Mol Neurobiol DOI 10.1007/s12035-014-8819-1

Enhanced Amyloidogenic Processing of Amyloid Precursor Protein and Cell Death Under Prolonged Endoplasmic Reticulum Stress in Brain Endothelial Cells Ana I. Plácido & Catarina R. Oliveira & Paula I. Moreira & Cláudia Maria F. Pereira

Received: 25 February 2014 / Accepted: 15 July 2014 # Springer Science+Business Media New York 2014

Abstract Cerebral amyloid angiopathy resulting from the deposition of misfolded amyloid beta (Aβ) peptide in the walls of brain’s blood vessels is exhibited by the majority of Alzheimer’s disease (AD) patients, suggesting that alterations in protein quality control contribute to AD-associated vascular dysfunction. The present work addressed the role of ER stress in the amyloidogenic amyloid precursor protein (APP) processing and subsequent Aβ generation in brain endothelial cells (ECs). For that purpose, the RBE4 cell line was exposed to the classical ER stressors thapsigargin or brefeldin A to mimic the altered ER homeostasis observed in AD. In treated cells, an increase in the levels of markers of ER stress (XBP1 and GRP78) and of the ER stress-induced apoptotic pathway (caspase-12, JNK, and CHOP) was observed concomitantly with the accumulation of reactive oxygen species. Under these conditions, a significant ER-to-mitochondria Ca2+ transfer was also found, which culminated in mitochondrial Ca2+ overload and activation of mitochondria-dependent apoptosis. Moreover, it was showed that prolonged ER stress induces intracellular APP accumulation, which colocalizes with the ER chaperone GRP78, and activation of β-secretase, leading to increased intracellular Aβ levels, together with a decrease in secreted Aβ. Finally, it was demonstrated that ER stressinduced changes in Aβ levels and apoptotic cell death can be ameliorated by a blocker of the mitochondrial Bax channel. These observations suggest that chronic ER stress triggers APP accumulation in early comportments along the secretory pathway in brain ECs and increases its amyloidogenic processing and Aβ generation leading to apoptotic cell death.

A. I. Plácido : C. R. Oliveira : P. I. Moreira : C. M. F. Pereira (*) Center for Neuroscience and Cell Biology, University of Coimbra, Rua Larga; Faculty of Medicine, Pólo I, 3004-504 Coimbra, Portugal e-mail: [email protected]

Keywords Endoplasmic reticulum stress . Unfolded protein response . Cerebral amyloid angiopathy . Amyloid beta peptide

Introduction Alzheimer’s disease (AD) is a chronic and fatal brain disorder characterized by the presence of extracellular senile plaques composed by amyloid beta (Aβ) fibrils, intracellular neurofibrillary tangles of hyperphosphorylated tau and synaptic and neuronal loss [1]. Recent evidences support that cerebrovascular dysfunction may precede cognitive decline and onset of neurodegeneration [2]. Accordingly, despite being considered two distinct disorders, pure cases of vascular dementia (VaD) without neurodegeneration are rare [3]. Furthermore, epidemiological studies point out that AD and VaD share several risk factors, suggesting inter-related pathogenic mechanisms [4]. In the AD brain, it is found decreased number of microvessels, reduction of capillaries’ diameter, atrophy of smooth muscle cells, rupture of vessel’s wall, deposition of amorphous material and presence of Aβ in microvasculature, demonstrating an imbalance between Aβ production, clearance and degradation in brain endothelial cells (ECs) [5]. The deposition of Aβ in the walls of cerebral vessels in addition to its deposition in the brain parenchyma, termed cerebral amyloid angiopathy (CAA), is exhibited by more than 80 % of AD individuals supporting that alterations in the vascular system might play a pivotal role in cognitive impairment [6]. Aβ damages both neurons and ECs leading to synaptic dysfunction and neuronal loss, which underlie cognitive deficits, and to vascular impairment due to morphological and functional alterations in ECs [7–13]. In vitro studies showed that Aβ suppresses the proliferative ability of brain ECs and also triggers apoptosis [7, 14, 15]. Despite the exact mechanisms responsible for Aβ-induced endothelial

