PKCε increases endothelin converting enzyme activity and reduces amyloid plaque pathology in transgenic mice

Abstract

Deposition of plaques containing amyloid β (Aβ) peptides is a neuropathological hallmark of Alzheimer's disease (AD). Here we demonstrate that neuronal overexpression of the ε isozyme of PKC decreases Aβ levels, plaque burden, and plaque-associated neuritic dystrophy and reactive astrocytosis in transgenic mice expressing familial AD-mutant forms of the human amyloid precursor protein (APP). Compared with APP singly transgenic mice, APP/PKCε doubly transgenic mice had decreased Aβ levels but showed no evidence for altered cleavage of APP. Instead, PKCε overexpression selectively increased the activity of endothelin-converting enzyme, which degrades Aβ. The activities of other Aβ-degrading enzymes, insulin degrading enzyme and neprilysin, were unchanged. These results indicate that increased neuronal PKCε activity can promote Aβ clearance and reduce AD neuropathology through increased endothelin-converting enzyme activity.

Alzheimer's disease (AD) is the most common cause of dementia in the elderly and is characterized pathologically by the deposition of senile plaques containing amyloid β (Aβ) peptides in the brain (1). Aβ is composed primarily of 40- and 42-aa peptides produced from the β-amyloid precursor protein (APP) through the combined proteolytic actions of β- and γ-secretase. In cultured cells, treatment with tumor-promoting phorbol esters that activate several PKC isozymes reduces Aβ levels through a mechanism that may involve α-secretase-mediated cleavage of APP (2, 3). However, although intraparenchymal administration of phorbol esters reduces Aβ levels and deceases amyloid plaque density in mice expressing an amyloidogenic variant of human APP, it does not increase α-secretase activity in the brain (4). This finding raises the possibility that PKC decreases Aβ levels in vivo by other mechanisms. Such mechanisms might involve reductions in the activity of amyloidogenic proteases such as BACE1 and γ-secretase (5).

Aβ accumulation is influenced not only by enzymes responsible for its production but also by mechanisms involved in its clearance (5). Several proteases that degrade Aβ in mice include insulin-degrading enzyme (IDE), neprilysin (NEP), and endothelin-coverting enzyme (ECE) 1 and 2 (6). Recently, Leissring et al. (7) demonstrated that transgenic overexpression of IDE or NEP in neurons significantly reduces Aβ levels and plaque-associated AD pathology. The in vivo function of ECE has been examined in ECE-1- and ECE-2-deficient mice and, in both cases, levels of Aβ were increased compared with WT mice, indicating that these ECEs are Aβ-degrading enzymes in vivo (8). However, it is not yet known whether stimulation of ECE activity can reduce Aβ levels or plaque-associated AD pathology, although in one association study, a genetic variant of human ECE1 (ECE1B C-338A) that shows increased promoter activity was associated with a reduced risk of sporadic AD in a French Caucasian population (9).

Several in vitro cell culture studies that have examined phorbol ester-stimulated secretion of APP fragments suggest that PKCα and PKCε stimulate α-secretase activity (10–14), although only PKCε also has been shown to reduce Aβ levels in vitro (15). Here we used mice that overexpress PKCε under control of a neuron-specific promoter to examine whether PKCε also can reduce Aβ production in vivo. Our results demonstrate that PKCε overexpression reduces the number of amyloid plaques and the level of Aβ in the brains of transgenic mice that express amyloidogenic variants of human APP. However, contrary to findings in cell culture systems, this reduction did not result from increased α-secretase-mediated cleavage of APP but rather was associated with an increase in brain ECE activity. These results provide evidence demonstrating that PKCε up-regulates ECE in the CNS and that increased ECE activity can reduce AD neuropathology.

Results

Generation of Mice That Overexpress PKCε in Neurons.

