通过使用DNA甲基化沉默基因进行筛选获得的FERMT基因新型合成致死组合

Synthetic lethality with a methylation-silenced gene is a promising strategy for cancer treatment. Here, we aimed to identify such a synthetic lethal combination using genome-wide DNA methylation data and the gene dependency database (DepMap) in 92 gastrointestinal cancer cell lines. First, we isolated 108 genes frequently methylation-silenced in the cell lines, using criteria of a gene with a promoter CpG island and frequent methylation among the cancer cell lines. Next, we searched for their synthetic lethal partners by calculating a correlation coefficient between the methylation level of each candidate gene and dependency of the gene upon 16,836 genes in the DepMap database. As a result, we identified 21 candidate combinations. Among them, to avoid false positive combinations, we focused on one paralog combination; kindlin1 dependency on a high methylation level of its paralog kindlin2 . The 92 cell lines could be classified into kindlin1 knock-down-sensitive and resistant groups by kindlin2 methylation. In addition, incorporation of methylation of kindlin3 , another paralog in the kindlin family, made the classification clearer. The synthetic lethality of the three genes was confirmed by rescue experiments in which overexpression of kindlin2 or kindlin3 decreased kindlin1 dependency. Mechanistically, suppression of all the three kindlin genes caused severe adhesion deficiency, which induced YAP inactivation through its re-localization from nucleus to cytosol, independent of its phosphorylation status. In mouse xenograft models, suppression of kindlin1 by inducible knockdown resulted in tumor regression and it was rescued by overexpression of one of other two paralogs. Methylation analysis of 927 clinical samples from TCGA database showed that 5-10% of gastrointestinal cancer had kindlin2 and kindlin3 methylation. Our research showed that kindlin1 is a new therapeutic target for kindlin2 and kindlin3 -methylated cancers and that a search for synthetic lethal combinations with methylation-silenced genes is a promising strategy.