Synaptic Activity: It’s Important, But How?
What exactly is Aβ doing inside neurons? Several lines of investigation are converging on some function near synapses. To start this section out with a hint from comparative pathology, consider the presentation by Rebecca Rosen, Lary Walker, and colleagues at Emory University in Atlanta. Using immunohistochemistry, this group compared the localization of intraneuronal Aβ in hippocampus and other brain areas in aging humans, chimpanzees, squirrel monkeys, and rhesus monkeys. All four species deposit Aβ with age, but only humans suffer neurodegeneration as seen in AD. The researchers found that of the four species, only humans had dense Aβ puncta in the dendritic compartments of their hippocampal pyramidal neurons. That’s where synapses reside, and several groups have begun focusing intensely on exactly what Aβ might be doing there.

Helen Hsieh in Roberto Malinow’s lab at Cold Spring Harbor Laboratory, New York, asked if increased levels of Aβ might affect the trafficking of receptor proteins. Previous work from this lab had shown that activity increases Aβ release, and that Aβ then depresses AMPA and NMDA receptor-related neurotransmission in hippocampal synapses (see ARF related news story). At the SfN conference, Hsieh showed two-photon laser scanning images of hippocampal neurons overexpressing APP, which showed that they were less densely packed with synapses. Next, Hsieh tested if Aβ might influence long-term depression (LTD), a mechanism of synaptic plasticity that results from endocytosis-mediated removal of the GluR2 receptor from the synapse. When Hsieh measured transmission in organotypic hippocampal slice cultures from neurons that overexpress APP or β-CTF, she found that transmission was reduced and LTD was decreased compared to wild-type neurons. Overexpression of an APP mutant that produces no Aβ had no effect on LTD. Thus, expression of APP or β-CTF mimicked and occluded LTD. Cotransfecting APP into neurons with a rectifying GluR2 resulted in less rectification, indicating synaptic removal of GluR2. Both sets of electrophysiology measurements suggest that Aβ changes GluR2 trafficking.

To address the fate of the receptors directly, Hsieh measured the level of GluR2 at the synapse and found that it was down in cells overexpressing APP. Furthermore, she reported that both the p38 MAP kinase inhibitor SB203580 and the phosphatase calcineurin, which block metabotropic glutamate receptor LTD, can attenuate the effect of APP. All these results indicate that Aβ increases endocytosis of GluR2 receptors mediated by clathrin-coated pits and in this way depresses synaptic transmission. This conclusion draws further support from Hsieh’s finding that the effects of APP and β-CTF disappear in cells expressing an endocytosis-resistant receptor.

How these experiments—conducted on young hippocampal slices from mice overexpressing APP—relate to AD is unclear at present. Also unclear is how Aβ might influence receptor trafficking. And glutamate receptors2 are not the only candidate victims of intraneuronal Aβ. GluR1 receptors have been reported to be down in Tg2576 mice (see Almeida et al., 2005), and LaFerla’s lab has reported that, in triple transgenic mice, α7 nicotinic acetylcholine receptors disappear from neurons in brain areas where intracellular Aβ oligomers accumulate (Oddo et al., 2005). These latter receptors are thought to interact with Aβ (see prior ARF SfN meeting report) and in Washington, Kelly Dineley from the University of Texas Medical Branch, Galveston, reported that deleting them in Tg2576 mice worsened the mice’s learning and memory deficits.

A fundamental question in this regard is whether Aβ strikes synapses from without or within. Food for thought on this question came from a presentation by John Cirrito in David Holtzman’s lab at Washington University, St. Louis, with colleagues at Lilly Research Laboratories in Indianapolis, and the University of Arizona at Tucson. Cirrito reported on experiments using a microdialysis probe to measure the amount of Aβ in the interstitial fluid (ISF) in the hippocampus (see ARF related news story). Their protocol allows them to take samples every half hour for up to 24 hours in awake, behaving mice. Cirrito reported that when hippocampal neurons in these mice were stimulated with electrical probes, the level of Aβ in the ISF shot up. When he used tetrodotoxin to attenuate normal neuronal activity, ISF Aβ went down. The experiments suggest that neurotransmission and release of Aβ are inextricably linked.

To probe this further, Cirrito treated some mice with tetanus toxin to block the release of neurotransmitter vesicles. This reduced the level of ISF Aβ by 80 percent within 8 hours, he reported. When he treated cultured brain slices with α-latrotoxin (which prompts release of synaptic vesicles) together with postsynaptic inhibitors, extracellular Aβ levels still increased, suggesting synaptic vesicle release alone could lead to increased Aβ release. This latter result, coupled with the fact that Aβ has not been found in synaptic vesicles, led Cirrito to speculate that the link between synaptic activity and ISF Aβ is indirect, perhaps related to vesicle recycling. In this regard, readers may want to revisit data by Brent Kelly, Robert Vassar, and Adriana Ferreira at Northwestern, who reported that Aβ decreases levels of dynamin 1, a protein needed for synaptic vesicle recycling (Kelly et al., 2005).—Gabrielle Strobel and Tom Fagan.

See also Introduction and Part 1, Introduction and Part 3, and Introduction and Part 4 of this series.