Expression and functional profiling of neprilysin, insulin degrading enzyme and endothelin converting enzyme in prospectively studied elderly and Alzheimer’s brain

Abstract

The brain steady state level of Aβ is determined by the balance between its production and removal, the latter through egress across blood and CSF barriers as well as Aβ degradation. The major Aβ degrading enzymes are neprilysin (NEP), insulin degrading enzyme (IDE) and endothelin converting enzyme (ECE-1). Although evidence suggests that NEP is down-regulated in AD, the role of IDE and ECE in the Aβ accumulation in aging and dementia remains less certain. In this study, we examined mRNA and protein expression, as well as biological activity of NEP, IDE and ECE-1 in human frontal cortex by real-time RT-PCR for mRNA, immunoblotting for protein and highly sensitive and specific fluorescence assays for activity. The relationships between Aβ degrading enzymes and pathologic measures and clinical features were also assessed. The results showed that NEP mRNA, protein level and activity were decreased in AD compared with normal controls with no cognitive impairment (NCI). In contrast IDE activity was unchanged, but there was higher expression of IDE mRNA, indicating a possible compensatory reaction due to deficits in activity. ECE-1 expression in AD brain showed no significant difference compared to age-matched controls. Correlation analyses suggested that NEP expression was correlated with Aβ accumulation and clinical diagnosis, being lower in AD than in NCI. In contrast, neither IDE nor ECE-1 correlated with Aβ or clinical diagnosis. These findings provide additional support for NEP as the major protease involved in Aβ degradation and suggest its possible therapeutic targeting in AD.

Introduction

Alzheimer disease is a progressive neurodegenerative disease of the elderly characterized by deposition of Aβ in the brain as amyloid plaques and amyloid angiopathy. Aβ is physiologically produced by proteolytic cleavage of amyloid precursor protein (APP) by the concerted action of β- and γ-secretases, and its pathologic accumulation in AD indicates an imbalance between Aβ biosynthesis and clearance (Eckman and Eckman 2005; Higuchi et al. 2005; Wang et al. 2006). Under normal conditions, as revealed in pulse-chase studies in mice, Aβ is rapidly removed from the brain after its synthesis (Cirrito et al. 2003). Deficiencies in Aβ clearance or enzyme-mediated Aβ degradation are thought to lead to accumulation and aggregation of Aβ, which in turn, triggers a complex multistep cascade leading to dementia. Therefore, strategies that lower Aβ levels in the brain may stop or delay onset of AD (Eckman and Eckman 2005; Higuchi et al. 2005; Wang et al. 2006; Miners et al. 2008c; Bates et al. 2009).

In the past decade multiple proteases have been shown to be capable of cleaving Aβ, including neprilysin (NEP), insulin-degrading enzyme (IDE) and endothelin-converting enzyme-1 (ECE-1), and a number of experimental and epidemiological studies implicate these enzymes in normal homeostasis of Aβ in the brain (Eckman and Eckman 2005; Higuchi et al. 2005; Wang et al. 2006; Miners et al. 2008c; Bates et al. 2009).

While NEP appears to be the dominant Aβ protease (Vardy et al. 2005; El-Amouri et al. 2008; Hersh and Rodgers 2008; Miners et al. 2008c), other catabolic enzymes such as ECE and IDE likely participate in regulating the steady-state levels of Aβ (Eckman et al. 2006; Qiu and Folstein 2006; Llovera et al. 2008). NEP is a type II integral membrane protein, known as zinc metallopeptidase, composed of 750 residues with an active site at its extracellular carboxyl-domain. NEP is capable of degrading monomeric and oligomeric forms of Aβ (El-Amouri et al. 2008). Neprilysin deficient mice have significantly elevated brain Aβ levels, and treatment with NEP inhibitors can cause a rapid increase of Aβ levels in the brain, associated with memory impairment (Apelt et al. 2003; Farris et al. 2007). By contrast, deposition of Aβ in NEP deleted mice can be reversed by injection of exogenous NEP (Iwata et al. 2001; Hemming et al. 2007). Over expression of NEP in brains of APP transgenic mice (Leissring et al. 2003; Marr et al. 2003; Miners et al. 2008c; Meilandt et al. 2009) or over-expression of NEP in the periphery (Guan et al. 2009) is associated with lower brain Aβ levels and amyloid plaque load. In addition to these investigations carried out in animal models, observations in human postmortem tissue suggest that NEP expression is inversely related to the extent of AD pathology (Akiyama et al. 2001; Eckman et al. 2006; Hellstrom-Lindahl et al. 2008).

