Is It the Pits? Aβ and Endosomes
The vesicle-Aβ link found by Cirrito was also the topic of a talk by Claudia Almeida in Gunnar Gouras’s lab. When these scientists reported earlier this year that neurons from Tg2576 transgenic mice had reduced surface expression of the GluR1 variant of glutamate receptor, they also noticed that these losses were accompanied by losses of PSD-95, a protein that anchors GluR1 to the synapse. In Washington, Almeida addressed how Aβ might influence cell surface receptors and proposed that it does so by impinging on the multivesicular body sorting pathway.

Multivesicular bodies, or late endosomes, deliver membrane proteins to lysosomes for degradation, and Aβ42 colocalizes strongly with markers of late endosomes. Early, recycling endosomes shuttle membrane proteins back to the cell membrane, and Aβ42 colocalizes weakly with them. To determine how Aβ42 affects the dynamics of these sorting bodies, Almeida measured trafficking of two markers. The transferrin receptor (TfR) is processed by the recycling of early endosomes, and the epidermal growth factor receptor (EGFR) is shuttled through late endosomes. She exposed neurons from Tg2576 mice to fluorescently labeled transferrin or EGF and then took time-lapse snapshots of their intracellular locale. EGF and Aβ began appearing together after 10 minutes, and did so substantially after an hour, whereas endocytosed TfR did not segregate to Aβ-labeled compartments.

Importantly, Almeida found that in cells expressing hAPP, the recycling EGF appeared to get stuck in late endosomes. In wild-type cells, it was all but gone from endosomes after an hour, whereas in transgenic neurons, both the size and the density of EGF-containing endosomes were significantly larger at this time point. Levels of the phosphorylated, endocytosed EGF receptor stayed up in transgenic cells for much longer than in wild-type. Almeida demonstrated the involvement of Aβ in this process by pretreating cells with the γ-secretase inhibitor DAPT. Under these conditions, phospho-EGFR was inactivated much faster. APP similarly retarded degradation of TrkB receptors through late endosomes. This line of research converges with research by Anne Cataldo and Randy Nixon on the endosomal system and the intraneuronal degradation process of autophagy (see prior SfN news story; Cataldo et al., 2004). Research by Michael Ehlers on how recycling endosomes supply AMPA receptors for synaptic plasticity provides a basic science context on the broader topic (Park et al., 2004).

Recycling of the EGFR requires an intact ubiquitin-proteasome system (UPS); indeed, activity-dependent turnover of many neurotransmitter receptors in dendritic spines is thought to occur through the proteasome (Ehlers, 2003). For this reason, the scientists wondered if Aβ might affect endosomal recycling through actions on the UPS. Almeida reported that neurons from transgenic mice show increased ubiquitination of the EGFR after treatment with EGF. Proteasome function was generally impaired in these neurons, but DAPT treatment restored it to near normal. Furthermore, treating wild-type cells with a proteasome inhibitor slowed the EGFR’s progression through the multivesicular bodies to a crawl, similar to what the scientists saw in APP-transgenic neurons. This suggests that Aβ may influence receptor degradation through actions on the proteasomal system, Almeida said. Other groups also reported hints that Aβ, perhaps its oligomers, impairs proteasome function, potentially slowing down the turnover of receptors and other synapse components.

The results on activity and Aβ available so far highlight a nagging paradox, noted Gouras. Data from Malinow’s, Holtzman’s, and his own lab all indicate that synaptic activity increases Aβ production/release, and increasing Aβ certainly is assumed to drive AD. Yet how does this fit in with the finding that neuronal output gradually declines in the run-up to AD, starting years before the disease is overt? PET imaging and other methods have established the brain’s waning activity. Put most simply, then, the question is: Is thinking good or bad? Epidemiological and other evidence suggests that education and challenging mental activity protect against AD, but it’s also true that one in 10 epileptic patients develop Aβ pathology in areas of excessive activity. Gouras himself is beginning to address the question of how synaptic activity and Aβ might interact to affect pathogenesis through his studies of intraneuronal Aβ. As a first step, his group reported at the SfN conference that when they activated cultured Tg2576 neurons, their intracellular Aβ levels dropped. Perhaps understanding the relationship between intraneuronal and extracellular Aβ could help clear up this mystery, Gouras said.

This relationship is not understood at all. Are the pools on either side of the cell membrane somehow linked, or are they independent of each other? Using the triple transgenic mice, Oddo and colleagues addressed this question with an extension of their previous immunotherapy experiments whereby an anti-Aβ antibody injected into the hippocampus removed the peptide from both the brain parenchyma and from the inside of neurons. In new work presented in Washington, D.C., the scientists showed that the extracellular pool shrinks within 12 hours of the injection, whereas the intracellular pool shrinks after 3 days. Once the antibody has dissipated, the intracellular pool becomes replenished first, and then extracellular Aβ reappears. Moreover, as the triple transgenic mice age, their intracellular Aβ load gradually decreases and extracellular Aβ increases. The authors interpret these data to mean that some sort of dynamic equilibrium connects the two pools of Aβ on either side of the neuronal cell membrane. This work will appear in the American Journal of Pathology next January. In a further suggestion along these lines, a poster presented by Erene Mina from Glabe’s laboratory suggested that autophagy gears up when cultured cells are exposed to extracellular Aβ oligomers from the outside. How this communication works remains an enigma. —Gabrielle Strobel and Tom Fagan.

See also Introduction and Part 1, Part 2, and Part 4 of this series.