The two strands of DNA’s double helix unwind to be copied, with a structure called a replication fork forming at the point of separation. The speed at which the replication fork progresses along DNA — and so the speed of replication — must be controlled to guarantee faithful duplication of the genome. In a paper in Nature, Maya-Mendoza et al.1 define a molecular network involved in the regulation of replication-fork speed. Changes to this network can cause that speed to increase above a safe threshold, causing DNA damage and genomic instability.
Replication forks that encounter damage in the genome sometimes temporarily stop, allowing DNA repair to occur before replication continues. Proteins of the poly(ADP)ribose polymerase (PARP) family, particularly PARP1, assist in the repair of breaks in single strands of DNA through a process called PARylation2, in which the proteins synthesize chains of ADP-ribose molecules that attract repair proteins to the damaged DNA. PARP inhibitors — drugs that block the PARylation activity of PARP proteins — are showing promise as therapeutics to treat various cancer types3. Previous models have proposed that, by preventing PARP activity, PARP inhibitors cause replication forks to stall for abnormally long periods, and eventually to collapse, when they encounter DNA damage4. This leads to accumulation of DNA damage owing to improper replication and death of the treated cells4.
Maya-Mendoza et al. challenge the idea that PARP inhibitors perturb the ability of replication forks to progress. The authors found that treating proliferating human cells with the PARP inhibitor olaparib in vitro led to aberrant acceleration of fork speed. They provide evidence that, if fork speed increases above a threshold speed of 40% faster than normal, there is insufficient time for the forks to recognize damaged DNA in need of repair. This leads to accumulation of DNA damage and reduced cell viability. Supporting this idea, the authors found that violation of the threshold speed led to the activation of proteins involved in a DNA-damage response, although the mechanism by which this occurs needs to be further investigated.
To uncover the pathway by which PARP inhibition speeds up replication forks, Maya-Mendoza et al. investigated the protein p21, which can inhibit DNA replication5. Expression of the p21 gene is controlled by PARP16. Moreover, the protein p53, which is a central player in maintenance of genome integrity, activates expression of p217 and is itself activated by PARylation8. The authors found that loss of p21 led to an increase in replication-fork speed similar to the acceleration caused by PARP inhibitors. Loss of p21 in addition to PARP inhibition increased fork speed more than either manipulation in isolation. Combining these observations, the authors propose the existence of a fork-speed regulatory network that has two interacting arms — the p53–p21 axis and PARylation. Each arm acts to keep fork speed below the threshold, with inhibition of either p21 or PARylation throwing the network out of balance and so increasing fork speed (Fig. 1). Several steps of this regulatory pathway will require further investigation. For example, exactly how the arms interact to properly control replication-fork speed is a key question to address.
PARP inhibitors are used to treat tumours that have deficiencies in a pathway called homologous repair that repairs double-stranded DNA breaks. These defects make the tumour cells particularly susceptible to PARP inhibitors. To explain, the single-stranded DNA breaks that accumulate owing to PARP inhibitors are converted to double-stranded breaks when the damaged strand is replicated. In normal cells, the breaks can then be repaired through homologous repair, but when this pathway is defective, the inability to repair these defects leads to cell death. Most notably, breast, ovarian and prostate cancers caused by mutations in the genes BRCA1 and BRCA2 are susceptible to PARP inhibition9,10.
Maya-Mendoza et al. found that PARP inhibition accelerates fork speed above the threshold in BRCA1-deficient cells. On the basis of these results, the authors suggest that the susceptibility of tumours harbouring BRCA mutations to PARP inhibitors might not be due to increased stalling and collapse of replication forks, as originally believed, but instead to aberrant acceleration that compromises the ability of forks to detect and repair DNA damage.
In summary, Maya-Mendoza et al. have provided a fresh view of why PARP inhibitors are toxic to cancer cells, and have outlined a previously unknown network that controls replication-fork speed. Their work has the potential to revolutionize current models of how cells cope with DNA damage — but it also raises several questions.
For example, do other PARP proteins help to control fork speed? The authors report that PARylation levels were not affected by PARP1 depletion. This observation implies that other members of the PARP family are involved in controlling replication-fork speed.
By what mechanism do PARP activity and p21 control replication-fork speed? PARP activity is crucial in the control of replication-fork reversal — a mechanism by which replication forks reverse their course when they face DNA breaks11,12, enabling the damage to be dealt with. Perhaps PARP activity and p21 affect fork speed by suppressing replication-fork reversal or other mechanisms used by replication forks to cope with DNA breaks.
There are other areas of interest for future research. For example, the effect of increased fork speed on polymerase enzymes, which carry out DNA replication, should be examined to determine whether the enzymes exacerbate the situation by introducing more errors into the newly replicated genome as a consequence of increased fork speed. Whether the toxic effects of PARP inhibitors on cancer cells are mainly linked to the fact that the forks do not detect damage when the threshold speed is violated remains to be confirmed.
Notably, PARP inhibitors have also been effectively used in combination with chemotherapeutic agents, which induce DNA damage by impairing the ability of replication forks to progress. It would therefore be interesting to determine whether the same mechanism underlies the effects of PARP inhibitors when used in combination with chemotherapy. Finally, Maya-Mendoza and colleagues’ findings will no doubt prompt many research groups to explore whether surpassing the threshold fork speed provides a more general way to explain the molecular basis of cancer and tumour sensitivity to chemotherapy.
Nature 559, 186-187 (2018)