In addition to helping the field settle on a role for APP, for which consensus remains elusive, the findings could also have implications for Alzheimer disease because the researchers report that Reelin reduces endocytosis of APP and promotes the non-amyloidogenic processing of the precursor. “I think for me it introduces a new element to the mechanism of APP signaling and cleavage, exposing the involvement of extracellular ligands, and may contribute to new therapeutic approaches for AD,” suggested Bernadette Allinquant, INSERM, France. Allinquant was not involved in the work but has studied the Reelin/APP interaction.

Reelin is perhaps best known because of the reeler mouse—a Reelin knockout that has stark neurodevelopmental problems. Reelin is predominantly produced by the Cajal-Retzius cells during development, and by interacting with members of the low-density lipoprotein receptor family, it activates the cytoplasmic protein disabled (Dab-1) and thereby regulates neuronal migration. “That whole pathway has been beautifully worked out, but what was not so clear is what happens when Reelin binds APP,” said Rebeck. For one, knocking out APP has a much milder effect than Reelin knockout (see Zheng et al., 1995), which suggests the APP/Reelin interaction is only part of the neurodevelopmental picture.

Rebeck said he was drawn to study Reelin because it shares some of the same protein partners as APP, including F-spondin (see ARF related news story), ApoE receptors (for a review, see Hoe and Rebeck, 2008), and Dab-1 (see Hoareau et al., 2008). Work from his lab showed that both Dab-1 and Reelin reduce β cleavage of APP (see Hoe et al., 2006) and this current paper delves into the Reelin/APP interaction in more detail.

First author Hyang-Sook Hoe and colleagues report that Reelin and APP interact with each other through specific domains—the E1 extracellular domain on APP, and domains 3-6 of Reelin. The two proteins co-immunoprecipitate from mouse brain lysates, primary hippocampal neurons, and COS7 cells expressing both proteins. Interestingly, Reelin expression is increased by about one-third in the brain of APP-negative mice and decreased in Tg2576 transgenic mice, which overexpress human APP. “The findings suggest an interaction between APP and Reelin is important for maintaining normal Reelin levels in the brain,” write the authors.

In turn, Reelin also influences APP dynamics. The researchers found that in several different systems (COS7 cells, primary cortical and hippocampal neurons, and Neuro2A cells), Reelin treatment led to increased APP on the cell surface. The authors reasoned that might be due to reduced endocytosis, and were able to show, using green fluorescent protein tagged APP, that ~40 percent less surface APP was taken up by hippocampal neurons treated with Reelin. The treatment also increased α-secretase processing of APP, decreasing release of Aβ40/42.

What is the biological significance of the Reelin/APP interaction? Both have been shown to promote neurite outgrowth (see Qiu et al., 1995 and Niu et al., 2004) and here Rebeck and colleagues show that they cooperate toward that end. In cultured hippocampal neurons, overexpression of APP increased dendritic neurite complexity in response to Reelin, while knocking down APP with interfering RNAs had the opposite effect, abolishing the Reelin response. “These data support the hypothesis that a Reelin-APP interaction is critical for induction of neurite outgrowth in primary hippocampal neurons,” write the authors. That interaction seems critical in vivo, too, because Hoe and colleagues found that in APP-overexpressing transgenic mice dendritic arborization was significantly increased, while it was significantly decreased in APP knockout mice.

Finally, the authors pieced together exactly how Reelin and APP are held together. Work from Dennis Selkoe’s group at Harvard Medical School showed that APP colocalizes with β1 integrin, a cell surface protein that interacts with extracellular matrix proteins (see Yamazaki et al., 1997), and that β1 integrin mediates neurite outgrowth induced by sAPPα, which is shed by α-secretase (see Young-Pearse et al., 2008). Hoe and colleagues confirmed the APP/β1 integrin colocalization in hippocampal neurons, and they also found that APP colocalizes with the α3 integrin, which forms a complex with its β1 partner. They also showed that APP, Reelin, and α3β1 integrin co-immunoprecipitate from mouse brain lysates, indicating that the three form a complex. In fact, the three proteins may be co-dependent because overexpression or loss of APP reduced and increased, respectively, the expression of α3β1 integrin, much like it did for Reelin. Furthermore, the researchers found that β1 integrin and Reelin synergistically cooperate to reduce endocytosis of APP and retain the protein in the cell surface. Lastly, Hoe and colleagues showed that an antibody against the α3β1 integrin prevents neurite outgrowth by both Reelin and APP.

All told, the work paints a triptych, suggesting that Reelin, APP, and α3β1 integrin work together to promote neurite outgrowth. This may not be entirely crucial for development, given the milder phenotype of the APP knockout, admits Rebeck, but he agreed there could be some redundancy between APP and its homologs, APLP1 and APLP2. Knocking out all three leads to lissencephaly, a severe developmental problem.

The complex may also have a role in the adult brain, which would be more germane to Alzheimer disease research. “We are used to thinking about proteins in development as being purely developmental, and Reelin fits into that category, but then they stick around your whole life,” said Rebeck. “The question is, What do they do all that time?” He suggested that while Reelin is involved in migration and growth of neuronal processes, it might help maintain those processed in the adult brain. “One of the things we always talk about is plasticity, which we depend on every day,” said Rebeck. “That has to be driven by the strengthening or weakening or elimination of synapses, or the formation of new synapses. That, of course, goes on a lot during development but also as long as we’re alive.” One way to address the role of Reelin in adulthood, he suggested, would be to make a conditional knockout, which has not yet been done.

Whether Reelin is involved in pathogenic mechanisms that lead to AD remains to be seen. But one interesting outcome from this work is that overexpressing APP can attenuate Reelin expression and lead to dendritic abnormalities in young mice, which lead the authors to suggest that some models of AD might exhibit behavioral problems that are independent of Aβ accumulation.—Tom Fagan.

Reference:
Hoe H-S, Lee KJ, Carney RSE, Lee J, Markova A, Lee J-Y, Howell BW, Hyman BT, Pak DTS, Bu G, Rebeck GW. Interaction of reelin with amyloid precursor protein promotes neurite outgrowth. Journal of Neuroscience 2009 June 10; 29:7459-7473. Abstract