Presenilin 1 is involved in maturation and trafficking of N-cadherin to the plasma membrane

One pathological characteristic of Alzheimer's disease (AD) is extensive synapse loss. Presenilin 1 (PS1) is linked to the pathogenesis of early onset familial Alzheimer's disease (FAD) and is localized at the synapse, where it binds N-cadherin and modulates its adhesive activity. To elucidate the role of the PS1/N-cadherin interaction in synaptic contact, we established SH-SY5Y cells stably expressing wild-type (wt) PS1 and dominant-negative (D385A) PS1. We show that the formation of cadherin-based cell-cell contact among SH-SY5Y cells stably expressing D385A PS1 was suppressed. Conversely, wt PS1 cells exhibited enhanced cell-cell contact and colony formation. Suppression of cell-cell contact in D385A cells was accompanied by an alteration in N-cadherin subcellular localization; N-cadherin was retained mainly in the endoplasmic reticulum (ER) and cell surface expression was reduced. We conclude that PS1 is essential for efficient trafficking of N-cadherin from the ER to the plasma membrane. PS1-mediated delivery of N-cadherin to the plasma membrane is important for N-cadherin to exert its physiological function, and it may control the state of cell-cell contact.     

Protein trafficking and Alzheimer's disease

Mutations in presenilin 1 (PS1) cause early-onset familial Alzheimer;s disease (FAD). Although FAD accounts for less than 5% of all cases of Alzheimer;s disease (AD), extensive analyses of PS1 function have elucidated an important neuronal mechanism underling AD pathogenesis. PS1 is considered to be an essential component of gamma-secretase, which cleaves amyloid precursor protein (APP) at the transmembrane region and releases amyloid beta (Abeta) peptide. In addition to this well-documented function, a growing amount of evidence suggests that PS1 is involved in the intracellular trafficking of selected membrane proteins (i.e. APP, nicastrin, trkB, telencephalin). Recently, we have also shown that PS1 is involved in the trafficking of N-cadherin from the endoplasmic reticulum to the plasma membrane via the microtubule network. N-cadherin is localized at the synaptic junctional complex, providing an adhesive force across the synaptic cleft, and the its regulation is crucial for the neuron to exert its specific function, i.e. synaptic activity. In a mature neuron, polarized targeting of proteins from the cell body to the axonal and dendritic processes is essential for its proper function, especially, for the maintenance of synaptic function. Alterations in protein transport caused by a dysfunction in PS1 could lead to a disturbance in synaptic transmission and finally to neurodegeneration. This article will review the current knowledge of PS1 function in protein trafficking and discuss its potential role in AD pathogenesis.     

Genes and mechanisms involved in β-amyloid generation and Alzheimer’s disease

Alzheimer's disease is characterized by the invariable accumulation of senile plaques that are predominantly composed of amyloid beta-peptide (Abeta). Abeta is generated by proteolytic processing of the beta-amyloid precursor protein (betaAPP) involving the combined action of beta- and gamma-secretase. Cleavage within the Abeta domain by alpha-secretase prevents Abeta generation. In some very rare cases of familial AD (FAD), mutations have been identified within the betaAPP gene. These mutations are located close to or at the cleavage sites of the secretases and pathologically effect betaAPP processing by increasing Abeta production, specifically its highly amyloidogenic 42 amino acid variant (Abeta42). Most of the mutations associated with FAD have been identified in the two presenilin (PS) genes, particularly the PS1 gene. Like the mutations identified within the betaAPP gene, mutations in PS1 and PS2 cause the increased generation of Abeta42. PS1 has been shown to be functionally involved in Notch signaling, a key process in cellular differentation, and in betaAPP processing. A gene knock out of PS1 in mice leads to an embryonic lethal phenotype similar to that of mice lacking Notch. In addition, absence of PS1 results in reduced gamma-secretase cleavage and leads to an accumulation of betaAPP C-terminal fragments and decreased amounts of Abeta. Recent work may suggest that PS1 could be the gamma-secretase itself, exhibiting the properties of a novel aspartyl protease. Mutagenesis of either of two highly conserved intramembraneous aspartate residues of PS1 leads to reduced Abeta production as observed in the PS1 knockout. A corresponding mutation in PS2 interfered with betaAPP processing and Notch signaling suggesting a functional redundancy of both presenilins. In this issue, some of the recent work on the molecular mechanisms involved in Alzheimer's disease (AD) as well as novel diagnostic approaches and risk factors for AD will be discussed. In the first article, we like to give an overview on mechanisms involved in the proteolytic generation of Amyloid beta-peptide (Abeta), the major pathological player of this devastating disease. In the second part of this article recent results will be described, which demonstrate an unexpected biological and pathological function of an AD associated gene.     

