We have transgenically engineered C. elegans to express, in body wall muscles and neurons, the human beta amyloid peptide (Abeta) associated with Alzheimer's disease. Induction of Abeta in muscle cells (via a
smg-1-dependent expression system developed in the Fire lab) leads to rapid paralysis. Using this Abeta toxicity model, we have previously identified Abeta-interacting proteins by co-immunoprecipitation/mass spectrometry, and characterized Abeta-dependent gene expression by microarray analysis. We are now undertaking a variety of approaches to sort through these identified proteins and genes to determine which are directly related to Abeta toxicity, and to understand how Abeta expression leads to cellular dysfunction. One informative approach has been to examine control transgenic strains that mimic aspects of Abeta-expressing strains. For example, strain XA1440 (kindly provided by Creg Darby) expresses the YpkA bacterial toxin using the same
smg-1-dependent expression system as the Abeta-expressing strains, and shows a paralysis phenotype upon transgene induction indistinguishable from that of induced Abeta animals. Using quantitative RT-PCR, we have asked which genes upregulated in Abeta-expressing animals are also up-regulated in induced XA1440, to identify genes that are generally responsive to muscle pathology and thus not specific for Abeta toxicity. Surprisingly, many genes upregulated before the onset of paralysis in Abeta-expressing animals are also induced in XA1440, suggesting that significant muscle pathology occurs in induced Abeta animals before any cellular changes are observable by light microscopy. We have therefore used high pressure freeze electron microscopy to examine induced Abeta animals before paralysis onset. We find that induced Abeta animals have a subset of ultrastructurally abnormal mitochondria 12 hrs after Abeta induction, preceding paralysis by at least 12 hrs. These abnormal mitochondria are only observed in (Abeta-expressing) body wall muscle cells, and not in other tissues or in any control strains, including induced XA1440 animals. This specific ultrastructural defect is particularly intriguing, given the extensive literature associating Alzheimer's disease with deficits in glucose utilization and mitochondrial function.