IDE is also a zinc metalloendopeptidase that is predominantly cytosolic (Duckworth et al. 1994), with less in the plasma membrane (Goldfine et al. 1984); it has a molecular weight of 110 kDa. As demonstrated by in vitro studies, IDE is not only a major soluble protease involved in the degradation of extracellular Aβ in the brain, but it also has the ability to remove the cytoplasmic products of APP (Bernstein et al. 1999; Edbauer et al. 2002). APP transgenic mouse that develop AD-like neuropathology with aging have elevated IDE mRNA and protein, and IDE protein expression is positively correlated with expression of full-length APP in cerebral cortex, indicating that IDE expression responds to Aβ accumulation (Vepsalainen et al. 2008). Mice with a homozygous deletion of IDE have elevated endogenous brain Aβ (Miller et al. 2003; Leal et al. 2006). Conversely, overexpression of IDE attenuates brain plaque formation, secondary pathology and prevents premature death in a transgenic AD mouse model (Leissring et al. 2003). In AD, in situ hybridization and western blot analyses demonstrate reduced neuronal IDE mRNA and protein levels in hippocampus (Cook et al. 2003), and immunohistochemistry studies show that neuronal IDE expression in hippocampus is also significantly reduced compared with controls (Miners et al. 2008c). Reduced levels of cytosolic IDE have also been reported in AD (Perez et al. 2000). In contrast, Zhao and co-workers found reduced hippocampal IDE protein and activity in membrane fraction, but not the cytosol (Zhao et al. 2007). Studies that focused on IDE in the cortex rather than hippocampus, showed different results; Caccamo and co-workers reported that IDE protein levels diminish with of age, while cortical IDE expression is elevated in AD (Caccamo et al. 2005). Kim and co-workers reported reduced IDE catalytic activity in AD lymphoblast lines without significant changes in IDE expression and suggested that this might involve genetic factors (Kim et al. 2007). The inconsistent results for IDE in AD remains to be explained, but may be due to methodological issues. Clearly, additional studies are needed to better understand the significance of changes in IDE in aging and AD.

As another set of type II integrated membrane zinc metalloendopeptidases, ECEs are primarily localized in endothelial cells, but have also been detected in neurons and glia (Davenport et al. 1998). Two ECE isoforms, ECE-1 and ECE-2, share 59% sequence homology and have similar catalytic activity; in the brain the predominant isoform is ECE-1 (Davenport et al. 1998; Miners et al. 2008c). In both animal models and postmortem human brain tissue, the role of ECE-1 in AD is far from established. Eckman and co-workers were the first to identify ECE-1 as a novel Aβ-degrading enzyme in cell models (Eckman et al. 2001). Later, they found that the level of brain Aβ was increased in ECE-1 deficient mice (Eckman et al. 2003), providing the first direct evidence that ECE-1 might play a role in Aβ metabolism in the brain. Over-expression of ECE-1 in APP presenilin 1 (PS1) bigenic transgenic mice with gene delivery mechanisms decreased Aβ accumulation in the cortex and hippocampus (Carty et al. 2008). While there is evidence for a role of ECE-1 in Aβ metabolism in experimental models, there is limited information on its role in AD. Only one study has investigated ECE-1 levels in AD brains, which showed increased neuronal expression in cerebral cortex and hippocampus in controls compared to AD (Funalot et al. 2004). The results need to be interpreted with caution because there is 59% identity between ECE-1 and ECE-2 (Barnes and Turner 1997), the possibility of cross-reactivity of antibodies with ECE-2 was not completely excluded.

We previously reported that NEP was selectively decreased in AD brain, and that there was an inverse correlation between NEP activity and vulnerability to AD pathology (Wang et al. 2005); however, there is still little comparative data exploring Aβ degrading enzymes simultaneously in AD and controls. To address this issue, in the current study we measured expression and activity levels of three Aβ degrading enzymes (NEP, IDE and ECE-1) in brain tissue of prospectively studied controls with no cognitive impairment (NCI), as well as those with mild cognitive impairment (MCI) and clinical AD. The findings provide further evidence connecting enzyme activity, Aβ accumulation and clinical features during pathogenesis of AD.

Materials and Methods

Case and clinical features data

Frozen brain tissue from frontal cortex was obtained from participants in the Religious Orders Study of the Rush Alzheimer Disease Center (P30AG10161) as described previously (Wang et al. 2007). Included in this study were 10 NCI, 10 MCI and 10 AD. All cases had undergone a uniform structured clinical evaluation that included a medical history, neurologic examination, neuropsychological performance testing, and review of a brain scan when available. The evaluation included the mini-mental status examination (MMSE) as an overall test of cognition and 20 other cognitive performance tests, 19 of which were used to create a global measure of cognition, as described previously (Bennett et al. 2002; Wilson et al. 2002). Characteristics of the three groups are summarized in Table 1.

