Cockayne syndrome is a rare human hereditary disease, characterized by growth failure, deficient neurological development and severe sensitivity to sunlight. It can arise from mutations in any one of five genes. The protein products of these genes are involved in different aspects of the repair of damaged DNA, and it has been far from clear how all these different mutations result in the same syndrome. Le Page and colleagues1, writing in Cell, now provide important clues to the answer. It seems that the common problem in cells from patients with Cockayne syndrome is a failure to repair oxidation-induced damage to DNA bases, specifically in the strands of DNA that are being transcribed into RNA.

Different DNA-repair pathways operate on different types of DNA lesions. Nucleotide-excision repair (NER), for example, is a ubiquitous cellular process by which short, single-stranded DNA segments, containing damaged nucleotides, are removed from duplex DNA. The gaps are then filled in by repair DNA synthesis, using the intact strand as a template (see ref. 2 for a review). Defects in NER underlie the hereditary disease xeroderma pigmentosum. This disease — like Cockayne syndrome — is characterized by severe sensitivity to sunlight. However, patients with xeroderma pigmentosum are several thousand times more likely than Cockayne syndrome patients to develop cancer in exposed areas of skin. Otherwise, mutations in most of the seven XP (xeroderma pigmentosum) genes needed for NER of photoproducts in DNA do not usually pose serious health problems. Another pathway, termed base-excision repair (BER), operates on the damage to bases produced by reactive oxygen species, ionizing radiation and some alkylating agents, as well as certain inappropriate bases (such as uracil) in DNA.

Yet another process, transcription- coupled repair (TCR), deals with a variety of DNA lesions that are thought to arrest the transcription of genes3. This process has been considered for some time to be a type of NER. If the ultraviolet wavelengths of sunlight cause damage to the strand of a DNA duplex that is being transcribed into RNA, TCR solves the problem. By contrast, lesions throughout the genome — including ultraviolet-light-induced damage to the non-transcribed strands of expressed genes — are repaired by global genomic NER.

Mutations in either of two non-essential genes — CSA or CSB — result in defective TCR, and are the genetic defect in over 90% of Cockayne syndrome patients. Certain mutations in XP genes also underlie a small number of Cockayne syndrome cases. Two of these genes, XPB and XPD, encode components of the general transcription factor TFIIH. This complex is needed to open up the DNA strands in preparation for the enzyme RNA polymerase II to begin transcription. It also opens up regions that include a DNA lesion, allowing NER to take place. The third XP gene so involved is XPG, which encodes a protein required to make the first of the two incisions in the DNA strand needed for NER.

Thus all the mutations that cause Cockayne syndrome have in common the property that they eliminate TCR of ultraviolet-damaged DNA. This explains the sensitivity to sunlight, but what about the developmental defects, which are unlikely to result from ultraviolet damage to DNA? Might the basis for these defects lie in the defective NER of similar damage caused by other agents? This seems unlikely, as mutations in the XPA gene (which is involved in lesion recognition) that totally eliminate both global genomic NER and the TCR of such damage do not result in Cockayne syndrome. A second hypothesis is that the disease is a ‘transcription syndrome’, in which certain groups of genes are deficiently expressed4. In this model, the mutations in CSA and CSB, like those in XPB and XPD , are envisaged to have direct effects on transcription itself. But it is hard to explain how the mutations in XPG affect transcription.

So what is the basis for Cockayne syndrome? Getting to the answer requires rethinking the relationship between TCR and NER. It seems that, far from being a subpathway of NER, TCR may in fact act upstream of both nucleotide- and base- excision repair.

A first step along the way to this answer was provided by the report5 that DNA damage produced by ionizing radiation (not ultraviolet radiation) is subject to TCR in normal human cells and in XPA mutant cells, but not in CSB mutant cells. DNA damage produced by ionizing radiation is generally thought to be remedied by BER, not NER. In addition, TCR of an oxidized base, thymine glycol, has been shown to be defective as a result of XPG mutations that result in Cockayne syndrome6, but not of other XPG mutations that result only in the symptoms of xeroderma pigmentosum7. Repair of thymine glycol is also achieved by BER. So, TCR can be linked to BER as well as to NER. These results are further evidence that Cockayne syndrome might result from defective TCR of oxidative lesions.

Le Page et al.1 have now finished testing the hypothesis that all patients with Cockayne syndrome — no matter which gene is mutated — should be deficient in the TCR of oxidative lesions, whatever their nature. Their results firmly establish that both the XPG protein and TFIIH have different roles in the TCR of oxidative damage and in NER.

Le Page et al. also turned their attention to another prominent oxidative lesion, 8-oxo-guanine. Using a shuttle vector, they introduced DNA containing 8-oxo-guanine residues into several human cell strains. They found that 8-oxo-guanine, like thymine glycol, is subject to TCR, but not significantly to NER — so it must be subject to BER instead. They also showed that 8-oxo-guanine (or 8-oxo-guanine plus a ligand) is an obstacle to the movement of RNA polymerase II, and that the arrested polymerase must be released to give the repair enzymes access to the site. So, not only is TCR missing in Cockayne syndrome, but the blocked RNA polymerase II prevents lesion recognition and repair by the global BER pathway as well.

Instead of being considered a subpathway of NER, it now seems that TCR is a process needed before either the BER or the NER pathway can repair oxidative damage to genes that are being transcribed, in which RNA polymerase is stalled at a lesion. TCR might be considered an ‘obstacle-recognition’ step, in which TFIIH and XPG may work together to assess the nature of the obstruction (Fig. 1). This obstruction might be removed only when the arrested RNA polymerase II has been moved away, which may be achieved by CSA and CSB, amongst other factors. TFIIH and XPG might then recruit the enzymes needed for BER or NER, as appropriate.

Figure 1: The DNA-repair pathways meet.
figure 1

The results of Le Page et al.1 lead to the following model for the repair of lesions in the transcribed strands of expressed genes. Top, RNA polymerase II transcribes DNA into RNA. If RNA polymerase II meets a lesion (asterisk) in the DNA strand, it will stop. Centre, repair of the lesion and further transcription is blocked until the polymerase is either released from the DNA or backed away from the lesion. Both TFIIH and XPG are then required before the lesions are repaired, perhaps to determine the nature of the obstruction. Bottom, enzymes of either the nucleotide-excision-repair (NER) or the base-excision-repair (BER) pathways are recruited, to remove the offending lesion and allow transcription to resume. The genes that can be mutated in Cockayne syndrome are required to remove the stalled polymerase, and to assess the nature of the damage.

It is clear that an impasse provided by an arrested RNA polymerase II would severely affect transcription. In that sense, Cockayne syndrome could, indeed, be considered a ‘transcription syndrome’. But the arrest of transcription also provides a strong signal for a cell-death (apoptosis) pathway8. So, Cockayne syndrome could also be characterized as a disease of excessive cell death by apoptosis — a disease that would affect rapidly metabolizing cells, such as neurons, that generate high levels of reactive oxygen species. The apoptosis model could also, in principle, explain the problems of stunted growth and neurological deterioration. It might also explain why Cockayne syndrome patients are not prone to skin cancer, even in the face of severe sunlight sensitivity — after all, dead cells do not form tumours.