|Home | About | Journals | Submit | Contact Us | Français|
Several large population-based or clinical trial studies have suggested that certain dihydropyridine (DHP) L-type calcium channel blockers (CCBs) used for the treatment of hypertension may confer protection against the development of Alzheimer disease (AD). However, other studies with drugs of the same class have shown no beneficial clinical effects. To determine whether certain DHPs are able to impact underlying disease processes in AD (specifically the accumulation of the Alzheimer Aβ peptide), we investigated the effect of several antihypertensive DHPs and non-DHP CCBs on Aβ production. Among the antihypertensive DHPs tested, a few, including nilvadipine, nitrendipine and amlodipine inhibited Aβ production in vitro, whereas others had no effect or raised Aβ levels. In vivo, nilvadipine and nitrendipine acutely reduced brain Aβ levels in a transgenic mouse model of AD (Tg PS1/APPsw) and improved Aβ clearance across the blood-brain barrier (BBB), whereas amlodipine and nifedipine were ineffective showing that the Aβ-lowering activity of the DHPs is independent of their antihypertensive activity. Chronic oral treatment with nilvadipine decreased Aβ burden in the brains of Tg APPsw (Tg2576) and Tg PS1/APPsw mice, and also improved learning abilities and spatial memory. Our data suggest that the clinical benefit conferred by certain antihypertensive DHPs against AD is unrelated to their antihypertensive activity, but rely on their ability to lower brain Aβ accumulation by affecting both Aβ production and Aβ clearance across the BBB.
Alzheimer disease (AD) is the major cause of dementia in the elderly in Western countries and is characterized by the progressive accumulation of in-tracellular neurofibrillary tangles, extra-cellular parenchymal senile plaques and cerebrovascular deposits (1). The principal component of senile plaques and cerebrovascular deposits is the 39–43 amino acid β-amyloid peptide (Aβ), which is proteolytically derived from the amyloid precursor protein (APP) (2). Although the central role of Aβ in AD remains to be proven in clinical trials, data accumulated during the past two decades place Aβ peptides at center stage as the main culprit in initiating the pathological cascade that eventually leads to AD (3–5). Aβ peptides are derived from the sequential proteolysis of APP by β- and γ-secretases. The major β-secretase is an aspartyl protease termed BACE-1 (β-site APP cleaving enzyme) (6). BACE-1 cleaves APP within the extracellular domain of APP, resulting in the secretion of the large ectodomain (sAPPβ) and generating a membrane-tethered C-terminal fragment CTFβ or C99 which serves as a substrate for γ-secretase. The multimeric γ-secretase complex cleaves at multiple sites within the transmembranous CTFβ, generating C-terminally heterogeneous Aβ peptides ranging between 38 to 43 amino-acid residues in length that are secreted (2). In addition to BACE-1 and γ-secretase, APP can be cleaved by α-secretase within the Aβ domain between Lys16 and Leu17, releasing APPsα and generating CTFα (C83) which is further cleaved by γ-secretase to generate an N-terminally truncated Aβ termed p3.
During the course of examining risk factors for AD, such as hypertension, it has become clear that certain antihypertensive compounds may be protective, not just against stroke-related dementia, but also independently against AD. For instance, in the Syst-Eur trial, which involved active treatment with the dihydropyridine (DHP) calcium channel blocker (CCB) nitrendipine in over 2,400 patients, there was a 55% reduction in the incidence of AD (7–9). Although nitrendipine is an antihypertensive and would therefore be expected to lower the incidence of stroke-related dementia, the protection against AD is unexplained by the known mechanism of action of nitrendipine. Small clinical studies of another DHP, nilvadipine, have shown stabilization of cognitive decline and reduced incidence of AD in hypertensive MCI patients, whereas the closely related antihypertensive DHP amlodipine failed to provide any benefit (10–12), suggesting that the beneficial effect of nilvadipine is not related to its antihypertensive activity. Although only a small number of study cases were taking DHPs in the Cache County antihypertensive study (13), a trend also was reported toward a lowered risk for AD. This was not observed for non-DHP calcium channel blockers. The Baltimore longitudinal study of aging (14) also showed a strong trend toward reduced relative risk of AD in DHP calcium channel blocker users, whereas no lowered risk was observed in the non-DHP calcium channel blocker using group. Additional evidence suggests that antihypertensive activity per se is not responsible for the protective effects against AD. For instance, several studies with antihypertensive medications have shown no evidence of prophylactic activity against AD. For example, although use of angiotensin converting enzyme (ACE) inhibitors has been shown to decrease the incidence of stroke-related dementia, no positive effect was demonstrated for AD in several studies (9,15–17). In addition, antihypertensive treatment with a thiazide did not protect against dementia in the Systolic Hypertension in the Elderly Program trial which again argues against the prevention of dementia by solely lowering blood pressure (18). Moreover, some studies have even suggested that certain antihypertensive DHPs may be detrimental, as far as risk for AD is concerned. For instance, older people who had a history of hypertension and who were taking the DHP nifedipine were more likely than subjects taking other antihypertensive agents to experience cognitive decline during the 5-year follow-up period of the Canadian Study of Health and Aging (19). Overall then, it is clear that any clinical signal from DHPs in the protection against AD is not a drug class effect and is not related to the antihypertensive effects of these drugs. However, specific DHPs like nitrendipine and nilvadipine do suggest clinical protective signals against AD whereas amlodipine and nifedipine do not. Given that central nervous system (CNS) accumulation of Aβ peptides are thought to be central to the AD process and that certain DHPs protect against AD, we investigated the effect of DHPs as well as non-DHP CCBs on the production of Aβ peptides.
