The FASEB Journal • Research Communication
Lipocalin 2: Novel component of proinflammatory signaling in Alzheimer’s disease Petrus J. W. Naudé,*,† Csaba Nyakas,㥋,¶ Lee E. Eiden,# Djida Ait-Ali,# Ragna van der Heide,* Sebastiaan Engelborghs,**,††,‡‡,§§,㥋㥋 Paul G. M. Luiten,*,† Peter P. De Deyn,‡,§,**,††,‡‡,§§,㥋㥋 Johan A. den Boer,† and Ulrich L. M. Eisel*,1 *Department of Molecular Neurobiology, University of Groningen, Groningen, The Netherlands; † Department of Biological Psychiatry, ‡Department of Neurology, and §Alzheimer Research Center, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands; 㛳 Neuropsychopharmacological Research Group, Semmelweis University, Budapest, Hungary; ¶ Hungarian Academy of Sciences, Budapest, Hungary; #Section on Molecular Neuroscience, Laboratory of Cellular and Molecular Regulation, National Institute of Mental Health, Bethesda, Maryland, USA; **Department of Neurology and ††Memory Clinic, Ziekenhuis Netwerk Antwerpen, Antwerp, Belgium; ‡‡Laboratory of Neurochemistry and Behavior, §§Reference Center for Biological Markers of Dementia, and 㛳㛳Biobank Antwerp, Institute Born-Bunge, University of Antwerp, Antwerp, Belgium Alzheimer’s disease (AD) is associated with an altered immune response, resulting in chronic increased inflammatory cytokine production with a prominent role of TNF-␣. TNF-␣ signals are mediated by two receptors: TNF receptor 1 (TNFR1) and TNF receptor 2 (TNFR2). Signaling through TNFR2 is associated with neuroprotection, whereas signaling through TNFR1 is generally proinflammatory and proapoptotic. Here, we have identified a TNF-␣-induced proinflammatory agent, lipocalin 2 (Lcn2) via gene array in murine primary cortical neurons. Further investigation showed that Lcn2 protein production and secretion were activated solely upon TNFR1 stimulation when primary murine neurons, astrocytes, and microglia were treated with TNFR1 and TNFR2 agonistic antibodies. Lcn2 was found to be significantly decreased in CSF of human patients with mild cognitive impairment and AD and increased in brain regions associated with AD pathology in human postmortem brain tissue. Mechanistic studies in cultures of primary cortical neurons showed that Lcn2 sensitizes nerve cells to -amyloid toxicity. Moreover, Lcn2 silences a TNFR2mediated protective neuronal signaling cascade in neurons, pivotal for TNF-␣-mediated neuroprotection. The present study introduces Lcn2 as a molecular actor in neuroinflammation in early clinical stages of AD.—
ABSTRACT
Abbreviations: A, -amyloid; AD, Alzheimer’s disease; CSF, cerebrospinal fluid; DMEM, Dulbecco’s modified Eagle medium; FBS, fetal bovine serum; FCS, fetal calf serum; HBSS, Hanks’ balanced salt solution; HRP, horseradish peroxidase; IPA, Ingenuity Pathway Analysis; Lcn2, lipocalin-2; MCI, mild cognitive impairment; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; NGAL, neutrophil gelatinase-associated lipocalin; PBS, phosphate-buffered saline; PTEN, phosphatase and tensin homolog deleted on chromosome 10, TNFR1, TNF receptor 1; TNFR2, TNF receptor 2; VMPC, ventromedial prefrontal cortex 0892-6638/12/0026-2811 © FASEB
Naudé, P. J. W., Nyakas, C., Eiden, L. E., Ait-Ali, D., van der Heide, R., Engelborghs, S., Luiten, P. G. M., De Deyn, P. P., den Boer, J. A., Eisel, U. L. M. Lipocalin 2: Novel component of proinflammatory signaling in Alzheimer’s disease. FASEB J. 26, 2811–2823 (2012). www. fasebj.org Key Words: mild cognitive impairment 䡠 TNF-␣ 䡠 neurodegeneration The main pathological features in the brain affected by Alzheimer’s disease (AD) are depositions of -amyloid (A) in the extracellular space and neurofibrillary tangles within the nerve cells. In particular, the neurofibrillary tangles accumulate in a predictable spatial pattern and correlate well with disease progression (1). Neuritic or amyloid plaques are rare in persons undergoing normal aging but begin to grow in number in the hippocampus and neocortex during mild cognitive impairment (MCI) and can subsequently develop into the characteristic pathology of AD. A plaques become more prevalent and spread to the parietal and frontal neocortical areas of the brain with the progression to AD (2). In the search for the biological etiology of AD, numerous studies during the past 2 decades revealed that an activated innate neuroimmune response is an integral component of the pathophysiology of this disease (3). This activated immune response leads to chronic production of inflammatory cytokines in the vicinity of AD-associated pathology in brain tissue 1 Correspondence: Department of Molecular Neurobiology, University of Groningen, Nijenborgh 7, 9747 AG Groningen, The Netherlands. E-mail:
[email protected]. doi: 10.1096/fj.11-202457 This article includes supplemental data. Please visit http:// www.fasebj.org to obtain this information.
2811
(4), which can also be detected in the cerebrospinal fluid (CSF) and in peripheral blood (5). A prominent role in the immune signaling has been assigned to TNF-␣ as a key proinflammatory cytokine that was found to be highly up-regulated in brain tissue associated with AD pathology (6, 7). It is unclear whether TNF-␣ signaling actively contributes to or limits neuronal injury in AD. TNF-␣ exerts balancing or offsetting biological functions in the nervous system via its receptors, TNF receptor 1 (TNFR1) and TNF receptor 2 (TNFR2) (8), especially with respect to cell survival in neuronal tissue exposed to traumatic insults. TNFR1-induced signaling initiates production of proinflammatory factors and can induce a proapoptotic environment in neuronal tissue, whereas TNFR2 stimulation promotes neuronal survival (9). More recently, in vivo studies using transgenic AD mouse models have shown that TNF-␣ functions as an important neuroprotective mediator and probably plays an important role in mitigating the pathological changes of AD (10, 11). Because of the complex cellular mechanisms of TNF, initiated by its two receptors, the ultimate role of TNF signaling in neurodegenerative disease may depend on production of different factors in the cellular environment and may be different at different stages of neurodegenerative disease. It was shown that the absence of TNFR1 in an AD mouse model caused reduced A plaque formation and prevented learning and memory defects (12). Targeting specific TNFR-mediated signaling is increasingly being viewed as important in the consideration of novel inflammation-modifying therapies for neurodegenerative diseases including AD (13). During an inflammatory response, TNF-␣ may trigger the expression and release of various inflammatory mediators in the brain, including some not yet specifically associated with brain inflammation; therefore, we used a gene microarray to identify all transcripts upregulated in rodent cortical neurons exposed to TNF-␣. One such transcript is that encoding the protein lipocalin-2 (Lcn2), also known as neutrophil gelatinase-associated lipocalin (NGAL), siderocalin, 24p3, or uterocalin. Lcn2 was discovered nearly 2 decades ago (14), and its expression in vivo has previously been associated with the acute-phase response (15). Of interest, Lcn2 mRNA is strongly up-regulated in cortical tissue from AD model mice, in which presenelin-1 was overexpressed but not in mice overexpressing presenilin-1 and amyloid precursor protein. This finding suggested that Lcn2 might be involved in an early response to inflammatory stimuli in these mice (16). However, the functions of Lcn2 in brain and brain disorders remain largely unknown. A study has shown that Lcn2 can decrease hippocampus dendritic spine numbers in mice subjected to stress (17). So far, 2 receptors have been identified for Lcn2: megalin (18) and 24p3R (19). In this article, the mechanism of TNF-␣-induced Lcn2 expression in primary murine cortical neurons, astrocytes, and microglial cells was investigated after specific stimulation of TNF-␣ and its receptors TNFR1 and TNFR2. Moreover, we investigated Lcn2 protein 2812
Vol. 26
July 2012
concentrations in serum and CSF from patients with MCI and AD and in AD human postmortem brain in comparison with control subjects without dementia. The mechanistic role of Lcn2 in the pathology of AD was further investigated in murine primary cortical neuronal cultures.
