Alzheimer's disease (AD) amyloid plaques are composed of amyloid-β (Aβ) peptides produced from proteolytic cleavage of amyloid precursor protein (APP). Isoprostanes, markers of in vivo oxidative stress, are elevated in AD patients and in the Tg2576 mouse model of AD-like Aβ brain pathology. To determine whether isoprostanes increase Aβ production, we delivered isoprostane iPF2α-III into the brains of Tg2576 mice. Although treated mice showed increased brain Aβ levels and plaque-like deposits, this was blocked by a thromboxane (TP) receptor antagonist, suggesting that TP receptor activation mediates the effects of iPF2α-III on Aβ. This hypothesis was supported by cell culture studies that showed that TP receptor activation increased Aβ and secreted APP ectodomains. This increase was a result of increased APP mRNA stability leading to elevated APP mRNA and protein levels. The increased APP provides more substrate for α and β secretase proteolytic cleavages, thereby increasing Aβ generation and amyloid plaque deposition. To test the effectiveness of targeting the TP receptor for AD therapy, Tg2576 mice underwent long-term treatment with S18886, an orally available TP receptor antagonist. S18886 treatment reduced amyloid plaques, insoluble Aβ, and APP levels, thereby implicating TP receptor signaling as a novel target for AD therapy.
Alzheimer's disease (AD) is pathologically marked by senile plaques, neurofibrillary tangles, inflammation, and oxidative stress, ultimately culminating in neuronal loss. Senile plaques are composed mainly of amyloid-β (Aβ) protein produced from the sequential proteolytic cleavage of amyloid precursor protein (APP). Cleavage of APP by α-secretase generates secreted APPα (sAPPα) ectodomain and C83 C-terminal stub, whereas β-secretase cleavage of APP generates secreted APPβ (sAPPβ) N-terminal and C99 C-terminal fragments. The C99 C-terminal fragment is then cleaved by γ-secretase, thereby generating Aβ peptides of varying lengths (Ling et al., 2003; Mattson, 2004; Vetrivel and Thinakaran, 2005; Hardy, 2006). Aβ 1–40 is the most commonly generated species, whereas Aβ 1–42 is less abundant but considered to be the more amyloidgenic and pathologenic species of Aβ. Alterations in APP processing, caused by, for example, the presence of APP gene mutations, can elevate Aβ levels thereby increasing amyloid deposition and the onset of AD (Selkoe, 1999; Wilson et al., 1999).
Oxidative stress is an early event in AD pathogenesis and often manifests itself as lipid peroxidation in the brain (Markesbery, 1997; Giasson et al., 2002). Arachidonic acid, a major component of cell membrane phospholipids, is particularly vulnerable to free radical attack and subsequent peroxidation. This reaction generates various isoprostane isoforms that are considered to be specific and sensitive markers of in vivo lipid peroxidation (Morrow et al., 1990; Greco et al., 2000). They are elevated in conditions associated with oxidative stress, including smoking, diabetes, heart disease, and AD. The isoprostane isoforms, iPF2α-III and iPF2α-VI, are elevated in AD patients (Montine et al., 1998; Pratico et al., 1998; Tuppo et al., 2001) as well as in patients with mild cognitive impairment, a prodromal phase of AD (Pratico et al., 2002). In addition, the Tg2576 mouse model of AD-like Aβ pathology (Hsiao et al., 1996), in which APP695 with the Swedish familial mutation (APPsw) is overexpressed, also shows elevated iPF2α-VI levels before the onset of amyloid plaque formation (Pratico et al., 2001). Because isoprostanes are elevated early in the disease process, they may play a mechanistic role in the onset or progression of AD.
Although isoprostanes are commonly used as markers of oxidative stress, they also exert biological activities. The isoprostane iPF2α-III (IsoP) can induce vasoconstriction and platelet aggregation (Morrow et al., 1992; Audoly et al., 2000), as well as promote atherogenesis (Tang et al., 2005), through activation of the thromboxane receptor, historically named the thromboxane-prostanoid (TP) receptor. The TP receptor is a seven transmembrane G-protein-coupled receptor that is present in many tissues, including cardiovascular and brain tissues (Borg et al., 1994; Gao et al., 1997; Kinsella et al., 1997; Huang et al., 2004).
