Organized by Israel Hanin, Abraham Fisher, and Ramon Cacabelos, the 6th International Conference on Alzheimer’s and Parkinson’s Diseases was a great success. The number of registered participants reached over 1,200, a significant increase in just two years that reflects the recent surge of interest in AD and PD. This year, the conference was held on 8-12 May in Seville, a lovely southern Spanish city that blends historical riches, including the king’s palaces right next to the conference hotel, with the vibrant life of a modern town. Queen Sofia of Spain presided over the opening ceremony, and participants met socially for an evening, among others, of flamenco, sangria, and tapas.

Our report is by no means comprehensive and, regrettably, focuses on AD as parallel sessions forced a choice between the two diseases. Even within AD, many excellent presentations could not be covered in a conference this large. We encourage colleagues to complement our spotty account by clicking the Submit Comment button above and writing in with their notes of presentations they found particularly thought-provoking.

Speakers included:

Gerard Schellenberg
Lars Lannfelt
Colin Masters
Peter St. George-Hyslop
Bart de Strooper
Christian Haass
Rita Reifegerster
Konrad Beyreuther
Tobias Hartmann
Dora Kovacs
Takeshi Tabira
Martin Citron
Fred Checler
Abraham Fisher

Gerard Schellenberg reviewed the function of tau mutations. In a simplified picture, he grouped the approximately 30 known mutations into two categories, namely, mutations that change tau protein function and mutations that alter tau splicing. Importantly, a number of mutations inside the coding region also may alter tau splicing by altering the binding to splice factors (see Reed presentation in ARF related news story). Schellenberg then presented data on the lab’s experiments using C. elegans as model animal. C. elegans has endogenous tau, which after development is expressed in only six neurons. After expressing human tau in GABAergic neurons, a dramatic decrease in thrashing (movement) was observed, and this decrease was further aggravated in roundworms transgenic for mutated tau. While this did not generate NFT, the scientists found, in collaboration with John Trojanowski, tau aggregation as well as neuronal degeneration. From this, Schellenberg concluded that tau can trigger neuronal degeneration before it forms insoluble aggregates, possibly as dimers or small oligomers.

Lars Lannfelt summarized the lab’s results on the arctic mutation. Together with the Dutch, Flemish, Iowa, and Italian mutations, the arctic mutation belongs to the group of Aβ mutations that cluster close to the α secretase cleavage site (see ARF related news story). The arctic mutation causes dementia with congophilic angiopathy and ring-like amyloid deposits without the typical senile plaque core. Even though strokes are commonly with cerebral amyloid angiopathy, arctic mutation carriers don’t suffer strokes, and Lannfelt ranked this dementia to fall clearly into the group of familial forms of AD. Citing his older data, he raised the point that they found an about equal decrease in total Aβ40 and Aβ42 level in the conditioned media from transfected SH-SY5Y cells. Since Aβ peptides with the arctic mutation show enhanced protofibril formation, Lannfeldt hypothesized that the reduction in secreted Aβ could result from impaired secretion and accordingly increased levels of intracellular Aβ. They are currently investigating this idea. An alternative idea would be that the pathology is due to the stronger toxicity of protofibrillar (arctic) Aβ, (see ARF related news story).

Colin Masters expressed the view that soluble Aβ is the toxic entity, while fibrillar Aβ is "wrapped up, sealed, and not a problem." What are the factors that drive the equilibrium between these states? Summarizing recent and earlier data, he finds that bivalent cations are the most important factor. Using clioquinol, a small hydrophobic antibiotic/chelator, in a clinical trial supported by the Australian biotech firm Prana, Masters and colleagues observed increased plasma Aβ levels in the control and decreased Aβ levels in the treated group. This effect was significant in the more severely affected AD patient group. Interestingly, clioquinol treatment also protected mice against the Parkinson’s disease-inducing agent MPTP, where it facilitated dissolution of α-synuclein aggregates (see related news story (see ARF related news story and related news story).

By contrast, Peter St. George-Hyslop highlighted the idea that vaccination designed to convert Aβ aggregates/protofibrils to soluble, less toxic Aβ may be beneficial. This was based on his finding that vaccinated transgenic mice that don’t show any decrease in brain Aβ levels show improved cognitive function. When he tested the respective antibodies in cell cultures and EM, he found that the antisera generated were actually active this way (main epitope 4-10).

Bart de Strooper showed that the PS Asp mutations are completely inactive in the double PS1/PS2 knockout background, and that inhibition of nicastrin maturation has no effect on γ-secretase activity. Therefore, glycosylation of nicastrin is only a marker for maturation of the protein. A simple sum of the four known γ-secretase components does not necessarily add up to the observed 250-440 kDa size of the mature complex, suggesting that more than one type of complex is active in cells. Indeed, already six different complexes can be obtained by different isoforms of the known components. This variation of protein components can also yield tissue specificity of γ-secretase activity. Chronic inhibition of γ-secretase in adult (PS1,2 +/-, -/-) mice results in increased β-catenin levels and appearance of nonmalignant keratosis, suggesting a specific link to WNT signaling. Similar data have previously been reported by Hui Zheng on PS1 knockout mice rescued through neuronal expression of human PS1 transgene (see ARF related news story). Severe autoimmune phenotype was also observed, likely attributable to deficient Notch-1 signaling (see also de Strooper section in ARF related news story).