Mol Neurobiol

dysfunction remain unknown, recent work revealed that Aβ activates apoptotic cell death in a blood brain barrier (BBB) in vitro model through an endoplasmic reticulum (ER)-dependent pathway [7]. Aβ is the product of the sequential amyloidogenic cleavage of amyloid precursor protein (APP) by β- and γ-secretases that occurs predominantly in the intracellular space. APP processing can occur in any organelle along the secretory pathway, namely, ER/intermediate compartment, Golgi apparatus, early, late or recycling endosomes and lysosomes, as well as at the plasma membrane [16–22]. Shin et al. [23] suggested that α- and β-secretases compete with each other for APP cleavage in the ER and that this competition regulates Aβ production. Under stress, APP is retained in the ER by GPR78 and activates the retrograde transport, which, in turn, potentiates the accumulation of APP in the early compartments of the secretory pathway [12, 24]. One function of the ER, the first organelle in the secretory pathway, is protein folding. The presence of misfolded or unfolded proteins activates the unfolded protein response (UPR) to restore homeostasis within the ER. The relief of the master chaperone GPR78 from the ER stress sensors protein kinase RNA-like ER kinase (PERK), activating transcription factor 6 (ATF6) and inositol-requiring factor 1 (IRE1) activates downstream signalling pathways. However, under sustained ER stress conditions, the activation of these sensors triggers apoptosis to eliminate stressed cells, which can occur through: (1) activation of the ER-resident caspase-12 that, in turn, activates the apoptosis’ executioner caspase-3; (2) upregulation of the transcription factor C/EBP homologous protein (CHOP) that downregulates the anti-apoptotic Bcl-2 and upregulates Ero-1 that promotes the production of reactive oxygen species (ROS); (3) phosphorylation of JNK [25]. The transfer of Ca2+ from the ER to mitochondria can be a potent apoptotic stimulus leading to mitochondrial Ca2+ overload, translocation of cytochrome c from mitochondria to cytosol and activation of a caspase cascade [26]. Changed levels of ER stress markers such as the ER chaperone GPR78 in the brain of AD patients suggest that UPR induction by prolonged ER stress can initiate the neurodegenerative process and/or that the deposition of amyloidogenic peptides such as Aβ is a consequence of ER dysfunction [27]. Takahashi et al. [28] suggested that APP is a mediator of ER stress-induced apoptosis. According with the author, ER stress increases mRNA expression and cleavage of APP resulting in AICD (APP intracellular domain) fragment release, which activates the transcription of CHOP and induces cell death. On the other hand, Kogel et al. [29] suggested that wild-type APP (APPwt) protects neurons from ER stress-mediated apoptosis, and this protective effect is abolished by APP mutations associated with familial AD. The present work was aimed to investigate the role of ER stress in the APP amyloidogenic processing and subsequent

accumulation of Aβ in brain ECs in order to clarify the molecular mechanisms involved in endothelial dysfunction and vascular deficits in AD. It was found that prolonged ER stress promotes APP accumulation, including in the ER, and activates β-secretase 1 (BACE1), leading to an increase in intracellular Aβ1-40 and Aβ1-42 levels. Taken together with our previous findings demonstrating Aβ-induced ER stressmediated apoptosis in brain ECs [7] these observations suggest that chronic ER stress plays a deleterious effect on brain ECs since it causes APP accumulation in the early compartments of the secretory pathway and increases its processing through the amyloidogenic pathway leading to Aβ generation, which, in turn, induces ER stress, promoting a deadly vicious cycle that culminates in endothelial dysfunction. Furthermore, these findings suggest that ER could be a promising target to prevent Aβ accumulation and toxicity in the cerebrovasculature and downstream cognitive impairment.

