Before neurons die and cognition fades, synapses wither and disappear—and amyloid-β peptides can trigger this synaptic breakdown. It therefore stands to reason that Aβ’s dirty work would involve factors that promote synapse formation or maintenance. Wnt signaling proteins function as such, and prior work has made a case for this pathway’s involvement in AD. A genomewide association study linked sporadic AD to a variant of a Wnt co-receptor (low-density lipoprotein receptor-related protein 6, or LRP6) that tones down Wnt signaling (ARF related news story on De Ferrari et al., 2007). Other research suggested that brain levels of Dkk1, which inhibits Wnt signaling, are elevated in AD patient biopsies and in AD mouse models, and that Aβ peptides turn on Dkk1 expression in cultured mouse cortical neurons (Caricasole et al., 2004; Rosi et al., 2010). Furthermore, excitatory neurotransmission in rat hippocampal neuronal cultures rose with exposure to Wnt3a and dropped in the presence of Dkk1 (Avila et al., 2010).

Now, joint first authors Silvia Purro and Ellen Dickins and colleagues at University College London confirm that Aβ induces Dkk1 expression, and they link Dkk1 directly with loss of synapses. Incubating mouse hippocampal slices with small Aβ oligomers prepared in vitro (Klein, 2002), the researchers saw the expected boost in Dkk1 mRNA and protein levels, along with synaptic loss marked by decreased colocalization of vGlut1 and PSD95 pre- and post-synaptic markers. Treatment with Dkk1 antibodies wiped out these effects, suggesting that Aβ cannot harm synapses without the Wnt pathway modulator. “We expected only a partial protection, thinking maybe other molecules are involved,” Salinas told ARF. On the contrary, “the data suggest Dkk1 is quite critical for Aβ’s effects on synapses.”

Dkk1 mimicked some of Aβ’s effects. When the researchers added Dkk1 to cultures of rat hippocampal neurons, fewer cells stained for presynaptic (VAMP2, synapsin-1, Bassoon, Cask) and postsynaptic (PSD-95, Grip) than in control cultures. The synaptic sites re-formed when Dkk1 was removed. However, Dkk1 did not induce cell death, as assessed by TUNEL staining, nor did it influence total levels of synaptic proteins on Western blots. These data suggest that Dkk1 decreases the number of synaptic sites by causing synaptic proteins to disperse rather than to degrade. The scientists further demonstrated Dkk1’s effect on synaptic disassembly through electron microscopy studies that measured shrinkage of synapses, and time-lapse recordings showing dispersal of synaptic proteins within a half-hour of Dkk1 treatment.

Whereas prior work indicated that Aβ upregulates Dkk1 and ablates synapses, this study examines whether Dkk1 directly triggers synaptic damage. “Apparently, it does. When you incubate neurons in the presence of Dkk1, synapses start to disassemble rapidly—very rapidly. That’s the main contribution,” said Giancarlo De Ferrari of Universidad Andres Bello in Santiago, Chile. De Ferrari hypothesized years ago that sustained loss of Wnt signaling function could lead to AD (see review by De Ferrari and Inestrosa, 2000).

Because Dkk1 goes up early in AD when neurons have not yet begun dying, it may be possible to use Dkk1 as a biomarker, Salinas said. And, “if we develop [drugs] targeting Dkk1, the hope is that we could stop disease progression at the early stages before loss of memory.” Her lab is collaborating with a small biotechnology company to find small brain-penetrant molecules that interfere with Dkk1 function. In the future, the scientists hope to test whether such molecules could slow disease in AD mouse models, Salinas said.

In the second paper, first author Tomomichi Watanabe and colleagues investigated an earlier stage of amyloidosis—production of the Aβ peptides themselves. They focused on F-box and leucine-rich repeat protein 2 (FBL2), a component of the ubiquitin-proteasome system that helps cells dispose of damaged or misfolded proteins. Malfunction of the ubiquitin-proteasome pathway is linked with neurodegenerative disease, and a microarray study found abnormally low FBL2 expression in AD brains (Blalock et al., 2004). Presuming FBL2 would influence hallmark protein aggregates in AD, and having discovered early on that it neither binds tau nor influences tau phosphorylation, the team turned its attention toward Aβ, Watanabe noted.

Overexpression of FBL2 lowered Aβ secretion—and FBL2 knockdown raised it—in several cell culture systems, including human embryonic kidney cells and mouse neuronal cell lines and primary cortical neurons. FBL2-expressing cells also had considerably less intracellular Aβ. Levels of β- and γ-secretases or Aβ-degrading enzymes in those in-vitro experiments did not change, though, suggesting that FBL2 should not influence rates of APP processing or Aβ clearance. Could FBL2 regulate the amount of APP? Further experiments in FBL2-transfected cells showed that FBL2 binds APP and promotes its ubiquitination and degradation. In addition, a greater proportion of APP lingers on the cell surface, and this membrane APP becomes internalized more slowly in FBL2-overexpressing cells relative to control cells. This suggests that ubiquitination by FBL2 impedes APP endocytosis as well. FBL2 ubiquitination may function as a dual regulator, the authors propose, sending intracellular APP on its way for proteasomal degradation while keeping cell-surface APP from getting endocytosed for β-secretase cleavage. The net effect would be less Aβ production, the authors suggest.

That scenario appears to be true in at least one animal model as well. AD transgenic mice overexpressing FBL2 had more ubiquitinated APP, fewer plaques, and less insoluble brain Aβ compared to control AD mice. The FBL2-overexpressing AD strain was generated by crossing AD mice that express mutant APP and presenilin-1 (Willuweit et al., 2009) with transgenic mice that overexpress human FBL2 ~16-fold in the brain. The researchers plan to test behavior and cognition in these mice, Watanabe noted.

“This is a really interesting and exciting paper,” M. Paul Murphy of the University of Kentucky, Lexington, wrote in an e-mail to ARF (see full comment below). “Although we’ve known for some time that the ubiquitin pathway is intimately involved with the development of AD at several levels, this is probably the most significant step forward in our understanding of the underlying mechanism in a long while.” Gunnar Gouras of Lund University, Sweden, said the data are “quite thorough and provide important new insights into APP metabolism.” (See full comment below.)

Activation of the FBL2 pathway may hold promise as a new therapeutic target because it slows APP processing into Aβ in two ways—speeding degradation of intracellular APP and keeping membrane APP from reaching endosomes, suggest the authors. However, because augmenting protein function is generally harder than blocking it, “targeting this pathway will be challenging,” Murphy said. But since it could potentially treat other diseases that involve defective protein turnover, not just AD, “the payoff could be huge,” he said.—Esther Landhuis.

References:
Purro SA, Dickins EM, Salinas PC. The Secreted Wnt Antagonist Dickkopf-1 Is Required for Amyloid-β-Mediated Synaptic Loss. J Neurosci. 2012 Mar 7;32(10):3492-3498. Abstract

Watanabe T, Hikichi Y, Willuweit A, Shintani Y, Horiguchi T. FBL2 Regulates Amyloid Precursor Protein (APP) Metabolism by Promoting Ubiquitination-Dependent APP Degradation and Inhibition of APP Endocytosis. J Neurosci. 2012 Mar 7;32(10):3352-3365. Abstract