Given that IsoP is elevated in human AD brains and has been shown to exert biological activity in other systems, we asked whether IsoP plays an active role in AD pathogenesis in the Tg2576 mouse model. Our in vivo and in vitro studies reported here suggest that IsoP promotes amyloid plaque formation by increasing Aβ levels through TP receptor activation, and that TP receptor inhibition may prove beneficial in the treatment of AD.
Materials and Methods
IsoP, 8-iso PGF3α (F3α), SQ29,548 (SQ), and [1S-[1α,2α(Z),3β(1E,3S*),4α]]-7-[3-[3-hydroxy-4-(4-iodophenoxy)-1-butenyl]-7 -oxabicyclo[2.2.1]hept -2 -yl] -5 -heptenoic acid (I-BOP) were purchased from Cayman Chemical (Ann Arbor, MI). S18886 was obtained from Institut de Recherches Servier (Suresnes, France). Cyclohexamide was purchased from Sigma-Aldrich (St. Louis, MO), and osmotic pumps were purchased from Alzet (Cupertino, CA). Tg2576 mice (Hsiao et al., 1996) were housed in accordance with University Institutional Animal Care and Use Committee (IACUC) guidelines, and all procedures were approved by the IACUC. Full-length pCMV-Notch1 construct was obtained from Dr. Robert Doms (University of Pennsylvania, Philadelphia, PA). TPα-pcDNA3.1 construct was obtained from Dr. Garret FitzGerald (University of Pennsylvania, Philadelphia, PA). Other constructs in pcDNA3.1 used here include wild-type APP cDNA (APPwt), APPsw, APPΔKK (endoplasmic reticulum targeted APP), and APP-TGN (trans-Golgi network targeted APP), and all have been described previously (Cook et al., 1997; Skovronsky et al., 2000; Wilson et al., 2002). To generate APPwt-pZeoSV2 and APPwt-pcDNA5/TO, APP-pcDNA3.1, pZeoSV2(+), and pcDNA5/TO were digested with HindIII and NotI, and APP insert was ligated into either pZeoSV2(+) or pcDNA5/TO.
Osmotic pump implantation and treatment.
Mice were treated with artificial CSF (aCSF) (1.3 CaCl2, 1.2 MgSO4, 3 KCl, 0.4 KH2PO4, 25 NaHCO3, and 122 NaCl, pH 7.35), IsoP at 1 μg/kg body weight, a biologically inactive isoprostane F3α at 1 μg/kg body weight, or with a TP receptor antagonist SQ at 2 mg/kg (Tang et al., 2005). One cohort was treated with IsoP and SQ together. Compounds were diluted from ethanol stocks into aCSF for infusion. Alzet pumps (100 μl) were filled with indicated compounds and kept in 0.9% saline at 37°C. Mice were anesthetized by intraperitoneal injection of ketamine hydrochloride (1 mg/10 g) and xylazine (0.1 mg/10 g). After being shaved on the head and back, a cannula was inserted into the right lateral ventricle using sterotaxic coordinates [frontal pole (F-P), 0.5 mm; mediolateral (M-L), 1.0 mm; dorsoventral (D-V), 2.7–3 mm]. Instant dental cement was used to secure the cannula to the skull, and an osmotic pump was placed subcutaneously on the back. After surgery, mice were housed individually. After 3 weeks of infusion, the pump was replaced with a fresh 100 μl pump filled with the indicated compound, and mice were infused for another 3 weeks before being killed.
Mice were anesthetized and prepared as described above. A volume of 5 μl of either I-BOP (5 μm in 2.5% EtOH in saline) or control (2.5% EtOH in saline) were drawn into a Hamilton microinjection syringe and injected into the hippocampus, using stereotaxic coordinates (F-P, 3.1 mm; M-L, 2.4 mm; D-V, 3.7–4 mm; 12° angle to vertical level), and the syringe was removed 10 min postinjection. Mice were killed 20 h after injection.
Seven-month-old female Tg2576 mice were treated with 5 mg/kg/d of S18886 dissolved in drinking water. Calculations were based on the assumption that a mouse drinks an average of 5 ml of water per day. Water was changed weekly for a total of 6 months. After treatment, mice were anesthetized and perfused with PBS as described below. One hemisphere of each brain was fixed in 4% neutral-buffered formalin (NBF) for 24 h, whereas the other hemisphere was dissected into cortex and hippocampus regions and frozen on dry ice for biochemical analysis.
