DNA double strand breaks do not play a role in heat-induced cell killing.

In their recent article, Takahashi et al. (1) suggest that heatinduced DNA double strand breaks contribute to heat-induced cell killing because heat treatment induces histone gH2AX-containing foci. Such foci have been associated with double strand breaks induced by ionizing radiation, other agents, and other stresses (2). However, the authors disregard the hyperthermia biology literature, which indicates that heat-induced DNA damage is not involved in heat killing. Heat probably does not induce DNA damage directly. The authors point out, however, that heat may induce DNA base damage indirectly via protein damage (1). For repair of ionizing radiation–induced base damage, it is thought that heat inhibits the excision step without impairing the incision step (3). For clustered base damage, such imbalance in incision and excision may result in conversion of base damage into DNA double strand breaks (3). By analogy, the authors speculate that heat-induced base damage may be converted into double strand breaks. However, the levels of base damage do not correlate with the extent of heat-induced cell killing (4). Moreover, although data using the neutral comet assay is presented (1), the use of a variety of sensitive biochemical and biophysical approaches that are more double strand break specific than the comet assay have failed to detect double strand breaks after heat (3). Moreover, Mre11, a protein directly involved double strand break repair that also forms foci immediately after ionizing radiation (5), is not found in foci after heat shock; rather, Mre11 exits from the nucleus after heat shock (6). This clearly shows that the gH2AX foci seen after heat shock are completely different from those seen after ionizing radiation and cannot be taken as evidence for heat-induced double strand break. Several other observations are also inconsistent with the notion that DNA damage is involved in heat-induced killing. Incorporation of BrdUrd into DNA, known to destabilize DNA and to enhance radiosensitivity, does not enhance heat killing. The potential lethal damage response associated with the ability to repair ionizing radiation–induced DNA damage has not been found for hyperthermia. Finally, if heat shock would induce DNA double strand breaks, either directly or indirectly, chromosomal aberrations, associated with the ionizing radiation–induced DNA double strand breaks, would be expected to occur also in cells heated in various phases of the cell cycle, which is not the case (3). The authors nevertheless state that a DNA double strand break repair–deficient cell line is also heat sensitive, arguing in favor of a role for double strand break in heat killing (1). However, a literature survey on 30 different mouse cell lines reveals no general correlation between heat sensitivity and radiosensitivity (Fig. 1A). Because the authors suggest that double strand break only partially determine heat killing, one could still explain the latter by assuming cell line–dependent differences in the sensitivity of other targets of heat (i.e., protein damage) that would mask this correlation. Therefore, we also analyzed data from the literature describing a number isogenic panels of radiosensitive cells deficient in DNA double strand break repair and their normal counterparts for their heat sensitivity. It is highly unlikely that DNA repair gene–complemented cell lines would simultaneously acquire an altered capacity to repair protein damage (the presumed main cause of heat-induced cell death). Thus, if heat would indeed induce DNA double strand breaks relevant to its toxicity, then this should be revealed in such pairs. However, as can be seen in Fig. 1B , there is no relation between double strand break repair deficiency and heat sensitivity in these isogenic panels. Thus, on the basis of these data, one must conclude that even if double strand breaks are induced by heat, they do not contribute to heatinduced killing. It is now known that gH2AX-containing foci can also be induced by other non–double strand breaks–inducing treatments, such as treatment with the methylating agent N-methyl-NV-nitro-N-nitrosoguanidine (7) or exposure to hypertonic buffers (8), showing that the gH2AX phosphorylation and foci formation is a general stress response, e.g., to changes in the chromatin (2), rather than to double strand breaks only. Heat is also known to induce major alterations in chromatin structure, not by inducing DNA damage but as a consequence of heat-induced protein denaturation or aggregation (3). In fact, the new and interesting observations by I2005 American Association for Cancer Research. doi:10.1158/1078-0432.CAN-05-0006 Figure 1. Absence of a correlation between radiation sensitivity and heat sensitivity. A, a cross-correlation of radiation sensitivity (expressed as the dose of X-ray required to kill 90% of the cells) with heat sensitivity (expressed as the equivalent time of heating at 44jC required to kill 90% of the cells) in 30 different mouse cell lines derived from the literature. B, comparison of heat sensitivity between various radiosensitive mutants (deficient in either nonhomologous end joining or in homologous recombination) and their isogenic repair–proficient counterparts. The average sensitivity of the groups (points with bars ) is not different; there is also no trend for increase or decrease in heat sensitivity in matched panels (lines ).