Age-related progressive synaptic dysfunction: the critical role of presenilin 1

Mutations in presenilin 1 gene (PS1) account for the majority of early-onset familial Alzheimer's disease (FAD) cases. The disease is characterized by intracellular neurofibrillary tangles and extracellular amyloid fibrils composed of amyloid beta peptides (Abeta). Two successive cleavages are necessary to free the Abeta peptide from the amyloid precursor protein (APP). Gamma-secretase catalyzes the final cleavage of APP to generate Abeta peptides. PS1 is a catalytic subunit of gamma-secretase and is also involved in the cleavage of many membrane proteins. PS1 also has functional interactions with many other proteins. The use of animal models of AD has initiated the deciphering of these molecular pathways and mechanisms. Transgenic mouse models are useful to study the features of FAD and to investigate the nature of the neural-tissue changes of the disease and their evolution during aging. When expressed alone, mutations in human PS1 do not induce any detectable lesions, although they do increase Abeta peptides. This absence has led to the criticism that PS1 mouse models are not valuable for the study of AD. In this review we present how studies using PS1 transgenic mice have raised new questions related to pathological mechanisms of AD and are useful models for the study of (1) progressive cognitive decline, (2) early-occurring synaptic dysfunction, and (3) mechanisms other than amyloidogenesis that can be involved in disease pathogenesis.     

Human wild presenilin-1 mimics the effect of the mutant presenilin-1 on the processing of Alzheimer's amyloid precursor protein in PC12D cells

Most familial early-onset Alzheimer's disease (FAD) is caused by mutations in the presenilin-1 (PS1) gene. Abeta 42 is derived from amyloid precursor protein (APP) and increased concentrations are widely believed to be a pathological hallmark of abnormal PS function. Thus, the interaction between PS1 and APP is central to the molecular mechanism of AD. To examine the effect of wild-type human PS1 on rat APP metabolism, we made several PC12D cell lines that expressed human wild or mutant PS1, and analyzed the processing of endogenous rat APP and the intracellular gamma-secretase activity. We found the ratio of Abeta 42/Abeta 40 increased in PC12D cells expressing wild-type human PS1. These changes were identical to those found in PC12D cells expressing human PS1 bearing the A260V mutation. These results suggest that APP metabolism is physiologically regulated by the PS1 and that loss of normal PS1 affects gamma-secretase activity.     

Pathological activity of familial Alzheimer's disease-associated mutant presenilin can be executed by six different gamma-secretase complexes

gamma-Secretase is a protease complex, which catalyzes the final of two subsequent cleavages of the beta-amyloid precursor protein (APP) to release the amyloid-beta peptide (Abeta) implicated in Alzheimer's disease (AD) pathogenesis. In human cells, six gamma-secretase complexes exist, which are composed of either presenilin (PS) 1 or 2, the catalytic subunit, nicastrin, PEN-2, and either APH-1a (as S or L splice variants) or its homolog APH-1b. It is not known whether and how different APH-1 species contribute to the pathogenic activity of gamma-secretase complexes with familial AD (FAD)-associated mutant PS. Here we show that all known gamma-secretase complexes are active in APP processing and that all combinations of APH-1 variants with either FAD mutant PS1 or PS2 support pathogenic Abeta(42) production. Since our data suggest that pathogenic gamma-secretase activity cannot be attributed to a discrete gamma-secretase complex, we propose that all gamma-secretase complexes have to be explored and evaluated for their potential as AD drug target.     

