Cancer therapy often involves treatment with cytotoxic drugs and/or ionising radiation. The potentially toxic DNA lesions induced can be counteracted and corrected by cellular DNA repair mechanisms. Consequently, DNA repair enzymes might be promising targets for anticancer drugs that could be used to enhance the biological effects of cytotoxic agents. We have characterised different DNA repair pathways in a long-term project, to provide better understanding of the cellular defence mechanisms against damage to the human genome.
A new DNA repair-based model for the cytotoxic action of the chemotherapeutic drug, 5-fluorouracil
Excision of uracil and derivatives occurring as aberrant bases in the mammalian genome demonstrates how DNA repair can militate against malignant transformation but also modulate chemotherapy based on DNA-damaging cytotoxic drugs. Mammalian cells contain two uracil-DNA glycosylases, the ubiquitous and highly conserved Ung enzyme, and Smug, which has only evolved in insects and vertebrates. We have previously reported that both Ung and Smug suppress spontaneous C→T mutability in mammalian cells - excising uracil from U:G lesions arising by hydrolytic deamination of cytosine - whereas Ung uniquely processes U:G mispairs generated enzymatically by AID (activation induced cytosine deaminase) during antibody gene diversification in activated B cells, with Ung-deficient gene-targeted mice developing B-cell lymphomas. We have shown that it is Smug, and not Ung, which excises the fluorine-substituted uracil derivative, 5-fluorouracil (FU), incorporated in DNA, such that Smug deficient cells are uniquely hypersensitive to this chemotherapeutic drug. FU is used in some two million patients a year, improving survival of breast, head-and-neck, aerodigestive and particularly colorectal cancer, where combination FU chemotherapy is a front-line treatment. Our data disprove a widely cited but unsubstantiated model by demonstrating that FU incorporation into DNA is a predominant cause of drug cytotoxicity. Furthermore, Smug over-expression decreases cellular FU sensitivity and up-regulation of Smug may provide a previously unrecognised mechanism of acquired drug resistance in tumours, acquired and de novo chemo-resistance being the major obstacle to successful FU-based cancer therapy. We now aim to analyse SMUG activity versus survival during FU treatment of patients and as a putative biomarker of initial drug response. These studies will be supported by micro-array gene-expression profiling of FU-responsive/resistant tumours (Elaina Collie-Duguid, Institute of Medical Sciences, University of Aberdeen), and collaborations with clinical oncologists at both Aberdeen Royal Infirmary and Mount Vernon Hospital.
A single-stranded by-product of lagging-strand DNA synthesis is degraded by the Trex1 DNA exonuclease in mammalian cells to prevent autoimmune disease
Trex1 is only found in mammals where it is the major 3'→5' DNA exonuclease. We generated TREX1 null mice which serve as a model of the human autoinflammatory disorder, Aicardi-Goutières syndrome (AGS), a recessive genetic mimic of acquired viral infection which can be due to mutations in TREX1 or any of the three genes encoding subunits of the RNaseH2 holoenzyme. The mechanism of Trex1-deficient disease remains unclear, although a link to endogenous retroelements has been suggested. However, we have shown that Trex1, ordinarily associated with the endoplasmic reticulum, relocalises to the nucleus in S phase and identify the in vivo substrate of Trex1 as a discrete 62mer single-stranded polynucleotide by-product of DNA replication that accumulates outside the nucleus, where it could be perceived as 'foreign' and so provoke the antivirallike autoimmunity characteristic of Trex1-deficient disease. Furthermore, we have now identified a common nucleic acid substrate for Trex1 and RNaseH2 during lagging-strand DNA replication that might underscore this genetically heterogeneous disease.
Repair of alkylated DNA by human ABH and FTO proteins
Methylating agents occur endogenously and in the environment, and also due to their cytotoxicity are used in cancer treatment. They methylate DNA bases generating lesions that block DNA replication or cause mutations. DNA repair activities that specifically revert some of these lesions reduce the cytotoxicity of the methylation damage. These activities are conserved from bacteria to human cells. We previously characterised E. coli AlkB and human ABH2 and ABH3 proteins that directly demethylate 1-methyladenine and 3-methylcytosine, major alkylation lesions that are generated in single stranded DNA. These enzymes are 2-oxoglutarate Fe2+-dependent dioxygenases that oxidise the aberrant methyl groups resulting in their destabilisation and release as formaldehyde. More recently we have identified an additional seven human homologues of AlkB (ABH1, ABH4-8 and FTO). Sequence variants in the first intron of the FTO gene have been strongly associated with obesity in several populations. In collaboration with the groups of Stephen O'Rahilly (Cambridge) and Christopher Schofield (Oxford), we have defined a biochemical activity of human FTO. The FTO protein is a dioxygenase that demethylates 3-methylthymine (but not 1-methyladenine) in single stranded DNA. 3-methylthymine is a rare DNA lesion. Using an assay that measures conversion of the cosubstrate 2-oxoglutarate to succinate, FTO was shown to specifically interact with pyrimidine nucleosides methylated at the 3¿ position, including 3-methyluridine.
A working hypothesis is that FTO regulates transcription or translation by demethylation of an as yet unspecified RNA. In an ongoing collaboration with Laurence Colleaux (Paris) and Stephen O'Rahilly, an Arab family carrying a mutation within the FTO coding sequence was identified. The mutated protein is inactive in our in vitro enzyme assays. Siblings that are homozygous for this variant have a severe developmental syndrome. These findings provide the first example of a human disorder due to a functional defect in an AlkB-related dioxygenase and reveal an unpredicted role for this protein in development.
Figure 1. Demethylation of 3-methylthymidine in single stranded DNA by FTO. FTO requires the cofactor Fe2+ and the co-substrates O2 and 2-oxoglutarate to catalyse hydroxylation of its substrate. The proposed intermediate degrades releasing the hydroxylated methyl group as formaldehyde and the unmodified nucleoside thymidine is directly regenerated.
For a list of refereed research papers, see Publications (in navigation on left).