To examine whether PKCε reduces levels of Aβ in vivo, we generated transgenic mice expressing human PKCε directed by the Thy-1.2 promoter (Fig. 1A), which drives strong constitutive expression of transgenes specifically in adult neurons (16, 17). We identified several founder lines by immunoblot analysis (Fig. 1B) whose cerebral PKCε levels in the brain were 3- to 5-fold higher than in nontransgenic controls. Immunostaining of brain sections for PKCε revealed a similar pattern of PKCε expression in transgenic and nontransgenic mice, with widespread labeling of the neuropil and neuronal cell bodies in all layers of the neocortex and hippocampus (Fig. 1C).

Fig. 1.

Overexpression of PKCε reduces amyloid plaque burden and inhibits Aβ accumulation in brain parenchyma. (A) Schematic of the human PKCε transgene. (B) Immunoblot of brain homogenates from wild-type mice (Wt) and from transgenic PKCε (Tg1 and Tg2) and nontransgenic (Ct1 and Ct2) littermates. (C) Similar distribution of PKCε-like immunoreactivity in hippocampal CA3 and CA1 regions and frontal cortex (Ctx) of Wt and PKCε_Tg1 mice. (D and E) Thioflavin-S staining (D) and anti-Aβ immunostaining (3D6 antibody) (E) showing fewer plaques and Aβ immunoreactive deposits in the hippocampus and neocortex in APP_Ind/PKCε_Tg1 mice than in APP_Ind mice. (Scale bars: 200 μm.) (F and G) Quantification of Thioflavin-S staining (F) and Aβ deposits (G) is expressed as the percentage of total surface area of the hippocampus and cortex in each section. Data are from 10 mice per genotype, averaged from three sections per mouse. ∗, P < 0.05 by two-tailed t tests.

Transgenic PKCε Decreases Plaque Burden and Aβ Deposition.

We crossed PKCε transgenic line 1 (PKCε_Tg1), which expresses 3-fold more PKCε in the brain than nontransgenic controls, with an APP transgenic line (H6) expressing familial AD (V717F, “Indiana”)-mutant APP (APP_Ind) in neurons directed by the platelet-derived growth factor β chain promoter (18). Singly transgenic APP_Ind mice showed some thioflavin-S-positive plaques by 9–10 months (data not shown) and had an average of 60 plaques per sagittal brain section at 12–18 months of age (Fig. 1 D and F). Notably, doubly transgenic APP_Ind/PKCε_Tg1 mice had dramatically fewer plaques at 12–18 months (Fig. 1 D and F). Immunostaining with the 3D6 antibody, which specifically recognizes human Aβ, also revealed formation of Aβ deposits at ≈9–10 months in APP_Ind mice (data not shown) and a striking reduction in Aβ deposits in APP_Ind/PKCε_Tg1 mice at 12 and 18 months of age (Fig. 1E). This result was quantified in 12- to 15-month-old animals as a 95% reduction in the area occupied by Aβ deposits in APP_Ind/PKCε_Tg1 mice compared with APP_Ind mice (Fig. 1G).

PKCε Decreases Plaque-Associated Astrocytosis and Dystrophic Neurites.

The decreased plaque load in APP_Ind/PKCε_Tg1 mice was associated with a prominent reduction in neuritic dystrophy and reactive astrocytosis. The neocortex of APP_Ind mice contained large clusters of glial fibrillary acidic protein (GFAP)-positive astrocytes surrounding amyloid plaques, whereas only faint GFAP labeling was seen in this region in APP_Ind/PKCε_Tg1 mice (Fig. 2A and B). Plaques in the hippocampus of APP_Ind mice also were associated with reactive astrocytes (Fig. 2C) and displacement of neurites (Fig. 2E), whereas these alterations were absent or much less prominent in APP_Ind/PKCε_Tg1 mice (Fig. 2 D and F). Thus, overexpression of PKCε prevents the accumulation of Aβ and the neurodegenerative changes associated with this process.

Fig. 2.