Table 1.

Clinical features of cases

	NCI (n=10)	MCI (n=10)	AD (n=10)
Age at death (yr)	80.5 ± 1.9	84.3 ± 1.6	87.3 ± 2.1
Sex (M:F)	7:3	4:6	3:7
Postmortem delay (hr)	6.2 ± 1.2	5.7 ± 0.97	7.3 ± 1.1
Education (yr)	17.9 ± 1.5	17.2 ± 1.2	17.2 ± 0.45
Braak NFT Stage	1.7 ± 0.38	2.8 ± 0.43	3.6 ± 0.50†
MMSE	27.3 ± 0.42	27.2 ± 0.50	19.4 ± 1.8*
Global z-score	0.49 ± 0.11	0.24 ± 0.0	−1.2 ± 0.20*

^*p<0.05 compared to NCI and to MCI;

^†p<0.05 compared to NCI

Chemicals and reagents

Oligonucleotides were synthesized by Gibco BRL Life Technologies (Gaithersburg, MD). TRIreagent, Omniscript Reverse Transcriptase was acquired from Invitrogen (Cincinnati, OH, USA) and Qiagen (Valencia, CA, USA), respectively. SYBR green PCR master mix was from AB applied biosystem. NEP, IDE and ECE-1 antibody were purchased from Chemicon, Abcam, and R&D Systems, respectively. Synaptophysin monoclonal antibody was a generous gift from Dr. William Honer, University of British Columbia, Vancouver, Canada. β-actin and α-actin antibodies were from Amershan and Sigma-Aldrich. Anti-rabbit and anti-mouse HRP-conjugated secondary antibody and anti-goat HRP-conjugated secondary antibody were from Cell Signaling and Santa Cruz. Mac-R-P-P-G-F-S-A-F-K(Dnp)-OH Fluorogenic peptide substrate came from R&D. An enhanced chemiluminescence (ECL) kit was obtained from Pierce. Other general chemicals and reagents were from Sigma-Aldrich.

RNA extraction and reverse transcription (RT)-PCR

Frontal cortical gray matter was carefully dissected from frozen tissue and analyzed blinded to clinical information. Total RNA was isolated from approximately 50 mg of tissue according to the TRI-Reagent protocol. RNA was quantified and assessed for purity by UV spectrophotometer. 1.5 μg of total RNA was subjected to reverse transcription with the Omniscript RT in a 20 μl reaction. The RT reaction was diluted five folds and 5 μl was amplified by Real-time PCR (qPCR) in 25 μl reaction mixture containing Master Sybr Green (ABI).

The primer sets were designed to span the intron-exon borders to distinguish amplified cDNA from genomic DNA by Primer Express software (Applied Biosystems, Foster City, CA). The sequences of primers used in current study were listed in Table 2. qPCR was performed on an ABI Prism 7500 real-time PCR system. The thermal cycling conditions comprised an initial denaturation step at 95°C for 2 min, then 40 cycles of two steps PCR including 95°C for 15s and 60°C for 1 min. Data were collected during the 60°C annealing step. Optimization of primers concentration was performed at Serial dilutions of the template cDNA made from PCRs within the linear range. The authenticity of the PCR products was confirmed with melting curve analysis using the Applied Biosystems software and electrophoreses in 2.0% agarose gel.

Table 2.

Primers used in qRT-PCR

Gene	Primer set
NEP	F 5′-GCC TCA GCC GAA CCT ACA AG	R 5′-AAT TTG CAC AAC GTC CTC AAG TT
ECE	F 5′-GAC GCC GAT GAG AAG TTC ATG	R 5′-GCA AAA CTT CCA GCG AGG AA
IDE	F 5′-GCC GAA GCC TTG TCT CAA CT	R 5′-CAA ATA GGC CAT GTT ACA GTG CAA
S26	S26 F 5′-CGC AGC AGT CAG GGA CAT TT	R 5′-TTC ACA TAC AGC TTG GGA AGC
β-actin	F 5′-CGC AGC AGT CAG GGA CAT TT	R 5′-TTC ACA TAC AGC TTG GGA AGC

Relative quantification of genes expression was carried out by comparative Ct method according to manufacture protocol (User Bulletin #2: ABI Prism 7500 Sequence Detection System). Briefly, the genes mRNA level was expressed in cycle threshold (Ct) value; the Ct values for each sample were averaged from duplicate. Base on the amplification efficiency, human β-actin for NEP, S26 for ECE and IDE were used as reference genes. Differences between the mean Ct values of NEP and reference genes were calculated as ΔCt^sample= Ct^NEP-Ct^β-actin for AD and MCI groups, and that of the delta Ct for the NCI group was set for calibrator (ΔCt^calibrator). Final results, the sample-calibrator ratio, expressed as N-fold differences of NEP expression in the AD or MCI groups compared with NCI group, were determined as 2 ^{−(ΔCtsample - ΔCtcalibrator)}. Similar method was used in ECE and IDE calculation of relative quantification of genes expression.

