Rudi Tanzi, Massachusetts General Hospital, Boston, opened the meeting with an extremely entertaining "after dinner" lecture on the status of the search for additional genes that may contribute to the development of AD. Aside from continuing to make the case for involvement of alpha-2 macroglobulin, the lecture included hilarious short movie clips illustrating the plight of the genetics researcher attempting to present evidence for the association of his/her favorite gene with AD. I did not know prior to the lecture that there are two distinct and not mutually exclusive possibilities. First, that there may be quite common gene variations that have a very weak effect on the risk of developing AD. Second, that there may be a large number of different gene variations, each occurring in only a small number of people but having more dramatic effects on the risk of developing the disease. Clearly, enough work remains in this area to keep geneticists busy for a while, and to guarantee more controversy in this field.

With some other research areas, it is clear that the major issue lies in separating primary from secondary effects. The involvement of cholesterol metabolism, oxidative stress, and microglial reactivity fit into this category. It is difficult to know whether the work has direct relevance to the pathogenesis of AD, or is describing consequences of the disease process. Opinions obviously differ sharply in these areas, but I remain undecided after the presentations at the meeting.

Perhaps the most provocative session dealt with axonal transport. Presentations by Larry Goldstein, University of California, San Diego, and Eva-Marie Mandelkow, Max-Planck-Unit for Structural Molecular Biology, Hamburg, suggested how tau and AβPP might interact in this area. The urgent need to find functions or mechanisms by which these two crucial proteins in AD might intersect produced some intriguing suggestions. Goldstein described work from his group (see related news, see related news) showing that AβPP may be intimately involved in axonal transport through binding to kinesin light chain. Goldstein suggests that AβPP might link other cargos moving on microtubules to kinesin I, the microtubule motor molecule. The data suggesting that a normal function of AβPP might be to serve as a docking molecule for transport of vesicles on the kinesin 1/microtubule system is quite compelling, and is supported by experiments in Drosophila as well as mammalian cells.

Eva-Marie Mandelkow and her colleagues had previously suggested (see related news) that excess tau, expressed in a variety of cells, would block microtubule binding of kinesin, blocking axonal transport. While it is hard to find evidence for increased expression of tau in AD, the suggestion was made during discussion that perhaps a normal function of tau was to bind microtubules and limit kinesin binding or movement, perhaps serving as a regulator of axonal transport. Many people in this field have expressed reservations about the sole function of tau being to stabilize microtubules. The high turnover rate of phosphate groups on tau, some of which greatly influence the binding to microtubules, appear inconsistent with a mere structural stability role. Why would a large, fully differentiated neuron want to dynamically regulate its microtubule assembly? Such a cell requires a highly ordered, stable system of microtubules to serve as a transport system to move components in axons (and probably dendrites). If the role of tau was to regulate the amount of material moving on the fast axonal transport kinesin 1 system, it would make good sense to be able to regulate this in a very dynamic fashion, and this may happen by tau phosphorylation. If the above is true, an interesting hypothesis is suggested: if abnormalities in AβPP (either mutations or altered cleavage) were to occur, fast axonal transport may be disrupted. One compensatory mechanism might be to increase the phosphorylation level of tau to reduce hindrance of the kinesin 1 movement. This phosphorylation would serve to dissociate tau from the microtubule, and render tau perhaps more susceptible to aggregation. If this was insufficient to maintain adequate axonal transport of synaptic components, a "dying-back" process may begin, in which the more distant synaptic terminals die first. Synapse loss is certainly a major feature of AD brain, and may be directly linked to the clinical dementia.

While highly speculative, this hypothesis has two virtues. First, it ties together abnormalities of AβPP processing and tau phosphorylation in a new way. Second, it is testable in a variety of systems from flies to mice. The idea that tau and AβPP may function in a coordinated way in normal cells to regulate axonal transport would suggest several scenarios for the development of Alzheimer's disease.

This is the most interesting new hypothesis I have heard in quite a while. The small meeting environment is especially conducive to the kinds of discussion that generate such ideas, and their expression is less intimidating in small groups than it can be in larger forms. A dinner cruise on the Elbe also helped getting the group to talk, although the topics of conversation there were less esoteric than at the conference. - Peter Davies, Albert Einstein College of Medicine, Bronx, New York.

Reference: Conference web page