PKCε decreases plaque-associated neuropathology. (A and B) Patches of GFAP-positive astrocytes were detected by confocal microscopy in the parietal cortex in singly transgenic APP_Ind mice but not in doubly transgenic APP_Ind/PKCε_Tg1 mice. (C and D) Dual-channel immunofluorescence images (anti-APP, green; GFAP, red) show astrocytosis and disruption of the pyramidal cell layer in the hippocampus of an APP_Ind mouse and no such changes in an APP_Ind/PKCε_Tg1 mouse. (E and F) Dual-channel immunofluorescence images (anti-APP, green; MAP-2, red) demonstrating plaques (arrows) and displacement of dendrites in the CA1 region of an APP_Ind mouse but not in an APP_Ind/PKCε_Tg1 mouse. (Scale bars: A–D, 100 μm; E and F, 50 μm.)

PKCε Promotes Aβ Clearance Independent of APP Mutation.

To confirm that the PKCε-dependent reduction in plaque pathology was not due to line-specific effects, we analyzed two additional lines of transgenic mice: PKCε transgenic line 2 (PKCε_Tg2; Fig. 1B) and APP_Swe,_Ind line J20 (18). Mice from the J20 line carry the Indiana (V717F) and the Swedish (K670N/M671L) APP mutations and develop plaques as early as 3 months of age. At 8–10 months, the hippocampal burden of thioflavin-S-positive plaques was 60% lower in APP_Swe,Ind/PKCε_Tg2 mice than in singly transgenic APP_Swe,_Ind mice (see Fig. 5, which is published as supporting information on the PNAS web site), suggesting that the plaque-reducing effect of PKCε overexpression is robust and largely independent of transgene insertion sites.

PKCε-Driven Decrease in Aβ Is Mediated by ECE Activity, Not by APP Processing.

To assess whether reduction of Aβ accumulation involves alterations in APP metabolism, we used Western blot analysis to measure the levels of α-secreted APP (sAPPs), full-length APP (flAPP), and C-terminal fragments (CTFs) generated by α- and β-secretase, respectively. APP_Ind mice and APP_Ind/PKCε_Tg1 mice did not differ significantly in sAPPs/flAPP ratios and in relative levels of CTFs (Fig. 3). At 12–18 months, hippocampal levels of Aβ1-x (total Aβ) and Aβ1–42 and Aβ1–42/Aβ1-x ratios were higher in APP_Ind mice than in APP_Ind/PKCε_Tg1 mice, consistent with their difference in plaque load (Table 1). However, at 1–3 months, when both groups of mice were devoid of plaques, no significant differences were detected in hippocampal levels of Aβ1-x and Aβ1–42 or in Aβ1–42/Aβ1-x ratios between APP_Ind mice and APP_Ind/PKCε_Tg1 mice, although there was a trend for lower values in the doubly transgenic mice (Table 1).

Fig. 3.

Overexpression of PKCε does not change levels of αAPPs and C-terminal fragments (CTFs) of APP. (A Upper) A representative immunoblot of sAPP and flAPP immunoreactivity from hippocampi of individual mice. (A Lower) The quantification of immunoreactive bands, indicating no difference between the genotypes (n = 7 for APP_Ind and n = 8 for APP_Ind/PKCε_Tg1) in relative sAPP levels, which were normalized to the level of flAPP. (B Upper) Representative immunoblots of CTF-α (C83) and CTF-β (C99) immunoreactivity. (B Lower) No difference in levels of these CTFs between APP_Ind and APP_Ind/PKCε_Tg1 mice (n = 5 for each genotype). CTF immunoreactivity was normalized to GAPDH immunoreactivity.

Table 1.

Aβ levels in the hippocampus of young and old mice

Age, mo	Genotype	n	Aβ1-x, ng/g	Aβ1-42, ng/g	Aβ1-42/Aβ1-x
1–3	APP	8	116.8 ± 12.9	27.3 ± 2.8	0.235 ± 0.003
	APP/PKCε	8	95.8 ± 9.8	20.9 ± 2.2	0.217 ± 0.003
12–18	APP	9	1,599.4 ± 630.6	1,350 ± 573.4	0.582 ± 0.010
	APP/PKCε	10	383.2 ± 268.8*	261.8 ± 186.2*	0.329 ± 0.08*

*, P < 0.05 versus age-matched APP mice (Mann–Whitney U test).