SDS-PAGE, Western blot and protein quantification

Brain tissue homogenates were prepared from dissected samples of unfixed frozen frontal cortex. Protease inhibitor cocktail (Sigma-Aldrich) and 1 mM PMSF (Sigma-Aldrich) were added to the lysis buffer (Pierce), 50 μl whole tissue lysates was added to 150 μl 1% SDS that contain protease inhibitor and 1 mM PMSF. After vortexing, the samples were incubated on ice for 30 min. From 1000 μg whole tissue lysates, 4x sample buffer was added and boiled at 95°C for 5min, cooled on ice and subsequently loaded to a 12% SDS-PAGE gel. Proteins were transferred to nitrocellulose membrane at 100V for 70 min. After blocking with 5% nonfat milk-TBST for 1hr at room temperature, membranes were incubated in 5% nonfat milk-TBST with mouse anti-Aβ (1:1000; Clone Ab9, a gift from Dr T. Golde, Mayo Clinic, Jacksonville, FL), rabbit anti-NEP antibody (1:2000; Chemicon, Temecula, CA, USA), mouse anti-IDE (1:2000; Abcam, Cambridge, MA, USA.), goat anti-ECE-1 (1:2000; R&D Systems, Minneapolis, MN), mouse anti-synaptophysin (EP10; 1:500), mouse monoclonal antibody BAN50 (recognizing the amino terminus of Aβ) (1:1000), mouse anti-β-actin (1:2000; Amersham, Piscataway, NJ, USA) or rabbit anti-α-actin antibodies (1:4000; Sigma-Aldrich, St. Louis, MO, USA), respectively, at 4°C overnight. After washing with TBS-T, membranes were incubated with HRP-conjugated anti-rabbit (1:8000; Cell Signaling, Danvers, MA, USA), or anti-mouse (1:8000; Cell Signaling,) or anti-goat (1:5000; Santa Cruz, CA, USA) secondary antibodies for 1 hour at room temperature, detected by the ECL plus Western blotting detection system and autoradiography film (Amersham, Piscataway, NJ, USA). The densities of target bands were measured with Image Quant 5.2 and expressed as relative level with respect to the normal control.

Measurement of Enzymatic Activity

NEP activity was determined by fluorescence resonance energy transfer (Johnson and Ahn 2000; Takaki et al. 2000; Wang et al. 2005). Human frontal cortex gray matter was homogenized in PBS (pH 7.4) containing protease inhibitor cocktail (Sigma, 200mg/mL), 1 mM PMSF and solubilized by sonication. Samples were then added to 0.1% TritonX-100 in PBS to final concentration 2.0 mg/mL, incubated on ice for 30 min to extract membrane proteins. 50 μl of fluorogenic peptide substrate (Mca-RPPGFSAFK-[Dnp]-OH; R&D Systems, Minneapolis, MN) [45] dissolved in HEPES buffer (pH 7.4) was added to 50 μl of homogenate from a normal case and then diluted to 2000, 1000, 500, 250, 125, 62.5, and 31.25 mg/ml to produce a standard curve. To determine the specificity of the assay, a set of samples were initially preincubated with 100 nM thiorphan (Sigma-Aldrich), a NEP specific inhibitor (Wang et al. 2003; Miners et al. 2008b), for 10 minutes. Specific NEP activity is optimal at pH 7.4 (Wang et al. 2005). Sample homogenates (1000 μg/mL) were pre-incubated with or without inhibitor prior to adding fluorogenic peptide substrate. Fluorescence was read after 30 minutes incubation at room temperature with excitation at 320nm and emission at 405nm on a fluorescent 96-well plate reader (Spectra Max Gemini; Molecular Devices, Sunnyvale, CA).