MATERIALS AND METHODS Materials Neurobasal medium, phosphate-buffered saline (PBS), Dulbecco’s modified Eagle medium (DMEM), Hanks’ balanced salt solution (HBSS), fetal bovine serum (FBS), fetal calf serum (FCS), B27 supplement, l-glutamine, and penicillin/ streptomycin were purchased from Invitrogen (Breda, The Netherlands). Complete mini protease inhibitor cocktail tablets were from Roche (Mannheim, Germany). Murine Lcn2 was purchased from R&D Systems (Minneapolis, MN, USA). VO-OHpic was from Biovision (Mountain View, CA, USA). A peptide1– 42 was purchased from EZBiolab (Carmel, IN, USA). Agonistic antibodies and murine TNFR1 agonist were from R&D Systems, and murine TNFR2 agonistic antibodies were from Hycult Biotechnology (Uden, The Netherlands). Primary antibodies for Western blot were anti-human lipocalin 2/NGAL (R&D Systems), 1:1000 dilution; anti-mouse Lcn2 (Abcam, Cambridge, UK), 1:1000 dilution; antibody for slc22a17 (LifeSpan BioSciences, Seattle, WA, USA), 1:2000 dilution; megalin (Santa Cruz Biotechnology, Santa Cruz, CA, USA), 1:1000 dilution; antibody specific for total PKB/Akt (Cell Signaling Technology, Beverly, MA, USA), 1:2000 dilution; anti p-Akt (Ser473; Cell Signaling Technology), 1:1000 dilution, anti-caspase 3 (Cell Signaling Technology), 1:2000 dilution; cleaved caspase 3 (Cell Signaling Technology), 1:1000 dilution; anti-BIM (Cell Signaling Technology), 1:500 dilution; anti-p65 (Ser536, Cell Signaling Technology), 1:2000 dilution; and monoclonal mouse antibody specific for actin as internal standard (MP Biomedicals, Irvine CA, USA), 1:2,000,000 dilution. The primary antibody for immunohistochemistry was rat monoclonal anti-human lipocalin 2/NGAL (R&D Systems). The secondary antibodies used were horseradish peroxidase (HPR)-conjugated donkey anti-rabbit (GE Healthcare, Chalfont St. Giles Buckinghamshire, UK), 1:5000 dilution; HRP-conjugated goat anti-mouse (Santa Cruz Biotechnology), 1:5000 dilution, HRP-conjugated goat anti-rat (Jackson ImmunoResearch, Soham, UK), 1:5000 dilution; and biotinylated donkey anti-rat (Jackson ImmunoResearch), 1:2000 dilution. All other materials were purchased from Sigma-Aldrich (Zwijndrecht, The Netherlands). Primary cortical neuron culture Primary cortical neurons were prepared from embryonic brains (embryonic day 14) of C57BL/6J mice. The cortices were carefully dissected, meninges were removed, and the neurons were separated by trituration. A single cell preparation was obtained by filtration though a gauge mesh filter (pore size 40 m). Cells were plated on poly-d-lysine-precoated plates at a density of 1 ⫻ 105 cells/well for 96-well plates and 2 ⫻ 106 cells/well for 6-well plates. Neurobasal medium supplemented with 2% (v/v) B27 supplement, 0.5 mM glutamine, and 1% (v/v) penicillin/streptomycin was used as a culture medium. After 48 h, neurons were treated with 10 M cytosine -d-arabinofuranoside hydrochloride for another 48 h to inhibit non-neuronal cell growth. Subsequently, the medium was completely exchanged. The neu-
The FASEB Journal 䡠 www.fasebj.org
NAUDÉ ET AL.
rons were used 7 d after preparation for experiments. Animal experimental procedures were approved according to the Dutch law by the local ethics committee (DEC 6103). Glial cell culture Cells were prepared from 1- to 3-d-old C57BL/6 mice cortices. In brief, cortices were removed under sterile conditions, and hippocampi were discarded. The cortices were cut with a scalpel blade into small pieces (⬍1 mm) and incubated in 5 ml of HBSS supplemented with 0.6% (w/v) glucose, 15 mM HEPES, and 1% (v/v) penicillin/streptomycin with 0.25% (v/v) trypsin and 1000 U/ml DNase for 20 min at 37°C. After a 20-min incubation, trypsin inhibitor was added (0.1 mg/ ml), and the tissue was triturated 10 times with a sterile Pasteur pipette. Supplemented DMEM (20 ml) containing 10% (v/v) FCS and 1% (v/v) penicillin/streptomycin was added, and the mixture was centrifuged at 200 g for 10 min. The supernatant was discarded, and the pellet was resuspended with a pipette in supplemented DMEM. The cell suspension was seeded into 75-cm2 culturing flasks (1.5 brain/flask) and incubated at 37°C in 5% CO2. The cells were grown to confluence (after ⬃1 wk). After confluence was reached, microglia cells were obtained by shaking the flasks for 2 h at 37°C on an orbital shaker at 180 rpm. Microglia cells were seeded in 6-well plates (1 ⫻ 105 cells/well) and used 24 – 48 h afterward for experimentation. The remaining astrocytes were incubated at 37°C in 5% CO2 with 5 mM l-leucine methyl ester in supplemented DMEM for 1 h to limit microglia survival. The astrocytes were subsequently trypsinized and seeded (1 ⫻ 105 cells/well) in a 96-well plate. All cells were used for experimentation 24 – 48 h after plating. Treatment of cells Primary cortical neurons were treated with TNF-␣ (100 ng/ ml) for 24 h, and mRNA was extracted for gene array. For the assessment of the kinetics of TNF-␣-induced Lcn2 protein expression, primary cortical neurons, astrocytes, and microglia cells were treated with TNF-␣ (100 ng/ml) for the times indicated. For TNFR-mediated Lcn2 production, primary cortical neurons, astrocytes, and microglia cells were treated with either TNF-␣ (100 ng/ml), TNFR1 agonistic antibody (2 M), or TNFR2 agonistic antibody (20 M) for 36 h. For the assessment of Lcn2 on TNFR-mediated signaling, primary cortical neurons were treated with either TNF-␣ (100 ng/ml), TNFR1 agonistic antibody (2 M), TNFR2 agonistic antibody (20 M), and Lcn2 (200 ng/ml) or TNF-␣ (100 ng/ml), TNFR1 agonistic antibody (2 M), TNFR2 agonistic antibody (20 M), and Lcn2 (200 ng/ml) for 24 h. A glutamateinduced excitotoxicity assay was performed whereby primary cortical neurons were incubated with 100 ng/ml TNF-␣, TNFR2 agonistic antibody (20 M), or 200 ng/ml Lcn2 or TNF-␣ (100 ng/ml) was coincubated with Lcn2 (200 ng/ml), or TNFR2 agonistic antibody (20 M) was coincubated with Lcn2 (200 ng/ml) for 24 h. After a 24-h incubation, 50 M (concentration that induced ⬇50% cell death) glutamate was added to the cells for 1 h; thereafter, medium was replaced, and neuronal cells were allowed to recover for another 24 h before determination of cell viability. For A-induced cellular toxicity, primary cortical neurons were incubated with 100 ng/ml TNF-␣ or 200 ng/ml Lcn2 or TNF-␣ (100 ng/ml) was coincubated with Lcn2 (200 ng/ml) for 24 h, A1– 42 (25 M) was added after 24 h, and a 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyl tetrazolium bromide (MTT) assay was performed 24 h after A1– 42 treatment. LIPOCALIN 2 IN ALZHEIMER’S DISEASE
Oligo microarray, hybridization, and data analysis Mouse cortical neurons were treated with TNF-␣ (100 ng/ml) for 24 h, and RNA was extracted using an RNeasy extraction kit from Qiagen (Valencia, CA, USA). Linear amplification of RNA was performed (Amino Allyl MessageAmp II aRNA Amplification Kit; Ambion, Austin, TX) before labeling with Cy3 and Cy5 fluorescent dyes (GE Healthcare) and microarray hybridization to oligonucleotide arrays consisting of the Qiagen mouse MV3 oligo set with oligonucleotide length generally ⬃70 nucleotides. General methods for data extraction, normalization, and filtration for quality and dye bias have been summarized in detail previously (21). Here data from three microarrays were analyzed in the National Cancer Institute mAdb microarray databasing system (http://madb. nci.nih.gov/) after standard Loess correction with an allowed quality ratio of 0.3. Significance of transcript up- or downregulation by treatment with TNF-␣ was evaluated using statistical analysis of microarray at a false discovery rate of 0.1. Data were filtered for average expression ratios ⱖ2, and the resulting gene list was exported to MS Excel for further manipulation. With use of the Ingenuity Pathway Analysis tool (IPA; http://www.ingenuity.com), the transcripts up-regulated 2-fold or more by TNF-␣ in mouse cortical neurons were clustered based on their functions. Genes were also clustered based on their distribution in significant canonical pathways. Both indirect and direct networks were built using the default settings to identify interacting gene products. RNA extraction and quantitative real-time PCR Approximately 2 g of total RNA was submitted to DNase I (Invitrogen, Carlsbad, CA, USA) digestion and reverse-transcribed using random hexamers pdN6 and SuperScript II RNase H⫺ reverse transcriptase (Invitrogen). Splign (http:// www.ncbi.nlm.nih.gov/sutils/splign/splign.cgi) was used to identify the exon and intron boundaries. Gene-specific forward and reverse primers were chosen