This is Part 1 of a two-part series. See also Part 2.
13 November 2009. One disease presents as electrical jolts that sporadically seize the brain, the other as forgetfulness and disorientation that progressively worsen. “At first pass, you’d think they are different universes,” said Helen Scharfman, Nathan Kline Institute for Psychiatric Research, Orangeburg, New York, in a symposium on epilepsy and Alzheimer disease at the Society for Neuroscience annual meeting held 17-21 October 2009 in Chicago. In the past two years, however, a closer look at these two conditions has revealed some fundamental similarities. “In both diseases, neuronal activity has gone awry in timing and synchronization,” Scharfman told the assembly of epilepsy and AD researchers. The symposium’s four presentations, as well as several posters on the topic, gave attendees plenty to chew on—highlighting links in electrophysiology, epidemiology, and animal models of the two diseases. “We believe there is some compelling clinical and experimental evidence suggesting that there could be an overlap between epilepsy and dementia, and that there could be a fruitful ground for collaboration for investigators in these areas,” said symposium organizer Lennart Mucke of the Gladstone Institute of Neurological Disease in San Francisco, California.
Mucke and colleagues provided the spark for this interface when they teamed up with epileptologist Jeffrey Noebels of Baylor College of Medicine in Houston, Texas, to reveal epileptiform activity in an AD mouse model (Palop et al., 2007 and ARF related news story). These abnormal discharges escape casual observation but readily showed up in electroencephalography (EEG) recordings done on freely moving J20 mice. “This confirmed our suspicion that, against what people expected, Aβ wasn’t shutting the network down,” Mucke told the SfN audience. Instead, Aβ was inducing peaks of overexcitation, which then trigger suppressive mechanisms, he suggested. “The system flips and flops between these states, resulting in an imbalance between overexcitation and inhibitory pathways in memory centers that we predict are key components of AD pathophysiology,” Mucke said.
His lab then asked whether this network instability affects adult neurogenesis. Using retrovirus to label newborn neurons in the dentate gyrus of J20 mice, the researchers followed the cells’ morphological and functional development from birth. The newborn granule cells seemed to develop on a fast track in the early stages, but later their maturation slowed down as compared to wild-type granule cells, Mucke reported. His team was able to restore normal neurogenesis by either blocking GABAergic signaling early on or inhibiting calcineurin at later stages. These findings are in press in Cell Stem Cell, Mucke said.
In addition to demonstrating that the network instability in AD mice has functional consequences, Mucke noted that his lab has found sodium channel abnormalities in J20 inhibitory interneurons. These findings, currently under review, could help reconcile the conundrum that Aβ, which is thought to dampen synaptic activity (Kamenetz et al., 2003), can in fact make networks hyperactive. If inhibitory interneurons are impaired, the end result is network disinhibition, Mucke told ARF in a post-symposium interview. Work published earlier this year by Dennis Selkoe, Brigham and Women’s Hospital, Boston, and colleagues suggests that Aβ-induced neuronal overstimulation can disrupt uptake of extracellular glutamate, making neurons more prone to long-term depression (Li et al., 2009 and ARF related news story).
A recent in-vivo calcium imaging study led by Arthur Konnerth at Technical University Munich in Germany also lent credence to the notion that Aβ can make some neurons hyperactive (Busche et al., 2008 and ARF related news story). In that analysis, researchers imaged cortical neurons from APP23xPS45 mice and showed that while a third of the neurons were less active, about a fifth of them actually became more active.
At SfN, Konnerth presented new data from his lab showing that these changes in activity status seem to have functional consequences. His team analyzed neurons from the primary visual cortex of wild-type and APP23xPS45 mice. They chose the visual system because the relationship between physical input (i.e., shape and orientations of images) and how the neurons typically respond to it is well described. Plus, the researchers had found amyloid plaques and impaired spontaneous neuronal activity in the mice’s visual cortex in their earlier study. In the new analysis, some cells in the APP23xPS45 visual cortex were hypoactive and some were hyperactive, as expected. But more than that, the silent neurons in the AD mice showed no response to sensory stimuli. This was in contrast to wild-type mice, where a proportion of the silent cells were still responsive. More serious problems showed up in the APP23xPS45 hyperactive neurons. These cells essentially lost their ability to respond precisely to specific orientations, compared with hyperactive cells in the wild-type visual cortex, Konnerth reported.
The Mucke lab presented several SfN posters with mechanistic data on the relationship between synaptic dysfunction, network hyperexcitability, and behavioral deficits. Julie Harris and colleagues tried to get a handle on where Aβ first acts within the entorhinal-hippocampal network to wreak havoc on networks and behavior. To that end, the scientists analyzed transgenic mice expressing human mutant APP primarily in layer 2/3 pyramidal cells of the medial entorhinal cortex. These neurons signal to granule cells of the dentate gyrus, which had no detectable APP expression. By the time they were six months old, the transgenic mice had molecular changes in the dentate gyrus, as well as defects in several behavioral assays, but with still hardly any Aβ deposition in the dentate gyrus. From these findings, they determined that Aβ acts trans-synaptically to induce molecular and functional impairments (see Harris et al. SfN abstract). In a separate poster (Palop et al. SfN abstract), Jorge Palop and colleagues showed that exacerbating Aβ-induced epileptiform activity in J20 mice further intensified the remodeling of hippocampal circuits and other molecular abnormalities they had characterized in their previous study (Palop et al., 2007). The data suggest to the authors that these changes are indeed a downstream consequence of the neuronal overexcitability.
Scientists in Finland have recently extended the observations of hyperactivity to another AD transgenic line, PSAPP (Minkeviciene et al., 2009 and ARF related news story). In an SfN poster (Leiser et al. SfN abstact), researchers at Pfizer Global Research and Development, Princeton, New Jersey, report epileptiform activity in that same strain as well as another, Tg2576. Steve Leiser—who has since moved to AstraZeneca in Wilmington, Delaware—and colleagues found hyperexcitability in PSAPP mice at 21, 34, and 47 weeks, and in Tg2576 mice at 27 and 34 weeks. Like the J20 mice, the Tg2576 and PSAPP animals showed freezing behavior but had no tremors, convulsions, or other visible signs of overt seizures. Both strains showed epileptiform activity only after Aβ deposition had begun but before measurable cognitive decline, Leiser wrote in an e-mail to ARF. The EEG recordings also revealed a shift from delta (slow-wave) toward theta activity in the transgenic mice, suggesting their brains are in a hyperexcited state. Leiser noted that this pattern parallels that of AD patients, who show an increase in theta power early in the disease, and later shift back to a high delta state. “These EEG features could indicate a hyperexcitable state predictive of seizures, and might serve as a biomarker for preclinical AD,” he wrote. For epidemiological and clinical data supporting the AD-epilepsy connection, see also Part 2.—Esther Landhuis.
This is Part 1 of a two-part series. See also Part 2.