Dogma holds that presenilins in the γ-secretase complex beget amyloid-β (Aβ) peptides by cleaving the membrane-bound β amyloid precursor protein (AβPP). But new results show there is (literally) another piece to the story. That piece is the AβPP intracellular domain (AICD), the cytosolic fragment of AβPP left after cleavage, which turns out to have the surprising ability to promote the destruction of Aβ. In a paper appearing in yesterday’s Neuron, Frederic Checler from Valbonne, France, and a multinational cohort of collaborators show that AICD activates transcription of the gene for the Aβ-degrading enzyme neprilysin. The ability of presenilins to coordinately affect Aβ production and degradation provides a neat physiologic mechanism for the tight regulation of Aβ levels, and may offer additional entry points for therapies that seek to lower those levels.

The starting point for first author Raphaelle Pardossi-Piquard and collaborators from France, Canada, Germany, Italy, and the US was an observation that fibroblasts or blastocysts from PS1 and PS2 double knockout mice failed to degrade Aβ peptides. From there, in an extensive series of well-controlled biochemical experiments, the researchers demonstrated that the cells specifically lacked neprilysin, but not other known Aβ proteases including insulin-degrading enzyme, and endothelin-converting enzyme. Transfecting the PS-/- cells with neprilysin fully restored Aβ degrading activity. The loss of neprilysin activity was accompanied by loss of neprilysin protein and mRNA, suggesting that somehow PS expression was regulating the transcription of the neprilysin gene. That regulation required γ-secretase activity, since several γ-secretase inhibitors mimicked the PS-/- phenotype in fibroblasts and in cultured neurons.

These results focused the researchers' interest on the AICD, since NICD, the intracellular domain of another γ-secretase substrate, Notch, had been shown to act as a transcription factor. Sure enough, when Pardossi-Piquard and colleagues expressed either a 50- or a 59-amino-acid AICD fragment in PS-/- cells, the cells regained neprilysin expression. The scientists went on to show that AICD increased activity of the neprilysin promoter in transactivation assays, that AICD could bind directly to the promoter DNA in mobility shift assays, and that it regulated neprilysin gene transcription via a pathway that involved the adaptor protein Fe65 and the histone acetytransferaseTip60, both proteins that were previously known to interact with AICD (see ARF related news story).

To show that endogenous AICD regulated neprilysin levels, the researchers checked enzyme activity in APP-deficient fibroblasts and in brain from APP knockout mice. In both cases, the investigators attained the expected result of decreased neprilysin activity. But activity was not null, which led them to check if intracellular domains from the APP-like proteins APLP1 and APLP2 also regulated neprilysin. They did, but Notch, e-cadherin, and n-cadherin intracellular fragments did not.

Finally, Pardossi-Piquard et al. determined that neprilysin expression and activity were higher in brain tissue from people with familial AD due to PS mutations compared to control tissue or brain from sporadic AD cases. This result is fully consistent with a physiologic role of PS and γ-secretase activity in neprilysin regulation.

“The experiments described here do more than simply confirm the previously and widely held suspicion that AICD, like NICD, might act as a signaling molecule involved in transcriptional activation,” the authors write. Instead, the demonstration that one product (AICD) of γ-secretase cleavage determines the lifetime of the other product (Aβ) reveals “an elegant and unusual method by which Aβ levels are controlled.” One ramification of this elegant linkage of γ-secretase and neprilysin activity is the possibility that γ-secretase inhibition could be self-defeating if inhibitors adversely affect degradation of Aβ. But on the other hand, (and this work proves there is always an other hand), since increasing neprilysin activity is a proven way to decrease brain Aβ, (Leissring et al., 2003), the results could point to a new way of lowering Aβ independently of γ-secretase using AICD or mimics to boost neprilysin levels.—Pat McCaffrey


  1. The manuscript by Pardossi-Piquard et al. (Neuron, 2005) provides an impressive array of experiments suggesting that the γ-secretase-derived intracellular domains (ICDs) of APP and APLPs serve to regulate the transcription of neprilysin (NEP). This paper is of high interest, first, because NEP is a key regulator of β-amyloid levels and amyloid pathology in vivo (Iwata et al., 2000; Iwata et al., 2001; Leissring et al., 2003) and second, because it adds strength to the evidence that APP may be a signaling molecule acting to regulate the transcription of genes through a mechanism that may or may not be analogous to Notch signaling (e.g., Cao and Sudhof, 2001; Leissring et al., 2002; Cao and Sudhof, 2004; Hass and Yankner; SfN abstract 146.4, 2004; among others).