using Primer 3 input (http:// frodo.wi.mit.edu/primer3/), in each case to amplify across an exon-exon junction to avoid generation of same-sized amplicons from contaminating genomic DNA as follows: 5=-TGGAAGAACCAAGGAGCTGT-3= and 5=-GGTGGGGACAGAGAAGATGA-3= for Lcn2 and 5=-GGCTGTATTCCCCTCCATCG-3= and 5=CCAGTTGGTAACAATGCCATGT-3= for actin. Quantitative realtime PCR was performed in a premade reaction mix in the presence of the transcribed cDNA and 10 nM concentrations of each specific primer, using SYBR Green chemistry and an iCycler real-time detection system (Bio-Rad Laboratories, Hercules, CA, USA). Fold changes in mRNA levels were determined by normalizing against a nonvariable control transcript, actin, using the ⌬⌬Ct method (20). Preparation of A oligomers for cell culture Oligomeric A1–24 was prepared as described by Granic et al. (21). Human samples Serum, CSF, and brain tissue were obtained from The Institute Born-Bunge Biobank (University of Antwerp, Antwerp, Belgium). Postmortem examination was performed by a neuropathologist. Plasma and CSF samples were from 26 age-matched control subjects without dementia, 28 patients with MCI, and 26 patients with AD. Human brain samples were from the hippocampus, entorhinal cortex, ventromedial prefrontal cortex (VMPC), occipital lobe, and cerebellum of 10 age-matched control subjects without dementia and 10 patients with diagnosed AD. The postmortem interval was 2813
comparable among all the samples, ranging from 2 to 12 h (see summarized information of patient demographics in Table. 1). None of the control subjects had any neurological diseases that could have led to neuroinflammatory activation. Because Lcn2 is being used as a biomarker for kidney disease, subjects with renal failure were excluded. Patients with MCI used in this study developed to AD. Brain tissue for immunohistochemistry was obtained from Semmelweis University and Hungarian Academy of Sciences (Budapest, Hungary). Brain tissues from 4 patients with AD and 5 control subjects without dementia were stained. Ethics approval for the use of postmortem brain tissue for immunohistochemistry was granted by the local ethics committee of Semmelweis University. The selected AD brain tissue was Braak stage V, and the selected nondementia control tissue was Braak stage III. Ethics approval for human sample collection of serum and CSF and use of human postmortem brain was granted by the local ethics committee of University of Antwerp (approvals 2805 and 2806). Preparation of brain protein extracts Whole-brain protein samples and membrane protein fractions were prepared as described previously (22). Phosphatase inhibitor was added to the homogenization buffers. Lcn2 measurement in serum and CSF by ELISA Quantification of Lcn2 from serum and CSF samples was measured via a constructed sandwich ELISA using human lipocalin 2/NGAL ELISA capture antibody (R&D Systems), recombinant human lipocalin 2/NGAL (R&D Systems) for the internal standard and biotinylated human lipocalin 2/NGAL detection antibody (R&D Systems). Serum was diluted 1:100, and CSF was undiluted. In brief, plates (96 wells, MaxiSorp; Nalge Nunc International, Rochester, NY, USA) were coated with the capture antibody (100 l; 2 g/ml) diluted in PBS (pH 7.4). After overnight incubation at room temperature, the coated plates were washed with TBS containing 0.05% Tween 20 (TBS/T), and nonspecific binding sites were blocked by incubation with 300 l of PBS containing 1% BSA (PBS/BSA) for 2 h at room temperature on a shaker. After washing, 100 l of either the standards (recombinant human Lcn2) or samples diluted in PBS/BSA were added to the plates, and the plates were incubated for 2 h at room temperature on a shaker. Plates were washed 6 times, and 100 l of biotinylated human lipocalin-2/NGAL detection antibody (100 ng/ml) diluted in PBS/BSA was added. After 2 h on a shaker at room temperature, plates were washed 6 times and incubated with 100 l of avidin-HRP (eBioscience, San Diego, CA, USA) in PBS/BSA (1:1000) for 20 min on a shaker at room temperature. Plates were then washed 6 times, and 100 l of substrate solution containing 1 mg/ml O-phenylenediamine in 0.05 M citric acid sodium phosphate buffer (pH 5.0) with hydrogen peroxide (0.06%) TABLE 1. Patient demographics
Parameter
Serum/CSF Gender (male/female) Age (yr) Brain tissue Gender (male/female) Age (yr)
2814
Vol. 26
July 2012
Control, n ⫽ 26
MCI, n ⫽ 28
AD, n ⫽ 26
13/13 79 ⫾ 6
14/14 79 ⫾ 4
13/13 78 ⫾ 5
5/5 78 ⫾ 6
— —
5/5 82 ⫾ 6
was added. The reaction was stopped by addition of 100 l of a 3 N HCl solution. The absorbance was determined at 492 nm with background subtraction at 620 nm using an ELISA reader (Asys UVM 340; Biochrom, Cambridge, UK). The quantity of Lcn2 was estimated from the calibration curve, which ranged from 78 to 5000 pg/ml. Immunohistochemistry of nondementia control and AD brain tissue The postmortem entorhinal cortex-hippocampus tissues from control subjects without dementia and patients with AD were fixed with Karnovsky’s fixative and sectioned to 30 m with a cryostat microtome. The free-floating sections were pretreated with 3% H2O2 followed by 5% normal donkey serum. As the primary antibody rat anti-lipocalin-2 (1:250; R&D Systems), was applied, and as the secondary antibody biotinylated donkey anti-rat antibody (1:500; Jackson ImmunoResearch) was used. The staining was completed with avidin-biotin complex (ABC elite kit; Vector Laboratories, Burlingame, CA, USA), followed by nickel-enhanced 3,3=-diaminobenzidine (DAB) and 0.0033% H2O2. Western blot analysis Isolation of proteins from cells was performed on crushed ice. Cells were washed twice with ice-cold PBS after treatment and subsequently lysed by the addition of 100 l of lysis buffer (50 mM Tris-HCl, pH 7.4; 1% Nonidet P-40; 0.25% sodium deoxycholate; 150 mM NaCl; 1 mM EDTA; and complete mini protease inhibitor cocktail tablet). Membranous (23) and nuclear protein extractions (24) were prepared as described previously. Protein concentrations obtained from cell lysates and brain tissues were quantified by the Bradford assay and adjusted to 1 mg protein/ml. The samples were centrifuged at 8000 g for 10 min at 4°C, and the supernatants were boiled for 5 min in Laemmli sample buffer. Proteins of interest fractionated by SDS-PAGE were transferred to PVDF membranes (Millipore, Billerica, MA, USA). Membranes were blocked for 1 h with 1% I-Block (Tropix, Bedford, MA, USA) in TBS containing 0.0625% Tween 20 and sequentially incubated overnight with primary antibodies at 4°C. The membranes were then washed with TBS containing 0.0625% Tween 20 and incubated with the appropriate HPR-conjugated secondary antibodies diluted to 1:5000 with TBS containing 0.0625% Tween 20 for 1 h. Immunoreactivity was detected using enhanced chemiluminescence (Pierce Biotechnology, Rockford, IL, USA). Protein bands were detected by imaging on a Kodak X-ray film (Eastman Kodak Company, Rochester, MN, USA) or by a computer analysis program using a ChemiDoc XRS system (chemiluminescence image analysis; Bio-Rad Laboratories, Hercules, CA, USA). For the detection of secreted Lcn2 protein, Western blot analysis of culture medium was performed. Actin served as an internal standard protein measurement to eliminate loading variations. Assessment of cell viability After treatment, MTT was added (0.5 mg/ml final concentration) to each well of a flat-bottom 96-well culture plate. Cells were further incubated for 2 h at 37°C in 5% CO2. After incubation, the medium was removed, and cells were lysed in 150 l of DMSO. The absorbance of each well was measured with an ELISA reader (Asys UVM 340) at 595 nm with background subtraction at 630 nm.
The FASEB Journal 䡠 www.fasebj.org
NAUDÉ ET AL.
TABLE 2. TNF-␣-mediated gene induction in primary cortical neurons Gene name
Chemokine (C-C motif) ligand 2 Lipocalin 2 Chemokine (C-X-C motif) ligand 10 Chemokine (C-C motif) ligand 7 CD14 antigen (Cd14) Chemokine (C-X-C motif) ligand 2 Immunoglobulin superfamily, member 6 Tissue inhibitor of metalloproteinase 1 Proteasome (prosome, macropain) subunit,  type 8 Lysozyme 1
Statistical analysis Data obtained from human CSF and serum Lcn2 analysis and in vitro studies was calculated using ANOVA, followed by Tukey’s multiple comparison test. An unpaired Student’s t test was used to calculate statistical significance of means of Lcn2 from postmortem human brain tissues and 24p3R expression. Results shown represent means ⫾ se. In vitro data represent ⱖ3 independent experiments. Values of P ⬍ 0.05 were considered to be statistically significant.