Because it seemed unlikely that such subtle effects on APP metabolism could account for the striking reduction in plaque pathology, we examined whether PKCε prevented plaque formation by activating proteases that degrade Aβ. We first examined insulin-degrading enzyme and neprilysin, because overexpression of these proteases decreases levels of Aβ (5), and found that their activities were similar in APP_Ind/PKCε_Tg1 and APP_Ind mice (Fig. 4A and B). In contrast, ECE activity was significantly higher in the frontal cortex (Fig. 4C) and hippocampus (Fig. 4D) of 9-month-old APP_Ind/PKCε_Tg1 mice compared with age-matched APP_Ind mice. These results indicate that neuronal overexpression of PKCε selectively increases ECE activity in mouse brain.

Fig. 4.

Effect of PKCε on Aβ-degrading enzymes. (A and B) IDE (A) and NEP (B) activity in the forebrain were similar in APP_Ind mice (n = 3) and APP_Ind/PKCε_Tg1 mice (n = 5). (C and D) ECE activity in the parietal cortex (C) and hippocampus (D) of APP_Ind mice (n = 7) and APP_Ind/PKCε_Tg1 mice (n = 8). (E) PKCε levels were determined by Western blot analysis (Left) and ECE (Right) activity in PKCε-transfected (n = 12) and mock-transfected EA.hy926 cultures (n = 12) after treatment with 16 nM phorbol 12-myristate 13-acetate for 24 h. ∗, P < 0.05 by two-tailed, unpaired t tests in C–E.

Phorbol esters that activate several PKC isozymes increase ECE1 activity in the human endothelial cell line EA.hy926 (19). Because our results suggested that PKCε regulates ECE activity in the brain, we investigated whether PKCε also mediates phorbol ester-stimulated ECE1 activity in this cell line. We found that after treatment with phorbol ester, ECE activity was 40% greater in PKCε-transfected cells than in mock-transfected cells (Fig. 4E).

Ablation of PKCε Does Not Alter Aβ Metabolism.

Because ECE activity was increased by overexpression of PKCε, we examined whether absence of PKCε is associated with decreased ECE activity. We found that ECE activity was similar in cerebral cortex and hippocampus of PKCε^−/− mice and PKCε^+/+ littermates (see Fig. 6, which is published as supporting information on the PNAS web site). We also examined whether absence of PKCε enhances plaque formation. Heterozygous transgenic APP_Ind mice from line H6 were crossed with PKCε^+/− mice (20), and then F1 APP_Ind/PKCε^+/− mice were bred to generate APP_Ind mice on a PKCε^+/+ versus PKCε^−/− background. There was no difference between these genotypes in hippocampal Aβ1-x levels at 2 and 6 months (data not shown) or in Aβ1-x levels and plaque load at 9 months of age (see Fig. 6).

Discussion

The current study identifies a previously uncharacterized role for PKCε in promoting Aβ clearance in vivo. Our findings indicate that overexpressing PKCε reduces amyloid plaques in two transgenic mouse models of AD. PKCε appears to act by increasing ECE activity, which we demonstrated was regulated by PKCε both in vitro and in vivo. The reduction in amyloid plaques does not result from increased activity of α-secretase, as predicted by prior in vitro work (15). Our findings likely explain why a previous study of mice expressing an amyloidogenic variant of human APP found that intraparenchymal administration of phorbol ester reduces Aβ levels and deceases amyloid plaque density without increasing α-secretase activity in the brain (4).