7W CHO (20) cells stably transfected with human APP751 were maintained in DMEM (ATCC, Manassas, VA, USA) medium containing 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA), 1× mixture of penicillin/streptomycin/fungizone mixture (Cambrex, Charles City, IA, USA) and 0.3% geneticin (Invitrogen) as a selecting agent. Cells were cultured in 96-well culture plates and treated for 24 h with different calcium channel blockers as indicated in the figure legend. All test compounds were diluted in dimethyl sulfoxide (DMSO) before being exposed to confluent Chinese hamster ovary (CHO) cells. Control wells received the same volume of DMSO, and the final DMSO concentrations in the culture medium for all treatment conditions were lower than 1%. Potential cytotoxicity of the different calcium channel blockers was evaluated using the Cytotoxicity Detection Kit (Roche Diagnostics, Mannheim, Germany) and no significant toxicity of the different calcium channel blockers was observed for the dose range studied (data not shown). Aβ1–40 and Aβ1–42 were analyzed in the culture medium by using commercially available sandwich enzyme-linked immunosorbent assays (ELISAs) (Invitrogen). All experiments were repeated 3–4×.
The impact of nilvadipine and amlodipine on APP processing was evaluated using 7W CHO cells as we published previously (21). Briefly, confluent 7W CHO cells were treated for 24 h with different doses of nilvadipine and amlodipine in 6-well plates. Cellular proteins were extracted with 150 μL of ice cold M-PER Reagent (Pierce Biotechnology, Rockford, IL, USA) containing 1 mmol/L phenylmethanesulfonyl fluoride, 1× of protease cocktail inhibitor (Roche Diagnostic Corporation, Indian-apolis, IN, USA) and 1 mmol/L sodium orthovanadate. Samples were sonicated, denatured by boiling in Laemmli buffer (Bio-Rad, Hercules, CA, USA) and resolved onto 4% to 20% gradient poly-acrylamide gels (Bio-Rad). After electro-transferring onto polyvinylidene difluoride membranes, Western blots were immunoprobed with an anti-APP C-terminal (751–770) antibody (EMD Chemicals Inc., Gibbstown, NJ, USA), with an antiactin antibody (Chemicon, Temecula, CA, USA) used as a reference antibody to ensure that an equal amount of proteins were electrotransferred. Additionally, sAPPα was detected by Western-blot in the culture medium surrounding 7W CHO cells using the antibody 6E10 (Signet Laboratories Inc., Dedham, MA, USA) which recognizes the amino acids 1–17 of Aβ, and sAPPβ was detected in the culture medium using an antihuman sAPPβ antibody (Immuno-Biological Laboratories Co. Ltd., Gunma, Japan).
β-Secretase activity was measured using human recombinant BACE-1 (Calbiochem, San Diego, CA, USA) with two commercially available kits, a FRET-based assay (Biovision Inc, Mountain View, CA, USA) and a chemoluminescent assay (DiscoveRx, Fremont, CA, USA) following the recommendations of the manufacturers. The β-secretase inhibitor IV (BACE IV) inhibitor was used as a positive control in both assays and was purchased from EMD Chemicals Inc. FRET and chemoluminescent signals were quantified on a HTS Synergy multiplate reader from Biotek (Winooski, VT, USA).