ECE activity was also analyzed by fluorogenic peptide substrate Mca-RPPGFSAFK(Dnp)-OH at RT in the presence or absence of the ECE inhibitor, phosphoramidon (1μM, Sigma) (Wang et al. 2003; Miners et al. 2008b). Phosphoramidon inhibition of ECE-1 is highly pH dependent (Ahn et al. 1998). The IC50 value of phosphoramidon is about 50-fold lower at pH 5.8 than at pH 7.2. In addition, NEP activity is also pH-dependent. There is almost no NEP activity at pH5.8 (Fahnoe et al. 2000). Thus, the inhibition by phosphoramidon at pH 5.8 can be considered the inhibition of ECE.

IDE specific activity was measured by using immunocapture based fluorometric assay (Miners et al. 2008a; Wang et al. 2009a). Frontal cortex brain tissues homogenates were mixed with 9 times volumes of sample buffer (0.5% Triton-X-100, 20 mM Tris pH 7.4, 10% sucrose (w/v) containing protease inhibitor cocktail (Sigma-Aldrich) for 30 min, ultrasonic homogenized and then centrifuged at 1300 rpm for 15 min at 4°C. Greiner 96-well high binding ELISA plate were coated with capture IDE antibody (Santa Cruz, sc27266, 50 mg/ml) diluted in carbonate buffer (pH 8.6) and left for 18 h at 4°C. The plates were washed with PBS containing 0.5% tween-20 for 3–5 times. After 30 min blocking with PBS-T containing 2% bovine serum albumin, 50 μl of brain homogenate was added and incubated at 4°C overnight. After 3–5 washes, 10 μM fluorogenic peptide diluted in 50 mM HEPES (pH 7.5) was added and incubated at 30°C in dark. Fluorescent readings were taken after 60 min. Control wells included in each plate contained PBS and fluorogenic peptide.

All the assays were triplicates, and the mean values were calculated for each sample.

Statistical analyses

All data were expressed as mean ± standard error of the mean. Statistical analyses were performed using the GraphPad Prism 4.0 software (GraphPad Software, Inc., San Diego, CA). One-way ANOVA analysis was used to evaluate differences in mean values among three groups, followed by post hoc tests of multiple comparisons when ANOVA were significant. Pearson correlation analysis was used to study relations between variables. Statistical significance level was set a p<0.05.

Results

Real-time RT-PCR

In an effort to understand the gene transcription of Aβ degrading enzymes participate in the regulation of Aβ accumulation in AD development, we examined NEP, IDE and ECE-1 level in frontal cortex from NCI, MCI and AD patients using qRT-PCR. NEP mRNA levels were decreased in MCI and AD groups with significance in AD compared with NCI (Figure 1A). In contrast to NEP, both ECE-1 and IDE mRNA increased in MCI and AD brains, and for IDE this reached significance in AD compared with NCI.

Figure 1. NEP, IDE and ECE levels in frontal cortex of NCI, MCI and AD patients.

(A): Alterations of NEP, IDE and ECE mRNA expression in frontal cortex of NCI, MCI and AD patients. Relative NEP, IDE and ECE mRNA levels in three groups were detected by qRT-PCR using the comparative cycle threshold method, assigning a value of 1 to the NCI group. NEP mRNA levels were normalized to β-actin in each sample and IDE and ECE levels were normalized to S26 in each sample. Bars show mean values and standard error of mean (n=10 in each group). Decrease NEP mRNA level was observed in AD group compared with NCI group. Increase IDE and ECE mRNA level in AD group compared with NCI group.

(B): NEP protein levels are significantly decreased in AD patient, while IDE and ECE protein expression stay steady in three groups. Western blot analysis on brain frontal cortex were performed, the actin bands confirm an equal loading of all samples. Data shown are representative of all brain biopsies studied. The densitometry of western blots showed the down regulation of NEP protein in AD group compared with NCI (p<0.05). NEP protein level in MCI group tend to decrease, but not statistically significant. IDE and ECE protein level in three groups did not changed significantly.

(C): NEP activity in AD frontal cortex is significantly decreased while IDE and ECE activity remains unchanged. Specific NEP and ECE activity in brain homogenate from all cases were measured based on high sensitive fluorogenic analysis, specific IDE activity is measured by using immunocapture fluorometric assay. Assays were repeated on 3 separate occasions. Bar show mean values and standard error of mean of 9 cases in AD group and 10 cases in NCI, MCI group, respectively. *:compare with NCI group, p<0.05

Western Blots

Many studies suggest that post-transcriptional regulation of Aβ-degrading enzymes may play a role in Aβ degradation and clearance (Wang et al. 2006; Miners et al. 2008c). To study the protein expression of these enzymes, immunoblotting was performed in postmortem brain samples. As shown in Figure 1B, NEP had a double band at 80- and 100-kDa molecular mass. Densitometric analyses of blots revealed a significant decrease in NEP in AD compared with NCI (p<0.05), in agreement with previous reports (Higuchi et al. 2005). Despite increases in mRNA observed for IDE and ECE-1, their protein levels did not show significant change.