Our study also suggests that increased ECE activity can reduce levels of Aβ and decrease the number of amyloid plaques in the brain. ECEs are a class of type II membrane zinc metalloproteases named for their activity to hydrolyze biologically inactive big endothelins (ETs) to the shorter active form (21). Two different ECEs have been cloned. Although ECE-1 is abundantly expressed in vascular endothelial cells (22), it is also expressed in nonvascular cells, including hippocampal and neocortical pyramidal neurons, cerebellar Purkinje cells, and astrocytes (23). ECE-2 is also expressed in the brain and is found in several subpopulations of neurons in the thalamus, hypothalamus, amygdala, and hippocampus (24). Whether endothelial, glial, or neuronal ECE activity is responsible for the results we observed in the present study is not yet known.

Although we are not yet certain which ECE is regulated by PKCε in the brain, our findings in EA.hy926 cells clearly indicate that PKCε regulates the expression of ECE-1. A single gene on chromosome 1 (1p36) encodes four isoforms of human ECE-1 (1a, 1b, 1c, and 1d), each from a distinct promoter (25). Luciferase reporter assays have shown that phorbol 12-myristate 13-acetate and coexpression of the transcription factor Ets-1 increase ECE-1a promoter activity in EA.hy926 cells (19). Therefore, our finding of increased ECE activity in PKCε-transfected EA.hy926 cells suggests that PKCε regulates ECE-1a expression via Ets-1. Consistent with this hypothesis is evidence that phorbol 12-myristate 13-acetate treatment increases Ets-1 expression-cultured endothelial cells (26). The precise mechanism by which PKCε regulates ECE-1a and, perhaps, other ECE-1 isoforms awaits further investigation.

Because overexpression of PKCε decreased levels of Aβ and plaque density, we also examined whether the absence of PKCε would decrease ECE activity, increase Aβ levels, and accelerate plaque deposition in mice expressing amyloidogenic variants of APP. PKCε^−/− and PKCε^+/+ mice showed similar ECE activity in cerebral cortex and hippocampus, suggesting that PKCε is not required for basal ECE activity in the brain. In addition, we found no additive Aβ accumulation in APP_Ind mice homozygous for the PKCε null mutation. This result is consistent with normal levels ECE activity in PKCε^−/− mice and with the existence of multiple Aβ-degrading enzymes in the brain (5), which likely reflects biological redundancy within this important metabolic network.

Our results indicate that PKCε regulates brain ECE activity and suggest that this regulation contributes to plaque clearance in mice that overexpress PKCε. However, it is possible that PKCε acts through additional mechanisms that also promote clearance of plaques. These mechanisms include circulating serine proteases that can also degrade Aβ, such as tissue plasminogen activator and urokinase-type plaminogen activator, and proteins that regulate Aβ transport, such as apolipoprotein E, α2 macroglobulin, lipoprotein receptor-related protein, and the receptor for advanced glycation end products (reviewed in ref. 27). We found that levels of apolipoprotein E were similar in APP_Ind and APP_Ind/PKCε_Tg1 mice (see Fig. 7, which is published as supporting information on the PNAS web site). Further work is needed to determine whether PKCε alters the abundance or function of other proteins that promote the clearance of Aβ.

In conclusion, although previous in vitro studies of AD pathogenesis have documented effects of PKC on Aβ production (13, 15), the current study has identified a previously undescribed role for PKCε in promoting Aβ clearance. Our results indicate that increasing neuronal PKCε activity can reduce the amyloid plaque burden in AD-related transgenic mouse models. This effect is likely mediated by increased activity of ECE, which is regulated by PKCε both in vitro and in vivo. Strategies to increase PKCε activity in the brain may increase Aβ clearance in the setting of pathologically increased Aβ accumulation and, thus, might prove useful in the treatment of AD.

Methods

Animals.