DAPT (10 mg/kg of body weight), nilvadipine (2 mg/kg), amlodipine (2 mg/kg), nitrendipine and nifedipine (2 mg/kg) were administered daily to 6-month-old Tg PS1/APPsw (22) mice intraperitoneally for 4 d. Vehicle-treated Tg PS1/APPsw mice received an i.p. injection of 50% DMSO in PBS for 4 d. Within 1 h after the last injection, animals were humanely euthanized, their brains and plasma were immediately frozen in liquid nitrogen until being an-alyzed for Aβ1–40 and Aβ1–42 levels using commercially available ELISA kits (Invitrogen). Briefly, brains were homogenized at 4°C in mammalian protein extraction reagent (MPER) containing 1 mmol/L PMSF and 1× protease inhibitor cocktail (protein research product from Pierce) and centrifuged at 15,000g for 30 min at 4°C. The supernatant containing soluble Aβ was collected and further diluted with the sample diluent provided in the Aβ ELISA kits and assessed for Aβ according to the manufacturer’s protocol. Protein concentrations were measured in the supernatant using the BCA method (Pierce) and results of brain soluble Aβ1–40 and Aβ1–42 were calculated in pg/mg of protein and expressed as a percentage of the values obtained in the vehicle treated animals for both Aβ1–40 and Aβ1–42. Plasma samples were diluted with the diluent provided in the ELISA kit and assayed according to the manufacturer’s recommendation. Plasma Aβ1–40 and Aβ1–42 values in the different treatment groups were expressed as a percentage of the values obtained in the vehicle treatment group. All experiments using animals were performed under protocols approved by the Institutional Animal Care and Use Committee of the Roskamp Institute.
Human brain microvascular endothelial cells (HBMEC), endothelial cell media (ECM), fetal bovine serum, penicillin/streptomycin solution, and endothelial cell growth supplement (ECGS) were purchased from Sciencell Research Laboratories (Carlsbad, CA, USA). Fibronectin solution was purchased from Sigma Chemical Co. (St. Louis, MO, USA). Fluorescein-Aβ1–42 (FlAβ1–42) was purchased from rPeptide (Bogart, GA, USA). The lyophilized peptide was dissolved to 1 mg/mL in 1,1,1,3,3,3-hexafluoro-2-propanol at 4°C to minimize the formation of β-sheet structures and promote α-helical secondary structure and monomerize the peptide. The peptide was allowed to air dry in a chemical fume hood for 1 h at room temperature, followed by further drying in a speedvac (Thermo-Savant, Manasquan, NJ, USA) for 30 min. The esulting clear film was resuspended in 100% dimethylsulfoxide (DMSO) to a concentration of 1 mmol/L, followed by aliquoting and storage at −80°C. The 24-well membrane inserts (translucent, 0.4 μm pore) and 24-well companion plates were purchased from Fisher Scientific (St. Louis, MO, USA). Aβ transcytosis in vitro was quantified as we described previously using HBMEC (23). Briefly, HBMEC were seeded at 50,000 cells/cm2 onto fibronectin-coated (4 μg/cm2), 24-well, 0.4 μm-pore, translucent membrane inserts (0.3 cm2/insert) to establish a polarized HBMEC monolayer representative of the BBB. The layer of HBMEC separates this system into apical (“blood”) and basolateral (“brain”) compartments. ECM containing 2-μmol/L FlAβ1–42 was placed in the basolateral (donor) compartment. The apical (receiver) side of the membrane was exposed to various concentrations of nilvadipine, amlodipine and nitrendipine (1, 5, and 10 μmol/L) in ECM. The donor compartment was sampled at time 0 to establish the initial concentration of FlAβ1–42 in each group. Following exposure of the insert to the well containing FlAβ1–42, samples were collected from the apical compartment at various time points up to 90 min to assess the movement of FlAβ1–42 across the HBMEC monolayer (basolateral-to-apical). The samples were analyzed (λex = 485 nm and λem = 516 nm) for FlAβ1–42 using a BioTek Synergy HT multi-detection microplate reader. The apparent permeability (Papp) of FlAβ1–42 was determined using the equation Papp = 1/AC0 * (dQ/dt), where A represents the surface area of the membrane, C0 is the initial concentration of FlAβ1–42 in the basolateral compartment, and dQ/dt is the amount of FlAβ1–42 appearing in the apical compartment in the given time period. The Papp of FlAβ1–42 in the presence of drug was compared with control (that is, no drug exposure) and expressed as a percentage. We corrected for permeability resistance associated with the blank membrane as reported recently (24).