Differential display analysis of presenilin 1-deficient mouse brains

Missense mutations in presenilin 1 (PS1) gene are the most common cause of early onset familial Alzheimer's disease (FAD). AD pathogenic PS1 mutations result in elevated gamma-secretase cleavage of APP and diminished S3-site cleavage of Notch. We have previously described a PS1-hypomorphic mouse line that could survive postnatally with markedly reduced gamma-secretase cleavage of APP and S3-site cleavage of Notch, resulting in a Notch developmental phenotype similar to PS1-null mice. This model was exploited to identify genes whose expression is altered due to the loss of PS1. A global gene expression study by differential display was performed on whole brains of PS1-hypomorphic mice and their wild type siblings. In total, more than 16,000 bands corresponding to cDNAs were compared between the mutant and wild-type brains. This analysis identified 19 cDNAs showing significantly altered expression resulting from PS1 deficiency. Four of the identified cDNAs corresponded to genes that could be associated with AD or presenilin function. Hypoxia inducible factor 1a (Hif1a), NPRAP (delta-catenin) and cell division cycle 10 (CDC10) showed significantly reduced expression in the PS1-hypomorphic compared to wild-type brains, whereas expression of nucleoside diphosphate kinase sub-unit A (NDPK-A) was markedly elevated in the respective brains. Clarification of the possible role of these genes in AD and the basis for their differential expression induced by PS1-deficiency may provide insight into the disease, presenilin function and consequences of its loss, as well as possible deleterious effects of AD therapeutics aimed at inhibiting PS1.     

Presenilin endoproteolysis is an intramolecular cleavage

Mutations in the presenilin genes (PS) account for most cases of familial Alzheimer's disease. PS contain the active site of the gamma-secretase complex that cleaves within the transmembrane domain of beta-amyloid precursor protein (APP). Full-length PS undergoes regulated endoproteolysis to produce fragments that comprise the active form of PS. The "presenilinase" responsible for endoproteolysis is unknown but may be the same presenilin-dependent gamma-secretase activity that cleaves APP. To investigate the mechanism of endoproteolysis, we examined sequence specificity at the cleavage site and tested whether PS dimers are important for endoproteolysis as well as gamma-secretase activity. No single point mutation, or a double mutation M292D/V293K, was able to completely abolish endoproteolysis and all mutants supported gamma-secretase activity. When wtPS1 was co-expressed with either M292D/V293K or D257A, it was unable to restore normal endoproteolysis to either mutant. Lack of transcleavage by wtPS1 suggests that PS1 endoproteolysis occurs via intramolecular cleavage and does not require dimerization.     

Analysis of alteration of p75NTR processing and signalling by PS2 mutation and gamma-secretase inhibition

The presenilins (PSs) were identified as causative genes in cases of early-onset familial Alzheimer's disease (AD) and current evidence indicates that PSs are part of the gamma-secretase complex responsible for proteolytic processing of type I membrane proteins. p75NTR, a common neurotrophin receptor, was shown to be subject to gamma-secretase processing. However, it is not clear if the p75NTR downstream signal is altered in response to gamma-secretase cleavage, and further there is a possibility that AD-related PS mutations may affect this cleavage, resulting in pathogenic alterations in signal transduction. In this study, we confirmed that p75NTR downstream signalling is altered by PS2 mutation or gamma-secretase inhibition in SHSY-5Y cells. The activity of the small GTPase RhoA is strongly affected by these treatments. This study demonstrates that gamma-secretase and PS2 play an important role in regulating neurotrophin signal transduction and either mutation of PS2 or inhibition of gamma-secretase disturbs this function.     

Mutations in APP have independent effects on Abeta and CTFgamma generation

Understanding the molecular mechanism of beta-amyloid (Abeta) generation is crucial for Alzheimer's disease pathogenesis as well as for normal APP function. The transmembrane domain (TM) of APP appears to undergo presenilin-dependent gamma-secretase cleavage at two topologically distinct sites: a site in the middle of the TM domain that is crucial for the generation of Abeta-peptides, and a site close to the cytoplasmic border (S3-like/epsilon site) of the TM domain that leads to production of the APP intracellular domain (CTFgamma/AICD). We demonstrate that, in contrast to the unique effect of familial Alzheimer's disease (FAD) mutations in APP on Abeta42 production, some but not all FAD mutations also affect CTFgamma generation. Furthermore, changes in total CTFgamma levels do not correlate with either an increase or a decrease of any Abeta species, and inhibition of Abeta-peptide formation starting from position +1 (Abeta1-x) does not affect CTFgamma production. These results suggest that cleavage at the gamma40/42- and the S3-like sites can be dissociated, and that APP signaling and Abeta production are not tightly linked.     