RESULTS TNF-␣-mediated induction of gene expression in primary cortical neurons To assess the effect of TNF-␣ on cortical neuronal gene transcription, we analyzed gene expression changes in primary cortical neurons after 24 h of treatment with TNF-␣ (100 ng/ml) using in-houseprinted Qiagen mouse oligonucleotide arrays. Transcripts that were significantly up-regulated (Pⱕ0.05, n⫽3/group) 2-fold or more by TNF-␣ treatment (Table. 2) were selected for bioinformatic analysis. The 10 transcripts identified shown to be up-regu-
Gene symbol
Ratio, n ⫽ 3
P
Ccl2 Lcn2 Cxcl10 Ccl7 Cd14 Cxcl2 Igsf6 Timp1 Psmb8 Lyz1
21.36 5.80 3.48 3.03 2.70 2.67 2.49 2.30 2.26 2.18
0.0004 0.0006 0.0231 0.0179 0.0017 0.0360 0.0458 0.0025 0.0001 0.0001
lated 2-fold or more by TNF-␣ in primary cortical neurons were tiled according to their function using IPA, and all fell into one single cluster related to inflammatory response. About one-third of these transcripts coded for chemokines such as Ccl2 and Ccl7, which are known to be increased in the CNS in numerous neurodegenerative diseases (25, 26). The microarray also showed that Lcn2 mRNA was significantly up-regulated 24 h after the treatment of neurons with TNF-␣ (100 ng/ml). We confirmed that Lcn2 protein was increased concomitantly with observed mRNA levels by Western blotting. Interaction networks for gene symbol lists of the up-regulated transcripts for mouse cortical neurons (10 transcripts) were obtained by using the IPA knowledge base, and a total of two networks were obtained. One cluster including the transcript encoding Igsf6 was filtered out on the basis of a reliance of the entire network on a single reference and for its low ranking. The highly ranked network included the rest of the 9 transcripts and showed, after enrichment for regulated transcripts within the network, that TNF-␣ induces a prominent NF-B-dependent gene network (Fig. 1).
Figure 1. Regulatory network for TNF-␣-dependent transcripts in mouse cortical neurons. Transcripts with an abundance of 2-fold or greater in TNF-␣-treated (24 h) compared with control treated primary cortical neurons by 2-color microarray analysis (10 transcripts; see Table 2) were used as the input for analysis of potential networks using the signal transduction knowledge environment of IPA. Network involves 9 TNF-␣-regulated transcripts and 4 additional components with IPA-confirmed direct or indirect links to them. TNF-␣-regulated transcripts are depicted in gray, linked components in white. Circular lines above symbols indicate autoregulation, connecting lines without arrows indicate direct protein interaction, and dashed and solid arrows indicate indirect (e.g., regulation of mRNA levels) and direct (e.g., enzymatic) activation.
LIPOCALIN 2 IN ALZHEIMER’S DISEASE
2815
TNF-␣-induced Lcn2 production in primary neuronal cells TNF-␣-mediated Lcn2 protein production was further investigated in primary cortical neurons, astrocytes, and microglia cells in culture. Cells were treated with murine TNF-␣ (100 ng/ml). To determine whether the kinetics of Lcn2 protein production correlated with the mRNA expression observed with the microarray, primary cortical neurons, astrocytes, and microglia cells were incubated with TNF-␣ at various time points. TNF-␣ induced a significant increase in intracellular Lcn2 expression 36 h after onset of treatment in neurons (Fig. 2A), astrocytes (Fig. 2B), and microglia cells (Fig. 2C), and expression normalized after 48 h. Qualitative observation of the secreted protein levels from the cell supernatant of all cell types showed that they followed the same pattern (Fig. 2A–C), but remained increased in the supernatant. To gain further insight into the underlying pathways of TNF-␣-induced
Lcn2 expression, the effect of TNFR1- or TNFR2mediated signaling was further investigated by specific receptor triggering via TNFR1 and TNFR2 agonistic antibodies. RNA samples extracted from cortical neurons of mice lacking TNFR1 or TNFR2 and treated with TNF-␣ (100 ng/ml) during 24 h indicated that Lcn2 mRNA up-regulation by TNF-␣ depends on the presence of the type 1 receptor (Supplemental Fig. S1A). Expression of Lcn2 protein production in these neurons was further validated (Supplemental Fig. S1B) by TNFR1 and TNFR2 stimulation using receptor-specific agonistic antibodies. Primary cortical neurons were treated with 100 ng/ml TNF-␣, 2 M TNFR1 agonistic antibody, or 20 M TNFR2 agonistic antibody for 36 h. A treatment time interval of 36 h for all cell types was used, based on results obtained in Fig. 2A–C. Intracellular Lcn2 expression was significantly increased by TNF-␣-mediated and more specifically by TNFR1-mediated stimulation in primary cortical neurons (Fig. 2D),
Figure 2. TNF-␣-mediated Lcn2 production in primary neuronal cells. A–C) Expression of intracellular and secreted Lcn2 protein levels in primary cortical neurons (A), astrocytes (B), and microglia (C) at different time intervals of TNF-␣ (100 ng/ml) treatment was determined by Western blot. D–F) Effect of TNFR1- and TNFR2-mediated signaling on Lcn2 expression in primary cortical neurons (D), astrocytes (E), and microglia (F) was assessed by Western blot by treating the cells with either TNF-␣ (100 ng/ml), TNFR1 agonistic antibody (2 g/ml), or TNFR2 agonistic antibody (20 g/ml) for 36 h. Lcn2 expression was normalized to actin before the percentage relative to control was calculated. Bars indicate the mean ⫾ se protein expression as a percentage relative to that for untreated controls. *P ⬍ 0.05; **P ⬍ 0.005; ***P ⬍ 0.0001. 2816
Vol. 26
July 2012
The FASEB Journal 䡠 www.fasebj.org
NAUDÉ ET AL.