Wild-type B6/SJL F1 mice (12 month-old, Jackson Laboratories, Bar Harbor, ME, USA) were anesthetized via inhalation using a 3% isoflurane/oxygen mix and maintained at 37°C using a homeothermic blanket system (Harvard Apparatus, Holliston, MA, USA). While under anesthesia, the mice were injected intraperitoneally with the vehicle only (50% DMSO in PBS) or with nilvadipine, nitrendipine and amlodipine at a dosage of 2 mg/kg. Five min after the intraperitoneal injection, the mice were stereotaxically injected with 3 nmol of human Aβ1–42 (Invitrogen) into the caudate putamen of the brain (3 mm posterior to the eye line and 2.0 mm lateral to the midline and 3 mm below the surface of the brain) using a sterile 27-gauge needle connected to a 10-μl luer-tip Hamilton syringe (Fisher Scientific). (In separate studies, we confirmed the site of delivery was in the caudate putamen by injecting Evans blue.) Ten min after the intracranial administration of human Aβ1–42, plasma samples were collected and analyzed for human Aβ1–42 by ELISA. The level of human Aβ1–42 in the plasma samples collected from the DHP-treated mice was compared with those receiving the vehicle alone. Control experiments were performed to determine whether the intracranial injection procedure and the administration of nilvadipine could affect the permeability of the BBB. Vascular permeability/BBB leakage was evaluated by measuring the extravasation of the Evans blue dye in the brain as described previously (25,26). Briefly, B6/SJL F1 mice were intraperitoneally injected with 100 μL of 50% DMSO/PBS (vehicle), with 2 mg/kg of nilvadipine or were injected intracranially as described above with 3 μL of vehicle, with 3 nmol of Aβ1–42 or went untreated (control mice). A group of mice also was subjected to a cold injury procedure to induce BBB leakage (positive control). A cylinder of aluminum (3 mm diameter) cooled in liquid nitrogen was applied for 15 s bilaterally at the surface of the skull (after removing the periosteum) in the occipital area of anesthetized mice. Five min after subjecting the animals to these different procedures, the Evans blue dye (Sigma) was injected intraperitoneally (400 mg/kg of body weight) and mice were euthanized 1 h later. The brains of the animals were collected, weighed and incubated at 37°C for 72 h in 2 mL of pure formamide (Sigma) to extract the Evans blue dye. Optical density of the extracted dye was measured at 620 nm, and values were reported per gram of brain. All experiments using animals were performed under protocols approved by the Institutional Animal Care and Use Committee of the Roskamp Institute.
Ten-month-old transgenic mice Tg PS1/APPsw (22) overexpressing APP695 containing the “Swedish” mutation and a mutant presenilin-1 (M146L) were fed with an irradiated powder diet (18% protein, Harlan Teklad, Madison, WI, USA) containing 0.03% (weight to weight) of nilvadipine (n = 8) formulation for oral dosage (which corresponds to an oral drug intake in mice of approximately 50 mg/kg/day, equivalent to a 2 mg/kg/day intraperitoneal dosage of nilvadipine according to pharmacokinetic data [data not shown]) or a placebo (0.03% of solid dispersion without nilvadipine) (n = 7) for 10 months. Tg PS1 control littermates were fed for 10 months as indicated above (n = 10 for nilvadipine treatment and n = 13 for placebo treatment). Additionally, 3-month-old Tg APPsw (27) were fed with the same formulation of nilvadipine (n = 10) or placebo (n = 12) for a period of 17 months. Solid dispersion formulations of nilvadipine and placebo were provided by Astellas Pharma Inc. (Osaka, Japan). Twenty-month-old mice were humanely euthanized and their brains were fixed in 4% paraformaldehyde for 24 h at 4°C before embedding in paraffin blocks using a Tissue-Tek (Sakura, Torrance, CA, USA). All experiments using animals were performed under protocols approved by the Institutional Animal Care and Use Committee of the Roskamp Institute.
Brains were cut sagitally into 6-μm-thick sections with a microtome (2030 Biocut, Reichert/Leica, Wetzlar, Germany) and Aβ burden determined as we published previously (21). Briefly, sections were mounted on slides and de-paraffinized in xylene (2 × 5 min) and hydrated in graded ethanol (2 × 5 min in 100%, 5 min in 85%, 5 min 70%) to water. The endogeneous peroxidase activity was quenched with a 20-min H2O2 treatment (0.3% in water) and after being rinsed, sections were incubated with blocking buffer (Protein Block Serum-free, Dako-Cytomation, Carpinteria, CA, USA) for 20 min. The monoclonal antibody 4G8 (Signet Laboratories) recognizing human Aβ (diluted 1:750) was applied onto the sections overnight at 4°C and was detected using the Vectastain ABC (avidin-biotin-peroxidase complex) Elite kit (Vector Laboratories, Burlingame, CA, USA). For each brain, 4 to 5 nonconsecutive and randomly selected sections containing the hippocampus were used to perform the quantification of Aβ burden. The stained area within particular regions (hippocampus, cortex or subfields of cortex) was quantified using the Image-Pro Plus software (Media Cybernetics, Bethesda, MD, USA). An average value was calculated for each area from individual mice. These averages were used to estimate the overall staining for each treatment group. Results of the Aβ burden were expressed as a percentage of Aβ area stained/total area examined.
After 10 months of treatment with nilvadipine or a placebo, 20-month-old Tg PS1/APPsw and their control littermates Tg PS1 mice were subjected to the following behavioral tests. A general test of activity and exploratory behavior was conducted by placing each animal in a large open field arena for 30 min. Exploratory behavior was monitored via an overhead video camera and the video signal was sent to a computer for analysis using EthoVision (Noldus Information Technology, Sterling, VA, USA) tracking software. Key dependent measures were total distance traveled in the entire arena, distance traveled along the outer 10-cm perimeter of the arena, and average distance from the center arena point. Dependent measures were calculated across time blocks and rates were compared across the duration of the open field session.
