he gene called p53 is the most frequently mutated gene in human cancers. It starts a process that usually results in programmed cell death, or apoptosis, which keeps potential cancers in check. The big question has been, what triggers this tumor-supression gene? What is it about rapidly dividing precancerous cells—but not normal cells—that brings p53 into play? A likely answer is provided by the two papers at #4 and #5. It seems to be something other than p53 alone, namely the DNA damage response.

Two groups, led by Jiri Bartek of the Institute of Cancer Biology in Copenhagen, Denmark, and by Thanos Halazonetis, now at the University of Geneva in Switzerland, obtained almost identical results implicating the DNA damage response. Bartek's group used immunochemistry to examine different stages from a variety of human tumors. High levels of proteins from four different genes—Chek2, ataxia telangiectasia mutated (ATM), phosphorylated histron (H2AX) and p53—characterized cells from early bladder, breast, colon, and lung cancers. Halazonetis' group found very similar results in lung cancer cells, with elevated levels of Chek2 and H2AX and localized spots of a p53-binding protein (PBP1). All of these genes are part of the DNA damage response.

Cells routinely check the DNA they have copied prior to division. One kind of damage they monitor is a double-stranded break in the DNA. There are two monitoring mechanisms at work. One prompts repair machinery to make good the damage. The other is a checkpoint rather than a proofreader. Some double-stranded breaks presumably cannot be repaired. The cell deals with these by committing suicide, and the effectiveness of this system is manifested in the amazing rarity of cancers, given the billions of cell divisions, with the attendant possibility of damage, that take place over the course of a lifetime. Chek2, ATM, H2AX and PBP1 are components of the DNA damage response, and they precede the activation of p53. These changes are not seen in rapidly dividing normal cells.

Both groups went beyond observation to experimentally induce the changes seen in precancerous cells. Bartek and his colleagues triggered over-expression of an oncogene in a cell culture. Halazonetis' team grafted human skin onto the backs of immune-deficient mice and then induced cancerous changes by injecting the mice with an excess of growth factors. In both cases, the cells showed a clear DNA damage response.

So, what is going on? Both groups propose that cancer will develop only if components of the DNA damage response are inactivated. This would include the ubiquitous p53, but not be limited to it. A consequence of this would be an increase in genetic instability, which in turn would increase the mutation rate and probably speed the development of the cancer. That too is borne out by observation. Both groups note that the double-stranded DNA breaks that they see in precancerous cells are preferentially associated with fragile sites, sequences that the normal cell machinery finds difficult to copy. But that still does not explain how the cell distinguishes the rapid division of precancerous cells from the equally rapid division of normal cells.

Halazonetis refers to "DNA replication stress." Bartek talks of "oncogenic stress." Both are referring to changes that would be associated with "abnormal" division and that would trigger the DNA damage response. But what is the nature of that stress? In a commentary on the two papers, Ashok Venkitaraman, of Cancer Research UK at Cambridge, speculates that a more complex understanding of cell replication than the simple on-off mechanism that dominates current thought might provide some answers (Nature, 434[7035]: 829-30, 14 April 2005). Perhaps, he speculates, cancer-cell cycles could alter the ratios of normal intermediates, such as single-stranded DNA or oxidative changes to the DNA. Or they might result in abnormal structures such as double-stranded breaks. Perhaps precancerous division could even somehow lower the threshold for eliciting the DNA damage response. The search for an understanding of replication stress presumably underlies some of the citation impact of the two papers.