Du, K.-P. | Search bioRxiv

bioRxiv2024

Break-induced replication drives large-scale genomic amplifications in cancer cells

DNA double-strand breaks (DSBs) are highly toxic lesions that underly the efficacy of ionizing radiation (IR) and a large number of cytotoxic chemotherapies1-3. Yet, abnormal repair of DSBs is associated with genomic instability and may contribute to cancer heterogeneity and tumour evolution. Here, we show that DSBs induced by IR, by DSB-inducing chemotherapeutics, or by the expression of a rare-cutting restriction endonuclease induce large-scale genomic amplification in human cancer cells. Importantly, the extent of DSB-induced genomic amplification (DIGA) in a panel of melanoma cell lines correlated with the degree of cytotoxicity elicited by IR, suggesting that DIGA contributes significantly to DSB-induced cancer cell lethality. DIGA, which is mediated through conservative DNA synthesis, does not require origin re-licensing, and is enhanced by the depletion or deletion of the methyltransferases SET8 and SUV4-20H1, which function sequentially to mono- and di-methylate histone H4 lysine 20 (H4K20) at DSBs to facilitate the recruitment of 53BP1-RIF1 and its downstream effector shieldin complex to DSBs to prevent hyper-resection4-11. Consistently, DIGA was enhanced in cells lacking 53BP1 or RIF1, or in cells that lacked components of the shieldin complex or of other factors that help recruit 53BP1 to DSBs. Mechanistically, DIGA requires MRE11/CtIP and EXO1, factors that promote resection and hyper-resection at DSBs, and is dependent on the catalytic activity of the RAD51 recombinase. Furthermore, deletion or depletion of POLD3, POLD4, or RAD52, proteins involved in break-induced replication (BIR), significantly inhibited DIGA, suggesting that DIGA is mediated through a RAD51-dependent BIR-like process. DIGA induction was maximal if the cells encountered DSBs in early and mid S-phase, whereas cells competent for homologous recombination (in late S and G2) exhibited less DIGA induction. We propose that unshielded, hyper-resected ends of DSBs may nucleate a replication-like intermediate that enables cytotoxic long-range genomic DNA amplification mediated through BIR.

cancer biology↗

bioRxiv2021

Transcriptome-wide Association Study and eQTL colocalization identify potentially causal genes responsible for bone mineral density GWAS associations

Genome-wide association studies (GWASs) for bone mineral density (BMD) have identified over 1,100 associations to date. However, identifying causal genes implicated by such studies has been challenging. Recent advances in the development of transcriptome reference datasets and computational approaches such as transcriptome-wide association studies (TWASs) and expression quantitative trait loci (eQTL) colocalization have proven to be informative in identifying putatively causal genes underlying GWAS associations. Here, we used TWAS/eQTL colocalization in conjunction with transcriptomic data from the Genotype-Tissue Expression (GTEx) project to identify potentially causal genes for the largest BMD GWAS performed to date. Using this approach, we identified 512 genes as significant (Bonferroni <= 0.05) using both TWAS and eQTL colocalization. This set of genes was enriched for regulators of BMD and members of bone relevant biological processes. To investigate the significance of our findings, we selected PPP6R3, the gene with the strongest support from our analysis which was not previously implicated in the regulation of BMD, for further investigation. We observed that Ppp6r3 deletion in mice decreased BMD. In this work, we provide an updated resource of putatively causal BMD genes and demonstrate that PPP6R3 is a putatively causal BMD GWAS gene. These data increase our understanding of the genetics of BMD and provide further evidence for the utility of combined TWAS/colocalization approaches in untangling the genetics of complex traits.

genomics↗