A mouse version of the Morris water maze (28) was used to assess spatial learning and memory. A 2-m polypropylene pool, filled with opacified water (24°C) and located in an approximately 3.5 m2 room rich in extra-maze cues was used. Testing began with a standard hidden platform reference memory protocol, where the pool contained a 20-cm diameter platform hidden 1.5 cm below the water surface in the center of one pool quadrant (NE, SE, SW or NW). Mice were placed in the pool from 1 of 4 entrance points (N, E, S or W) and allowed 90 s to locate the hidden platform. Four consecutive acquisition trials were given per day across 9 d. Mice were allowed to remain on the platform 30 s prior to subsequent trials, and any mouse that did not locate the hidden platform within 90 s was guided to the platform using an extraction tool as a beacon. During the acquisition trials the platform location remained constant for a given mouse. During all water maze trials, an overhead video camera captured the image of the mouse in the water maze and it was digitally tracked using the EthoVision software. All experiments using animals were performed under protocols approved by the Institutional Animal Care and Use Committee of the Roskamp Institute.
Statistical analysis were performed by analysis of variance (ANOVA) using SPSS Version 12.0 for Windows and post hoc analysis were carried out using Bon-ferroni correction. P < 0.05 was considered significant.
In this study, we compared the effects of eight commonly used antihypertensive compounds known to inhibit L-type calcium channels on Aβ1–40 and Aβ1–42 production using 7W CHO cells overexpressing human APP751 (20), including seven clinically used antihypertensive DHPs (felodipine, nifedipine, amlodipine, nilvadipine, nitrendipine, nimodipine and isradipine) and two non-DHPs (verapamil and diltiazem). Among the L-type calcium channel blockers tested, only nilvadipine and amlodipine lowered both Aβ1–40 and Aβ1–42 for doses between 1 and 10 μmol/L significantly (Figure 1). Nitrendipine appears to marginally inhibit Aβ1–40 at 5 and 10 μmol/L but does not affect Aβ1–42 production significantly. At higher doses, nitrendipine inhibits both Aβ1–40 and Aβ1–42 production (data not shown). Felodipine, nifedipine, diltiazem and verapamil have no significant impact on Aβ1–40 and Aβ1–42 production for the dose range studied. Nimodipine appears to dose dependently stimulate both Aβ1–40 and Aβ1–42 levels whereas isradipine stimulates Aβ1–42 levels without significantly affecting Aβ1–40 values (see Figure 1).
As L-type calcium channel blockers also may impact other types of calcium channels (for instance amlodipine significantly blocks N-type and P/Q-type calcium channels ), we also investigated the effect of different toxins (agatoxin TK, agatoxin IVa, conotoxin GVIA and conotoxin MVIIC) known to selectively block N, P and Q-type calcium channels. Blockade of N, P or Q-type calcium channels with these toxins did not significantly impact Aβ1–40 or Aβ1–42 production (data not shown) in 7W cells. In addition, blockade of potassium, sodium and chloride channels with glyburide, lidocaine N-ethyl bromide (QX-314), protopine, 4′-([1–(2-[6-Methyl-2-pyridinyl] ethyl)-4-piperidinyl] carbonyl) methanesulfonanilide (E4031), 5-Nitro-2-(3-phenylpropylamino) benzoic Acid (NPPB) or 4,4′-Diisothiocyanostilbene-2,2′-disulfonic Acid, 2Na (DIDS) did not affect Aβ production (data not shown). Also, neither FPL64176 (methyl 2,5 dimethyl-4[2-(phenylmethyl)benzoyl]-1H-pyrrole-3-carboxylate) nor Bay-K 8664, two potent agonists of L-type calcium channels, significantly affected Aβ1–40 or Aβ1–42 levels (data not shown). Altogether these data suggest that selective DHPs are able to lower Aβ production in vitro independently of their inhibitory action on L-type calcium channels.