Activity Assays

NEP enzymatic activity was measured in frontal cortex with a fluorometric assay. As shown in Figure 1C, NEP catalytic activity was significantly decreased in AD compared with NCI and MCI. On the other hand, IDE and ECE activity tended to increase in AD compared to MCI and NCI, but this did not reach statistical significance. It is reasonable to speculate that deficits in NEP expression and activity in AD may lead to compensatory increases in IDE and ECE-1 expression and possibly to activity.

Synaptophysin, a pre-synaptic marker, is not altered with AD development

NEP and ECE-1 is integral transmembrane protease that is primarily expressed in presynaptic terminals of neurons (Fukami et al. 2002), while IDE is expressed in neurons and other cell types. Neuron loss should potentially affect expression of these enzymes. In order to clarify that decreased NEP level in AD is not merely due to neuronal or synaptic loss, western blots of a presynaptic marker, synaptophysin, were performed in the same frontal cortex homogenates used for western blots of Aβ degrading enzymes. As shown in Figure 2, synaptophysin level was less in MCI and AD, but the difference did not reach significance compared to NCI, probably due to the location of the tissue (frontal cortex, which is has neuronal loss later than medial temporal lobe in AD pathogenesis), the fact that AD and MCI had relatively mild cognitive deficits, and the relatively small sample size. After adjusted by synaptophysin, NEP protein in AD cases remained significantly lower in AD compares to NCI, indicating that NEP reduction is not merely due to neuronal and synaptic loss.

Figure 2.

Neuronal loss makes limited contribution to NEP protein down-regulation. Synaptophysin, pre-synaptic marker in neuron, was measured by western blot (fig 2A). AD and MCI group has very slightly decline synaptophysin expression compare to NCI, the difference did not reach statistic significance. NEP protein levels were significantly decreased in AD cases by normalizing with synaptophysin (fig 2B). *: compare to NCI, p<0.05.

NEP protein level, but not ECE-1 or IDE, is closely correlated with Aβ accumulation

To understand the relationship between Aβ degrading enzymes and Aβ accumulation, we conducted western blot to measure the content of Aβ by full length Aβ antibody in human postmortem brain tissue. As expected we observed a marked increase in Aβ in AD compared with NCI (Figure 3). Further exploration of the functional association between the Aβ degrading enzyme and Aβ accumulation indicated that NEP immunoreactivity was inversely correlated with Aβ in all cases (Figure 4. R²= −0.37, p=0.0004). Similar analyses performed in subsets showed negative correlation in NCI (R²= −0.32, p=0.08) and MCI (R²= −0.49, p=0.02) group. However, there was no association in the AD group (R²= 0.26, p=0.13) (Figure 4). No significant correlations between IDE or ECE immunoreactivity and Aβ content were observed (data not shown).

Figure 3.

Western blot analysis conform Aβ deposition in AD patients. Total Aβ protein content in frontal cortex homogenate was detected by full length Aβ antibody in three groups. As a reference protein, β-actin indicates equal loading of all samples. Results show Aβ increases dramatically in AD group compared with NCI. **: p<0.01 as compared with NCI group.

Figure 4.

Correlation of NEP protein level and Aβ accumulation. The negative correlation between NEP protein and Aβ accumulation was observed (r2=0.37, p=0.0004) when combined all the tree groups together suggesting that NEP is a key factor in Aβ clearance during AD development. However, when correlation analysis were performed in sub-group, the positive correlation between NEP protein level and Aβ content was found in AD group (r2=0.26, p=0.13). These results implied that the capacity of NEP clearance in AD development is not unlimited.

NEP, not ECE-1, or IDE, activity is directly related with clinical manifestations

Two clinical neuropsychological parameters, MMSE and global neuropsychological z-scores were examined with respect to Aβ degrading enzymes (Figure 5). The MMSE is widely used to assess the severity of cognitive impairment, and there was a significant correlation between MMSE and NEP activity (p=0.016), and a even more significant correlation between NEP activity and z-score (p=0.006). In contrast, there was no correlation between IDE and ECE-1 in either mental state scores or the neuropsychologic summary measure of cognitive function.

Figure 5.

Correlation of NEP enzymatic activity and AD clinical parameters. The positive correlation between NEP activity and MMSE, NEP activity and Global z-score were observed when all cases were pooled together, indicating that NEP activity may be an essential factor in AD clinical physiological features development.

