A wealth of data from Mucke’s and other labs has established that Aβ triggers downregulation of NMDA (see Palop et al., 2005; Palop et al., 2007; Snyder et al., 2005) and AMPA (see Hsieh et al., 2006; Almeida et al., 2005) receptors and associated signaling molecules, eventually causing dendritic spines to shrink and die (see also ARF related news story). However, “the mechanisms for how this may happen have been pretty elusive,” Mucke said in an interview with ARF. “There are speculations that a receptor-mediated process could be involved, but it’s unclear whether Aβ binds directly to NMDA receptors.”

First author Moustapha Cissé and colleagues pondered whether Aβ might stymie glutamatergic transmission by acting through EphB2, a master regulator of NMDA function. EphB2 controls expression of genes important for learning and memory by modulating calcium flow through NMDA receptors (Takasu et al., 2002), and, indeed, EphB2-knockout mice show reduced LTP in hippocampal and dentate granule neurons (Henderson et al., 2001). Furthermore, Aβ oligomers cause loss of EphB2 expression in cultured neurons (see ARF related news story on Lacor et al., 2007), and hippocampal levels of Eph receptors dipped prior to memory loss in several AD mouse strains, and in postmortem brain tissue of people with newly diagnosed AD (Simón et al., 2009).

In the current study, the Gladstone researchers addressed two issues—the mechanism by which Aβ causes EphB2 depletion, and the consequences of manipulating EphB2 expression in mice. To look at mechanism, Cissé and colleagues ran pull-down assays and co-immunoprecipitation on transfected cells and primary neurons, reporting that Aβ oligomers bind the fibronectin repeat domain of EphB2. When spiked into primary neuronal cultures from wild-type rats, Aβ oligomers caused EphB2 to drop within three days, and this reduction was blocked by lactacystin (a proteasome inhibitor) but not bafilomycin (an inhibitor of endosomal acidification). They found that three- to four-month-old APP transgenic mice (J20 strain) also had a dearth of hippocampal EphB2 compared to control mice, as reported previously (see Simon et al., 2009). Taken together, the biochemical experiments suggest that “Aβ interacts with EphB2, leading to its proteasomal degradation,” Mucke said. “Our data suggest that this process could trigger the well-established depletion of synaptic NMDA receptors in the presence of Aβ.”

In the second set of studies, the scientists use lentiviruses to alter EphB2 expression in the mouse dentate gyrus, a brain area important for memory formation. They wanted to see if knocking down the receptor would cause trouble for wild-type mice, and if supplying extra EphB2 could preserve cognition in AD transgenic mice.

To address the former, Cissé and colleagues injected lentiviral vectors expressing short-hairpin (sh) EphB2 RNA into the dentate gyrus of four- to five-month-old wild-type mice. The constructs also expressed green fluorescent protein (GFP) to flag transduction efficiency, which was 50 to 74 percent, typical for this procedure. The procedure worked; infected neurons had expressed less EphB2, and the reduction mattered functionally. Compared to cells expressing scrambled control short-hairpin RNA, EphB2-depleted neurons expressed fewer NMDA NR1 subunits, and failed to upregulate the immediate early gene Fos in response to EphB2 activation. Moreover, EphB2 knockdown impaired NMDA (but not AMPA) receptor function in wild-type hippocampal slices, causing LTP deficits similar to those from untreated J20 mice that overexpress human APP with Swedish and Indiana mutations. These mice develop progressive amyloid deposition and memory loss.

For the rescue experiments, Mucke’s team injected two-month-old J20 mice with EphB2 lentiviral constructs, bringing their EphB2 to wild-type levels, and analyzed the animals’ electrophysiology and behavior two months later. Boosting dentate gyrus EphB2 levels protected these mice from LTP deficits, rescued synaptic function through NMDA receptors, and improved cognition as judged by tests of spatial memory, object recognition memory, and passive avoidance memory. Based on these findings, Mucke proposed that rescuing EphB2 levels restores glutamatergic transmission by countering the Aβ-induced loss of synaptic NMDA receptors. “That’s our favorite working hypothesis,” he said, acknowledging that “EphB2 is a complicated molecule that also affects other pathways.”

Along those lines, Pascale Lacor, a research assistant professor working with William Klein at Northwestern University in Evanston, Illinois, noted that it may be hard to determine if other receptors besides EphB2 are also mediating Aβ’s ill effects at synapses (see full comment below). A recent study by the Klein lab showed Aβ oligomers co-immunoprecipitating with EphB2 alongside other synaptic proteins, including NR2, GluR1, and the metabotropic glutamate receptor mGluR5 (see ARF related news story on Renner et al., 2010).

Still, scientists say the new data could point to new avenues of therapy for AD. “Strategies to manipulate EphB2, either by increasing receptor levels or by designing drugs to activate EphB2 signaling pathways, may help in the treatment of synaptic dysfunction associated with AD,” wrote Joaquin Del Rio and Diana Frechilla in an e-mail to ARF (see full comment below). They and coworkers at the University of Navarra, Pamplona, Spain, reported that EphB2 levels were down in postmortem hippocampal tissue from patients with incipient AD (Simón et al., 2009)—a result confirmed in the present study as data not shown.

Cissé and colleagues want to test next whether EphB2 benefits might hold up in a more physiological context, such as using small molecules to block Aβ oligomer-EphB2 binding and/or prevent EphB2’s degradation. In addition, the scientists are considering a reversion experiment in which they would test J20 mice at six months to quantify their cognitive impairment, then inject EphB2 lentiviral vectors and retest the mice to check if they get better. The current report describes a preventative paradigm, as the transgenic mice had no deficits when injected with EphB2 at two months of age.

Lending hope to future efforts, the present data suggest that even partial normalization of EphB2 had benefits. “We only infected about 60 percent of neurons in the dentate gyrus,” Mucke said. “This shows that one doesn’t necessarily have to normalize molecular alterations in every cell. The brain regions don’t need a full set of neurons to function pretty normally, as long as there are enough neurons to form functional networks.”—Esther Landhuis.

Reference:
Cissé M, Halabisky B, Harris J, Devidze N, Dubal DB, Sun B, Orr A, Lotz G, Kim DH, Hamto P, Ho K, Yu GQ, Mucke L. Reversing EphB2 depletion rescues cognitive functions in Alzheimer model. Nature. 2010 Nov 28. Abstract