Genome-wide screening of tumor suppressor genes in oncogenic transformation of normal lung epithelial cells and preliminary validation in sample databases

To identify new human lung cancer suppressors in a genome-wide screening library during neoplastic transformation of normal human lung epithelial cells using CRISPR/Cas9-mediated genome editing, and preliminarily validate the results in non-small cell lung cancer (NSCLC) sample databases.

Human normal lung epithelial BEAS-2B cell line with stable expression of Cas9 protein was constructed, and then infected with the Brunello lentiviral library encompassing 77 441 single-guide RNAs (sgRNAs) targeting 19 114 genes. Then the obtained cells were transplanted subcutaneously into nude mice. Genomic DNA of tumor cells (Tumor) and pre-transplantation cells (Input) were extracted. The extracted DNA and plasmid library were amplified for deep sequencing. By analyzing the proportion of sgRNA reads in Tumor cells compared to the Input cells and the Log Fold Change (LFC) of sgRNA reads, tumor suppressor genes that can induce oncogenic transformation of BEAS-2B cells after knockout were screened, and preliminary validation were performed using public data platforms.

Library infections in BEAS-2B cells stably expressing Cas9 protein and subcutaneous tumorigenesis in nude mice were successfully established. Only cells infected with the library formed tumors at the inoculation sites. Sequencing results showed that only a small part of sgRNAs were detected in tumor cells. Here, 38 genes with sgRNA reads accounting for more than 1% of the total reads or with 2 or more significantly enriched sgRNA were selected as effective candidate human lung tumor suppressor genes, including known tumor suppressor genes such as NF2 and PTEN etc, and some candidate tumor suppressor genes (AP2M1 and PSENEN) had not been previously reported in lung cancer. Enrichment analysis revealed that these candidate tumor suppressor genes were enriched in key biological pathways involved in cancer, such as Notch and Hippo signaling pathways. It is confirmed in the clinical NSCLC sample databases that 38 candidate tumor suppressor genes were mutated in NSCLC samples. Among the 17 genes with mutation rates greater than 2%, 14 genes were co-mutated with KRAS or TP53, and 3 ones were associated with the prognosis of NSCLC patients. There were 30 candidate genes with their expression in the samples of NSCLC different from those in para-tumor normal tissues, and the expression of 17 candidate genes were correlated with the prognosis of NSCLC patients.

By using an in vivo genome-wide CRISPR/Cas9 screening library, 38 human lung cancer candidate suppressor genes are identified and preliminarily verified in the clinical NSCLC sample databases.