The effects of nilvadipine and amlodipine (the two most potent Aβ-lowering DHP identified in vitro among the CCB tested) on APP-CTF, APPsα and APPsβ secretion using 7W CHO cells were investigated. In whole cell lysate, neither amlodipine nor nilvadipine modified the level of full length APP (Figure 2). Following 24 h of treatment with amlodipine, a dose-dependent increase in APP-CTFα and decrease in CTFβ levels were observed, whereas a slight stimulation of APP-CTFα was noted for nilvadipine (see Figure 2). However, no significant increase in sAPPα production was observed in the culture media following nilvadipine or amlodipine treatments suggesting that these compounds do not stimulate the α-secretase cleavage of APP. A decreased APPsβ production was observed following 24 h of treatment with amlodipine and nilvadipine suggesting that these compounds impact the β-cleavage of APP (see Figure 2). A decreased APPsβ secretion also was observed following 6 h of treatment with nilvadipine and amlodipine (data not shown). BACE-1 (β-secretase) specific FRET and chemoluminescent assays were implemented using recombinant human BACE-1 to determine a possible direct inhibition of BACE-1 activity by nilvadipine and amlodipine. No noticeable inhibition of BACE-1 activity was observed with either nilvadipine or amlodipine in either assays ruling out the possibility that amlodipine and nilvadipine are direct BACE-1 inhibitors for the dose range studied (Figure 3).
The impact of nilvadipine and amlodipine on the γ-secretase pathway also was tested by investigating Aβ production in SHSY cells overexpressing the APP-CTFβ (C99), the sole substrate of γ-secretase leading to Aβ production. Neither nilvadipine nor amlodipine inhibited Aβ production in C99 over -expressing cells (Figure 4). In addition, no alteration of Notch cleavage (data not shown) was observed after treatment with nilvadipine or amlodipine, further suggesting that these compounds do not impact the γ-secretase pathway significantly.
We tested nilvadipine, amlodipine, nifedipine, nitrendipine and DAPT for their ability to acutely affect Aβ1–40 and Aβ1–42 levels in the brains of Tg PS1/APPsw mice. The known functional γ-secretase inhibitor DAPT was used as a positive control. Nifedipine, which was unable to lower Aβ production in vitro, was used as a negative control. Although nitrendipine shows only a weak inhibition of Aβ production in vitro, this compound was included because it shows a clear prophylactic activity against AD (7). Nilvadipine, amlodipine, nitrendipine and nifedipine were dosed at 2 mg/kg/d (a physiologically relevant dose conferring antihypertensive activity) whereas DAPT was used at a dosage of 10 mg/kg/d for a period of 4 d. An analysis of brain Aβ levels following 4 d of treatment shows that DAPT, nilvadipine and nitrendipine significantly reduce soluble Aβ levels in the brains of Tg PS1/APPsw mice whereas amlodipine and nifedipine were ineffective (Figure 5). The failure of amlodipine to lower Aβ levels in vivo despite its ability to lower Aβ in vitro may be related to its inability to cross the blood-brain barrier (BBB) compared with highly lipophilic DHPs such as nilvadipine and nitrendipine which accumulate in the brain (30). In addition, these data further confirm that the Aβ-lowering activity of nilvadipine and nitrendipine in vivo is not related to their antihypertensive activity, since both amlodipine and nifedipine were unable to reduce brain Aβ levels.
Surprisingly, an increased plasma Aβ level was observed in animals that received the acute treatment with nilvadipine and nitrendipine despite a reduction in brain Aβ levels (see Figure 5), suggesting that these compounds may facilitate the clearance of Aβ across the BBB or may prevent Aβ degradation/elimination in the periphery. We therefore investigated the effect of nilvadipine, nitrendipine and amlodipine in an in vitro model of the BBB, employing human brain microvascular endothelial cells. Both nilvadipine and nitrendipine increased Aβ clearance from the brain to the peripheral side of the in vitro BBB model, whereas amlodipine was ineffective (Figure 6). In addition, nilvadipine and nitrendipine appear to facilitate the clearance of human Aβ across the BBB in vivo in wild-type mice that were injected intracranially with human Aβ, whereas amlodipine was ineffective (Figure 7), further confirming that the elevation in plasma Aβ observed in Tg PS1/APPsw are likely reflective of an increased clearance of Aβ across the BBB. We performed additional control experiments to determine whether nilvadipine or the intracranial Aβ injection procedure were impacting the leakiness of the BBB using Evans blue injections to monitor BBB permeability (25,26). The cold injury procedure which induces a rapid breakdown of the BBB (31) was used as a positive control and lead to a significant extravasation of the Evans blue dye in the brain (see Figure 7). We did not observe a significant effect of nilvadipine or of the intracranial injection procedure on the leakiness of the BBB (see Figure 7), suggesting that nilvadipine is affecting the transport of Aβ across the BBB and is not increasing the permeability of the BBB nonspecifically.
We also investigated the effect of chronic treatment with nilvadipine on β-amyloid burden in Tg APPsw and Tg PS1/APPsw mice. Ten-month-old Tg PS1/APPsw mice (already presenting a β-amyloid pathology) and their control littermates were fed a powder diet containing 0.03% of nilvadipine (solid dispersion formulation for oral dosage) or a placebo (solid dispersion formulation without nilvadipine) for 10 months. In addition, 3-month-old Tg APPsw (Tg2576; prophylactic treatment initiated prior to Aβ deposits) were fed with nilvadipine or a placebo as indicated above until 20 months of age. Data show that chronic nilvadipine treatments reduced plaque burden by approximately 40% for different areas of the brains of Tg PS1/APPsw mice and by approximately 50% in the brains of Tg APPsw mice (Figure 8).
