XPF-ERCC1 protects liver, kidney and blood homeostasis outside the canonical excision repair pathways

Loss of the XPF-ERCC1 endonuclease causes a dramatic phenotype that results in progeroid features associated with liver, kidney and bone marrow dysfunction. As this nuclease is involved in multiple DNA repair transactions, it is plausible that this severe phenotype results from the simultaneous inactivation of both branches of nucleotide excision repair (GG- and TC-NER) and Fanconi anaemia (FA) inter-strand crosslink (ICL) repair. Here we use genetics in human cells and mice to investigate the interaction between the canonical NER and ICL repair pathways and, subsequently, how their joint inactivation phenotypically overlaps with XPF-ERCC1 deficiency. We find that cells lacking TC-NER are sensitive to crosslinking agents and that there is a genetic interaction between NER and FA in the repair of certain endogenous crosslinking agents. However, joint inactivation of GG-NER, TC-NER and FA crosslink repair cannot account for the hypersensitivity of XPF-deficient cells to classical crosslinking agents nor is it sufficient to explain the extreme phenotype of Ercc1(-/-) mice. These analyses indicate that XPF-ERCC1 has important functions outside of its central role in NER and FA crosslink repair which are required to prevent endogenous DNA damage. Failure to resolve such damage leads to loss of tissue homeostasis in mice and humans. Author summary The integrity of DNA is essential for life so even the most primitive life forms have evolved a DNA repair 'toolkit' to detect and fix different types of DNA damage. XPF-ERCC1 is an enzyme that can cut DNA and a key player in many of these DNA repair transactions. Consistent with this, inactivating mutations of XPF-ERCC1 in humans and mice lead to a dramatic premature ageing phenotype associated with liver, kidney and bone marrow dysfunction. Here, we ask which of the many functions of XPF-ERCC1 are required to protect tissues from endogenous DNA damage. To do this, we generated cells and mice lacking two of the best characterised functions of XPF-ERCC1: nucleotide excision repair and inter-strand crosslink repair. Surprisingly, neither the cells nor mice lacking these two repair pathways behave like the XPF-ERCC1 mutants, in fact the mice are remarkable in their lack of phenotype. Our work suggests that there are functions of XPF-ERCC1 outside of the canonical repair pathways which are important for DNA repair and the homeostasis of multiple organs.