This is Part 2 of a two-part story. See Part 1.
6 December 2007. At this year’s Society for Neuroscience annual meeting, held 3-8 November in San Diego, California, a special ancillary symposium called “Function of APP Gene Family Members and Clues to AD Pathogenesis: Studies from Worms to Mammals” explored the normal and pathological roles of APP. An emerging theme was that different parts of the APP molecule might play different roles in the development of the nervous system (see Part 1 of this news story). Curiously, different parts of the APP molecule also seem to have non-developmental roles, both within and outside of neurons.
One APP domain that may play a developmental role is the NPTY motif that occurs in the C-terminal of the protein. An APP construct lacking this motif rescues APP knockout deficits but also reduces turnover of holo APP, increasing cell surface expression and reducing production of Aβ (see Part 1 of this story). This NPTY motif may have pleiotropic effects, since it binds Dab1 (see Part 1 of this story) and other key players in APP biology, including Fe65 (see Part 1). One of those effects could be to regulate intracellular transport. Stefan Kins and colleagues at the University of Heidelberg, Germany, recently showed that the NPTY motif is crucial for transport of synaptic markers in Drosophila (see Rusu et al., 2007). However, Kins’s data suggest it may not be important for the trafficking of APP itself, and with that addressed a debated question in the field (see ARF related conference story and crosslinks). In San Diego, Kins reported that a variety of different APP constructs, including those lacking the C-terminus, are transported to axons just fine in cultured neurons. Even the speed of transport is unaffected by removing the C-terminus.
Kins also reported that APP transport may be linked with that of other synaptic proteins and also with activity of the small GTPase Rab3. Using APP antibodies to immunoisolate proteins from the low-density membrane fraction of wild-type mouse brain, he showed that synaptic proteins snap25, syntaxin1b, and synapsin, Rab3 GTPase, its activating protein Gap (both p130 and p150 subunits), and Rab3 partners RIM and Munc13-1 are all pulled down. However, in a testament to the specificity of the antibody, none of these proteins could be isolated from APP knockout tissue using the same approach. The experiments suggest that all these proteins coexist with APP in cellular vesicles. So could transport of APP and Rab3, which is abundant in synaptic vesicles, be intertwined? Kins asked. He demonstrated that APP anterograde transport was diminished in genetically modified cells expressing Rab3 locked in the GTP bound state. Further, he found that Kinesin-1, normally present in APP-immunoisolations, was lost in APP-membrane isolations from mouse brains that predominantly contain Rab3 locked in a GTP bound state. All together, these findings link transport of synaptic proteins to APP and interconnect for the first time APP targeting and transport mechanisms. Interestingly, other Rab family members have been linked to APP processing (see, for example, Laifenfeld et al., 2007) and, via reduced isoprenylation, the ability of statins to lower Aβ production (see ARF related news story).
Angels Almenar from Larry Goldstein’s lab at University of California, San Diego, also addressed the role of APP in axonal transport, particularly the idea that it may couple cargo to kinesin. Almenar has taken a direct approach to this question, purifying synaptic vesicles and examining their contents. Using mass spectroscopy, Almenar has identified 280 potential candidates that might co-transport with APP and are currently being analyzed.
“The analyses of APP transport is still an important issue for AD research as it allows insights in the complex regulation of APP processing and the putative function of APP,” Kins told ARF after the symposium. “Interestingly, although Angels Almenar and I characterized very different APP membrane compartments, both analyses established a connection between APP with synaptic vesicles. This argues that APP dysfunction in AD may directly contribute to altered synaptic transmission and neurodegeneration,” he said.
The transport theme was echoed in a slightly different way by Hui Zheng from Baylor College of Medicine in Houston, Texas. Zheng suggested that one of APP’s primary roles is to ensure that a different kind of transporter, the high-affinity choline transporter (CHT), is appropriately expressed in presynaptic terminals. Because CHT recycles choline back to presynaptic neurons for re-synthesis into acetylcholine, loss of CHT could explain cholinergic deficits that occur in Alzheimer disease (see ARF related news story).
To study the relationship between APP and CHT, Zheng’s group has focused on the neuromuscular junction as a model system. In this peripheral, cholinergic synapse, neuronal synaptophysin normally is perfectly juxtaposed to muscle-bound αBungarotoxin, which has high affinity for the acetylcholine receptor. In APP and APLP2 double knockout mice, however, synaptophysin and the toxin only partially colocalize, as synaptophysin is diffusely distributed along axons. Zheng demonstrated that the same pattern emerges in CHT knockout animals, suggesting that APP and CHT may somehow cross paths.