    Beyond these important aspects, however, it is interesting to consider the results of Pardossi-Piquard and colleagues in relation to the finding that APP and APLP ICDs are rapidly and efficiently degraded by insulin-degrading enzyme (IDE) both in vitro (Edbauer et al., 2002; Walsh et al., 2003) and in vivo (Farris et al., 2003; Miller et al., 2003). Of course, IDE has additional numerous substrates besides the ICDs of APP and APLPs, so the following comments should be interpreted in this context. Nonetheless, several points should be made with regard to the interplay between IDE, ICD levels, and NEP.

    First, NEP levels were found to be unchanged in IDE knockout mice (Farris et al., 2003), which themselves had significantly elevated levels of non-phosphorylated ICDs (Farris et al., 2003; Miller et al., 2003). It is notable that phosphorylated forms of APP ICD were found to be unchanged in IDE knockout mice (Farris et al., 2003); however, this finding does not appear to explain the apparent conflict with the work of Pardossi-Piquard and colleagues, since these investigators found significant upregulation of NEP upon overexpression of cDNAs encoding the ICDs of APP or APLPs, which presumably would not be phosphorylated to any significant extent. Perhaps the discrepancy relates to the absolute magnitude of the changes in ICD levels or to other factors that differ in the in vitro versus the in vivo paradigms.

    Second, other in vivo work from our group (Leissring et al., 2003) lends some support to the notion that IDE can influence NEP expression levels, but it is not clear that this involves the mechanism proposed by Pardossi-Piquard and colleagues. In this work, transgenic overexpression of IDE to modest levels (~100 percent increase) in neurons did, in fact, lead to decreases in endogenous NEP levels and activity levels. However, it should be pointed out that we did not directly determine APP ICD levels in the IDE transgenic mice, and the decrease in NEP levels, while discernable in some animals, was variable and did not achieve statistical significance. Interestingly, however, in that same study, transgenic overexpression of NEP (by ~700 percent) led to significant downregulation of IDE. Taken together, these findings seem to suggest that some "extracellular" substrate common to the two proteases (perhaps β-amyloid itself; see Mohajeri et al., 2002) was responsible for their cross-regulation, since NEP, a type II membrane-bound protease, does not act on intracellular substrates. In light of the work of Pardossi-Piquard and colleagues, this topic deserves further careful study.

    Third, if ICD levels can indeed upregulate NEP levels, then it raises the curious possibility that inhibition of "intracellular" IDE—by increasing ICD levels—might represent a therapeutic strategy for combating Alzheimer disease. Inhibition of all pools of IDE non-selectively would, of course, not be viable, given the fact that IDE knockout mice show elevated, rather than reduced, β-amyloid levels (Farris et al., 2003; Miller et al., 2003). Moreover, IDE appears to be the major protease involved in the degradation of extracellular β-amyloid in cultured primary neurons (see Fig. 1B in Farris et al., 2003), suggesting that extracellular pools of IDE are more important to the regulation of β-amyloid than are intracellular pools (albeit the cellular locus of β-amyloid degradation by IDE is presently unclear).

    In summary, IDE knockout mice show clear elevations in ICD levels, yet do not show significant elevation in NEP levels, a result that tends to cast some doubt on the idea that increasing ICD levels per se will lead to upregulation of NEP in the brain in the intact animal. On the other hand, transgenic overexpression of IDE in neurons did lead to a decrease in NEP levels in some animals, though this change was variable and not significant across all animals examined. These results can perhaps be reconciled if we assume that levels of Tip60 and Fe65—essential cofactors for the ICD-mediated regulation of downstream genes (Cao and Sudhof, 2001)—are limiting in the normal mouse brain, making reductions—but not increases—in NEP levels possible through this mechanism. Even if ICD regulation turns out to be a tough target for combating Alzheimer disease, the tantalizing findings of Pardossi-Piquard and colleagues recommend closer scrutiny of the physiologic consequences of ICD regulation by IDE.