Mice were perfused intracardially with PBS plus heparin after being lethally anesthetized by an intraperitoneal injection of ketamine hydrochloride (1 mg/10 g) and xylazine (0.1 mg/10 g), in accordance with protocols approved by the University of Pennsylvania. The left hemisphere was fixed in 4% NBF for 24 h, whereas cortex and hippocampus from the right hemisphere were dissected and frozen on dry ice for biochemical analysis. NBF-fixed brain hemispheres were dehydrated and embedded in paraffin blocks as described previously (Trojanowski et al., 1989). Tissue sections (6 μm) were cut and mounted onto 3-aminopropyltriethoxysilane-coated slides. Nonserial sections were stained for amyloid plaques using 4G8, an anti-Aβ antibody, with standard avidin-biotin-peroxidase methods with diaminobenzidine as chromogen. For quantification of amyloid plaque burden, images of stained sections from the sensory cortex and hippocampus were analyzed using Image-Pro Plus software to determine percentage of area occupied by amyloid plaques. Analysis was performed in a coded manner.
Brain extractions and sandwich ELISA.
Dissected cortex and hippocampus were homogenized separately and sonicated in either 4 ml of high-salt buffer (50 mm Tris base, pH 7.6, 10 mm EGTA, 5 mm MgSO4 0.75 m NaCl, 0.02 m NaF) for every 150 mg of tissue or 1 ml of radioimmunoprecipitation assay (RIPA) buffer (0.5% sodium deoxycholate, 0.1% SDS, 1% NP-40, 5 mm EDTA, pH 8.0, 1 μg/ml protease inhibitor mixture and 0.5 mm PMSF) for every 150 mg of tissue. Samples were centrifuged at 100,000 × g for 30 min at 4°C, and supernatant was collected. Pellets from high-salt extraction were sonicated in RIPA buffer as described above. Pellets following RIPA extraction were sonicated in 70% formic acid (FA) at 2 μl/mg tissue, and centrifuged at 100,000 × g for 30 min at room temperature (RT). FA samples were diluted 20-fold in 1 m Tris base buffer. Sandwich ELISA was performed as described previously (Gravina et al., 1995). Briefly, high-salt or RIPA soluble fractions were diluted onto plates coated with Ban50 as the capture antibody. After incubation overnight at 4°C, plates were washed and incubated with HRP-conjugated BA27 (Aβ40 specific) or BC05 (Aβ42 specific) antibodies for 4 h at room temperature followed by chemiluminescence detection. Values were calculated by comparison with standard curves of Aβ40 and Aβ42 synthetic peptides (Bachem, Torrance, CA).
Measurements were performed as described previously (Pratico et al., 1998). Briefly, high-salt extracted brain fractions and fractions that were spiked with a standard amount of iPF2α-III underwent standard lipid extraction. The extracts were then purified by thin-layer chromatography and finally analyzed by gas chromatography/mass spectrometry. For iPF2α-III biodistribution and clearance experiments, wild-type mice were injected in the lateral ventricle with 9 μCi of H3-8-iso-prostaglandin F2α (Cayman Chemical) evaporated under nitrogen gas and resuspended in PBS. Mice were killed at 2 or 24 h after injection, and indicated brain regions were dissected and monitored for radioactivity.
Cell culture experiments.
QBI293 cells [a subclone derived from human embryonic kidney 293 (HEK293) cells], HEK293-APP stably transfected cells, and COS-1 cells were grown in DMEM 4.5 g/L glucose (Invitrogen, Carlsbad, CA), 10% fetal bovine serum, l-glutamine, and penicillin/streptomycin at 37°C and 5% CO2 according to standard protocols. Cells were transfected at 90% confluence with lipofectamine 2000 (Invitrogen) following the manufacture protocol. Where indicated, cells were treated with 50 nm I-BOP, 10 μm IsoP, 1 μm SQ, 1 μm S18886, 50 μg/ml cyclohexamide, 50 μm chloroquine, or 50 mm NH4Cl. Primary hippocampal neuronal cultures were prepared from mouse embryonic day 17 (E17) C57Bl/6 mice, as described previously (Banker, 1998). Briefly, dissociated cells were plated at a density of 300,000–500,000 cells/well in poly-lysine-coated 12-well dishes and were maintained in Neural basal media supplemented with B27 (Invitrogen) for 5–7 d before treatment with IsoP or I-BOP as described above. T-Rex-293 cells stably expressing the tet repressor were purchased from Invitrogen. For RNA stability experiments, cells were transfected with APPwt-pcDNA5/TO and TPα and treated overnight with 1 μg/ml tetracycline (Sigma-Aldrich) to induce expression. After induction, cells were washed four times with tetracycline-free media over 1 h and then stimulated with control or I-BOP as described above. APP mRNA was harvested at the indicated time points.
