This work represents another tour de force by ShiDu Yan and her laboratory. In this paper, the investigators isolated synaptic and non-synaptic mitochondria from transgenic mice overexpressing mutant human APP. As compared to the non-synaptic mitochondria, the synaptic mitochondria contained more β amyloid (Aβ) and were more impaired. Synaptic mitochondria were smaller and less likely to migrate through axons. They had lower coupled oxygen consumption, reduced cytochrome oxidase activity, and produced more reactive oxygen species. Functional perturbations were seen well before plaque accumulation. In related experiments, Aβ was shown to directly impair mitochondrial movement through axons. The authors conclude Aβ toxicity is mediated through its effects on mitochondria.

The ability of Aβ to affect mitochondrial function, at least under in vitro conditions, was demonstrated well over a decade ago. The idea that mitochondria mediate Aβ's cell toxicity was first shown by Sandra Cardoso (see Cardoso et al., 2001). Colocalization of...  Read more

This work represents another tour de force by ShiDu Yan and her laboratory. In this paper, the investigators isolated synaptic and non-synaptic mitochondria from transgenic mice overexpressing mutant human APP. As compared to the non-synaptic mitochondria, the synaptic mitochondria contained more β amyloid (Aβ) and were more impaired. Synaptic mitochondria were smaller and less likely to migrate through axons. They had lower coupled oxygen consumption, reduced cytochrome oxidase activity, and produced more reactive oxygen species. Functional perturbations were seen well before plaque accumulation. In related experiments, Aβ was shown to directly impair mitochondrial movement through axons. The authors conclude Aβ toxicity is mediated through its effects on mitochondria.

The ability of Aβ to affect mitochondrial function, at least under in vitro conditions, was demonstrated well over a decade ago. The idea that mitochondria mediate Aβ's cell toxicity was first shown by Sandra Cardoso (see Cardoso et al., 2001). Colocalization of APP and Aβ with mitochondria was shown soon after this, and this phenomenon has now been confirmed by several groups studying both transgenic mice and human AD subject brains. Subsequent work evaluating the specific aspects of the mitochondria-Aβ nexus suggest Aβ may have protean effects on the mitochondria. Indeed, Yan's group previously reported that Aβ interacts with the amyloid-binding alcohol dehydrogenase (ABAD) and cyclophilin D.

Like all good papers, this one is notable for raising as many questions as it addresses. Did the increase in synaptic mitochondrial Aβ (as opposed to cell body mitochondrial Aβ) reflect a specific functional or structural difference between these two different mitochondrial populations? Did it arise as a consequence of differences in the cell body and synapse environment? Was it simply due to proximity of the synaptic mitochondrial pool to areas of Aβ production? Also, these experiments utilized a primary amyloidosis model. The Moraes lab has addressed the opposite question of whether altered mitochondrial function influences APP processing in transgenic mice that overproduce Aβ, and found that it does. In that work, knockout of a gene needed to assemble a functional cytochrome oxidase actually reduced plaque formation (Fukui et al., 2007). The bigger picture view of the findings mentioned above will certainly make more sense as the context around them grows. In the meantime, more and more would agree the mitochondria-Aβ nexus has become an important and fruitful area of AD research. This is consistent with multiple lines of investigation that increasingly suggest a central, and in some cases possibly very upstream, role for mitochondria in AD.

View all comments by Russell Swerdlow