Conserved "PAL" sequence in presenilins is essential for gamma-secretase activity, but not required for formation or stabilization of gamma-secretase complexes

Generation of A beta from the beta-amyloid precursor protein (APP) requires a series of proteolytic processes, including an intramembranous cleavage catalyzed by an aspartyl protease, gamma-secretase. Two aspartates in presenilins (PS) are required for gamma-secretase activity (D257 and D385 of PS1), suggesting that PS may be part of this protease. Little is known concerning the importance of other sequences in PS for activity. We introduced point mutations (P433L, A434D, L435R) into a completely conserved region C-terminal to transmembrane domain eight of PS1. The P433L mutation abolished PS1 endoproteolysis as well as gamma-secretase cleavage of APP and Notch in PS1/2 K/O cells. In HEK cells, expression of PS1/P433L reduced A beta production and caused accumulation of APP C-terminal stubs. When the P433L mutation was introduced into the non-cleavable Delta exon 9 (Delta E9) variant of PS1, it abolished gamma-secretase cleavage of APP and Notch. The P433L holoprotein is stable and incorporated into the high molecular weight gamma-secretase complex, arguing that P433 is not necessary for formation or stabilization of the gamma-secretase complex. Other non-conservative mutations in the invariant P(433)A(434)L(435) sequence also result in a phenotype that is indistinguishable from the aspartate mutants, suggesting a direct involvement of this sequence in gamma-secretase activity.     

Endoproteolysis of the ER stress transducer ATF6 in the presence of functionally inactive presenilins

Presenilin (PS) proteins facilitate endoproteolysis of selected type I transmembrane proteins such as the Alzheimer's disease (AD) associated beta-Amyloid precursor protein (beta APP) and Notch. beta APP is cleaved within its transmembrane domain by an aspartyl protease activity termed gamma-secretase, which may be identical with PS1 and PS2. Notch also undergoes a PS-dependent intramembraneous proteolysis. A similar gamma-secretase-like cleavage may also occur with IRE1 and ATF6, two signaling molecules of the unfolded protein response (UPR) that may require PSs for their activation. Here, we have analyzed whether ATF6 cleavage requires a PS-dependent gamma-secretase activity and whether inhibition of gamma-secretase activity would affect the UPR. Endoproteolysis of ATF6 was observed in the presence of the highly potent gamma-secretase inhibitor L-685,458. ATF6 processing also occurred in the presence of functionally inactive dominant negative mutants of PS1 (PS1 D385N) and PS2 (PS2 D366A) that do not support endoproteolysis of beta APP and Notch. Our results therefore demonstrate that ATF6 is not a substrate for PS mediated gamma-secretase-like endoproteolysis. This finding indicates that gamma-secretase inhibitors, which are currently developed as therapeutic agents to lower the A beta burden in brains of AD patients, do not interfere with the UPR response.     

Twenty-nine missense mutations linked with familial Alzheimer's disease alter the processing of presenilin 1.

1. Full-length form of human presenilin 1 (PS1) is processed and an N-terminal fragment (28 KD) and C-terminal fragment (19 KD) are generated. To elucidate the possible role of presenilin mutations in Alzheimer's disease (AD), the authors analyze the effects of AD-linked mutations on PS1 processing in cultured cells. 2. Complementary DNAs encoding genes for human PS1 harboring twenty-nine missense mutations linked with familial Alzheimer's disease (FAD) were introduced into PC12 cells. Human PS1 exogenously expressed in the cells was detected by immunoblotting using a monoclonal antibody that recognized the N-terminal region of human PS1. The amounts of full-length form (48 KD) and N-terminal fragment (28 KD) of PS1 was quantified by densitometrical analysis. 3. The ratio of the N-terminal fragment to total PS1 was reduced by twenty-nine mutations. The specific effects on PS1 processing varied according to mutation. 4. These results suggest that AD-linked missense mutations of PS1 are involved in neurodegeneration via inhibition of PS1 processing.     

