27 February 2009. Considering how nasty amyloid plaques can be to nearby neurons, it should come as no surprise that these clumps of brain Aβ also fluster astrocytes, the abundant support cells that surround neurons. The shocker—as reported in today’s issue of Science—lies in how far-reaching those astrocytic effects seem to be. By quantitatively imaging calcium in brain cells in vivo, researchers led by Brian Bacskai at Massachusetts General Hospital, Boston, showed that astrocytes in an Alzheimer disease mouse model have elevated resting calcium levels and are hyperactive—independent of neuronal activity or plaque proximity. What’s more, most of the AD mice exhibited intercellular calcium waves that started from astrocytes near Aβ plaques and spread radially across large distances. “The meat of our story is that the focal appearance of individual senile plaques has a long-range network effect on astrocytes,” Bacskai told ARF. The findings unleash further questions about how these global astrocytic changes impact neurophysiology, and whether they exacerbate, or result from, Aβ deposition.
“It's very interesting,” Kim Green, University of California, Irvine, said of the new work. “They have a fantastic setup that's really been able to look at calcium in ways that haven't been possible in mouse models before.”
First author Kishore Kuchibhotla and colleagues are not the first to use a calcium-sensitive fluorescent dye with multiphoton microscopy to measure calcium signaling in astrocytes in vivo. A few labs around the world use these methods. However, a key limitation in these studies is the inability of the fluorescence intensity to distinguish pre-existing intracellular calcium from dye that entered the cells during the procedure. “Therefore, most applications have used only dynamic changes in fluorescence intensity to monitor signaling events,” Bacskai explained.
For the new study, his team adapted fluorescence lifetime imaging microscopy (FLIM) to measure calcium levels independent of dye. This allowed the researchers to compare resting-state calcium concentrations in cells from wild-type and AD transgenic (APP/PS1) mice. Using the small-molecule calcium dye Oregon-Green BAPTA-1 (OGB), they found nearly twofold higher resting calcium levels in astrocytes of plaque-laden mice, compared with wild-type cells. Furthermore, time-lapse calcium imaging revealed that about 28 percent of astrocytes in APP/PS1 mice were hyperactive, whereas only 8 percent of wild-type astrocytes showed this sort of spontaneous activity. Curiously, the abnormalities did not appear to depend on how close the plaques were to the astrocytes. This issue may require further investigation, though. The AD mice used in this study undergo such robust amyloidosis that most astrocytes were fairly close to a plaque (i.e., no farther than 100 micrometers away) at the time they were studied, the authors note.
If not plaque proximity, the researchers wondered whether the neurons close at hand—which can become hyperactive in response to local Aβ (Busche et al., 2008)—were somehow inducing calcium dysregulation in surrounding astrocyte networks. This did not appear to be the case, since the astrocyte hyperactivity persisted even after they blunted neuronal activity with the sodium-channel blocker tetrodotoxin. Moreover, the calcium changes in astrocytes of AD mice were synchronized across long distances—about four times greater than what was observed in wild-type astrocytes. In fact, in six of eight transgenic mice studied—but none of the five wild-type animals—the researchers observed intercellular calcium waves similar to those seen previously in culture (Scemes and Giaume, 2006).
“We believe the waves themselves are somehow related to pathophysiology,” Bacskai told ARF. His contention finds support in the fact that astrocyte hyperactivity only appeared in AD mice with Aβ plaques, and not in younger animals that had no pathology yet. The hyperactivity of these waves and their uncoupling from neuronal activity “might explain or contribute to abnormal neural circuit function and to memory and cognitive impairment,” suggested Ben Barres of Stanford University, Palo Alto, California, in an e-mail to ARF. Given that neurons, astrocytes, and blood vessels coexist in neurovascular units, it is conceivable that the calcium disturbances in astrocytes could in turn affect neuronal signaling and alter local blood flow. “This abnormal vasoregulation might cause or aid the neurodegenerative process or at least impair neural circuit function and cognition,” Barres speculated.
At the same time, other scientists caution that the use of double mutant APP/PS1 animals may make the new findings difficult to interpret. The presenilin (PS1) mutation is “known to increase endoplasmic reticulum calcium release from ER stores, which could contribute to these waves and to the basal calcium in astrocytes,” Green told ARF. He was the lead author of a recent study showing that presenilins interact with sarco-endoplasmic reticulum Ca2+-ATPase (SERCA) pump proteins, and that SERCA activity influences Aβ generation (see ARF related news story).
Nevertheless, the authors, as well as other researchers interviewed for this story, agree on this much—the neurodegenerative process seems to include a specific network-wide response in astrocytes. “The question becomes, is this the chicken or the egg?” Bacskai said. “Does the altered calcium signaling lead to increased deposition of senile plaques, or is it a response to it?” His group will begin addressing this in young AD mice that do not yet show amyloid pathology. “We'd like to watch a new plaque form while imaging calcium,” he said.
Tackling the chicken-or-egg question could have ramifications for future AD treatments.
“It will be interesting to determine if inhibition of intercellular calcium waves in astrocytes of APP/PS1 mice results in improved behavioral performance in memory tasks, as it would suggest that blocking calcium waves in astrocytes may hold therapeutic potential for alleviating some of the symptoms in human AD patients or even in treating AD,” wrote Ilya Bezprozvanny of University of Texas Southwestern Medical Center, Dallas, in a forthcoming Perspective on the new study. Bezprozvanny’s commentary will appear in Science Signaling, Science magazine’s online journal for cell signaling research, in preview form on 3 March and in full publication on 24 March. (For a review on the role of astrocyte calcium in neurophysiology, see Agulhon et al., 2008.)
The new data also brings to light a great disparity between how Aβ deposition affects neurons and astrocytes. In terms of neurons, studies in cell cultures, tissue slices, and more recently in vivo (see ARF related news story), have shown that Aβ oligomers shrink dendritic spines. Last summer, Bacskai and colleagues reported that Aβ-induced calcium overload may precede this morphological degeneration (see ARF related news story), and it’s long been known that the pathogenic process ultimately sends neurons to their deathbed.
Astrocytes, by comparison, seem more resistant to excitotoxicity and degeneration, suggested Beth Stutzmann of Rosalind Franklin University in North Chicago, Illinois, in an email to ARF (see additional comments below). “Perhaps there is something to this network of calcium signals that serves a neuroprotective role,” she wrote, noting that in other cell types, the frequency of calcium oscillations can encode transcription of protective proteins such as NFκB. “Maybe something like this is happening in astrocytes—the calcium signals are communicating a protective call across a network,” she wrote. “The scene from Lord of the Rings comes to mind, when all the alarm lights are lit across the entire kingdom.” A similar theme cropped up in a recent study showing how astrocytes use gap junction proteins to assemble into metabolic networks that keep neurons sated with lactate (Rouach et al., 2008).—Esther Landhuis.
Kuchibhotla KV, Lattarulo CR, Hyman BT, Bacskai BJ. Synchronous Hyperactivity and Intercellular Calcium Waves in Astrocytes in Alzheimer Mice. Science. 2009 Feb 27;323(5918):1211-15. Abstract