sgRNA complex.
"Whereas traditional sgRNA sequencing can only provide information about the relative abundances of each sgRNA, capture sequencing enables high-resolution analysis of individual indel variants for clonal analysis of tumor heterogeneity," they wrote.
In order to examine the potential for AAV-CRISPR-mediated mutagenesis in other organs, the team performed MIP capture sequencing on all genomic material, including non-liver tumors and a variety of tissues without detectable tumors under a fluorescent dissecting scope.
Of the genes that were mutated in at least one sample, the majority had multiple mutated sites. The sequencing therefore revealed a heterogeneous mutational landscape, indicating that several genes in a mouse's tumor suppressor gene library can function as tumor suppressors.
The team generated hundred of variants in the top 49 tumor suppressors driving liver tumorigenesis in immunocompetent mice and examined their role in tumor development. Analysis revealed that frameshift insertions and frameshift deletions comprised the majority of total variant reads, consistent with the belief that frameshift mutations are expected to cause loss of function in genes.
In order to investigate synergistic effects between different genes that might have correlated with each other, the researchers then performed a co-mutation analysis. They picked a top co-occurring pair of genes (B2m and Kansl1), two moderate co-occurring pairs (Pik3r1 and Pten, and Pik3r1 and Stk11), and a non significantly co-occurring pair (Arid2 and Kdm5c). Combinatorially knocking out either B2m or Kansl1 led to significantly faster liver tumorigenesis than knocking either out individually, while the rest of the genes did not show any synergistic effects.
"The validation results were consistent with our analysis of the screening data, indicating that the in vivo AAV-CRISPR screen is an effective approach to identify potent [tumor suppressor genes] and potential synergistic drivers of liver tumorigenesis to immunocompetent mice," the authors wrote.
The team did note that the study was partially limited because it chose to exclude oncogenes as potential mutational factors. Because alterations in proto-oncogenes occur through multiple different mechanisms, modeling proto-oncogenes required more sophisticated methods than modeling tumor suppressors. However, the team also noted that adapting the tools used in the study would allow for high-throughput oncogene screens, providing variant mapping of oncogenes as well.
In addition, the researchers believe that the study's approach could potentially be extended to identify genetic factors that affect a variety of cancer types and other human diseases. While the study only examined tumor suppressors in liver cancer, the team has previously applied the approach to glioblastoma and found that "single and co-occurring driver mutations differ between the mouse liver and brain."
The authors concluded, "Application of this approach ... will enable precision preclinical testing for rapid identification of effective compounds against specific mutant genotypes, paving new ways for cancer target discovery."