Western blot analysis.
Cells were washed with PBS and lysed in RIPA buffer followed by brief sonication and centrifugation. Protein amount was determined by BCA assay. Equal amounts of protein were loaded and run on a 7.5% SDS-PAGE tris-glycine gel, transferred to nitrocellulose membrane, and probed with indicated antibodies. Immunoreactive bands were identified either by enhanced chemiluminescence or by [125I] labeled goat anti-mouse secondary antibodies when careful quantification was required. For detection of full-length APP, we used a rabbit polyclonal antibody raised to the C terminus of APP (5685) (Lee et al., 2005). For detection of Notch 1, we used a goat polyclonal Notch1 antibody (Santa Cruz Biochemicals, Santa Cruz, CA). For tau protein, we used a rabbit polyclonal antibody raised to recombinant tau (17025) (Ishihara et al., 1999). To detect HA-tagged Akt, we used anti-HA antibody, 12CA5 (Roche Applied Science, Indianapolis, IN). To verify consistent loading, the bottom half of the blot was probed with an anti-α-tubulin antibody (Sigma-Aldrich). For detection of secreted sAPPα and sAPPβ, conditioned media were collected and supplemented with PMSF and protease inhibitors and was either immunoprecipitated with a goat anti-APP antibody (Karen) (Turner et al., 1996) and protein A/G beads overnight at 4°C or run directly on 7.5% SDS-PAGE. For secreted sAPPβsw measurements, immunoprecipitates were run on 7.5% SDS-PAGE gel and immunoblotted with an antibody specific to the C terminus of sAPPβsw (#54) (Lee et al., 2005), whereas for sAPPα measurements, media run directly on the gel was probed with Ban50 antibody.
RNA extraction and Northern blot analysis.
RNA was extracted from cells using Trizol (Invitrogen) per the manufacture instructions. Between 4 and 6 μg of RNA were run on a 1% agarose denaturing formaldehyde gel and transferred to nylon membrane by capillary action. Membrane was UV-crosslinked and probed overnight at 42°C with an APP cDNA probe. APP cDNA probe was generated by gel extraction of EcoRI digested APP-PCDNA3.1 and labeled by random priming using [32P] α dCTP. After probe hybridization, blots were washed two times for 5 min with 2× SSC, 0.1% SDS at RT, 0.2× SSC, 0.1% SDS one time for 5 min at RT, and then one time for 15 min at 42°C. Blots were rinsed with 2× SSC and exposed to phosphoimager plate overnight. After hybridization with APP cDNA probe, blots were stripped with 1% SDS at 80°C two times for 15 min and probed with an actin cDNA probe (Clontech, Mountain View, CA) using the same protocol described above.
Data are expressed as mean ± SEM or ± SD as indicated in the figures. Results were statistically analyzed using GraphPad Prism 4.0 software (GraphPad Software, San Diego, CA) using ANOVA followed by Bonferroni's posttest or Student's unpaired t test where appropriate. In addition, linear regression was performed where appropriate. Statistical significance was set to a minimum of p < 0.05.