astrocytes (Fig. 2E), and microglia cells (Fig. 2F), whereas TNFR2 stimulation had no effect on Lcn2 expression. TNF-␣ and specifically TNFR1 stimulation also increased secreted Lcn2 levels (Fig. 2D–F). We further found that TNF-␣-mediated Lcn2 expression requires NF-B in primary cortical neurons, as shown in Supplemental Fig. S1C. Analysis of Lcn2 from human tissues Lcn2 was not significantly elevated in serum of patients with MCI (152⫾17 ng/ml) or AD (169⫾19 ng/ml) compared with that for control subjects (132⫾13 ng/ ml) (Fig. 3A). However, CSF Lcn2 levels were significantly lower in the CSF of the MCI (1340⫾119 pg/ml; P⫽0.0027) and AD groups (1494⫾172 pg/ml; P⫽ 0.036) compared with the normal age-matched control group (2079⫾172 pg/ml) (Fig. 3B). Lcn2 levels were measured in brain tissue by Western blotting. Lcn2 is ubiquitously expressed throughout the human brain (Fig. 3C). It is significantly increased in AD brain tissue, specifically in the entorhinal cortex and hippocampus (Fig. 3C), both regions in which AD neuropathology is prominent (27–30). Evaluation of Lcn2 in the hippocampus by immunohistochemistry revealed robust, punctate, and diffuse staining for Lcn2 in pyramidal neurons in the subiculum-CA1 region in AD brain tissue, also observable, albeit at much lower density, in control brain tissue (Fig. 3D). A punctate distribution of Lcn2 was observed only in the pyramidal neurons and their axons and dendrites but not in other cell types. Effect of Lcn2 on TNF-␣ signaling pathways Previous reports clearly indicated that neuroprotective signaling of TNF-␣ is mediated by TNFR2 stimulation, which activates the PKB/Akt pathway, leading to the phosphorylation of p65 and therefore to the activation of NF-B, which in turn initiates up-regulation of prosurvival factors (31). Because it has been shown that Lcn2 can reduce TNF-␣-mediated effects (32, 33), we further investigated whether Lcn2 had an influence on TNF-␣-mediated signaling, and, in particular, TNFR2mediated prosurvival signaling pathways in primary cortical neurons. Cultured neurons were treated with either TNF-␣, TNFR1 agonistic antibody, or TNFR2 agonistic antibody and also were coincubated with Lcn2 for 24 h. In these conditions, we measured the effect of Lcn2 on TNF-␣-induced phosphorylated-AktSer473 and nuclear extract of phosphorylated-p65Ser536. Lcn2 significantly impaired TNF-␣- and TNFR2-mediated phosphorylation of Akt (Fig. 4A) and p65 (Fig. 4B) when Lcn2 was coincubated with TNF-␣ and TNFR1 or TNFR2 agonistic antibodies. Furthermore, a significant increase in the proapoptotic factor, cleaved caspase 3, was induced only when cells were coincubated with Lcn2 and TNF-␣ or the TNFR2 agonistic antibody (Fig. 4C). Even though cleaved caspase 3 was increased, no effect on cell viability of the primary cortical neurons LIPOCALIN 2 IN ALZHEIMER’S DISEASE
was found (Supplemental Fig. S2A). NF-B activation via TNF-␣ leads to a down-regulation of the phosphatase and tensin homolog deleted on chromosome 10 (PTEN) protein, which in turn is a strong inhibitor of the PI3K/Akt pathway (34). It has been shown previously that Lcn2 can up-regulate PTEN activity in cancer cells (35). Therefore, we further explored the inhibitory effect of Lcn2 on TNF-␣-induced PI3K/Akt pathway activation by determining whether Lcn2 affected PTEN expression in primary cortical neuron cells. Primary cortical neurons were treated with Lcn2 (200 ng/ml) at the indicated time intervals (Fig. 4D). Lcn2 caused a significant up-regulation of PTEN after 2 h, which remained increased for 24 h after Lcn2 treatment (Fig. 4D). Lcn2 also impaired TNF-␣-mediated and specifically TNFR2-mediated down-regulation of PTEN (Fig. 4E). We then verified whether Lcn2 causes an up-regulation of cleaved caspase 3 observed (Fig. 4C) as a result of an increase in PTEN. To that purpose neurons were pretreated with or without the PTEN inhibitor VO-OHpic (300 nM; ref. 36) for 30 min and then treated with Lcn2 (200 ng/ml) coincubated with TNF-␣ (100 ng/ml) or TNFR2 agonistic antibody (20 M). VO-OHpic significantly inhibited the increase in cleaved caspase 3 when Lcn2 was coincubated with TNF-␣ or TNFR2 agonistic antibody (Fig. 4F). Lcn2 abrogates TNF-␣ protection against ADassociated A and glutamate-induced neurotoxicity TNF-␣ via a TNF type 2 receptor signaling pathway can be protective against glutamate-induced (31) and Ainduced neuronal cell death (37). In view of the effects of Lcn2 on suppression of the TNFR2-mediated cellular prosurvival cascade, the effect of Lcn2 on TNF-␣mediated protection against glutamate- and A-induced toxicity was explored. As reported previously (31), at a concentration of 100 ng/ml soluble TNF-␣ was found to be protective against A because of its rather low affinity to TNFR2 (refs. 38, 39 and Supplemental Fig. S2B). As expected, TNF-␣ had a significant protective effect on A-induced (Fig. 5A) and glutamate-induced (Fig. 5B) toxicity. Coincubation of Lcn2 together with either TNF-␣ or TNFR2 agonistic antibody diminished the protective effect exerted by TNF-␣ via TNFR2 against A-induced (Fig. 5A) and glutamate-induced (Fig. 5B) toxicity. Lcn2 on its own induced significant cell death when coincubated with A in a dose-dependent manner (Fig. 5C). Finally, because the Lcn2 receptor 24p3R exerts proapoptotic functions (19), the effect of TNF-␣ on 24p3R expression was assessed by incubating primary cortical neurons with TNF-␣ (100 ng/ml) for 24 h. TNF-␣ significantly increased 24p3R expression (Fig. 5D). However, postmortem brain tissue from patients with AD showed no changes in 24p3R levels compared with that for age-matched control subjects (Supplemental Fig. S3). 2817
Figure 3. Expression of Lcn2 in human samples. A) Serum Lcn2 concentrations among control (n⫽26), MCI (n⫽28), and AD (n⫽26) groups. B) Differences in CSF Lcn2 concentrations among control (n⫽26), MCI (n⫽28), and AD (n⫽26) groups. Bars indicate mean ⫾ se protein concentrations in the different study groups. *P ⬍ 0.05; **P ⬍ 0.005. C) Lcn2 expression in homogenated cerebellum, VMPC, occipital lobe, entorhinal cortex, and hippocampus of control (n⫽10) and AD (n⫽10) brain tissues, determined by Western blot analysis. Lcn2 expression was normalized to actin before the percentage relative to control was calculated. Bars indicate mean ⫾ se ratio of protein expression in nondementia control and AD brain tissue. *P ⬍ 0.05; **P ⬍ 0.005. D) Immunohistochemical staining of Lcn2 of pyramidal neurons in the CA1 hippocampus region of a patient with AD compared with that of a control subject without dementia.
DISCUSSION Recent studies have provided convincing evidence for TNF-␣ as a mainly neuroprotective agent. We therefore performed a gene array on primary cortical neurons treated with TNF-␣ to investigate whether expression of genes associated with protective cellular mechanisms was altered. Of interest, TNF exposure yielded upregulated gene expression for some chemokines that was highest for Ccl2 and second highest for Lcn2. Lcn2 attracted our interest because of recent findings demonstrating that this protein is involved in activating and sensitizing astrocytes and microglia to apoptosis (40, 41) and that increased Lcn2 gene expression was found in an AD mouse model (16) although its underlying mechanisms are largely unknown. Microglia, astrocytes, and also neurons are known to produce inflammatory cytokines (42). Lcn2 has been described as an acutephase protein that is induced via various proinflammatory stimuli, including TNF-␣ (15). It was further shown that TNF-␣ is a strong stimulating factor of Lcn2 in astrocytes (41). Although TNF-␣ is known not to be the sole inducing factor of Lcn2, a recent study has shown that a monoclonal antibody (infliximab), which prevents TNF-␣ from binding to its receptors, significantly reduced peripheral Lcn2 concentrations in patients with Crohn’s disease (43). It was also shown that TNF-␣ is up-regulated in the vicinity of neuropathological hallmarks in the brain of patients with AD (6, 7). These 2818
Vol. 26
July 2012
findings therefore indicate that TNF-␣ may act as a key regulator of Lcn2 production in neurodegenerative diseases such as AD. The present study shows that TNF-␣ induced significantly increased expression of intracellular Lcn2 in astrocytes, microglia cells, and primary cortical neurons with 36 h of exposure and that secreted Lcn2 levels corresponded well with the intracellular levels of this protein. Because TNFR1 and TNFR2 are expressed on neurons (44), astrocytes (45), and microglia (46), the effect of the distinct signaling pathways exerted by individual receptor stimulation was determined by treating cells with receptor-specific agonistic antibodies. TNFR1-mediated signaling was found to be the sole triggering pathway for Lcn2 production in all three cell types. Given the usual signaling events triggered by TNFR1, Lcn2 production is an astoundingly late response. In principle, this late response can be dependent on either an unexplored delayed TNFR1-associated signaling event or can be due to a cascade of events triggered by TNFR1-induced gene transcription, which finally leads to increased Lcn2. The very late response to TNFR1 triggering makes the last hypothesis more likely. In fact, we found that TNF-␣ mediated up-regulation of Lcn2 is dependent on NF-B signaling (Fig. 1C). Whether the Lcn2 gene, however, is directly activated through NF-B or by downstream gene products remains to be shown. The relevance of TNFR1 signaling in AD was recently substantiated by a study of Cheng et al. (22), who showed increased
The FASEB Journal 䡠 www.fasebj.org
NAUDÉ ET AL.
Figure 4. Influence of Lcn 2 on the TNFR2-mediated signaling pathway in primary cortical neurons. Primary cortical neurons were incubated with TNF-␣ (100 ng/ml), TNFR1 agonistic antibody (2 M), or TNFR2 agonistic antibody (20 M) and also coincubated with recombinant murine Lcn2 (200 ng/ml) for 24 h. Western blotting was used to quantify protein expression. A) Neurons were analyzed for PKB/Akt phosphorylation at the Ser473 site. B) NF-B activation was determined via analysis of phosphorylation of nuclear p65 at Ser536. C) Caspase 3 activation was determined by the ratio of cleaved caspase 3 levels to caspase 3 levels. Lcn2 mediated PTEN up-regulation. D) Intracellular PTEN expression in primary cortical neurons treated with murine Lcn2 (200 ng/ml) at different time intervals was determined via Western blot. E) Effect of Lcn2 on TNF-␣-mediated PTEN regulation. F) Role of PTEN in Lcn2-mediated TNFR2 cosignaling-induced caspase 3 activation was assessed by pretreating the indicated neurons with PTEN inhibitor VO-OHpic for 1 h and subsequent incubation with either TNF-␣ (100 ng/ml) or TNFR2 agonistic antibody (20 M) with or without Lcn2 (200 ng/ml) for 24 h. Measured protein expressions were normalized to actin before the percentage relative to control was calculated. Bars indicate mean ⫾ se protein expression as a percentage relative to that of untreated control subjects. *P ⬍ 0.05; **P ⬍ 0.005; ***P ⬍ 0.0001.
