We study the molecular mechanisms of autophagic cell death on a very advantageous model, the protist Dictyostelium discoideum. In D. discoideum, autophagic cell death is developmental, but can be mimicked under in vitro monolayer conditions.

The approach is mainly genetics, random insertional mutagenesis, taking advantage of this haploid organism, and now leading to a definition of the molecular mechanisms involved. It is likely that the main features of these mechanisms are conserved in higher eukaryotes.



Dictyostelium cells in monolayers in vitro lend themselves well to a study of autophagic cell death (ACD). There is no apoptosis machinery in the protist Dictyostelium, no caspase nor Bcl-2 family members (except a para- caspase whose inactivation does not alter cell death), thus there is no apoptosis that could interfere with, or substitute for, nonapoptotic cell death. Also, Dictyostelium, a eukaryote, has a haploid genome, which facilitates random insertional mutagenesis.

Autophagy often protects cells from cell death. However, in a growing number of cases of cell death there are not only concomitant signs of autophagy, but also a causative relationship between autophagy and cell death: Preventing autophagy prevents cell death. How can autophagy in some cases protect from cell death, but in other cases cause it? In other words, what is the relationship between protective autophagy and destructive ACD? It is often believed that this relationship is quantitative: A little autophagy protects, whereas a lot leads to death. However, at least in the Dictyostelium model, the relationship is qualitative: On top of autophagy, a second signal, quite distinct from what causes autophagy, leads to cell death. We believe such a qualitative difference between autophagy and ACD, namely such a second signal (not necessarily molecularly but functionally similar) might exist also in higher eukaryotes. There, it may still be hidden in complexity, or it may already be exposed for all to see, as may be, for instance, the case for hormones regulating ACD in metamorphosis.
In the Dictyostelium model, the second signal is DIF-1, a main differentiation factor of known structure. Dictyostelium cells, when starved in the presence of cAMP, undergo autophagy and become sensitive to DIF-1. Addition of exogenous DIF-1 then pushes the starved cells to undergo ACD, that is, to show paddle cell formation, actin depolymerization and rounding, cellulose shell formation, vacuolization and plasma membrane rupture. Starvation and autophagy do not lead to death by themselves, but they sensitize the cells to the DIF-1 second signal. DIF-1 is unable to trigger death of nonstarving cells, but it triggers death of starved cells undergoing autophagy.

A number of cell death “second signal” mutants are being obtained by random insertional mutagenesis and selection on the basis of resistance to cell death. These mutants have common characteristics: They are conditional (showing a mutant phenotype not during vegetative growth, but upon induction of death), they do not impair autophagy, but they impair ACD, they are causatively organized (each mutation prevents all downstream events, and these mutations can be ordered according to the step blocked in the ACD cascade), and they exert control rather than direct effector activity (blocking the subcellular lesions of ACD, without directly being part of their machinery). More of such mutants are being obtained, to enrich our knowledge of the corresponding molecular machinery, which may at least in part be conserved in higher eukaryotes.