Once a cell has received a death signal and makes the molecular decision to commit suicide, the killing is carried out in a stepwise fashion by, in most but not all cases, members of the Bcl-2 family, release of cytochrome c and other factors from the mitochondria and, in all cases, activation of the caspase family of proteases. Caspases dismantle the cell and also activate other proteases to aid in the execution. Once the deed is done, the dead cell's neighbors engulf the cell corpse.

Engulfment in C. elegans is regulated by a set of seven genes, some expressed by the engulfing cell, others by the suicidal one. Mutation of any one of these genes results in a pile-up of cell corpses or the production of "undead" cells - those that die in wild-type, but survive in mutant, animals. The link between engulfment and "undeadness" has recently been revealed by two papers [1,2] showing that the engulfing cells do not simply dispose of the corpse but actively participate in the killing. Even more striking is the observation that a small percentage of target cells in several engulfment mutants, such as ced-7, ced-6 and ced-1, show some morphological signs of apoptosis but then apparently change their mind and survive as if they had never contemplated suicide.

In order to investigate the extent to which a cell can go down the apoptotic path and then revert, the Hengartner group [1] described the stages of apoptosis as progressing from the "ring" to the "erythrocyte" to the "lentil" stage, which is followed by engulfment. The death defects in engulfment gene mutants are varied – some of the cells do actually die and persist as corpses, whereas others progress to the ring or erythrocyte stages but then revert. Still others never initiate the apoptotic program at all, suggesting that the genes involved in engulfment signal to the target to either initiate or terminate the death process.

It appears that the apoptotic machinery might also signal the engulfer, for there is an increased number of corpses in animals that have a defective caspase (ced-3), and this is even higher in animals with mutations in both ced-3 and in an engulfment gene. Moreover, a few of the cells in a single ced-3 mutant persisted as corpses, whereas others progressed through the erythrocyte stage and then doubled back and survived. How a cell can activate and then inactivate caspases is a mystery.

The obvious, but unprecedented, conclusion from these studies is that there is communication between the suicidal cell and the engulfer early on in the death process that actually determines the apoptotic fate. It is suggested that this serves as a sort of cellular peer-pressure applied by the engulfer to assure that the suicidal cell completes its mission. This is supported by the observation that, in a very few ced-3 and double ced-3 ced-7 mutants, cells were engulfed before they reached the lentil stage, indicating that the "eat-me" signal had been received and processed by the engulfer. The details of the conversation between the dying and engulfing cell are still unclear, and it is not known whether similar relationships exist in higher animals, but the potential for "cell-death-reversion therapies" ensures that this topic will be the focus of much attention.