TNFR1 and decreased TNFR2 expression in AD brain, in addition to a significantly higher binding affinity of TNF-␣ to TNFR1 than to TNFR2 in the AD brain. A recent study by Choi et al. (47), however, showed a significant increase in Lcn2 concentrations in plasma of MCI compared with control and AD groups. The lack of significant differences in our study might be ascribed to the small number of patients used. The finding of lowered Lcn2 in CSF of individuals with MCI and AD, however, was a robust finding in our study and suggests that Lcn2-mediated neuronal sensitivity to toxicity in AD progression may be due in part to impaired clearLIPOCALIN 2 IN ALZHEIMER’S DISEASE
ance of this molecule from the brain, as has also been suggested for A1– 42. These results point to the potential dynamic regulation of Lcn2 in various cellular compartments through the course of the disease. Upregulation of Lcn2 expression in the choroid plexus, presumably within epithelial cells, has been reported previously (48). Atrophy of these epithelial cells occurs during AD (49) and might explain the reduced Lcn2 production in the CSF. Secondly, megalin is one of the known receptors responsible for the transport of Lcn2 across cell membranes (18). Patients with AD have reduced levels of megalin in the choroid plexus (50). 2819
Figure 5. Effect of Lcn2 on TNF-␣-mediated neuronal protection against A and glutamate-induced neuron toxicity. A) Primary cortical neurons were preincubated for 24 h with TNF-␣ (100 ng/ml), TNFR2 agonistic antibody (20 M), and Lcn2 (200 ng/ml) or TNF-␣ (100 ng/ml) and TNFR2 agonistic antibody (20 M) together with Lcn2 (200 ng/ml). After the 24-h incubation period, A1– 42 (25 M) was added, and cell viability was determined 24 h later with the MTT assay. B) Neurons were also pretreated as indicated with TNF-␣ (100 ng/ml), Lcn2 (200 ng/ml), and TNFR2 agonistic antibody (20 M) and/or TNF-␣ (100 ng/ml) and TNFR2 agonistic antibody (20 M) together with Lcn2 (200 ng/ml) for 24 h and then challenged with 50 M glutamate (concentration that induced ⬇50% cell death) for 1 h. Neuronal viability was determined 24 h after the glutamate challenge using the MTT assay. C) Primary cortical neurons were preincubated with Lcn2 at different concentrations for 24 h, and A1– 42 (25 M) was then added, after which neuronal viability was determined using the MTT assay. 24p3R expression was normalized to actin before the percentage relative to control was calculated. Bars indicate mean ⫾ se cell viability as a percentage relative to that for untreated control subjects. *P ⬍ 0.05; **P ⬍ 0.005; ***P ⬍ 0.0001. D) Expression of the Lcn2 receptor 24p3R in primary cortical neurons after a 24-h incubation with TNF-␣. Bars indicate mean ⫾ se protein expression as a percentage relative to that for the untreated control subjects. **P ⬍ 0.005.
This decreased megalin expression may also contribute to lower Lcn2 into the CSF and thereby to the lower concentrations found. A large number of studies have indicated that one of the most consistent markers in AD is decreased A1– 42 in CSF of patients with MCI who develop AD (51). Interestingly, megalin also partakes in A1– 42 clearance through the blood-brain barrier and blood-CSF barrier at the choroid plexus (52, 53). The fact that Lcn2 is significantly lower in CSF of patients with MCI and AD suggests that whatever the mechanism, dysregulation of Lcn2 expression appears as an early event in AD. Increased Lcn2 expression in the entorhinal cortex and hippocampus brain regions correlates strongly with region-specific AD pathology as classically described (2) and also with TNF-␣ expression in AD brain tissue (6, 7). A marked increase in intracellular Lcn2 staining in the pyramidal neurons of AD human brain tissue was observed compared with that in control brain tissue. This observation correlates with the finding that cyclooxygenase-2, which is up-regulated at sites of inflammation, is also primarily up-regulated in these neurons (29). Lcn2 mRNA expression was also found to be up-regulated in area CA1 of the hippocampus in rat after oxidative stress in vivo (54), indicating that Lcn2 is up-regulated in hippocampus tissue during neuronal injury. At the cellular level, Lcn2 can act as a proapoptotic or prosurvival factor, depending on the cell type and cellular environment (19, 40, 41, 55). Although the 2820
Vol. 26
July 2012
presence of Lcn2 in brain tissue has been shown (15), its main functions in neuronal tissue are largely unknown, especially during disease conditions. In this study, we found that Lcn2 silenced TNFR2-mediated PI3K/Akt activation and NF-B activation. Lcn2 alone did not have an effect on proapoptotic cleavage of caspase 3, but when it was coincubated with TNF-␣ or TNFR2 agonistic antibodies, a significant increase in cleaved caspase 3 but no neuronal cell death was observed. Silencing of TNFR2-mediated PKB/Akt signaling might have shifted the protective signaling cascade to a proapoptotic pathway. Our data further indicate that the inhibitory effect of Lcn2 on TNFR2 signaling can occur via the up-regulation of PTEN. Lcn2 exhibits its proapoptotic properties via the increase in proapoptotic BIM (Bcl-2-interacting mediator of cell death) expression in certain cancerous cell types (19) as well as in astrocytes (41). However, BIM was not involved as a proapoptotic factor in Lcn2-mediated cell death of microglia cells (40). BIM is spliced into the 3 apoptotic isoforms, BimEL, BimL, and BimS, on activation (56). Because we did not find any substantial increase in BimEL expression in primary cortical neurons (Supplemental Fig. S4), we speculate that Lcn2 might execute proapoptotic signaling differently in different cell types. Previous in vitro studies have shown that Lcn2 can activate and sensitize astrocytes and microglia to apoptosis and increase cell death to toxic substances (40, 41). We show here that sole incubation with Lcn2 had no effect on cell death, but incubation
The FASEB Journal 䡠 www.fasebj.org
NAUDÉ ET AL.
Figure 6. Lcn2-mediated sensitivity of neurons toward glutamate and A1– 42 toxicity. Lcn2 expression is induced via TNFR1-specific signaling in primary cortical neurons, astrocytes, and microglia. Lcn2 can silence TNFR2-mediated phosphorylation of Akt and phosphorylation of p65 via the up-regulation of PTEN and therefore inhibit TNFR2-mediated neuroprotection against the AD-associated excitotoxic factors.
with Lcn2 significantly increased cell death of primary cortical neurons in a dose-dependent manner when it was coincubated with A1– 42. Moreover, the finding that Lcn2 diminished the neuroprotective effect of TNF-␣ against glutamate- and A-mediated toxicity in neurons correlates with the mechanistic data described. It is known that Lcn2 can promote apoptosis via its receptor 24p3R (19, 57). Although no differences in 24p3R expression in human brain between control subjects and patients with AD was found, TNF-␣ increased 24p3R in primary cortical neurons in vitro. Because Lcn2 can act in an autocrine manner (40), our in vitro findings lead to the speculation that TNF-␣ not only increases Lcn2 production and secretion, but also increases 24p3R expression, thereby enhancing Lcn2mediated signaling via 24p3R in a proinflammatory environment. In accordance with our findings, Rathore et al. (58) recently reported that the absence of Lcn2 in Lcn2 knockout mice results in enhanced neuronal and tissue survival and reduced proinflammatory cytokines and chemokines after spinal injury. In summary, in vitro results suggest first that Lcn2 can contribute to the pathophysiology of AD by sensitizing LIPOCALIN 2 IN ALZHEIMER’S DISEASE
neuronal cells against A through silencing of essential immunological neuroprotective pathways (Fig. 6), therefore making it a rational target for future pharmacological intervention. Second, our data suggest that Lcn2 is a promising CSF marker that merits further investigation a possible candidate to improve diagnosis for early symptoms of AD. Third, Lcn2 appears to be an important determinant of the balance of effect of TNF-␣ via type 1 and type 2 receptor signaling in the brain and a potential target for progressive neurodegeneration associated with inflammation in human brain. The authors thank Dr. Abdel Elkahloun [Manager, National Human Genome Research Institute/National Institute of Mental Health (NIMH)/National Institute of Neurological Disorders and Stroke, U.S. National Institutes of Health (NIH) Microarray Core] for invaluable assistance with microarray analysis and data capture; Babru Samal [NIMH– Intramural Research Program (IRP) Bioinformatics Core Specialist] for help with statistical analysis; David Huddleston (Biologist, Section on Molecular Neuroscience, NIMH-IRP) for processing of RNAs for microarray analysis; Dennis van der Meer and Erin van Buel for helping with cell culture and 2821
Western blotting; Dr. Amalia Dolga for assisting with cell culture, treatment, and mRNA isolation; and the Antwerp Biobank at the Institute Born-Bunge of the University of Antwerp for providing the biosamples of control subjects, patients with MCI, and patients with AD. This study was funded by De Cock Stichting (project 635155 to P.J.W.N.) and the European Union (grant FP6 NeuropromMiSe LSHMCT-2005-018637 to U.L.M.E.) and by a stipend from the School of Behavioral and Cognitive Neurosciences to P.J.W.N. The authors acknowledge the support of NIMH–IRP (project Z01-MH002386-23). In addition, this work was supported by the Research Foundation Flanders (FWO), Interuniversity Poles of Attraction (IAP Network P6/43) of the Belgian Federal Science Policy Office, Methusalem excellence grant of the Flemish Government, agreement between Institute Born-Bunge and University of Antwerp, the Medical Research Foundation Antwerp, the Thomas Riellaerts research fund, and Neurosearch Antwerp.