Baculoviruses expressing the human familial Alzheimer's disease presenilin 1 mutation lacking exon 9 increase levels of an amyloid beta-like protein in Sf9 cells

Presenilin 1 (PS1) plays a pivotal role in the production of the amyloid-beta protein (Abeta) that is central to the pathogenesis of Alzheimer's disease. PS1 regulates the intramembranous proteolysis of a 99-amino-acid C-terminal fragment of the amyloid precursor protein (APP-C99), a cleavage event that releases Abeta following a reaction catalyzed by an enzyme termed 'gamma-secretase'. The molecular mechanism of PS1-mediated, gamma-secretase cleavage remains largely unresolved. In particular, controversy surrounds whether PS1 includes the catalytic site of the gamma-secretase protease or whether instead PS1 mediates gamma-secretase activity indirectly, perhaps by regulating the trafficking or presentation of substrates to the 'authentic' protease, which may be a molecule distinct from PS1. To address this issue, the baculovirus expression system was used to co-express: (i) APP-C99; (ii) a pathogenic, constitutively active mutant form of PS1 lacking exon 9 (PS1DeltaE9); (iii) nicastrin and (iv) tropomyosin in Spodoptera frugiperda (Sf9) cells. Cells infected with APP-C99 alone produced an Abeta-like species, and levels of this species were enhanced by the addition of baculoviruses bearing the PS1DeltaE9 mutation. The addition to APP-C99-infected cells of baculoviruses bearing nicastrin, also a transmembrane protein, had a neutral or inhibitory effect on the reaction; tropomyosin viruses had the same effect as nicastrin viruses. These results suggest that PS1DeltaE9 molecules expressed in Sf9 cells retain the ability to modulate Abeta levels. Baculoviral-expressed PS1DeltaE9 provides a source of microgram quantities of bioactive molecules for use as starting material for purifying and reconstituting gamma-secretase activity from its individual purified component parts.     

FAD Mutations in Presenilin-1 or Amyloid Precursor Protein Decrease the Efficacy of a γ-Secretase Inhibitor: Evidence for Direct Involvement of PS1 in the γ-Secretase Cleavage Complex

To investigate the mechanism of regulation of Aß production by familial Alzheimer's disease (FAD)-linked presenilin 1 (PS1), we used a cell-free system that allows de novo Aß generation to examine whether PS1 participates directly in the γ-secretase reaction. Optimal Aß generation in vitro was achieved at mildly acidic pH and could be inhibited by the aspartyl protease inhibitor pepstatin A, consistent with the suggestion that γ-secretase is an aspartyl protease. Dominant negative mutations of the critical transmembrane aspartates in PS1 or full deletion of PS1 did not alter the maturation of APP in the secretory pathway. Instead, PS1 had a direct effect on the inhibition of Aß production by a designed peptidomimetic inhibitor: the inhibition was significantly less effective in cells expressing FAD-causing mutations in either APP or PS1 than in cells expressing the wild-type proteins. Taken together, these findings suggest that PS1 participates physically in a complex with APP during the γ-secretase cleavage event.     

The catalytic core of gamma-secretase: presenilin revisited

Mutations in the presenilin 1 (PS1) gene are the major cause of familial Alzheimer s disease (AD). They effect an increased production of the highly neurotoxic 42 amino acid variant of the amyloid-beta peptide (Abeta), which is believed to initiate the disease. Abeta is the product of two consecutive cleavages of the beta-amyloid precursor protein (APP) by two proteases, beta-secretase and gamma-secretase. The latter enzyme has been identified as an intramembrane-cleaving multiprotein complex that apart from APP cleaves a large number of other type I transmembrane proteins. PS1 and its homologue PS2 are essential for gamma-secretase cleavage and more than a decade after their discovery it is now firmly established that they function as catalytic subunits of gamma-secretase. This review recapitulates the findings that led to this conclusion as well as the further progress made on the function of PS as gamma-secretase since then.     

Assembly, trafficking and function of gamma-secretase

Gamma-secretase catalyzes the final cleavage of the beta-amyloid precursor protein to generate amyloid-beta peptide, the principal component of amyloid plaques in the brains of patients suffering from Alzheimer's disease. Here, we review the identification of gamma-secretase as a protease complex and its assembly and trafficking to its site(s) of cellular function. In reconstitution experiments, gamma-secretase was found to be composed of four integral membrane proteins, presenilin (PS), nicastrin (NCT), PEN-2 and APH-1 that are essential and sufficient for gamma-secretase activity. PS, which serves as a catalytic subunit of gamma-secretase, was identified as a prototypic member of novel aspartyl proteases of the GxGD type. In human cells, gamma-secretase could be further defined as a heterogeneous activity consisting of distinct complexes that are composed of PS1 or PS2 and APH-1a or APH-1b homologues together with NCT and PEN-2. Using green fluorescent protein as a reporter we localized PS and gamma-secretase activity at the plasma membrane and endosomes. Investigation of gamma-secretase complex assembly in knockdown and knockout cells of the individual subunits allowed us to develop a model of complex assembly in which NCT and APH-1 first stabilize PS before PEN-2 assembles as the last component. Furthermore, we could map domains in PS and PEN-2 that govern assembly and trafficking of the complex. Finally, Rer1 was identified as a PEN-2-binding protein that serves a role as an auxiliary factor for gamma-secretase complex assembly.     