In support of this idea, Zheng’s group demonstrated that CHT is absent from presynaptic terminals in APP/APLP2 double knockout mice, suggesting that APP may somehow help target CHT to its normal location (see Wang et al., 2007). To test this, her lab has made both pre- and post-synaptic APP knockouts where the protein is absent from either neurons or the neuromuscular junction, respectively. Zheng reported that the phenotype of the post-synaptic knockout is more severe. How can this be? Zheng speculated that an intercellular interaction between APP and CHT localizes the latter to the neuromuscular junction.
A symposium on APP can hardly get by without mention of Aβ. On that note, Orly Lazarov, who has established her own laboratory at the University of Illinois at Chicago, outlined some of her work linking APP processing with environmental enrichment (Lazarov et al., 2005). In that work, she demonstrated that mice exposed to enriched environments have fewer plaques and less Aβ40 and 42 than do animals kept in normal, boring lab cages. The reduction in Aβ species may be thanks to enhanced clearance, because Lazarov and colleagues found that environmental enrichment increased the levels of the Aβ-degrading enzyme neprilysin. In San Diego, Lazarov reported that she is currently conducting microarray analysis on environmentally enriched mice to better understand crosstalk between APP processing, synaptic activity, and transcriptional activation. In related news, Sam Sisodia presented data on the involvement of presenilin mutations and microglia in the response of mice to living in a more stimulating environment (see ARF related news story).
And last, but by no means least, Huaxi Xu from the Burnham Institute of Medical Research, La Jolla, California, took the audience into uncharted waters by studying the role of APP in non-neuronal tissue. Though APP is widely expressed in mammalian tissues, relatively scant attention is being paid to its role outside the nervous system. It may be prudent to take off the blinkers and look to extraneuronal roles of APP, given that loss of presenilin activity has been linked to skin lesions and tumorigenesis. Could this be related to APP processing? Earlier this year, Xu and colleagues reported that expression of the epidermal growth factor receptor (EGFR) goes up in fibroblasts lacking presenilin or nicastrin (see Zhang et al., 2007). Interestingly, normal EGFR levels are restored by adding back the relevant proteins, i.e., presenilin or nicastrin, but also by adding AICD, the APP intracellular domain (but not NICD, the Notch intracellular domain) to APP/APLP2 double knockouts which also have an increase in EGFR. These results suggest that production of AICD may attenuate EGFR expression. Indeed, Xu reported pulse-chase experiments showing that it was synthesis of the receptor and not its degradation that was altered by the various knockout and rescue experiments.
How does AICD regulate EFGR expression, Xu asked? It appears to work, in concert with Fe65, on the receptor’s promoter. Xu demonstrated chromatin immunoprecipitation experiments showing AICD bound to the EGFR promoter, while AICD together with Fe65 could attenuate elevated receptor levels in APP/APLP2 double knockout fibroblasts.
Taken together, these experiments suggest that APP processing may play a role in preventing tumorigenesis, since many cancers are driven by EGFR-mediated responses.
On a similar note, Xu introduced a new AICD-related protein called 168, which induces apoptosis. The protein decreased the membrane potential in mitochondria, which is an indication of mitochondrial dysfunction, and it increases activity of the apoptotic caspase-3. These effects seem to be related to AICD because APP/APLP2 double knockout cells exhibit less apoptosis than wild-type cells when transfected with 168 constructs, and 168-mediated caspase-3 cleavage was enhanced on addition of AICD. The results, if confirmed, indicate that in certain tissues AICD may protect against tumorigenesis in more ways than one, decreasing EGFR expression and enhancing programmed cell death.
Overall, the symposium demonstrated that while a lot has been learned in recent years about the non-amyloidogenic roles of APP, broad consensus is still elusive. Progress in mammalian research has been complicated by the presence of different isoforms, but with double and even triple knockouts and rescue experiments at hand, researchers are finally getting a grip on the various roles the different cytosolic and extracellular domains of APP are playing. Perhaps the major gaping hole in APP knowledge is how the extracellular domain exerts its effects.—Tom Fagan.
This is Part 2 of a two-part story. See Part 1.