14.
15. 16.
17.
18.
19.
REFERENCES 1.
2.
3. 4.
5. 6.
7.
8. 9.
10.
11.
12.
13.
2822
20.
Nelson, P. T., Braak, H., and Markesbery, W. R. (2009) Neuropathology and cognitive impairment in Alzheimer disease: a complex but coherent relationship. J. Neuropathol. Exp. Neurol. 68, 1–14 Small, G. W., Kepe, V., Ercoli, L. M., Siddarth, P., Bookheimer, S. Y., Miller, K. J., Lavretsky, H., Burggren, A. C., Cole, G. M., Vinters, H. V., Thompson, P. M., Huang, S.-C., Satyamurthy, N., Phelps, M. E., and Barrio, J. R. (2006) PET of brain amyloid and tau in mild cognitive impairment. N. Engl. J. Med. 355, 2652– 2663 Lee, Y. J., Han, S. B., Nam, S. Y., Oh, K. W., and Hong, J. T. (2010) Inflammation and Alzheimer’s disease. Arch. Pharm. Res. 33, 1539 –1556 Eikelenboom, P., Veerhuis, R., Scheper, W., Rozemuller, A. J., van Gool, W. A., and Hoozemans, J. J. (2006) The significance of neuroinflammation in understanding Alzheimer’s disease. J. Neural Transm. 113, 1685–1695 Rosenberg, P. B. (2005) Clinical aspects of inflammation in Alzheimer’s disease. Int. Rev. Psychiatry 17, 503–514 Rao, J. S., Rapoport, S. I., and Kim, H. W. (2011) Altered neuroinflammatory, arachidonic acid cascade and synaptic markers in postmortem Alzheimer’s disease brain. Transl. Psychiatry 1, e31 Zhao, M., Cribbs, D. H., Anderson, A. J., Cummings, B. J., Su, J. H., Wasserman, A. J., and Cotman, C. W. (2003) The induction of the TNF␣ death domain signaling pathway in Alzheimer’s disease brain. Neurochem. Res. 28, 307–318 Wajant, H., Pfizenmaier, K., and Scheurich, P. (2003) Tumor necrosis factor signaling. Cell Death Differ. 10, 45–65 Fontaine, V., Mohand-Said, S., Hanoteau, N., Fuchs, C., Pfizenmaier, K., and Eisel, U. (2002) Neurodegenerative and neuroprotective effects of tumor necrosis factor (TNF) in retinal ischemia: opposite roles of TNF receptor 1 and TNF receptor 2. J. Neurosci. 22, 216RC Chakrabarty, P., Herring, A., Ceballos-Diaz, C., Das, P., and Golde, T. E. (2011) Hippocampal expression of murine TNF␣ results in attenuation of amyloid deposition in vivo. Mol. Neurodegener. 6, 16 Montgomery, S. L., Mastrangelo, M. A., Habib, D., Narrow, W. C., Knowlden, S. A., Wright, T. W., and Bowers, W. J. (2011) Ablation of TNF-RI/RII expression in Alzheimer’s disease mice leads to an unexpected enhancement of pathology: implications for chronic Pan-TNF-␣ suppressive therapeutic strategies in the brain. Am. J. Pathol. 179, 2053–2070 He, P., Zhong, Z., Lindholm, K., Berning, L., Lee, W., Lemere, C., Staufenbiel, M., Li, R., and Shen, Y. (2007) Deletion of tumor necrosis factor death receptor inhibits amyloid  generation and prevents learning and memory deficits in Alzheimer’s mice. J. Cell Biol. 178, 829 –841 Nijholt, I. M., Granic, I., Luiten, P. G., and Eisel, U. L. (2011) TNFR2—target for therapeutics against neurodegenerative diseases? Adv. Exp. Med. Biol. 691, 567–573
Vol. 26
July 2012
21.
22.
23. 24.
25.
26.
27.
28.
29. 30.
31.
32.
Kjeldsen, L., Johnsen, A. H., Sengelov, H., and Borregaard, N. (1993) Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase. J. Biol. Chem. 268, 10425–10432 Liu, Q., and Nilsen-Hamilton, M. (1995) Identification of a new acute phase protein. J. Biol. Chem. 270, 22565–22570 Wu, Z.-L., Ciallella, J. R., Flood, D. G., O’Kane, T. M., BozyczkoCoyne, D., and Savage, M. J. (2006) Comparative analysis of cortical gene expression in mouse models of Alzheimer’s disease. Neurobiol. Aging 27, 377–386 Mucha, M., Skrzypiec, A. E., Schiavon, E., Attwood, B. K., Kucerova, E., and Pawlak, R. (2011) Lipocalin-2 controls neuronal excitability and anxiety by regulating dendritic spine formation and maturation. Proc. Natl. Acad. Sci. U. S. A. 108, 18436 –18441 Hvidberg, V., Jacobsen, C., Strong, R. K., Cowland, J. B., Moestrup, S. K., and Borregaard, N. (2005) The endocytic receptor megalin binds the iron transporting neutrophil-gelatinase-associated lipocalin with high affinity and mediates its cellular uptake. FEBS Lett. 579, 773–777 Devireddy, L. R., Gazin, C., Zhu, X., and Green, M. R. (2005) A cell-surface receptor for lipocalin 24p3 selectively mediates apoptosis and iron uptake. Cell 123, 1293–1305 Livak, K. J., and Schmittgen, T. D. (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 25, 402–408 Granic, I., Masman, M. F., Kees Mulder, C., Nijholt, I. M., Naude, P. J., de Haan, A., Borbely, E., Penke, B., Luiten, P. G., and Eisel, U. L. (2010) LPYFDa neutralizes amyloid--induced memory impairment and toxicity. J. Alzheimers Dis. 19, 991–1005 Cheng, X., Yang, L., He, P., Li, R., and Shen, Y. (2010) Differential activation of tumor necrosis factor receptors distinguishes between brains from Alzheimer’s disease and nondemented patients. J. Alzheimers Dis. 19, 621–630 Egan, B. S., Lane, K. B., and Shepherd, V. L. (1999) PU.1 and USF are required for macrophage-specific mannose receptor promoter activity. J. Biol. Chem. 274, 9098 –9107 Joo, M., Hahn, Y. S., Kwon, M., Sadikot, R. T., Blackwell, T. S., and Christman, J. W. (2005) Hepatitis C virus core protein suppresses NF-B activation and cyclooxygenase-2 expression by direct interaction with IB kinase . J. Virol. 79, 7648 –7657 Conductier, G., Blondeau, N., Guyon, A., Nahon, J. L., and Rovere, C. (2010) The role of monocyte chemoattractant protein MCP1/CCL2 in neuroinflammatory diseases. J. Neuroimmunol. 224, 93–100 Renner, N. A., Ivey, N. S., Redmann, R. K., Lackner, A. A., and MacLean, A. G. (2011) MCP-3/CCL7 production by astrocytes: implications for SIV neuroinvasion and AIDS encephalitis. J. Neurovirol. 17, 146 –152 Haroutunian, V., Purohit, D. P., Perl, D. P., Marin, D., Khan, K., Lantz, M., Davis, K. L., and Mohs, R. C. (1999) Neurofibrillary tangles in nondemented elderly subjects and mild Alzheimer disease. Arch. Neurol. 56, 713–718 Haroutunian, V., Perl, D. P., Purohit, D. P., Marin, D., Khan, K., Lantz, M., Davis, K. L., and Mohs, R. C. (1998) Regional distribution of neuritic plaques in the nondemented elderly and subjects with very mild Alzheimer disease. Arch. Neurol. 55, 1185–1191 Braak, H., and Braak, E. (1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. 82, 239 –259 ¨ onönen, M., Mikkonen, M., Laakso, M. P., Frisoni, G. B., K Beltramello, A., Geroldi, C., Bianchetti, A., Trabucchi, M., Soininen, H., and Aronen, H. J. (2000) Hippocampus and entorhinal cortex in frontotemporal dementia and Alzheimer’s disease: a morphometric MRI study. Biol. Psychiatry 47, 1056 – 1063 Marchetti, L., Klein, M., Schlett, K., Pfizenmaier, K., and Eisel, U. L. (2004) Tumor necrosis factor (TNF)-mediated neuroprotection against glutamate-induced excitotoxicity is enhanced by N-methyl-d-aspartate receptor activation. Essential role of a TNF receptor 2-mediated phosphatidylinositol 3-kinase-dependent NF-B pathway. J. Biol. Chem. 279, 32869 –32881 Jin, D., Zhang, Y., and Chen, X. (2011) Lipocalin 2 deficiency inhibits cell proliferation, autophagy, and mitochondrial biogenesis in mouse embryonic cells. Mol. Cell. Biochem. 351, 165–172
The FASEB Journal 䡠 www.fasebj.org
NAUDÉ ET AL.
