A Case of “Heads Win, Tails Don’t Lose”
The progress in uncovering the function of amyloid precursor protein (APP) has remained frustratingly slow due to the presence of APP-related proteins APLP1 and APLP2 and also because APP is continuously processed into three major fragments (paradoxically, intense focus on pathology of Aβ may also have contributed to this situation). However, the tide now seems to be shifting as more and more in vivo studies are shedding light on APP function in animal models ranging from C. elegans to mice (1). Employing an elegant approach, the Muller group used reverse genetics to express APP fragments in mice in which endogenous APP had been deleted (APP-KO). The “knock-in” strategy they used is superior to regular transgene-expression techniques since the endogenous regulatory elements still remain in place. The authors knocked-in either APPsα or APP lacking the last 15 residues of the cytoplasmic domain and subjected the resultant transgenic animals to a battery of behavioral tests. The authors report (2) that expression of APPsα,...  Read more

A Case of “Heads Win, Tails Don’t Lose”
The progress in uncovering the function of amyloid precursor protein (APP) has remained frustratingly slow due to the presence of APP-related proteins APLP1 and APLP2 and also because APP is continuously processed into three major fragments (paradoxically, intense focus on pathology of Aβ may also have contributed to this situation). However, the tide now seems to be shifting as more and more in vivo studies are shedding light on APP function in animal models ranging from C. elegans to mice (1). Employing an elegant approach, the Muller group used reverse genetics to express APP fragments in mice in which endogenous APP had been deleted (APP-KO). The “knock-in” strategy they used is superior to regular transgene-expression techniques since the endogenous regulatory elements still remain in place. The authors knocked-in either APPsα or APP lacking the last 15 residues of the cytoplasmic domain and subjected the resultant transgenic animals to a battery of behavioral tests. The authors report (2) that expression of APPsα, which lacks the transmembrane and the cytoplasmic domain of APP “…grossly attenuated or completely rescued the prominent deficits of APP-KO mice.” Although, the single APP-KO mice show only modest phenotypic changes (compared to the triple KO animals lacking all three proteins) and the behavioral testing can be affected by mixed genetic background of KI mice (ES cells from 129/Ola were injected in C57BL/6 mice), this paper clearly shows that expression of APP ectoplasmic domain rescues the deficits examined here. Does it mean that the APP cytoplasmic domain is dispensable and not relevant to APP function?

The answer to this question is a “no”! The cytoplasmic domain of APP represents the most conserved region of the protein through the evolution and shares a very high degree of homology with that of APLP1 and APLP2. The “GYENPTY” motif in APP cytoplasmic domain which allows binding of Fe65 (and several other adapter proteins) to APP is completely conserved in all APP family members. Ring, Muller, and colleagues expressed APP fragments in single APP-KO mice which expressed endogenous APLP1 and APLP2. The high degree of homology in the cytoplasmic region and conserved protein-protein interactions make it almost certain that the function of the “missing” APP cytoplasmic domain was replaced by the counterparts of APLP1 and APLP2. Indeed, APLP2 alone is capable of replacing APP, and APLP1 functions if postnatal lethality, observed in triple KO animals, is used as a functional readout. A crucial experiment to be performed, as the authors concede, will be to test whether APPsα rescues the deficits observed when all three APP family members are deleted. These experiments, although not trivial, must be on the drawing board.

While we await results from such experiments, it should be noted that the present results do make two important contributions to our knowledge base. It has long been believed (3) that APP ectodomain has neuroprotective and growth-promoting properties, and this notion has been supported by numerous in vitro tissue culture studies. The present studies provide an in vivo proof of concept by showing that APPsα was able to restore normal brain weight. Secondly, the results of Ring et al. argue strongly against a physiological role for Aβ as has been previously suggested (4). There is no Aβ made in APPSα-KI mice, and there is no likelihood that an “Aβ”-like fragment could mask the deficiency of Aβ function; the Aβ region, unlike the cytoplasmic domain, is completely dissimilar in APP and APLP1/APLP2. So, while the field may continue to debate the role of Aβ in pathology, the question of a “normal” function of Aβ may be put to rest.

References:
1. Zheng H, Koo EH. The amyloid precursor protein: beyond amyloid. Mol Neurodegener. 2006 Jul 3;1:5. Abstract

2. Ring S, Weyer SW, Kilian SB, Waldron E, Pietrzik CU, Filippov MA, Herms J, Buchholz C, Eckman CB, Korte M, Wolfer DP, Muller UC. The secreted beta-amyloid precursor protein ectodomain APPs alpha is sufficient to rescue the anatomical, behavioral, and electrophysiological abnormalities of APP-deficient mice. J Neurosci. 2007 Jul 18;27(29):7817-26. Abstract

3. Saitoh T, Sundsmo M, Roch JM, Kimura N, Cole G, Schubert D, Oltersdorf T, Schenk DB. Secreted form of amyloid beta protein precursor is involved in the growth regulation of fibroblasts. Cell. 1989 Aug 25;58(4):615-22. Abstract

4. Kamenetz F, Tomita T, Hsieh H, Seabrook G, Borchelt D, Iwatsubo T, Sisodia S, Malinow R. APP processing and synaptic function. Neuron. 2003 Mar 27;37(6):925-37. Abstract