PKCε transgenic mice were generated with a fusion gene consisting of a FLAG (DYKDDDDK)-tagged human PKCε cDNA placed under the regulatory control of the murine Thy-1.2 promoter (16). To generate PKCε transgenic mice, a Kozak sequence (CCCGCCGCCACC) was added to a FLAG-tagged human PKCε cDNA and cloned into a XhoI site of the vector. The resulting NotI–PvuI fusion gene (8.7 kb) was microinjected into the male pronucleus of C57BL/6J single-cell embryos, which were implanted into pseudopregnant CD1 foster dames. Founder lines were identified from tail DNA and further characterized by Southern blotting to confirm that each transgene was integrated at a single locus. Routine genotyping was performed by PCR with a primer set: 5′-AGTAGGTGGCCATGAACTTGTG-3′, 5′-CTGAGTGGCAAAGGACCTTAGG-3′. These primers generated fragments of 700 bp. The generation of platelet-derived growth factor β chain-APP_Ind line H6 and platelet-derived growth factor β chain-APP_Swe,Ind line J20 has been described in ref. 18. The transgenic APP and PKCε lines were maintained on a C57BL/6J background. Heterozygous singly transgenic PKCε mice and APP mice were crossed to yield four genotypes: PKCε, APP, PKCε/APP, and nontransgenic (non-Tg).

PKCε^−/− mice were generated by homologous recombination as described in ref. 20. PKCε^+/− mice were maintained on an inbred 129/S4 background and crossed with C57BL/6J mice to generate F₁ generation C57BL/6J × 129/S4 hybrid PKCε^+/− mice. These mice were crossed with H6 mice on a mixed C57BL/6J × DBA/2J background. APP/PKCε^+/− offspring from this cross were intercrossed to yield APP/PKCε^+/+ and APP/PKCε^−/− littermates for experiments.

Mice were housed in standard Plexiglas cages with rodent chow and water available ad libitum. The colony room was maintained on a 12 h light/dark cycle with lights on at 0600 hours. Animal care and handling procedures were approved by the Gallo Center Institutional Animal Care and Use Committee in accord with National Institutes of Health guidelines.

Histology and Immunohistochemistry.

Mice were anesthetized with chloral hydrate and killed by transcardiac saline perfusion. Brains were removed rapidly and drop-fixed in phosphate-buffered 4% paraformaldehyde at 4°C for 48 h. Postfixed brains were cut sagitally into 40-μm sections with a vibratome (VT1000S; Leica Instruments, Nussloch, Germany). Thioflavin-S staining was performed as described in ref. 28. For each mouse, thioflavin-S-positive plaques were counted in five to six sections per hemibrain at a magnification of ×40. Individual images also were captured electronically and transferred to a Bioquant image analysis system (R & M Biometrics, Nashville, TN), and the number and area of plaques were measured as described in ref. 28. Images were pseudocolored in photoshop (Adobe Systems, San Jose, CA).

For immunohistochemistry, sections were first incubated with 3 μg/ml monoclonal anti-Aβ_1–5, (3D6; Elan Pharmaceuticals, South San Francisco, CA), then horse anti-mouse biotinylated antibodies, followed by ABC reagents (Vector Laboratories). Sections were viewed by light microscopy, and images were captured electronically with a Spot-2 camera (Diagnostic Instruments, Sterling Heights, MI). Images were transferred to a Bioquant image analysis system to determine the relative areas occupied by immunoreactive structures.

For immunofluorescence, sections were incubated in 50% alcohol for 20 min, rinsed in PBS, and then incubated in PBS containing 10% normal donkey serum (Jackson ImmunoResearch, West Grove, PA) for 24–48 h with primary antibodies at 4°C. This step was followed by incubation in secondary antibodies for 2 h. Primary antibodies were as follows: monoclonal anti-human APP (8E5, 1:10,000 dilution; Elan Pharmaceuticals), polyclonal anti-MAP2 (1:500 dilution; Chemicon International, Temecula, CA), and monoclonal anti-GFAP (1:1,000 dilution; Sigma-Aldrich). We used the TSA Biotin System (PerkinElmer) for double labeling with two monoclonal antibodies, GFAP (1:1,000 dilution; Sigma), and anti-human APP (8E5, 1:50,000 dilution; Elan Pharmaceuticals). Secondary fluorescent antibodies were donkey anti-rabbit Cy3 and donkey anti-mouse FITC or Texas red (1:250 dilution; Jackson ImmunoResearch). Sections were mounted on gelatin-coated slides, air-dried, coverslipped with Vectashield mounting media (Vector Laboratories), and examined with a Zeiss LSM 510 laser confocal microscope.

