We seek to answer two questions: how do neurons become connected during development, and why do they become disconnected during neurodegenerative disease? We have developed a Drosophila model of mammalian neurodegenerative disease by inactivating the protein kinase Cdk5. This is the fly homolog of one of the two main proteins responsible for phosphorylating tau into the form found in the neurofibrillary tangles that are characteristic of many forms of human neurodegeneration. This year, we published data documenting many phenotypic similarities between the neurodegenerative syndrome in this endogenous process of flies and that observed in humans and mice, thus validating our Drosophila mutant as a valuable model of the mammalian disease process. We also identified two pathological processes that occur early in disease progression in our fly model that had not previously been described in studies of the mammalian diseases. One is failure to segregate the portion of the neuron where action potentials initiate (the axon initial segment), the other is altered regulation of neurite integrity, as assayed by defects in the rate and onset of developmentally programmed dendritic remodeling. Both of these are good candidates for processes that could play an important role in mammalian disease. A key question for us, therefore, is to determine their importance in the overall process of neurodegeneration. As a first step we have now begun to dissect the molecular mechanisms underlying each of these phenotypes. A key component of the axon initial segment is the scaffolding protein ankyrin. We find that one isoform of the neuron-specific ankyrin, localizes to the AIS and that Drosophila mutants lacking Cdk5 activity fail to show this localization. We also find that ankyrin is apt to be important for Cdk5-dependent AIS specification, since lowering or raising Ank2 expression mimics, respectively, the Cdk5 gain and loss of function AIS phenotypes. This strongly suggests that the effect of Cdk5 on AIS formation and maintenance is apt to be mediated, in part, through its regulation of Ankyrin 2. Regarding dendritic stability, a central question is to establish at what point in the process of developmental remodeling Cdk5 activity plays a regulatory role. Dissolution of the microtubule cytoskeleton is the earliest known, and probably the rate-limiting, event in dendrite pruning. Our data now show that Cdk5 acts upstream of microtubule disassembly in dendrite pruning. This would make Cdk5 activity the earliest, initiating event in dendrite disassembly. A key test of this hypothesis is to demonstrate whether microtubule stability is indeed epistatic to Cdk5 activity in remodeling. To this end, we have developed and published a novel method that supports developmental remodeling of the developing Drosophila brain in an organotypic culture. By allowing both live monitoring of the remodeling process and access for pharmacological manipulation of the brain, this technique provides us with unparalleled temporal resolution for dissecting the sequence of events before and during Cdk5-associated neurite disassembly. We are currently applying this method to validate or falsify the hypothesis that regulation of dendrite integrity is a central, direct function of Cdk5 in vivo.