View all comments by Sanjay W. Pimplikar

This is a carefully crafted and cautiously interpreted study showing that the α-secretase derived extracellular domain of APP (APPsα) is sufficient for rescue of the APP knockout (KO) mouse phenotypes that have been described so far. The beauty of this work is that a knock-in strategy was used, thus Ring et al. ensured that their conclusions would not be confounded by ectopic and/or excessive expression of APPsα and all isoforms of APP can be generated. Importantly, APPsα can rescue learning deficits of aged mice in the Morris water maze test and impaired long-term potentiation (LTP) in the CA3/CA1 pathway of hippocampal slices obtained from aged APP knockout mice to the same extent as an APP allele lacking the C-terminal 15 amino acid residues. Furthermore, the absence of the APP C-terminus did not affect behavior in the probe test, suggesting that retention of the learned behavior does not require the APP C-terminus.

Although these data do not preclude a role for the intracellular domain of APP and its related family members, APLP1 and APLP2, in normal brain development,...  Read more

This is a carefully crafted and cautiously interpreted study showing that the α-secretase derived extracellular domain of APP (APPsα) is sufficient for rescue of the APP knockout (KO) mouse phenotypes that have been described so far. The beauty of this work is that a knock-in strategy was used, thus Ring et al. ensured that their conclusions would not be confounded by ectopic and/or excessive expression of APPsα and all isoforms of APP can be generated. Importantly, APPsα can rescue learning deficits of aged mice in the Morris water maze test and impaired long-term potentiation (LTP) in the CA3/CA1 pathway of hippocampal slices obtained from aged APP knockout mice to the same extent as an APP allele lacking the C-terminal 15 amino acid residues. Furthermore, the absence of the APP C-terminus did not affect behavior in the probe test, suggesting that retention of the learned behavior does not require the APP C-terminus.

Although these data do not preclude a role for the intracellular domain of APP and its related family members, APLP1 and APLP2, in normal brain development, they do show that the APP C-terminus is not essential for APP-dependent learning and memory in the adult mouse brain. However, recently Ma et al. (PNAS 2007) reported that the γ-secretase derived intracellular domain of APP (AICD) was the only APP fragment that correlated with enhanced memory observed in the Morris water maze test for mice that overexpress WT APP. We can conclude from these two studies that the APP ectodomain in the form of APPsα is essential for learning, that the APP C-terminus is not required for retention of the learned behavior, and that AICD may play a role in enhancing memory of the learned behavior.

It has also previously been hypothesized that Aβ, on the basis of its ability to depress excitatory synaptic transmission and the ability of neuronal activation to increase secreted Aβ, may be part of a negative feedback loop for neuronal activity (Kamenetz et al., 2003). Since APPsα can completely rescue the learning deficits of the APP KO mice, the authors conclude that Aβ deficiency does not contribute to this phenotype. Although the Ring et al. study does not negate the proposed role for Aβ in regulation of neuronal activity and it remains possible that future studies will reveal a phenotypic manifestation for Aβ deficiency that is related to the proposed effects of Aβ on neuronal activity, there currently are no obvious consequences of Aβ deficiency for spatial learning and memory in mice.

View all comments by Suzanne Guenette

I stand in awe of this painstaking, comprehensive study from Ulrike Müller and coworkers. This is probably as close as we can get to the function of APP and the amyloid peptides in vivo—and to the problem of their physiological function and pathological role. The latter is evident and indisputable from what we have learned over the years since Glenner and Wong identified the amyloid peptides in 1984, but the normal function of APP and Aβ remains controversial. The data collected and documented in our own mouse models corroborated those of others (e.g., Kamenetz et al, 2003), and it led us to compare the amyloid peptides to totally unrelated but equally real and mysterious objects with unknown functions, i.e., the pentagonal dodecaeder from Gallo-Roman times (see comment by Dewachter and Van Leuven, 2005).

I fully agree that this study leaves little to the imagination about the amyloid peptides exerting major physiological...  Read more

I stand in awe of this painstaking, comprehensive study from Ulrike Müller and coworkers. This is probably as close as we can get to the function of APP and the amyloid peptides in vivo—and to the problem of their physiological function and pathological role. The latter is evident and indisputable from what we have learned over the years since Glenner and Wong identified the amyloid peptides in 1984, but the normal function of APP and Aβ remains controversial. The data collected and documented in our own mouse models corroborated those of others (e.g., Kamenetz et al, 2003), and it led us to compare the amyloid peptides to totally unrelated but equally real and mysterious objects with unknown functions, i.e., the pentagonal dodecaeder from Gallo-Roman times (see comment by Dewachter and Van Leuven, 2005).

I fully agree that this study leaves little to the imagination about the amyloid peptides exerting major physiological actions in brain. Perhaps one could or should examine platelets and their related (patho)physiology in these novel mice, but I suspect that few researchers will wage money on that horse!

The AICD, on the other hand, remains a subject of discussion, although this study strengthens the case against it being essential for neuronal well-being. To untangle that knot, the possible experimental approaches in vivo are rather limited. Further, as they must involve APLP1 and APLP2, they are fraught with technical obstacles and “project-breaking” pitfalls. We hope that the group of Ulrike Müller can continue along the path they are on, since they appear alone to have the needed tools—and courage!

View all comments by Fred Van Leuven