Discussion

Although the physiological functions of Aβ remain to be elucidated, imbalance in its production and clearance contributes to its accumulation in brain, which is one of the most characteristic features of AD. Considerable evidence has emerged indicating that proteolytic degradation of Aβ is a major process involved in Aβ clearance. To date, most attention has been focused on NEP and its homologues IDE and ECE-1. A range of different studies has produced conflicting results for NEP and IDE in AD (El-Amouri et al. 2008; Miners et al. 2008c; Bates et al. 2009). Except for experimental models, little is known about ECE-1 especially in human tissue (Weeraratna et al. 2007; Miners et al. 2008c). In the current study, we measured mRNA, protein and enzyme activity for these Aβ degrading enzymes in frontal cortex of a series of well characterized patients and correlated these findings to Aβ levels and clinical and clinical features.

While some previous studies have shown increases in NEP and IDE levels in AD that correlated with disease severity (Dorfman et al. 2008; Miners et al. 2009), in the present study we found mRNA, protein and activity of NEP were decreased in AD, which is in agreement with our previous studies (Wang et al. 2003; Wang et al. 2005) as well findings of independent investigators using comparable methods (Akiyama et al. 2001; Miners et al. 2006). Previous experiments in cultured neuronal cells indicated that NEP mRNA, protein and activity are upregulated by treatment with Aβ (Wang et al. 2009b). In the context of the present findings, it suggests that any increase in NEP in the human brain in response to Aβ may be relatively early in the disease process. With the advent of amyloid imaging, it is becoming clear that Aβ deposition in the frontal cortex is an early pathologic process (Jack et al. 2009). Compensatory increases in NEP would be expected in clinically normal individuals with cortical amyloid deposits, a process referred to as pathological aging, and this is what was found in our previous study (Wang et al. 2005).

NEP is known as an integral membrane protein and to be expressed predominantly in axonal and synaptic membranes (Fukami et al. 2002). Given this localization, it is important that any decreases in NEP mRNA, protein or catalytic activity, be assessed in light of neuron or synaptic loss in order to exclude the possibility that decreases are merely secondary to synaptic loss. In previous studies, neuron specific enolase has been used (Miners et al. 2009), while we used the presynaptic marker synaptophysin in this and previous studies (Wang et al. 2005). We found a trend for decreasing synaptophysin immunoreactivity in AD and MCI compared to NCI, but this did not reach statistical significance. Synaptophysin loss in AD shows regional anatomical differences (Masliah et al. 1994), with frontal cortex involved later than medial temporal lobe, which may explain the subtle changes we observed. NEP protein level, after adjusting for synaptophysin level, still showed significant decreases in AD compared to NCI. The results suggest that neuron or synaptic loss cannot explain the decreases we observed in NEP in AD.

Experimental studies have shown that ECE-1 can degrade Aβ (Eckman et al. 2001; Eckman et al. 2003; Eckman et al. 2006), but the relative contribution of ECE-1 to Aβ degradation in the human brain is unknown. The current study is the first systematic investigation of ECE-1 levels and catalytic activity in AD. A previous study reported on ECE-1 protein expression in cortex and hippocampus, but did not measure enzyme activity (Funalot et al. 2004). We found a trend for increases in ECE-1 mRNA, protein and activity in AD compared with NCI, although none of these reached statistical significance. Our findings support the hypothesis that increased ECE-1 in AD is compensatory to increases in Aβ and fit with our in vitro findings that SH-SY5Y treated with Aβ had significantly increased ECE-1 expression and activity (Wang et al. 2009a).

Previous studies of IDE regulation in AD brain have yielded inconsistent results; some studies have shown reduction of IDE in AD (Perez et al. 2000; Cook et al. 2003; Caccamo et al. 2005; Zhao et al. 2007), while others have shown significant increases (Caccamo et al. 2005; Dorfman et al. 2008). The studies were not directly comparable, given that different cellular components (cytosolic versus membrane) and different brain regions (hippocampus, neocortex, or cerebellum) were studied. In the current studies we found higher IDE mRNA in AD compared to NCI, with IDE protein and catalytic activity also tending to be higher in AD frontal cortex, but not reaching statistical significance, consistent with findings from the original report (Bernstein et al. 1999). Our in vitro data also support our observation in AD cases (Wang et al. 2009a) and provide support for the hypothesis that IDE is up-regulated in response to increasing levels of brain Aβ, similar to ECE-1.