33.
34.
35. 36. 37.
38.
39.
40.
41. 42. 43.
44.
45.
46.
Guo, H., Jin, D., Zhang, Y., Wright, W., Bazuine, M., Brockman, D. A., Bernlohr, D. A., and Chen, X. (2010) Lipocalin-2 deficiency impairs thermogenesis and potentiates diet-induced insulin resistance in mice. Diabetes 59, 1376 –1385 Kim, S., Domon-Dell, C., Kang, J., Chung, D. H., Freund, J.-N., and Evers, B. M. (2004) Down-regulation of the tumor suppressor PTEN by the tumor necrosis factor-␣/nuclear factor-B (NF-B)-inducing kinase/NF-B pathway is linked to a default IB-␣ autoregulatory loop. J. Biol. Chem. 279, 4285–4291 Shi, H., Gu, Y., Yang, J., Xu, L., Mi, W., and Yu, W. (2008) Lipocalin 2 promotes lung metastasis of murine breast cancer cells. J. Exp. Clin. Cancer Res. 27, 83 Mak, L. H., Vilar, R., and Woscholski, R. (2010) Characterisation of the PTEN inhibitor VO-OHpic. J. Chem. Biol. 3, 157–163 Patel, J. R., and Brewer, G. J. (2008) Age-related differences in NFB translocation and Bcl-2/Bax ratio caused by TNF␣ and A42 promote survival in middle-age neurons and death in old neurons. Exp. Neurol. 213, 93–100 Grell, M., Douni, E., Wajant, H., Lohden, M., Clauss, M., Maxeiner, B., Georgopoulos, S., Lesslauer, W., Kollias, G., Pfizenmaier, K., and Scheurich, P. (1995) The transmembrane form of tumor necrosis factor is the prime activating ligand of the 80 kDa tumor necrosis factor receptor. Cell 83, 793–802 Grell, M., Wajant, H., Zimmermann, G., and Scheurich, P. (1998) The type 1 receptor (CD120a) is the high-affinity receptor for soluble tumor necrosis factor. Proc. Natl. Acad. Sci. U. S. A. 95, 570 –575 Lee, S., Lee, J., Kim, S., Park, J. Y., Lee, W. H., Mori, K., Kim, S. H., Kim, I. K., and Suk, K. (2007) A dual role of lipocalin 2 in the apoptosis and deramification of activated microglia. J. Immunol. 179, 3231–3241 Lee, S., Park, J. Y., Lee, W. H., Kim, H., Park, H. C., Mori, K., and Suk, K. (2009) Lipocalin-2 is an autocrine mediator of reactive astrocytosis. J. Neurosci. 29, 234 –249 Chopra, K., Misra, S., and Kuhad, A. (2011) Neurobiological aspects of Alzheimer’s disease. Expert Opin. Ther. Targets 15, 535–555 Bolignano, D., Della Torre, A., Lacquaniti, A., Costantino, G., Fries, W., and Buemi, M. (2010) Neutrophil gelatinase-associated lipocalin levels in patients with Crohn disease undergoing treatment with infliximab. J. Investig. Med. 58, 569 –571 Viel, J. J., McManus, D. Q., Smith, S. S., and Brewer, G. J. (2001) Age- and concentration-dependent neuroprotection and toxicity by TNF in cortical neurons from -amyloid. J. Neurosci. Res. 64, 454 –465 Fernandes, A., Barateiro, A., Falcão, A. S., Silva, S. L.-A., Vaz, A. R., Brito, M. A., Marques Silva, R. F., and Brites, D. (2011) Astrocyte reactivity to unconjugated bilirubin requires TNF-␣ and IL-1 receptor signaling pathways. Glia 59, 14 –25 Veroni, C., Gabriele, L., Canini, I., Castiello, L., Coccia, E., Remoli, M. E., Columba-Cabezas, S., Aricò, E., Aloisi, F., and Agresti, C. (2010) Activation of TNF receptor 2 in microglia
LIPOCALIN 2 IN ALZHEIMER’S DISEASE
47. 48.
49. 50. 51. 52.
53.
54.
55.
56.
57. 58.
promotes induction of anti-inflammatory pathways. Mol. Cell. Neurosci. 45, 234 –244 Choi, J., Lee, H.-W., and Suk, K. (2011) Increased plasma levels of lipocalin 2 in mild cognitive impairment. J. Neurol. Sci. 305, 28 –33 Marques, F., Rodrigues, A. J., Sousa, J. C., Coppola, G., Geschwind, D. H., Sousa, N., Correia-Neves, M., and Palha, J. A. (2008) Lipocalin 2 is a choroid plexus acute-phase protein. J. Cereb. Blood Flow Metab. 28, 450 –455 Serot, J. M., Bene, M. C., and Faure, G. C. (2003) Choroid plexus, aging of the brain, and Alzheimer’s disease. Front. Biosci. 8, s515–521 Alvira-Botero, X., and Carro, E. M. (2010) Clearance of amyloidbeta peptide across the choroid plexus in Alzheimer’s disease. Curr. Aging Sci. 3, 219 –229 Prvulovic, D., and Hampel, H. (2011) Amyloid  (A) and phospho-tau (p-tau) as diagnostic biomarkers in Alzheimer’s disease. Clin. Chem. Lab. Med. 49, 367–374 Zlokovic, B. V., Martel, C. L., Matsubara, E., McComb, J. G., Zheng, G., McCluskey, R. T., Frangione, B., and Ghiso, J. (1996) Glycoprotein 330/megalin: probable role in receptor-mediated transport of apolipoprotein J alone and in a complex with Alzheimer disease amyloid  at the blood-brain and bloodcerebrospinal fluid barriers. Proc. Natl. Acad. Sci. U. S. A. 93, 4229 –4234 Hammad, S. M., Ranganathan, S., Loukinova, E., Twal, W. O., and Argraves, W. S. (1997) Interaction of apolipoprotein J-amyloid -peptide complex with low density lipoprotein receptorrelated protein-2/megalin. J. Biol. Chem. 272, 18644 –18649 Wang, X., Pal, R., Chen, X. W., Kumar, K. N., Kim, O. J., and Michaelis, E. K. (2007) Genome-wide transcriptome profiling of region-specific vulnerability to oxidative stress in the hippocampus. Genomics 90, 201–212 Miharada, K., Hiroyama, T., Sudo, K., Danjo, I., Nagasawa, T., and Nakamura, Y. (2008) Lipocalin 2-mediated growth suppression is evident in human erythroid and monocyte/macrophage lineage cells. J. Cell. Physiol. 215, 526 –537 O’Connor, L., Strasser, A., O’Reilly, L. A., Hausmann, G., Adams, J. M., Cory, S., and Huang, D. C. S. (1998) Bim: a novel member of the Bcl-2 family that promotes apoptosis. EMBO J. 17, 384 –395 Li, P. T., Lee, Y. C., Elangovan, N., and Chu, S. T. (2007) Mouse 24p3 protein has an effect on L929 cell viability. Int. J. Biol. Sci. 3, 100 –107 Rathore, K. I., Berard, J. L., Redensek, A., Chierzi, S., LopezVales, R., Santos, M., Akira, S., and David, S. (2011) Lipocalin 2 plays an immunomodulatory role and has detrimental effects after spinal cord injury. J. Neurosci. 31, 13412–13419 Received for publication January 23, 2012. Accepted for publication March 12, 2012.
2823