Aβ ELISA.

Homogenization of snap-frozen hippocampi and cortices in 5 M guanidine buffer and ELISA measurements of human Aβ1–x (approximates total Aβ) and Aβ1–42 were performed as described in ref. 29. For the detection of Aβ1–x, a sandwich ELISA consisting of capture antibody 266 (anti-human Aβ3–28) and biotinylated reporter antibody 3D6 (anti-human Aβ1–5) were used. Hence, this assay measures levels of Aβ species ranging from amino acid positions 1–28 or longer. Aβ1–42 levels were measured with capture antibody 21F12 (anti-human Aβ33–42) and biotinylated detection antibody 3D6. All antibodies were from Elan Pharmaceuticals.

ECE Activity in Brain.

Hippocampal and cortical tissues were analyzed as described in ref. 30. Briefly, brain tissues were washed twice in PBS, and membranes were prepared by homogenization in 4 ml of homogenization buffer (50 mM Tris/maleate, pH 6.8/0.1 mM PMSF), followed by centrifugation at 100,000 × g for 30 min. This step was repeated, and final membrane pellets were dissolved in 500 ml of 50 mM Tris/maleate (pH 6.8) containing 1% (wt/vol) N-octyl glucoside (Sigma-Aldrich) for 1 h on ice, before centrifugation at 20,000 × g for 60 min. Supernatant was stored at 4°C and used directly for enzyme assays. Protein concentration was determined by the Bio-Rad Protein Assay by using BSA as a standard. Fifty milligrams of membrane protein was incubated with rat big ET-1 (100 ng) for 4 h at 37°C in 250 ml of a reaction mixture containing 50 mM Tris·HCl buffer, pH 7.1. The reaction was stopped by adding 600 ml of cold ethanol (−20°C). After centrifugation at 10,000 × g for 10 min, the resulting supernatant was lyophilized in a benchtop lyophilizer (Laboratory Equipment Co., Hayward, CA). Dried residues were reconstituted with 250 ml of assay buffer, and the production of ET-1 in each sample was measured in duplicate 100-ml samples by an ELISA for ET-1 with a 96-well plate reader (Amersham Pharmacia Biosciences). Blanks consisted of all reagents minus enzyme, and blank OD values were subtracted from each sample OD value. A cubic-spline curve was fitted to the standards, and activity in each sample was interpolated from the standard curve.

ECE Activity in Cultured Cells.

Human umbilical vein endothelial cells (EA.hy926) expressing endogenous ECEs (31) were transfected with a pCDNA3 vector encoding human PKCε by using FuGene 6 (Roche Molecular Biochemicals, Indianapolis). The cells were subsequently treated with 16 nM phorbol 12-myristate 13-acetate (Sigma) in growth medium containing 0.25% FBS for 24 h before measuring ECE activity (8). Cells were collected and homogenized in 1 ml of 20 mM Tris·HCl (pH 7.0)/5 mM MgCl₂/0.1 mM PMSF/20 mM pepstatin A/20 mM leupeptin by using a polytron homogenizer. The homogenate was centrifuged at 1,300 × g for 10 min, and the resulting supernatant was further centrifuged at 100,000 × g for 60 min. The crude membrane pellet was solubilized for 60 min in homogenization buffer (250 ml) containing 0.5% (wt/vol) Triton X-100 to give a final protein concentration of 2–3 μg/ml. The sample was centrifuged at 100,000 × g for 60 min, and the resulting supernatant was used as the solubilized membrane fraction and stored at 4°C until use. Cell membrane fractions (20 mg of protein) were incubated for 20 min at 37°C with 0.1 μM big ET-1 (Sigma) in 0.1 M sodium phosphate buffer (pH 6.8) containing 0.5 M NaCl, 1 mM PMSF, 100 mM leupeptin, and 20 mM pepstatin. Reactions were stopped by the addition of EDTA (5 mM), and mature ET-1_1–21 peptide was quantified in duplicate by sandwich ELISA (Amersham Pharmacia Biosciences) as described above.