From the above, we speculate that IDE and ECE-1 might be second line defense to increasing Aβ related to decreases in NEP. On the other hand, decreased catalytic capacity of NEP is not adequately compensated by the ECE-1 and IDE, as Aβ progressively accumulates in the brain. Moreover, these enzymes most likely act independently with respect to Aβ removal from the brain, in agreement with experimental data from mice (Qiu and Folstein 2006).

In the brain, NEP, IDE and ECE show a distinct pattern of expression. NEP is mainly expressed in presynaptic terminals of neurons (Fukami et al. 2002), IDE is predominately a cytosolic protein (Yfanti et al. 2008) and ECE exhibits a dual localization in the cell (Barnes and Turner 1997). The region and subcellular expression of these enzymes may explain their functions during AD development.

There has been strong evidence that Aβ accumulation is central to the pathogenesis of AD. There is, however, little evidence for increased Aβ production in the majority of late onset AD. Deficiencies in Aβ clearance or enzyme-mediated Aβ degradation are increasingly thought to be responsible for Aβ accumulation, particularly in late onset AD (Wang et al. 2006; Miners et al. 2008c). Post-mortem AD brain showed NEP mRNA and protein reduced in high plaque areas (Yasojima et al. 2001b; Yasojima et al. 2001a; Wang et al. 2003; Caccamo et al. 2005; Wang et al. 2005). Reductions were most prominent in regions most vulnerable to AD pathology, such as hippocampus, but not in other brain areas such as cerebellum or in peripheral organs (Yasojima et al. 2001b; Caccamo et al. 2005). Similar to findings in AD, NEP expression is reduced in the hippocampus and cortex of aged mice compared to young mice (Iwata et al. 2002; Apelt et al. 2003). Conversely, over-expression of NEP reduces Aβ levels in a dose-dependent manner (Iwata et al. 2002; Marr et al. 2004), and protects neuronal cells from Aβ toxicity in vitro (El-Amouri et al. 2007). Consistent with the previous studies by our lab (Wang et al. 2005) and above mentioned reports by others, present data showed that a decrease in the overall level of NEP mRNA, protein and activity in AD brains, was observed. Importantly, NEP protein expression was inversely correlated with Aβ levels. A strong correlation was also found between NEP enzymatic activity and clinical cognitive scores, MMSE and Global82 (Figure 5), regardless of whether or not there were cortical amyloid deposits, which is in agreement with previous findings (Funato et al. 1998; Morishima-Kawashima et al. 2000; Fukumoto et al. 2004; Wang et al. 2005). These data support the hypothesis that decreased NEP contributes to Aβ deposition in AD.

Interestingly, negative correlation between NEP protein and Aβ level was found in MCI (R²=0.49, p=0.02) and NCI (R²=0.32, p=0.08) in this study, which is reasonable and in agreement with previous findings. In contrast to NCI and MCI, however, we observed no significant correlation between NEP levels and Aβ in AD brains. The levels of NEP were uniformly low and the Aβ levels were high providing a narrow dynamic range for correlations. In addition, pathological induction and modification of NEP may occur in AD. In AD transgenic mice (TgCRND8), over-expressing human APP, relative NEP levels are greater in the transgenic than in control mice, and NEP immunoreactivity is localized to a subpopulation of plaques. Aβ deposition correlated with total NEP immunoreactivity. This suggests that NEP may have been pathologically induced because of Aβ deposition (Sato et al. 1991). Our previous study indicated, acutely administrated Aβ decreased NEP catalytic activity, which consequently up-regulated NEP mRNA and protein levels. Oxidative modification of NEP plays a crucial role in enzyme inactivation [57]. It indicates that NEP may also be up-regulated under the induction of increased Aβ during AD development. Due to oxidative modification of NEP, increased NEP is insufficient to clear the excessive Aβ that eventually results in the brain Aβ deposition and plaque formation.

Unlike Zhao and coworkers who found correlation between IDE immunoreactivity and the levels of Aβ in hippocampus and occipital cortex (Zhao et al. 2007), we found no significant correlations between IDE or ECE-1 and Aβ. These results suggested NEP is a key Aβ degrading enzyme in the aging and AD. The decrease in NEP expression in AD may be induced by accumulation of Aβ with pro-oxidant state favoring oxidative modification and inactivity of NEP. Thus, low NEP may promote Aβ formation and be part of a negative feedforward process whereby increased Aβ leads to greater deficiencies in NEP activity (Farris et al. 2007).

In summary, our finding suggests that NEP is a key protease in Aβ degradation, though ECE and IDE may act as a second line defense. The strategies aimed at promoting NEP expression or enzymatic activities in the brain may help prevent progression of AD through mechanisms involving the clearance of the Aβ peptides from the brain.