Christian Haass extended the above finding by that PS Asp mutations do not give raise to proteolytically active PS-complexes by showing that, in the yeast background, these mutations prevent the generation of PS-NTF/CTF fragments, as well. Deletion scanning the nicastrin protein rendered it trypsin-sensitive, inhibited maturation, and did not result in the formation of an active γ-secretase complex. From this, Haass concluded that nicastrin undergoes a conformational change during maturation, and that this change was independent of glycosylation. Evaluating the regulation of the expression level of the four identified γ-secretase complex proteins, he found that whenever one component is missing, all three remaining proteins are downregulated. Expression of the four components is sufficient to obtain in vitro Aβ and AICD generation in yeast, which normally lacks γ-secretase activity (see also Edbauer et al., 2003).

In collaboration with Roger Nitsch, Rita Reifegerster examined the effect of AβPP, BACE, and presenilin (PS), and combinations thereof, in GAL4-transgenic Drosophila melanogaster. When expressed in the retina, AβPP yields a 5.5 kDa fragment, which is also known as Aβ, and is generated by γ-secretase and δ-secretase cleavage of a few amino acids N-terminal of the β-cleavage site. When coexpressed with BACE, dimeric Aβ peptides are observed and Aβ disappears. Ubiquitous expression results in wing vein formation and a drastic reduction in viability. This can be altered by PS FAD mutations, reduced Drosophila PS levels, or exposure to γ-secretase inhibitors.

Konrad Beyreuther showed that the use of the Aβ42/Aβ40 ratio in CSF results in a highly sensitive and specific predictor for the diagnosis of AD. Following the observation that it is possible to generate covalently linked Aβ dimers from covalently linked AβPP, Beyreuther concluded that the γ-secretase complex must be present as a dimer complex in which each subunit performs one of these cleavages. Reporting on a recent double-blind, placebo-controlled prospective pilot clinical study, he showed data suggesting that a six-month course of high-dose simvastatin in patients with mild AD results in an increase in cognitive performance and a parallel decrease in CSF Aβ levels and 24S-hydroxycholesterol, or cerebrosterol (see ARF related news story).

Tobias Hartmann presented data suggesting that the β- and γ-secretase cleavages are independently regulated by cholesterol levels. When the Hartmann lab studied lipid homeostasis in ’s PS double knockout cells, they found that presenilin 1 is an essential protein for the regulation of cellular lipid levels. PS2 was fully able to replace PS 1 in this function, at least for certain individual lipids such as ceramide, glucosylceramide, and lactosylceramide. Hartmann concluded that if one of the biological functions of PS is related to lipid homeostasis, this might provide the missing functional link between lipid effects on Aβ production, epidemiological data on statin use, the clinical trial mentioned above, and AD pathology. (See also ARF live discussion).

Dora Kovacs extended her findings on the dependence of AβPP processing on the balance between membrane cholesterol and cellular cholesterylesters (CE). After an in-depth analysis that resulted in the discovery of a CE-dependent cleavage site in AβPP, she concluded that Aβ reduction is due to bypassing of the α, β-, or γ-secretase cleavage pathways. Possibly, this is achieved through the binding of an unknown protein to the N-terminus of AβPP, in a membrane cholesterol-CE dependent manner. To validate the in-vivo relevance of these studies, Kovacs and colleagues tested an ACAT inhibitor that is toxic in humans, but not in mice. Together with Manfred Windisch at the JWS research labs in Graz, Austria, they found a drastic decrease in Aβ levels in AβPP-transgenic mice and a trend toward improved cognition (scroll to Kovacs entry in ARF related news story).

Takeshi Tabira analyzed the composition of cotton wool plaques in the brains of six AD patients with PS1 mutations who also presented symptoms of spastic paraparesis (i.e., weakness and mild paralysis of the legs). Aβ40 was largely absent, replaced by Aβ42, but without forming characteristic amyloid fibrils. The unusually large size of cotton wool plaques is primarily due to the accumulation of synaptic structures, mainly presynaptic organelles. This suggests, again, that PS1 has a role in synaptic function. Indeed, transgenic mice expressing PS1(L286V) develop an abnormal dendritic tree, and also present altered short- and long-term potentiation. In spite of an increase in the number of synapses, these mice show symptoms of decreased synaptic function, perhaps due to elevated levels of Aβ42.

Martin Citron summarized the current knowledge about BACE 1 inhibition as a therapeutic approach for AD. Although BACE 1 cleaves at least two more substrates in addition to AβPP, no sign of pathology or altered gene expression is detectable in BACE KO mice. Interestingly, even aged BACE KO mice are healthy, despite complete absence of Aβ generation. Based on these data, he suggests that BACE is currently a suitable target. The principal problem is the generation of specific BACE inhibitors.

Fred Checler showed data suggesting that exposure of the compound JLK2 to cells results in reduced Aβ levels without an accompanying decrease in mouse Notch 1 processing (Petit et al., 2002). To confirm this result, Checler exposed C. elegans to JLK2 and did not observe developmental alterations.

Abraham Fisher showed the results of his highly M1-selective agonists on Aβ production in vitro and in vivo. At high concentrations, both the FDA-approved drug AF102 and its more effective later derivatives show a significant reduction of Aβ produced from primary hippocampal neurons. This reduction was dependent on γ-, but not α-secretase activity. A possible link was indicated by the finding that AF molecules affect GSK3α. Extending this finding, he also observed a reduction of tau phosphorylation. To show the in vivo relevance of this data, Fisher and his collaborators used the rabbit model and found a reduction of the Aβ signal with AF267B (see ARF related news story).