The ��-bromoacryoyl moiety has been proposed to be important since it reacts opposite with GSH, and reactive drug-GSH intermediates may then modify the DNA by mechanisms not yet fully understood (Geroni et al, 2002; Cozzi, 2003). DNA interaction data reported in the present study suggest that the distamycin A backbone drives the brostallicin molecule towards the AT regions present in the minor groove of the DNA. In addition, brostallicin binds covalently to DNA through interaction with the GSH/GST system. Brostallicin covalently binds to DNA with a completely different sequence specificity than tallimustine. One hypothesis for the different behaviour of brostallicin against MMR status is that this covalent interaction is not substrate for MMR, whereas the alkylated DNA by tallimustine is recognised by MMR.

It should be noted that no direct interaction between MMR and the GSH/GST system is known, and that the GSH/GST status of the cell lines under study does not matter for the experiments because the cell lines are quasi-isogenic, that is, they differ only in their MMR status and the extra chromosomes. Moreover, as reported for PNU-151807, the bromoacryloyl moiety seems to be relevant for cell cycle checkpoint control (Marchini et al, 1999). The identity of mediators for signalling between DNA damage and downstream effectors is not clear. One possibility is that the DNA damage is recognised by one or several members of BASC (BRCA1-associated genome surveillance complex), a multiprotein complex including BRCA1, ATM, MMR proteins, and other proteins implicated in DNA repair (Wang et al, 2000).

Our data, however, show that deficiency in ATM or DNA-PK did not affect brostallicin sensitivity in p53-deficient cells, arguing against a role of these kinases in these cells. Since these kinases are activated upon DNA double-strand breaks introduced by radiation or radiomimetic drugs (Jackson, 1997; Smith et al, 1999), ��-bromoacryoyl derivatives seem unlikely to produce this type of lesion. Although the cytotoxic effect of tallimustine and PNU-151807 has been shown not to be dependent on the p53 status (Marchini et al, 1999), the data for these kinases obtained in p53-deficient cells may not be conclusive for p53-proficient cells. There is an apparent higher sensitivity to brostallicin of the DNA-PK data set compared to the ATM data set, but this is likely due to the use of two assays that differ in their sensitivities. Mutations in the p53 Cilengitide tumour suppressor gene are found in a large fraction of human cancers (Hollstein et al, 1991) and this may be the genetic basis underlying failure to respond to chemotherapy (Ferreira et al, 1999).