7 December 2012. Clinical failures have all but slammed the door on direct γ-secretase inhibition as a strategy to curb Aβ accumulation, and left researchers hunting for new approaches. Now, a study in the December 2 Nature Medicine suggests a novel way to put the brakes on the membrane protease. Researchers led by Bart De Strooper at KU Leuven, Belgium, report that the scaffolding protein β-arrestin 2 (βARR2) interacts with γ-secretase to promote Aβ production. Postmortem AD brains contain elevated levels of βARR2 mRNA, suggesting the protein plays a role in the disease. Moreover, in a mouse model of AD, deleting β-arrestin 2 slashed soluble Aβ. The arrestin might provide a safer, more selective therapeutic target compared to direct inhibition of γ-secretase, the authors propose.
Other scientists were intrigued by the findings. “This study may open up a whole new front in the fight against the accumulation of Aβ in AD,” said Ahmad Salehi at Stanford University, Palo Alto, California. However, he noted that many unanswered questions remain, such as how β-arrestin 2 becomes elevated in AD brains, and whether targeting the protein would improve cognition. It is also not clear how other γ-secretase substrates might be affected by targeting arrestin 2, though knocking out the gene seems to have little effect on mouse behavior or viability.
Aβ is released when γ-secretase snips the β C-terminal fragment (CTF) left behind by β-secretase cleavage of amyloid precursor protein (APP). In clinical trials, however, γ-secretase inhibitors seem to worsen cognition in AD patients (see ARF related news story and ARF news story). Scientists are now focusing on γ-secretase modulators (GSMs), which change how the enzyme cuts CTFs specifically while allowing γ-secretase to cleave other substrates as usual. GSMs seem to have fewer side effects (see ARF related news story; ARF related news story).
De Strooper and colleagues approached γ-secretase from a different direction, through G protein-coupled receptors (GPCRs). They previously showed that the G protein-coupled receptor 3 (GPR3) amplifies γ-secretase production of Aβ, although the mechanism was not clear (see ARF related news story). Similarly, a Chinese group reported that another G protein receptor, the β2-adrenergic receptor (β2-AR), stimulates γ-secretase activity and Aβ generation (see Ni et al., 2006). Classically, GPCRs signal through G proteins, but in recent years researchers have uncovered an independent signaling pathway through β-arrestin. This protein acts as a scaffold for assembling signaling complexes of downstream kinases. Scientists are now developing ways to selectively target β-arrestin signaling versus G protein signaling (see, e.g., Whalen et al., 2011). β-arrestins have not previously been linked to Alzheimer’s disease, although expression of arrestin 2 and 3 may be elevated in people with Parkinson’s disease dementia (see ARF related news story).
To look at the role of β-arrestin in AD, first author Amantha Thathiah expressed the protein in a human kidney cell line that also expresses APP. She found that the presence of β-arrestin 2 pumped up Aβ40 and Aβ42 production by 50 percent, while silencing β-arrestin 2 by RNA interference cut Aβ levels in half. Likewise, neurons from βARR2 knockout mice produced less Aβ compared to wild-type cells.
The authors wondered if β-arrestin 2 interacted with the GPCRs previously implicated in γ-secretase activation, GPR3 and β2-AR. They used a two-hybrid approach to show a direct physical interaction between the scaffold protein and each receptor. But is this interaction necessary for Aβ production? To investigate this, the authors mutated the βARR2 binding sites in GPR3 and found that this abolished the enhanced Aβ production in these cultures.
What is the link between γ-secretase and GPCRs/arrestin? Since GPCRs and γ-secretase are typically found in lipid-rich areas of the cell membrane, the researchers wondered if arrestin regulated their localization. When overexpressed in cell lines, β-arrestin 2 sequestered about 50 percent more γ-secretase in lipid rafts, and also increased activity of the secretase two- to threefold, the authors report. Conversely, silencing βARR2 kept the secretase out of lipid compartments. Moreover, the intact γ-secretase complex co-immunoprecipitated with βARR2, indicating a direct physical association. When the authors dissociated the secretase, only the Aph-1a subunit co-immunoprecipitated with β-arrestin 2, suggesting that this is the subunit that binds the protein.
Does silencing βARR2 actually inhibit γ-secretase? One finding in the paper casts doubt on this. While Aβ dropped, surprisingly, the authors saw very little increase in β-CTFs, which normally accumulate when γ-secretase is inhibited. Some researchers thought this finding curious. However, De Strooper and colleagues report that if they also blocked protein degradation, then β-CTFs built up in cells lacking β-arrestin 2. The authors suggest that silencing βARR2 does inhibit γ-secretase and leads to more β-CTF fragments, but also stimulates their turnover. If this can be independently confirmed, it would make β-arrestin 2 a more attractive therapeutic target, since several studies have linked the β-CTF fragments to toxicity and worsening of cognition in mice (see ARF related news story; ARF news story; and ARF news story). Some researchers suggest this may be part of the reason for the negative effects of γ-secretase inhibitors in clinical trials.
Turning to in-vivo models, the authors crossed APP/PS1 mice (see Radde et al., 2006) with β-arrestin 2 knockouts. In three-month-old βARR2-negative offspring, soluble Aβ plummeted by 75 to 90 percent compared to APP/PS1 controls. The authors plan to test cognition in these mice once the animals have aged sufficiently, Thathiah told Alzforum.
The authors also found evidence of β-arrestin dysfunction in human disease. On average, more βARR2 mRNA appeared in hippocampal and entorhinal cortex samples from 18 postmortem AD brains compared to 20 age-matched controls. In a separate sample series, the researchers saw higher βARR2 mRNA in the frontal cortex of 14 brains at Braak stage 5 or 6 of AD compared to 21 brains at stage 0-2. Thathiah noted the need to validate these results in more tissue samples. She is also looking for a good antibody to β-arrestin 2 that will allow her to look at protein levels and immunohistochemically localize the protein.
All told, the work suggests that β-arrestin 2 drives γ-secretase processing of Aβ, slows degradation of β-CTFs, and may do this in humans with Alzheimer's. However, Salehi wondered whether increased β-arrestin 2 is specific to AD, or if it might occur in other neurodegenerative diseases as well. Other commentators noted that it would be nice to see what happens when the protein is overexpressed in mice, given that β-arrestin 2 rises, rather than falls, in people with AD.
In future work, Thathiah plans to examine the substrate specificity of βARR2. Does it affect the cleavage of γ-secretase substrates Notch and cadherin, for example? She will also look at whether and how β-arrestin 2 affects γ-secretase complex formation. The answers to these questions will provide more data about how safe and specific inhibition of this pathway might be.
“I think this is great biology and a very thorough and complete study,” Todd Golde at the University of Florida, Gainesville, wrote to Alzforum. “It raises a lot of forward-looking questions about targeting various GPCRs in order to alter neuronal signaling and APP processing. As with many strategies that target Aβ, the devil will be in the details. Is it really safe enough for prophylaxis in humans or early intervention in preclinical stage 1 AD?”
Intriguingly, another recent paper reports that β-arrestin 1 plays a role in assembling the γ-secretase complex, and that βARR1 deletion lowers Aβ production, strengthening the idea that arrestins play a role in the disease (see Liu et al., 2012).—Madolyn Bowman Rogers.
Thathiah A, Horré K, Snellinx A, Vandewyer E, Huang Y, Ciesielska M, De Kloe G, Munck S, De Strooper B. β-arrestin 2 regulates Aβ generation and γ-secretase activity in Alzheimer’s disease. Nat Med. 2012 Dec 2. Abstract