Behavioral testing was conducted in Tg PS1/APPsw mice and control litter-mates (Tg PS1; these mice do not show cognitive impairment and perform similarly to wild-type mice in the open field and the Morris water maze ) fed with nilvadipine and the placebo. In the open field test (Figure 9), Tg PS1/APPsw mice show a hyperactivity and do not habituate to the arena compared with their Tg PS1 littermate controls. Nilvadipine treatment significantly increases the habituation of Tg PS1/APPsw to the arena, suggesting an improvement in memory (see Figure 9). In the Morris water maze, the acquisition of place learning was impaired in Tg PS1/APPsw mice compared with Tg PS1 mice (Figure 10). Nilvadipine treatment improved the learning functions of Tg PS1/APPsw mice in the Morris water maze to values comparable to their control littermates during the acquisition trials, showing improvement in learning (see Figure 10). The reversal training was used to distinguish goal- directed navigation from search strategy. With the platform moved to the opposite quadrant of the pool, path length to the platform during the reversal trials was reduced for Tg PS1, Tg PS1 treated with nilvadipine and Tg PS1/APPsw treated with nilvadipine compared with placebo treated Tg PS1/APPsw (see Figure 10), showing that nilvadipine improved spatial learning in Tg PS1/APPsw mice. Probe trials (in which the platform was removed) were conducted after a 5-day period of rest to assess the impact of nilvadipine on long-term memory. Memory of the platform localization (retention) was reduced significantly in Tg PS1/APPsw mice compared with their control littermates (Tg PS1 mice) (see Figure 10). Nilvadipine significantly improved retention in Tg PS1 mice and Tg PS1/APPsw compared respectively to Tg PS1 and Tg PS1/APPsw mice of the placebo group. Overall, nilvadipine significantly reduced the impaired learning and memory deficits which characterize the cognitive dysfunction observed in transgenic mouse models of AD.
The use of particular antihypertensive medications, specifically the DHPs nitrendipine and nilvadipine, have been associated with a reduced risk of developing dementia, including AD (7,8,10,16,33). We therefore investigated the effects of several clinically used anti-hypertensive DHPs as well as non-DHP CCBs on Aβ production as this peptide is believed to be central to the disease process. Our data reveal that the DHPs am-lodipine and nilvadipine are the most potent of the DHPs that we tested for inhibiting Aβ1–40 and Aβ1–42 production in vitro whereas other DHPs and non-DHP CCBs appear inefficient or even raise the levels of Aβ1–40 and Aβ1–42 in vitro significantly. Nilvadipine and amlodipine were unable to inhibit Aβ production in a cell line overexpressing the APP-CTFβ fragment and did not impact the γ-secretase cleavage of Notch, suggesting that these compounds do not significantly impact the γ-secretase pathway. The in vitro Aβ-lowering activity of nilvadipine and am-lodipine appear to be mediated by an inhibition of the β-cleavage of APP since a reduction in sAPPβ production was observed. However, no direct effect of nil-vadipine and amlodipine was observed on β-secretase activity in two different cell free assays, suggesting that these drugs are indirect inhibitors of BACE-1 activity. The effect of nilvadipine and amlodipine on Aβ levels is independent of their calcium channel blocking activity, as other antihypertensive DHPs that display similar activity toward L-type calcium channels do not display the same ability to inhibit Aβ production.