IsoP increases amyloid plaque formation and Aβ levels in a TP receptor-dependent manner
To determine whether IsoP might play a causative role in promoting AD pathogenesis, 10-month-old Tg2576 mice were continuously infused with IsoP or aCSF for 6 weeks. After treatment, mice were killed, and one hemisphere of the brain was fixed for immunohistochemistry analysis, whereas the other hemisphere was used for biochemical experiments. Mice treated with IsoP exhibited significantly increased amyloid plaques in the sensory cortex compared with CSF-treated mice (Fig. 1). This area of the cortex was analyzed because of its relative proximity to the IsoP injection site. There was a trend toward an increased plaque burden with IsoP treatment in the hippocampus (p = 0.09). Because IsoP is known to exert biological effects through activation of the TP receptor, we asked whether TP receptor inhibition could block the increase in amyloid plaques seen with IsoP treatment. Mice were treated with TP receptor antagonist SQ alone, IsoP+SQ together, or with F3α, a biologically inactive isoprostane that does not activate the TP receptor. In the IsoP+SQ treatment group, SQ significantly blocked the increase in amyloid plaques induced by IsoP in the cortex (Fig. 1). F3α did not produce any significant changes in amyloid plaques compared with control CSF (Fig. 1), further supporting TP receptor activation as the mechanism for increased amyloid plaque formation.
For biochemical analysis, we focused on the hippocampus region. Unlike the cortex, which when extracted for biochemical analysis contains areas both proximal and distal to the site of infusion, the hippocampus is likely to get a higher and more constant exposure to the compound. To validate this hypothesis, we injected H3-iPF2α-III into the lateral ventricle and monitored radioactivity in different brain regions 2 h after injection. We found that the levels of the compound were highest in the hippocampus (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). Consistent with these results, we saw significant increases in both soluble and insoluble Aβ 1–40 and Aβ 1–42 in this brain region after IsoP treatment (Fig. 2 A). These increases in Aβ were significantly blocked by SQ, and nonbiologically active F3α did not produce any changes in Aβ levels compared with aCSF (Fig. 2 A).
Mice treated with IsoP also showed elevated endogenous IsoP in hippocampal homogenates 1 week after the completion of treatment (Fig. 2 B), reflecting the effects of elevated amyloid in these treated mice. This increase is not attributable to residual levels of exogenous IsoP from the treatment, because pharmacokinetic studies using H3-iPF2α-III suggest complete elimination of this compound by 24 h (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). Moreover, there was no elevation of iPF2α-III levels in the IsoP+SQ treated group, further ruling out the possibility that the IsoP measured is from the treatment itself. These data support a feedforward model where an early event before the onset of plaque deposition causes oxidative stress and increases IsoP production, which in turn promotes amyloid pathology that leads to further increases in oxidative stress and IsoP generation.
TP receptor activation increases steady-state APP and APP cleavage products
To determine a mechanism of how TP receptor activation could increase Aβ production and promote amyloid plaque formation, we cotransfected TP receptor and APPsw cDNAs into QBI293 cells and stimulated the cells with IsoP or I-BOP, a potent TP receptor agonist. IsoP and I-BOP treatment significantly increased the levels of steady-state APPsw, sAPPβsw, Aβ1–40, and Aβ1–42 by twofold to threefold (Fig. 3 A–C,E). Levels of sAPPα were increased >20-fold (Fig. 3 A,D). SQ significantly blocked the increase in APP and APP cleavage products seen with IsoP treatment (Fig. 3 A–E), further supporting the hypothesis that IsoP functions via TP receptor activation in this paradigm. Similar results were observed when QBI293 cells were transiently transfected with TP receptor and APPwt. Also, treatment with SQ alone did not produce any significant effects (data not shown). Therefore, IsoP treatment appears to increase APP levels, thereby providing more substrate for α and β secretase cleavages, resulting in an increase in Aβ levels that augments Aβ plaque pathology. To test the specificity of this effect for APP, we coexpressed full-length Notch1 (Notch1FL), tau, and Akt cDNAs with TP receptor. Stimulation with I-BOP did not increase Notch1FL, Notch1 120 kDa cleavage product (Notch 120), tau, or Akt protein levels, suggesting the effects of TP receptor activation on APP are quite specific. Furthermore, the I-BOP-stimulated increase in APP levels and sAPPα secretion is independent of the promoter used in our transfection experiments, because both endogenous APP and APP expressed under control of the SV40 promoter (Fig. 4 A) showed similar increases in QBI293 cells. These data demonstrate that the effect of TP receptor activation on APP levels most likely occurs posttranscriptionally.
