A dysfunctional harlequin

Mitochondrial dysfunction has been implicated in several neurodegenerative disorders, including Alzheimer's and Parkinson's diseases. The changes in mitochondrial function can lead to increased energy production and oxidative stress, and subsequent apoptosis. Reporting in Nature, Susan Ackerman and co-workers now describe a genetic model for neurodegeneration that is mediated by oxidative stress. And, in so doing, they highlight a vital role for a key pro-apoptotic molecule — the apoptosis-inducing factor (AIF) — in neuronal survival.

The authors were investigating a late-onset neurodegenerative mouse model called harlequin (Hq). The mutant mice showed a progressive loss of granule cells from the cerebellum, and TUNEL staining showed various characteristics of apoptosis in these cells, including nicked DNA, chromatin condensation and blebbing. There was also a progressive degeneration of retinal cells in the Hq mutant mice.

Ackerman and co-workers genetically mapped the Hq mutation, and showed that it is due to a proviral insertion in the Aif gene. This leads to an 80% reduction in Aif messenger RNA and protein relative to wild-type levels. Under normal physiological conditions, AIF is found in the mitochondrial intermembrane space. Here it acts as an oxidoreductase, a group of molecules that have been implicated in maintaining free-radical homeostasis. So the authors compared the levels of antioxidant enzymes (catalase and glutathione), lipid peroxidation and DNA oxidative damage between Hq and wild-type mice. They observed increases in the levels of all of these factors in the Hq mice, indicating that a loss of AIF function might lead to increased oxidative stress.

As the crystal structure of AIF is similar to that of glutathione reductase — an enzyme that is involved in recycling of glutathione, and a potent scavenger of hydrogen peroxide (H₂O₂) — the authors examined the H₂O₂-sensitivity of Hq granule cells. They found that these cells were more sensitive than wild type to both exogenously and endogenously generated H₂O₂. But when mutant granule cells were infected with retrovirus containing wild-type Aif sequences, the susceptibility to H₂O₂ was rescued. Moreover, overexpression of AIF in wild-type neurons resulted in a decreased sensitivity to H₂O₂.

Finally, abnormalities in the cell cycle have been observed in degenerating neurons, so the authors examined cell-cycle control in the Hq mice. A series of experiments showed that granule and retinal cells from Hq mice re-enter the cell cycle aberrantly, before they die by apoptosis, supporting the idea of a link between cell-cycle re-entry and oxidative stress.

How do these results square with AIF's known function as a pro-apoptotic molecule? The authors propose that AIF normally acts, either indirectly or directly, as a free-radical scavenger in the mitochondrial membrane — it prevents oxidative stress by mopping up H₂O₂ in particular. Under conditions that induce apoptosis, however, AIF translocates to the cytoplasm and nucleus, where it promotes chromatin condensation and other features of apoptosis. As well as shedding light on the functions of AIF, then, this study has pinpointed the first in vivo model for studying how oxidative stress might affect cell-cycle re-entry and apoptosis.