In vivo acute dosage studies in Tg PS1/APPsw mice reveal that nilvadipine and nitrendipine reduce brain levels of soluble Aβ, whereas other antihypertensive DHPs such as amlodipine and felodipine are inactive, further confirming that the in vivo Aβ-lowering activity of nilvadipine and nitrendipine is independent of their antihypertensive activity. Surprisingly, an increased level of Aβ in the plasma was observed following an acute dosage with nilvadipine and nitrendipine, which prompted us to explore whether these drugs were facilitating the clearance of Aβ across the BBB. We observed, both in vitro and in vivo, that nilvadipine and nitrendipine effectively improved the clearance of Aβ across the BBB whereas amlodipine was ineffective. We used the Evans blue methodology to determine whether nilvadipine was impacting the overall permeability of the BBB. Compared with peripheral organs such as the liver, kidney and heart, we observed that the extravasation of the Evans blue dye was approximately ten times lower in the brain (data not shown), illustrating the restrictive permeability of the BBB compared with peripheral tissues as analyzed previously in detail for peptides and polar solutes (34,35). We found that nilvadipine does not increase the permeability of the BBB to Evans blue, suggesting that nilvadipine is impacting the transport of Aβ across the BBB. The molecular mechanisms responsible for the transport of Aβ across the BBB have been studied extensively. The receptor for advanced glycation end products (RAGE) has been identified as the main receptor responsible for the influx of circulating Aβ to the brain whereas the lipoprotein receptor (LRP-1) has been identified as the receptor mediating the efflux of Aβ from the brain into the circulation (36,37). We have confirmed that LRP-1 and RAGE are the main receptors mediating the transport of Aβ in our in vitro model of the BBB employing HBMEC (23). Our data showing that nilvadipine can selectively improve the transport of Aβ from the brain side to the peripheral side of this in vitro BBB model suggest that nilvadipine is either inhibiting RAGE or is facilitating LRP-1-mediated Aβ transport. Nilvadipine has been shown to inhibit nuclear factor-κB (NFκB)-dependent transcription (38) and, as expected for an NFκB inhibitor, nilvadipine users have reduced plasma cytokine levels (39). Interestingly, both RAGE and BACE-1 expression are regulated by NFκB (40–43). In particular, inhibition of NFκB signaling decreases RAGE expression in endothelial cells (44) and reduces Aβ production and BACE-1 expression levels (21,45,46). The inhibition of NFκB by nilvadipine therefore may represent a plausible mechanism responsible for both the inhibition of Aβ production and for the increased clearance of Aβ across the BBB observed following nilvadipine treatment.
We next investigated the effect of a chronic nilvadipine administration in transgenic mouse models of AD. Chronic oral dosage with nilvadipine resulted in an approximately 40% to 50% reduction in β-amyloid burden in Tg PS1/APPsw and Tg APPsw mice. In addition, improvement of cognitive behavior was observed after nilvadipine administration in Tg PS1/APPsw mice showing that nilvadipine can reduce memory deficits associated with the β-amyloid load in Tg PS1/APPsw mice.
We have shown previously that nilvadipine increases cerebral blood flow in a transgenic mouse model of AD (47). In addition, nilvadipine has been shown to improve spatial memory in a rat model of AD (48), consistent with our findings in transgenic mouse models of AD. Interestingly, nilvadipine, but not amlodipine, has been shown to improve cerebral blood flow and cognitive function, and to reduce the rate of conversion to AD in patients with mild cognitive impairment (10). Moreover, in an open-label safety and tolerability trial with nilvadipine in AD patients, we observed that cognition was stabilized in AD patients treated with nilvadipine, whereas untreated subjects showed cognitive decline (49).
The proteases that generate Aβ directly are considered major targets for AD. However, direct inhibitors of these proteases with a clinically appropriate profile have been difficult to discover so far. Brain Aβ accumulation theoretically is dependent on the balance between Aβ production and Aβ clearance. The routes of Aβ elimination involve both the catabolism of Aβ (50) and its transport across the BBB (37,51,52). The clearance of Aβ across the BBB plays a critical role in brain Aβ accumulation (53,54) and has been suggested to represent the major pathway leading to brain Aβ elimination (55). Recently, a study measuring both Aβ production and clearance rates in AD and healthy controls has revealed that the production of Aβ is similar in AD and controls, whereas the clearance rate of Aβ is decreased by around 30% in AD compared with healthy patients, showing that Aβ clearance mechanisms are defective in AD brains (56,57). Our study identifies and characterizes two clinically used antihypertensive DHPs with a proven safety record in elderly patients, nilvadipine and nitrendipine, as dual modulators of APP processing and of the clearance of Aβ across the BBB. To our knowledge, these two DHPs represent the first small molecules identified so far that are able to stimulate the clearance of Aβ across the BBB. These data may explain why nilvadipine and nitrendipine have shown beneficial effects in AD patients (7,8,10,49). Since the efficacy of nilvadipine and nitrendipine toward Aβ accumulation is independent of their antihypertensive effect, as other antihypertensive DHPs are inefficient, we hypothesize that compounds derived from the chemical structure of Aβ-lowering DHPs could be optimized to negate their antihypertensive activity, and at the same time, enhance their Aβ-lowering properties, providing a new class of AD medications.
We thank Dr. Michael Wolf (Harvard Medical School, Boston, Massachusetts, USA) for providing the 7W CHO APP overexpressing cells. The authors are grateful to Mr. and Mrs. Roskamp for their generous support which helped make this work possible. The authors are thankful to Dr. Minoru Otsuka from Astella Pharmaceutical (Osaka, Japan) for providing nilvadipine, the oral formulation of nilvadipine and the placebo used in this study, as well as, for helpful discussions.
Several authors have financial interests in the commercialization of nilvadipine and related DHPs as clinical treatments for the prevention and mitigation of Alzheimer disease. These authors are D Paris, G Ait-Ghezala, F Crawford and MJ Mullan.
Online address: http://www.molmed.org