131 corpses and a Nobel prize

As seen in Hilary Ellis' original genetic screen. Top panel, C.elegans with active ced-3; bottom panel, C.elegans with mutant ced-3. Programmed cell deaths occur only with active ced-3 (arrows in top panel). Arrowheads indicate marker cells in both strains. Image courtesy of H. R. Horvitz.

Pruning of cells helps to shape organs, remove a tadpole's tail and carve out fingers or toes. In essence, cell death is necessary to develop gracefully — even for worms. With only 959 cells, a glass-like appearance and a wriggling body, it might seem short of sophisticated features, but over the past two decades the nematode Caenorhabditis elegans has wormed its way into the textbooks by showing how to kill a cell and dispose of the body efficiently.

Scientists had previously noticed dying cells that looked different from necrotic cells in various tissues. On the basis of shared morphological features among these deaths, Alastair Currie, John Kerr and Andrew Wyllie proposed the existence of a conserved, endogenous cell-death programme, which they dubbed 'apoptosis'. But it wasn't until the mid-1980s that Robert Horvitz and colleagues opened their cans of worms to test this idea comprehensively. By that time, John Sulston had mapped out cell lineages in C. elegans (see Milestone 10), and he observed that in every worm, out of 1,090 newborn cells, the same 131 cells die during development, resulting in a nematode of 959 cells exactly.

To identify genes that control the fate of the 131 doomed cells, Horvitz and colleagues screened for worms that — after mutagenesis of their genome — contained 'un-dead' cells (that is, cells that should have died, but survived instead). Tracing the mutations in these worms led them to two genes, ced-3 and ced-4 (called ced for cell death abnormal), which were both essential for cell death, and one, ced-9, which prevented death in cells that needed to survive. These experiments established a genetic basis for programmed cell death. Moreover, they showed that most developmental deaths are cell-autonomous — thereby establishing suicide rather than murder as the cause of death.

The ced-4 gene was found to encode a protein without any known direct relative, but the cloning of ced-3 provided some immediate insights into the cell's suicide gear. The CED-3 protein was a homologue of a newly discovered mammalian protease, indicating that it caused programmed cell death by clipping other proteins. Along with others, Horvitz and colleagues reported that ced-9, which blocks cell suicide, was a functional homologue of Bcl-2 — a gene that was found to be deregulated in follicular lymphoma cells. By now, Horvitz and colleagues were persuaded that programmed cell death, and its molecular machinery, was conserved between nematodes and mammals, and therefore of very ancient origin.

After this, the research on apoptosis snowballed markedly. It soon became clear that in most, if not all, worm cells, the CED-9 protein deters CED-4 from activating CED-3, thereby avoiding accidental deaths. CED-3 was just the first of a series of proteases — collectively called caspases — that were found to be indispensable for efficient apoptosis of animal cells. Also, CED-9 turned out to have many siblings, some of which prevent, whereas others promote, programmed cell death. Many more proteins since have owned up to a role in apoptosis. In a relatively short time we have moved from worm to clinic, as too little or too much apoptosis can result in pathology — not only developmental defects, but also neurodegeneration, autoimmune disease or cancer. Scientific progress frequently happens by serendipity, but here a visionary must have been at play.