Material and Methods Fetal bovine serum (FBS), geneticin (G480), HAM’s F-10 and MEM-alpha medium with Glutamax-1 (α-MEM) were purchased from Gibco-Invitrogen (Grand Island, NY, USA). Fura2-acetoxymethyl ester (Fura2-AM) and 2′-7′dichlorodihydrofluorescein diacetate (H2DCF-DA) were purchased from Molecular Probes-Invitrogen (Grand Island, NY, USA). Aβ1-40 and Aβ1-42 enzyme-linked immunosorbent assay (ELISA) kits, SuperscriptTM III first-strand synthesis system, forward and reverse sets of RT-PCR primers, and Taq polymerase were purchased from Invitrogen (Grand Island, NY, USA); anti-β-actin antibody, basic Fibroblast Growth Factor (bFGF), anti-α-tubulin antibody, anti-APP antibody, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT), thapsigargin, and brefeldin A were obtained from Sigma (St. Louis, MO, USA). Rneasy Mini Kit® was purchased from Qiagen (Germantown, MD, USA) and PstI was obtained from New England Biolabs (Hitchin, UK). Rattail collagen was purchased from Roche Diagnostics (Mannheim, Germany). Anti-TOM 20 and anti-CHOP antibodies were obtained from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA). Anti-GRP78 and anti-cytocrome c antibodies, N-acetyl-Asp-Glu-val-Asp-P-milnoanilid (Ac-DEVDpNA) Ac-Leu-His-Asp-P-milnoanilid (Ac-LEHD-pNA) and Bax channel blocker [(±)-1-(3,6-dibromocarbazol-9-yl)-3piperazin-1-yl-propan-2-ol, bis TFA] were obtained from Calbiochem, Merck KGaA (Darmstadt, Germany). Anticaspase-12, anti-Bax, anti-JNK, and anti-phospho-JNK were obtained from Cell Signaling Technology (Danvers, MA, USA), anti-caspase-9 was obtained from Abcam (Cambridge, USA), and anti-BACE was obtained from Cell Signaling Technology. All the other chemicals were of the highest grade of purity that is commercially available.

Mol Neurobiol

Cell Culture and Treatments The rat brain endothelial cell line (/RBE4) was provided by Dr. Jon Holy (University of Minnesota, Duluth, MN, USA). RBE4 is a continuous, immortalized cell line that retains a stable phenotype reminiscent of BBB endothelium in vitro. RBE4 cells were maintained as monolayer cultures in collagen-coated T-75 flasks in a humidified atmosphere of 5 % CO2/95 % air at 37 °C, in an α-MEM:Ham’s F-10 nutrient mixture (1:1), supplemented with streptomycin/ penicillin (100 U/ml), 10 % (vol/vol) FBS, 1 ng/ml bFGF, and 300 μg/ml geneticin [30]. For plating, the number of viable cells in suspension was quantified by counting trypan blue, excluding cells in a hemocytometer chamber. Cells were used for experiments 1 day after plating. RBE4 cells (80 % confluency) were treated for 3–24 h with thapsigargin (2 μM) or brefeldin A (2 μM), which were added from a 5 mM stock, prepared in DMSO or ethanol, respectively. When tested, the Bax channel blocker (50 nM) was pre-incubated during 30 min before exposure to thapsigargin or brefeldin A.

Yvon Inc., Edison, NJ, USA). Results were expressed relatively to baseline and were normalized to control values. The accumulation of ROS was also evaluated using the Amplex™ Red-horseradish peroxidase assay kit, as previously described [33]. Once in the cell, the horseradish peroxidase catalyzes H2O2-dependent oxidation of non-fluorescent Amplex™ Red into fluorescent Resorufin Red. Briefly, control and treated RBE4 cells were loaded with 100 μM Amplex™ Red reagent and 0.2 U/ml peroxidase in Krebs buffer for 30 min at 37 °C, protected from light. After that, the fluorescence signal, corresponding to H2O2 generation, was monitored for 30 min at 530/560 nm excitation and 590 nm emission, using a temperature-controlled SPEX 1681 Fluorolog spectofluorometer (HORIBA, Jobin Yvon Inc.). Results were expressed relative to baseline and were normalized to control values. Measurement of Caspase-3- and Caspase-9-Like Activities

Cell viability was analyzed using the MTT assay, which measures the ability of metabolic active cells to form formazan through cleavage of the tetrazolium ring of MTT [31]. Control and treated cells were washed with Krebs buffer [(in mM): 132 NaCl, 4 KCl, 1.2 NaH2PO4, 1.4 MgCl2, 6 glucose, 10 HEPES, and 1 CaCl2 (pH 7.4)] and incubated with MTT (0.5 mg/ml) for 3 h at 37 °C. The blue formazan crystals formed were dissolved in an equal volume of 0.04 M HCl in isopropanol and quantified spectrophotometrically by measuring the absorbance at 570 nm using a microplater reader (SpectraMax Plus 384, Molecular Devices, San Francisco, CA, USA). Results were expressed as the percentage of the absorbance determined in control cells.

