Defects in homologous recombination repair in mismatch repair -deficient tumour cell lines

MMR -deficiency increases the rate of mutations and often sensitizes cells to DSB-inducing agents (e. g. camptothecin and etoposide) as well as MMC (Jacob et a/., 2001 and Fiumicino et al., 2000). MMR -deficient tumour cell lines are also sensitive to the cytotoxic effects thymidine (Mohindra et al., 2002). This sensitivity is not a direct consequence of MMR -deficiency or alterations of DNA precursor metabolism. Instead, the results described in the present study suggest that MMR -deficient cells are sensitive to thymidine as a result of defects in HRR. The ScNeo recombination reporter substrate was used to determine the integrity of the HRR pathway in several MMR -proficient and -deficient tumour cell lines. Four MMR -deficient tumour cell lines were defective in the production of neo+ recombinants by homology based recombination following the transient expression of a site specific break. Furthermore, all MMR -deficient tumour cell lines tested were sensitive to the cross-linking agent MMC; an effect that is consistent with cells being deficient in HRR (including XRCC2, XRCC3 and BRCAI). To determine the alterations responsible for such HRR defects, genes known to be required for this pathway were screened for mutations in eight tumour cell lines. This revealed a heterozygous frameshift mutation within the RAD51 paralog, XRCC2, (342deIT) in SKUT-1 cells. 342delT was introduced into HRR proficient cells containing the ScNeo substrate. In SW480/SN. 3 transfectants, expression of 342delT conferred sensitivity to thymidine and MMC and suppressed HRR induced at the recombination reporter by thymidine but not by DSBs. In the MRC5VA/SN. 13 transfectants, expression of 342delT was accompanied by a decreased level of the full-length XRCC2. These cells were defective in the induction of HRR by either thymidine or DSBs. Thus 342delT suppresses recombination induced by thymidine in a dominant negative manner while recombination induced by DSBs appears to depend upon the level of wild-type XRCC2 as well as the expression of the mutant XRCC2 allele. These results suggest that HRR pathways responding to stalled replication forks or DSBs are genetically distinguishable. They further suggest a critical role for XRCC2 in HRR at replication forks, possibly in the loading of RAD51 onto gapped DNA.