Presenilin Proteins Undergo Heterogeneous Endoproteolysis between Thr291and Ala299and Occur as Stable N- and C-Terminal Fragments in Normal and Alzheimer Brain Tissue

Humans inheriting missense mutations in thepresenilin(PS)1 and -2 genes undergo progressive cerebral deposition of the amyloid β-protein at an early age and develop a clinically and pathologically severe form of familial Alzheimer's disease (FAD). Because PS1 mutations cause the most aggressive known form of AD, it is important to elucidate the structure and function of this multitransmembrane protein in the brain. Using a panel of region-specific PS antibodies, we characterized the presenilin polypeptides in mammalian tissues, including brains of normal, AD, and PS1-linked FAD subjects, and in transfected and nontransfected cell lines. Very little full-length PS1 or -2 was detected in brain and untransfected cells; instead the protein occurred as a heterogeneous array of stable N- and C-terminal proteolytic fragments that differed subtly among cell types and mammalian tissues. Sequencing of the major C-terminal fragment from PS1-transfected human 293 cells showed that the principal endoproteolytic cleavage occurs at and near Met₂₉₈in the proximal portion of the large hydrophilic loop. Full-length PS1 in these cells is quickly turned over (T_1/2 ≈ 60 min), in part to the two major fragments. The sizes and amounts of the PS fragments were not significantly altered in four FAD brains with the Cys410Tyr PS1 missense mutation. Our results indicate that presenilins are rapidly processed to N- and C-terminal fragments in both neural and nonneural cells and that interference with this processing is not an obligatory feature of FAD-causing mutations.     

Amyloid precursor protein gene analysis in familial Alzheimer's disease cases: a lack of mutations in exons 16 and 17

The beta-amyloid precursor protein (APP) gene (on chromosome 21), Presenilin 1 (PS1) gene (on chromosome 14) and Presenilin 2 (PS2) gene (on chromosome 1) are responsible for autosomal dominant early-onset Alzheimer's disease (EOAD). Missense mutations in these genes cause abnormal APP processing with subsequent overproduction of amyloidogenic and toxic A beta (42 peptide. A mutational analysis of APP, PS1, and PS2 genes can be used for both symptomatic and presymptomatic genetic testing and counselling in familial Alzheimer's disease (FAD). To contribute to our knowledge on genetic background of Alzheimer's disease in Poland, we screened APP mutations in a sample of familial EOAD cases from Poznan region. We did not find pathogenic mutations within exons 16 and 17 of the APP gene. Our study confirmed that APP gene mutations account only for a very small portion of FAD.     

Presenilin-dependent gamma-secretase activity modulates neurite outgrowth

Demonstration that cleavage of both APP and Notch are dependent on the product of the early onset Alzheimer's disease gene, presenilin-1 (PS1), has raised the possibility that Notch function may be altered in AD. This finding also suggests that Notch may be affected by APPgamma-secretase inhibitors under development for the treatment of Alzheimer's disease, as these target PS1. Data that address these questions have been lacking, due to inability to specifically modulate PS1 activity in a system directly relevant to the adult human brain. Using novel highly specific inhibitors of PS1/gamma-secretase, we demonstrate that modulation of PS1 activity in human CNS neurons not only affects Abeta generation, but also has unanticipated effects on Notch and its activity. We demonstrate that intracellular trafficking of Notch in human CNS neurons is altered by inhibition of PS1 and is accompanied by dramatic changes in neurite morphology, consistent with inhibition of Notch activity. These data, together with immunohistochemical evidence of elevation of Notch pathway expression in AD brain, suggest that Notch dysregulation may contribute to the neuritic dystrophy characteristically seen in Alzheimer's disease brain. In addition, they raise the possibility that inhibition of gamma-secretase/PS1 may have clinically beneficial effects on the neuritic pathology of AD, in addition to its expected effect to reduce amyloid burden.