Activation of caspase-3 and caspase-9 was monitored using a colorimetric method described by Cregan and collaborators [34]. Control and treated cells were washed twice with phosphate buffered saline (PBS) and were then lysed in cold lysis buffer. Cells were harvested by scraping and frozen/defrozen three times. The lysates were centrifuged for 10 min at 20,800×g (2-16 K Sigma-Aldrich Co., St. Louis, MO, USA) at 4 °C. The supernatant was collected and protein concentration was measured using the Bio-Rad protein dye assay reagent. Cell extracts (50 μg of protein) were incubated at 37 °C for 2 h in 25 mM HEPES, pH 7.5 containing 0.1 % (wt/vol) CHAPS, 10 % (wt/vol) sucrose, 2 mM DTT, and 40 μM DEVD-pNA (caspase-3 substrate) or 40 μM Ac-LEHDpNA (caspase-9 substrate). Caspase-like activities were determined by measuring substrate cleavage at 405 nm in a microplate reader (SpectraMax Plus 384, Molecular Devices). The results were expressed relatively to the control values.

Measurement of Intracellular Reactive Oxygen Species (ROS)

ELISA-Based Measurement of Intracellular and Secreted Aβ1-40 and Aβ1-42

The accumulation of reactive oxygen species (ROS) was evaluated using DCFH2-DA, a non-fluorescent compound that crosses the plasma membrane. Once in the cell, esterases hydrolyze its acetyl moieties to produce DCFH2 that is susceptible to oxidation, generating the fluorescent product DCF that can be monitored [32]. Control and treated RBE4 cells were loaded with 5 μM DCFH2-DA in Krebs buffer for 30 min at 37 °C. After that, cells were washed with Krebs buffer and the fluorescence signal, corresponding to intracellular ROS, was monitored for 1 h at 502 nm excitation and 550 emission wavelengths, using a temperature-controlled SPEX 1681 Fluorolog spectrofluorometer (HORIBA, Jobin

Aβ1-40 and Aβ1-42 levels were determined in total cell extracts (intracellular Aβ), which were prepared as described above, and in the extracellular medium (secreted Aβ). Briefly, conditioned medium from control and treated RBE4 cells was centrifuged at 20,800×g (2-16 K Sigma-Aldrich) during 10 min at 4 °C and supernatant was collected. Aβ1-40 and Aβ1-42 levels were quantified in the supernatant and in cell lysates with mouse ELISA kits according to the manufacturer’s instructions. The fluorescence intensity corresponding to intracellular Aβ was normalized to micrograms of protein used in the assay, and results were then normalized to control values.

Assessment of Cell Viability

Mol Neurobiol

Cell Death Analysis Using Hoechst 33342/Propidium Iodide Staining

Reverse Transcription (RT)-PCR

The number of viable, apoptotic, and necrotic cells was analyzed by nuclear morphology as previously described [7]. Briefly, untreated and treated cells were washed twice with PBS at 37 °C, incubated with 15 μg/ml Hoechst 33342 and 3 μg/ml propidium iodide (PI) in PBS for 5 min in the dark and washed again twice with PBS. Images were acquired in a photo-activated localization microscope (Zeiss axioskop2, Zeiss, Jena, Germany) with a LD-PlanNeofluar objective (20×, 0.4 korr NA). Viable cells display a normal nuclear size without PI staining. Cells with normal nuclear size and PI staining were scored as necrotic cells. Scored apoptotic cells included cells that displayed pyknotic nuclei with condensed or fragmented chromatin, with or without PI staining (primary and secondary apoptotic cells). All experiments were performed twice in duplicate, and a minimum of 150 cells was scored for each well. The number of viable, apoptotic or necrotic cells was expressed as the percentage (%) of the total number of cells in all microscope fields analysed for each well.