    . Identification of the major Abeta1-42-degrading catabolic pathway in brain parenchyma: suppression leads to biochemical and pathological deposition. Nat Med. 2000 Feb;6(2):143-50. PubMed.

    . Metabolic regulation of brain Abeta by neprilysin. Science. 2001 May 25;292(5521):1550-2. PubMed.

    . Enhanced proteolysis of beta-amyloid in APP transgenic mice prevents plaque formation, secondary pathology, and premature death. Neuron. 2003 Dec 18;40(6):1087-93. PubMed.

    . A transcriptionally [correction of transcriptively] active complex of APP with Fe65 and histone acetyltransferase Tip60. Science. 2001 Jul 6;293(5527):115-20. PubMed.

    . A physiologic signaling role for the gamma -secretase-derived intracellular fragment of APP. Proc Natl Acad Sci U S A. 2002 Apr 2;99(7):4697-702. PubMed.

    . Dissection of amyloid-beta precursor protein-dependent transcriptional transactivation. J Biol Chem. 2004 Jun 4;279(23):24601-11. PubMed.

    . Insulin-degrading enzyme rapidly removes the beta-amyloid precursor protein intracellular domain (AICD). J Biol Chem. 2002 Apr 19;277(16):13389-93. PubMed.

    . gamma-Secretase cleavage and binding to FE65 regulate the nuclear translocation of the intracellular C-terminal domain (ICD) of the APP family of proteins. Biochemistry. 2003 Jun 10;42(22):6664-73. PubMed.

    . Insulin-degrading enzyme regulates the levels of insulin, amyloid beta-protein, and the beta-amyloid precursor protein intracellular domain in vivo. Proc Natl Acad Sci U S A. 2003 Apr 1;100(7):4162-7. PubMed.

    . Amyloid-beta peptide levels in brain are inversely correlated with insulysin activity levels in vivo. Proc Natl Acad Sci U S A. 2003 May 13;100(10):6221-6. PubMed.

    . Abeta 42-induced increase in neprilysin is associated with prevention of amyloid plaque formation in vivo. J Biol Chem. 2002 Sep 20;277(38):35460-5. PubMed.

    . A transcriptionally [correction of transcriptively] active complex of APP with Fe65 and histone acetyltransferase Tip60. Science. 2001 Jul 6;293(5527):115-20. PubMed.

  2. The results reported by Raphaelle Pardossi-Piquard and colleagues are most interesting and also provocative. The implication is that the amyloid peptides (any or all) exert a physiologic function (which I have not seen demonstrated), and that this function is indirectly controlled or terminated by its "co-metabolite" AICD via induction of increased NEP expression.

    Questions that arise are plenty: Does that mean control of a "constant stream" of Aβ or is its production modulated or even triggered? By what signals? In which brain regions? And how indirect is control by AICD/NEP, that is, what is the timeframe or delay between generation of Aβ and induction of NEP?

    On the other hand, if production of ALID1 and ALID2, evidently without co-production of Aβ from APLP1 or APLP2, also induce NEP, the regulatory connection or relation among Aβ/AICD/NEP is not evident or at least much more indirect.

    A final note concerns a recent report that NEP is selective for Aβ42 degradation in vivo (Saito et al., 2005). Does that imply that only Aβ42 is physiologically functional, or rather that different control mechanisms control the different Aβ peptides?

    As stated above, the study is provocative in generating these and several other questions.


    . Somatostatin regulates brain amyloid beta peptide Abeta42 through modulation of proteolytic degradation. Nat Med. 2005 Apr;11(4):434-9. Epub 2005 Mar 20 PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Aβ's Shadowy Sibling—What Becomes of the Intracellular Domain?

Paper Citations

  1. . Enhanced proteolysis of beta-amyloid in APP transgenic mice prevents plaque formation, secondary pathology, and premature death. Neuron. 2003 Dec 18;40(6):1087-93. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Presenilin-dependent transcriptional control of the Abeta-degrading enzyme neprilysin by intracellular domains of betaAPP and APLP. Neuron. 2005 May 19;46(4):541-54. PubMed.