IDE and NEP Assays.

Frozen cerebral hemispheres were homogenized by using a Teflon glass homogenizer in 8 vol of 0.25 M sucrose in 50 mM Tris·HCl (pH 7.4 at 4°C) containing 1:100 dilution of protein phosphatase inhibitor 1 mixture (Sigma-Aldrich). The homogenate was centrifuged at 1,000 × g for 10 min to remove nuclei and debris, and then the supernatant was centrifuged at 100,000 × g for 1 h at 4°C. The supernatants were used for IDE assay as described in ref. 7. The pellets were resuspended in the same volume of ice-cold PBS with phosphatase inhibitor mixture and centrifuged at 100,000 × g again for 1 h. The pellets were resuspended in 50 mM Tris·HCl (pH 7.4) with phosphatase inhibitor mixture and sonicated by using a 10-s pulse from a Fisher sonic dismembranator at 60% power. The resuspended pellets were used for NEP assay as described in refs. 7 and 32.

Western Blot Analysis.

The ratio of sAPP and flAPP was measured by Western blot analysis with anti-Aβ_1–17, 6E10 (Signet Laboratories, Dedham, MA), and anti-APP_444–592, 8E5 (Elan Pharmaceuticals), respectively. Half of each hippocampus was homogenized in buffer containing 50 mM Tris·HCl (pH 8.0), 150 mM NaCl, 5 mM EDTA, and protease inhibitor tablet (Roche Molecular Biochemicals, Indianapolis). The homogenates were centrifuged at 14,000 × g at 4°C for 10 min. SDS sample buffer (5×) was added to the supernatant to final 1×. The proteins were separated on 4–12% Bis-Tris gels. Proteins in gels were transferred to nitrocellulose membranes, which were incubated for 2 h at 4°C in blocking buffer containing 5% BSA and 0.05% Tween 20 in Tris-buffered saline (10 mM Tris, pH 7.5/100 mM NaCl). The membranes were incubated overnight at 4°C with primary antibodies against 6E10 (1:2,000 dilution) and 8E5 (1:2,000 dilution).

The levels of CTF-α and CTF-β were measured by Western blot analysis from samples prepared above (33). The membranes were incubated overnight at 4°C with anti-C-terminal APP antibody (1:4,000 dilution; Sigma) and anti-GAPDH antibody (1:1,000 dilution; Abcam, Inc., Cambridge, MA). For measurement of ApoE, parallel blots were incubated with anti-ApoE antibody (1:2,000 dilution; provided by Dr. Ya Dong Huang, University of California, San Francisco) and anti-GAPDH antibody. Membranes were then washed three times for 10 min, each in blocking buffer, and incubated with horseradish peroxidase-linked secondary anti-rabbit antibody (1:1,000 dilution; Cell Signaling Technology, Bedford, MA) for 30 min at room temperature. Membranes were washed four times with blocking solution, and immunoreactivity was detected by enhanced chemiluminescence. Immunoreactive bands on autoradiograms were quantified by using image j (http://rsb.info.nih.gov/ij).

Statistical Analysis.

All data are presented as mean ± SE values. Data were analyzed by Kolmogorov–Smirnov tests for normality, and then means were compared by two-tailed t tests or Mann–Whitney U tests as appropriate and considered significantly different where P < 0.05.

Abbreviations

Aβ

amyloid β

AD

Alzheimer's disease

APP

amyloid precursor protein

CTF

C-terminal fragment

ECE

endothelin-converting enzyme

ET

endothelin

flAPP

full-length APP

GFAP

glial fibrillary acidic protein

IDE

insulin-degrading enzyme

NEP

neprilysin

SAPP

α-secreted APP.