Splicing in XBP-1 mRNA was investigated by RT-PCR as previously described [36]. Briefly, total RNA from control, thapsigargin- and brefeldin A-treated RBE4 cells was extracted with Rneasy Mini Kit®. Total RNA yields were measured by ND-1000 Nanodrop Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and the purity was evaluated by measuring the ratio of optic density at 260 and 280 nm. RNA (2 μg) was reverse-transcribed into cDNA with the SuperscriptTM III first-strand synthesis system and then a PCR was done in an iQ5 thermocycler (Bio-Rad, Hercules CA, USA) using Taq polymerase (0.05 U/μl). Target genes for amplification, their primer sets and thermal conditions are listed in Table 1. The PCR cycles consisted of denaturation at 94 °C for 1 min, annealing at 62 °C for 1 min and extension at 72 °C for 1 min, for 33 cycles. PCR products were digested with the PstI restriction enzyme (0.25 U/μl) during 16 h at 37 °C, then separated on a 1 % (wt/vol) agarose gel with ethidium bromide and visualized under ultraviolet (UV) illumination. The ratio between XBP1s and GAPDH mRNA was calculated and normalized to control values.

Colocalization of GRP78 with APP by Immunocytochemistry

Western Blot Analysis

The colocalization of GRP78 with APP was analyzed by immunocytochemistry. Briefly, untreated and treated cells were washed twice in PBS and then fixed with 4 % (w/v) paraformaldehyde at 37 °C for 30 min. After washed twice with PBS, cells were permeabilized with 0.2 % (vol/vol) Triton X-100 for 2 min and incubated in blocking solution containing 3 % (wt/vol) bovine serum albumin (BSA) in PBS to prevent nonspecific binding and then incubated during 2 h at 37 °C with anti-GRP78 (1:100) and anti-APP (1:250) antibodies. Thereafter, cells were washed in PBS and incubated for 1 h at room temperature (RT) with the secondary antibody anti-rabbit IgG labelled with Alexa Fluor 488 (1:200) diluted in PBS with 1 % (wt/vol) BSA or anti-mouse IgG labelled with Alexa Fluor 594 (1:200). Following additional rinses in PBS, cells were stained in the dark for 5 min, at RT, with 15 μg/ml Hoechst 33342 prepared in PBS and mounted using Glycergel mounting medium. Images were collected using a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss, New York, USA) with a plan-apochromat objective (63×, 1.4 NA) by using the v3.2 software. All experiments were performed twice in triplicate and a minimum of 30 cells was scored for each coverslip. Colocalization between APP and GRP78 was quantified in tresholded images with the JACoP plug-in of the ImageJ software (Wayne Rasband, Research Services Branch, National Institute of Mental Health, Bethesda, MA, USA), according to Bolte and Cordelières [35]. Data were expressed as percentage of colocalization between both proteins.

For preparation of whole cell extracts, RBE4 cells were washed twice with PBS, scraped and resuspended in icecold lysis buffer (25 mM HEPES-Na, 2 mM MgCl2, 1 mM EDTA, and 1 mM EGTA) supplemented with 0.1 M PMSF, 0.2 M DTT, and 1:1,000 of a protease inhibitor cocktail (containing chymostatin, pepstatin A, leupeptin, and antipain, 1 μg/ml). Extracts were then frozen and thawed three times to favor disruption. The amount of protein content in each sample was quantified using the Bio-Rad protein dye assay reagent. For the isolation of cytosolic and mitochondrial fractions, control and treated RBE4 cells were harvested and subcellular fractions were prepared using the ProteoExtract Subcellular Proteome extraction kit. Protein content was measured using the Pierce’s BCA Protein Assay kit. Samples were resolved by electrophoresis in 8–12 % sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred to polyvinylidene fluoride (PVDF) membranes. Nonspecific binding was blocked by gentle agitation in 5 % (wt/vol) BSA and 0.1 % (vol/vol) Tween in tris-buffered saline (TBS) for 1 h at room temperature (RT). Membranes were subsequently incubated overnight at 4 °C with gentle agitation with a mouse monoclonal antibody (mAB) anti-GRP78 (1:750), rabbit pAB anticaspase-12 (1:1,000), mouse mAB anti-cytochrome c (1:500), rabbit mAB anti-APP (1:500), rabbit mAB anti-BACE (1:500), rabbit pAB anti-XBP1 (1:500), rabbit mAB antiJNK (1:1,000), rabbit pAB anti-Bax (1:1,000), mouse mAB anti-phospho JNK (1:1,000), mouse mAB anti-CHOP (1:300), rabbit mAB anti-caspase-9 (1:1,000), rabbit mAB