TP receptor activation increases both endogenous and neuronal APP levels
To provide evidence that the mechanism we observe also applies to primary neuronal cells, we treated dissociated wild-type mouse primary hippocampal neurons with either IsoP or I-BOP and measured APP protein levels by Western blot (Fig. 4 B). In response to these agonists, endogenous mouse APP levels were significantly increased with both IsoP and I-BOP treatment. In addition, an increase in endogenous APP was also observed after I-BOP stimulation in COS-1 cells (data not shown), which express endogenous levels of the TP receptor (Hirata et al., 1991). Finally, we injected I-BOP or vehicle into the left hemisphere hippocampus of 5-month-old Tg2576 male mice. APP levels in hippocampal homogenates obtained 24 h postinjection were assessed by [125I] labeled secondary antibodies for quantitative Western blot (Fig. 4 C,D) and were moderately, but significantly, elevated in the injected but not in the uninjected side of the hippocampus.
TP receptor activation elevates APP levels through increased mRNA stability
To provide further insights into how TP receptor activation increases APP levels, we tested whether TP receptor activation stabilized APP by altering APP protein degradation. We measured APP protein stability in HEK293-APP stable cells transfected with vector or TP receptor. To measure APP degradation rates, cells were stimulated overnight with either control (vehicle) or I-BOP and then were treated with cycloheximide to block protein synthesis. APP protein levels were monitored over the indicated time points. Although there was an increase in APP levels at the initial time point with TP receptor overexpression and a further increase with additional I-BOP activation, there was no significant change in APP protein stability over the times tested (Fig. 5 A,B). In addition, I-BOP stimulation increased APP levels even when APP constructs were used that resulted in retention of the protein in the ER or Golgi (supplemental Fig. 2A, available at www.jneurosci.org as supplemental material), indicating that TP receptor activation increases APP levels early in APP maturation. This increase in APP was also not attributable to reduced lysosomal degradation, because I-BOP was still able to increase APP levels under conditions of inhibition of lysosomal degradation (supplemental Fig. 2B, available at www.jneurosci.org as supplemental material). Therefore, TP receptor activation does not appear to affect APP at the posttranslational level.
Because elevated APP protein levels after I-BOP treatment cannot be explained by altered APP protein degradation, we examined APP mRNA levels after TP receptor activation. In stably transfected HEK-APP cells, TP receptor activation increases APP mRNA (Fig. 5 C). Through careful time course studies, it appears that APP mRNA is elevated in parallel with APP protein levels (Fig. 5 D,E). sAPPα is also elevated over the 8 h time course with I-BOP stimulation and accumulates in the media (Fig. 5 D). Because of these parallel changes in both mRNA and protein and because of the promoter independence, we hypothesized that the increase in APP seen with TP receptor activation occurs posttranscriptionally, presumably through increased APP mRNA stability. To directly measure changes in APP mRNA stability with TP receptor activation, a tetracycline-inducible APPwt construct was cotransfected with TPα into T-Rex-293 cells. APP expression was induced with tetracycline, and APP mRNA was monitored at time points after tetracycline withdrawal in the presence or absence of I-BOP stimulation. I-BOP stimulation significantly increased APP mRNA stability and increased APP mRNA half life from 6.1 to 13.1 h (Fig. 5 F,G).
Treatment with S18886, an orally active TP receptor antagonist, reduced amyloid plaques and insoluble Aβ
Because we observed increased amyloid pathology after TP receptor activation in vivo, we wanted to test whether long-term TP receptor inhibition in an established mouse model of AD-like pathology would prevent activation of the receptor by endogenous IsoP, resulting in decreased amyloid plaques and Aβ levels. For these studies, we used an orally active TP receptor antagonist (S18886) that, like SQ, is able to block I-BOP-induced increases in APP protein, APP mRNA, sAPPα, and secreted Aβ in QBI293 cells expressing both APP and TP receptor (Fig. 6 A,B). Seven-month-old Tg2576 mice were treated with S18886 for 6 months and, compared with controls, S18886 treatment significantly decreased amyloid plaques in the hippocampus and caused a trend toward a decrease of amyloid plaques in the cortex (Fig. 7 A,B). Although there was no significant changes in soluble Aβ (data not shown), insoluble Aβ40 and Aβ42 levels were significantly reduced after S18886 treatment (Fig. 7 C,D). In addition, mice treated with S18886 also exhibited a significant decrease in APP protein levels (Fig. 7 E,F) as assessed by quantitative [125I] Western blot. These data demonstrate that S18886 and other TP receptor antagonists may prove beneficial in preventing amyloid plaque formation, one of the major hallmarks of AD.
