21 February 2007. The current issue of Neuron contains a scientific debate that bears watching. It is about one of the great mysteries of amyloid-β precursor protein (APP), that is, its physiological role. Scientists know a fair amount about APP cleavage and release of Aβ at the N-terminal end of the protein, but why that happens in the brain is less clear. The tail end of the molecule has been little help in solving the problem—it remains almost completely mysterious. One intriguing theory is that the APP intracellular domain (AICD) transduces signals, much like the Notch intracellular domain (NICD) does. In fact, work from Frederic Checler and colleagues at the CNRS, Valbonne, France, suggested that AICD operates a type of feedback loop, inducing expression of the Aβ-degrading enzyme neprilysin to counteract β- and γ-secretase cleavage of APP (see ARF related news story). Several labs have since called that theory into question, including a report in the February 15 Neuron by Allen Chen and Dennis Selkoe at Brigham and Women’s Hospital, Boston, which is countered by another paper in the same journal by Checler and colleagues. The reports pose a dilemma that is as old as science itself, namely, how to reconcile conflicting results from different reputable laboratories.
Matthew Hass and Bruce Yankner at Children’s Hospital, Boston, were the first to offer evidence that γ-secretase activity may not regulate neprilysin, after all. In 2005, they reported that they were unable to find a correlation between γ-secretase cleavage of APP and neprilysin (Hass and Yankner 2005). They showed that while there is clonal variation in neprilysin activity in mouse blastocysts, that activity does not correlate with presenilin 1 (PS1) level. Months later, Bart de Strooper and colleagues at the Flanders Interuniversity Institute for Biotechnology in Leuven, Belgium, reported that in their hands AICD failed to regulate several putative transcriptional targets, including neprilysin (see Hebert et al. 2006). In their present Neuron paper, Chen and Selkoe also found no correlation between presenilin genotype and blastocyst neprilysin activity.
In their response, Raphaelle Pardossi-Piquard and colleagues at the Checler lab and elsewhere counter that blastocysts are not the ideal cell type for studying this relationship because their constitutive neprilysin level is high and does not readily respond to alterations in presenilin expression. They also note that several of their results are confirmed by Chen and Selkoe’s study: modest reductions in neprilysin in PS1/PS2 double knockout (dKO) mouse blastocyst cells; dramatically reduced neprilysin expression in PS1/PS2 dKO mouse embryonic fibroblasts (MEFs); and modest or no change in neprilysin activity in single KO cells. In regard to the last point, Chen and Selkoe found dramatic increases in neprilysin in PS1 KO cells, which they also attribute to clonal variation.
Pardossi-Piquard and colleagues also offer some new evidence in support of their theory, including lack of neprilysin activity in nicastrin-negative MEF cells (nicastrin, like presenilin, is a key component of γ-secretase) and increased neprilysin in brain extracts from Fe65/AICD transgenic mice. In contrast, Chen and Selkoe found that transforming blastocysts with AICD, Fe65, and Tip60, which are thought to form an active regulatory complex, left neprilysin activity unchanged.
More important for Alzheimer disease is perhaps what happens in differentiated tissue in the brain. Here, too, the data diverge. In their original paper, Pardossi-Piquard and colleagues reported that neprilysin activity drops to about half in the brains of APP knockout mice. Chen and Selkoe were unable to reproduce this deficit, finding instead that neprilysin activity in both APP-/- and APLP2-/- is the same as in wild-type. Both Selkoe and Checler told ARF that they have agreed to swap mouse brain samples to try to resolve this issue.
When there is conflicting data in science, further studies often go on to reveal that both sides of the issue were partially right. It may not be long before we find out if that is true in this case. As Joachim Herz, UT Southwestern, Dallas, writes in an accompanying Neuron correspondence, “the conceptual and biomedical importance of the mechanism Pardossi-Piquard and colleagues have proposed will ensure that the current controversy will be resolved quickly and in due course through independent work of other laboratories that will build on these findings and conclusions.” We will keep you posted.—Tom Fagan.
Chen AC, Selkoe DJ. Response to: Pardossi-Piquard et al., "Presenilin-Dependent Transcriptional Control of the Abeta-Degrading Enzyme Neprilysin by Intracellular Domains of betaAPP and APLP." Neuron 46, 541-554.
Neuron. 2007 Feb 15;53(4):479-83. No abstract available.
Pardossi-Piquard R, Dunys J, Kawarai T, Sunyach C, Alves da Costa C, Vincent B, Sevalle J, Pimplikar S, St George-Hyslop P, Checler F. Response to Correspondence: Pardossi-Piquard et al., "Presenilin-Dependent Transcriptional Control of the Abeta-Degrading Enzyme Neprilysin by Intracellular Domains of betaAPP and APLP." Neuron 46, 541-554.
Neuron. 2007 Feb 15;53(4):483-6. No abstract available.
Herz J. Overview: The long and winding road to understanding Alzheimer’s disease. Neuron. 2007 February 15;53:477-479. Abstract