Mol Neurobiol Table 1 Primer sequences and reaction conditions for RT-PCR

Target genes

Sequence

Product size (bp)

Annealing temperature (°C)

No. of cycles

GAPDH

F: 5_-GCAAGTTCAACGGCACAG-3_ R: 5_-GCCGTAGACTCCACGACAT-3_ F: 5_-AAACAGAGTAGCAGCGCAGA CTGC-3_ R: 5_-GGATCTCTAAAACTAGAGGC TTGGTG-3_

140

62

33

XBP-1

anti-TOM-20 (1:500), rabbit mAB anti-β-actin (1:5,000), rabbit mAB anti-tubulin (1:1,000), and mouse mAB antiGAPDH (1:1,000), diluted in 5 % of BSA. Then, membranes were washed three times, during 15 min, with Tris buffer containing 0.1 % Tween (TBS-T) and were incubated with secondary antibodies [anti-rabbit (1:20,000) and anti-mouse (1:20,000)] for 2 h at RT with gentle agitation. After three washes with TBS-T, specific bands of immunoreactive proteins were visualized after membrane incubation with ECF for 5 min in a VersaDoc Imaging System 3000 (Bio-Rad), and the density of protein bands was calculated using Quantity One 1D Analysis Software Version 4.6.5 (Bio-Rad). The ratios between pro-caspase-12, caspase-12 and actin, the ratio between p-JNK, JNK and actin, the ratios between GRP78, BACE, CHOP, XBP1, cytochrome c in cytosol, Bax in cytosol or APP, and actin, the ratio between Bax in mitochondria or cytochrome c in cytosol and TOM-20 were calculated and normalized to control values. According with the manufacturer’s data sheet information, cytochrome c might be present in the cytoplasm in polymeric forms appearing as a 58-60 kDa band, rather than in its monomeric form, which migrates at a lower molecular weight of 15 kDa. Measurement of Calcium Levels Reticular Ca2+ levels were measured using a spectrofluorometric method described by Nutt and colleagues [37] with some modifications. Briefly, control and treated RBE4 cells were washed twice in Krebs and loaded with 10 μM Fura2AM supplemented with 0.2 (vol/vol) pluronic F-127 and 1 % (wt/vol) BSA in Krebs buffer for 45 min at 37 °C. Thereafter, cells were washed in Ca2+-free Krebs buffer and incubated for 30 min in medium free of Ca2+ and of dye to allow the hydrolysis of the acetoxymethylester. After fluorescence baseline stabilization, cells were stimulated with thapsigargin (5 μM) in the absence of extracellular Ca2+, to empty the ER store. Fura2 fluorescence was recorded at 340/380 nm excitation and 512 nm emission. The peak amplitude of Fura2 fluorescence (ratio 340/380) was used to evaluate ER Ca2+ levels. To monitor mitochondrial Ca2+ content, cells were incubated with 10 μM of the fluorescent membrane permeable probe Rhod2-AM (552 nm excitation and 581 nm emission). To assure a selective accumulation of Rhod2 into

600

mitochondria, probe loading was performed at low temperature followed by incubation at 37 °C. Mitochondrial maximal Ca2+ uptake was assessed by challenging mitochondria with the subsequent addition of Ca2+ ionophore A23187 (5 μM) as described previously [38]. Results were normalized to control values. Determination of BACE Activity BACE activity was determined in total cellular extracts from control and treated cells, which were prepared as described above, using a β-secretase activity assay kit, according to the manufacturer’s instructions. The fluorescence intensity was normalized to micrograms of protein used in the assay. Data Analysis Data are means±SEM of at least three independent experiments performed in duplicate. Statistical significance was obtained using the unpaired one-tailed t test or the one-way analysis of variance (ANOVA) test followed by Dunnett’s post hoc test in the GraphPad prism software (San Diego, CA, USA). The differences were considered significant for p values

Enhanced amyloidogenic processing of amyloid precursor protein and cell death under prolonged endoplasmic reticulum stress in brain endothelial cells.

Cerebral amyloid angiopathy resulting from the deposition of misfolded amyloid beta (Aβ) peptide in the walls of brain's blood vessels is exhibited by...
3MB Sizes 0 Downloads 7 Views