Numerous studies have demonstrated that isoprostanes are biomarkers for AD and mild cognitive impairment (Grossman et al., 2005; Montine et al., 2005; de Leon et al., 2006), and that this elevation appears to be somewhat specific for AD compared with other neurodegenerative diseases like Parkinson's disease (Pratico et al., 1998) and frontotemporal dementia (Yao et al., 2003). It is still unknown, however, if isoprostanes play an active role in promoting the disease process. In this study, iPF2α-III increased amyloid plaque formation and soluble and insoluble Aβ1–40 and Aβ1–42 levels in a TP receptor-dependent manner. TP receptor activation posttranscriptionally stabilized APP mRNA, providing more APP substrate for α and β secretase cleavage, thus increasing Aβ production and amyloid plaque levels. An orally active TP receptor antagonist, S18886, decreased APP levels and significantly lowered amyloid plaques and insoluble Aβ.
Although much of the AD literature has focused on modulating Aβ generation at the level of APP processing, increasing attention is focusing on the regulation of APP mRNA and protein. For example, familial AD mutations have been identified in the APP promoter region that increase APP mRNA expression and are associated with increased risk for AD (Theuns et al., 2006), demonstrating that even modest changes in APP expression levels can have profound effects on AD pathogenesis. Also, APP gene duplications have been identified in families with autosomal dominant early onset AD (Cabrejo et al., 2006). In addition, individuals with Down's syndrome have a triplication of chromosome 21, where the APP gene is located, and as a result, exhibit fourfold to fivefold overexpression of APP (Beyreuther et al., 1993). These individuals develop extracellular plaques 20–30 years earlier than normal individuals (Wisniewski et al., 1985). Additionally, APP mRNA and protein levels have been shown to be elevated in sporadic AD, correlating with elevated sAPPα and soluble Aβ (Matsui et al., 2007). Certain APP mRNA splice variants have also been shown to be specifically elevated in AD brains (Rockenstein et al., 1995; Preece et al., 2004). These data show that Aβ levels and amyloid plaque load, in both familial and sporadic AD, are extremely sensitive to alterations in APP mRNA and protein levels, and that small changes can accumulate over time to profoundly effect disease.
Currently, the precise mechanism of TP receptor-induced increase in APP mRNA is still unknown. Although the TP receptor is expressed in the brain (Borg et al., 1994; Gao et al., 1997), its function is not well understood. In oligodendrocytes, TP receptors have been shown to mediate cell proliferation, survival, and gene expression (Blackman et al., 1998; Lin et al., 2005). Additional studies have provided evidence of functional TP receptors on hippocampal neurons (Hsu et al., 1996; Nishihara et al., 2000). In non-neuronal systems, the TP receptor has been shown to couple to multiple G-proteins and activates numerous signaling pathways in a cell type-dependent manner, including phopholipase C activation with IP3 generation and associated increase in calcium (Brass et al., 1987), cAMP generation, (Cracowski et al., 2000; Muja et al., 2001), and activation of the mitogen-activated protein kinase pathway (Gao et al., 2001; Miggin and Kinsella, 2002). Activation of some of these signal transduction pathways has shown similar effects on regulating APP levels. Activation of the prostaglandin E2 receptor induces APP expression at the mRNA and protein level via cAMP (Lee et al., 1999; Pooler et al., 2004). In addition, increases of cAMP by other receptors, such as adrenergic receptors, elevates APP expression (Lee et al., 1997). Also, hepatocyte growth factor receptor stimulation increases APP protein levels via activation of extracellular signal-regulated kinase (ERK) 1/2 (Liu et al., 2003).
mRNA is regulated tightly at the posttranscriptional level. ERK activation has been shown previously to modulate APP mRNA decay (Westmark and Malter, 2001) as does interaction of the 5′ and 3′ untranslated regions of APP mRNA with different RNA binding proteins. For our experiments, however, we are still able to see increases in APP using APP constructs that are deficient in 5′ and 3′ UTRs. TP receptor activation induces a number of downstream signaling pathways that could promote binding of cellular proteins to the coding region of APP mRNA, altering mRNA stability. Multiple RNA binding proteins have been shown in other systems to regulate mRNA stability by binding coding regions (Noubissi et al., 2006; Prechtel et al., 2006; Yarovinsky et al., 2006). In the case of APP, the fragile-X mental retardation protein was recently found to bind to the coding region of APP and promote APP translation (Westmark and Malter, 2007). Additional research is needed to define the precise target downstream of TP receptor activation that affects APP mRNA stability.
Our data show sAPPα levels are elevated more dramatically than APP, sAPPβ, or Aβ species. One possibility for this result is that α-cleavage occurs in higher frequencies than β-cleavage in the cell culture model systems used here. sAPPα accumulates in the media faster and in much greater levels, therefore magnifying the effect shown with TP receptor activation. As shown in Figure 5 D, sAPPα is increased at 8 h but not to the same magnitude that we see with overnight I-BOP stimulation, suggesting that accumulation, rather then generation, is responsible for the larger increases in sAPPα seen with I-BOP stimulation. It is also possible, however, that sAPPα secretion is independently stimulated by TP receptor activation. Studies have shown that protein kinase C (PKC) activation stimulates sAPPα secretion (Slack et al., 1993; Jolly-Tornetta and Wolf, 2000) and TP receptor activation can activate the PKC pathway (Gao et al., 2001). Therefore, we cannot rule out the possibility that sAPPα levels may be increased to such a dramatic extent because of two mechanisms, increase in APP precursor as well as stimulation of secretion.
Given that isoprostanes are elevated before amyloid plaque formation, it appears that there is a feedforward mechanism where isoprostane and amyloid plaque formation both propagate each other (Fig. 8). Although our studies show that IsoP can increase Aβ, Aβ can also induce isoprostane generation (Mark et al., 1999; Pratico et al., 2001) in an age-dependent manner (Brunetti et al., 2004). Our studies also show an increase of IsoP (Fig. 2 B) that is reflective of the increased amyloid load in IsoP-treated mice, further supporting this feedforward model. Increased oxidative insults or decreased antioxidant defenses associated with aging can lead to increased isoprostane levels and promote amyloid plaque formation, generating more oxidative stress and elevating isoprostanes, ultimately leading to AD.
Our studies highlight a novel pathogenic pathway that promotes Aβ generation and enhances amyloid plaque formation in a mouse model of AD amyloidosis. This pathway provides a new avenue for AD drug discovery. Indeed, we found that the orally available TP receptor antagonist S18886 was successful in preventing amyloid plaque formation and decreasing insoluble Aβ in our AD mouse model. S18886 also decreased APP levels in the mouse brain, providing evidence that the mechanism we observe in cell culture is similar to what occurs in vivo, and again demonstrating that small changes in APP levels can greatly affect AD pathology over time. Although our studies show that TP receptor antagonism attenuates the development of amyloid plaque pathology, it is also important to test whether TP receptor antagonism would be beneficial in reversing amyloid plaque formation in mice further along in the disease process. This study is the first of its kind to suggest the TP receptor signaling pathway as a mediating pathway in AD pathogenesis. Although existing TP receptor antagonists have been developed primarily for cardiovascular indications, TP receptor antagonists that readily gain access to the brain may prove beneficial for the treatment of AD.
This work was supported by National Institutes of Health Grant AG11542, the Coins for Alzheimer's Research Trust Fund, Cell and Molecular Biology Training Grant T32-GM07229, and the Marian S. Ware Program for Drug Discovery. We thank Drs. John Trojanowski, Kurt Brunden, and Peter Klein for their advice and helpful suggestions on this manuscript; Drs. Linda Kwong and Matthew Winton for their assistance with experiments; Daniel Martinez, Ingrid Solano, and Ian Mills for excellent technical support; and Dr. Garret FitzGerald for reagents. V.M.-Y.L. is the John H. Ware III Professor for Alzheimer's Disease Research.
- Correspondence should be addressed to Virginia M.-Y. Lee, Center for Neurodegenerative Disease Research, Department of Pathology and Laboratory Medicine, Maloney Building, 3rd Floor, 3600 Spruce Street, Philadelphia, PA 19104-4283.