© The Author 2015. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

1

Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, 1–11doi:10.1093/gerona/glu231

Original Article

Original Article

Long-term Neuroglial Cocultures as a Brain Aging Model: Hallmarks of Senescence, MicroRNA Expression Profiles, and Comparison With In Vivo ModelsElisabetta Bigagli,1 Cristina Luceri,1 Tania Scartabelli,2 Piero Dolara,1 Fiorella Casamenti,1 Domenico E. Pellegrini-Giampietro,2 and Lisa Giovannelli1

1Department of Neuroscience, Psychology, Drug Research and Child Health (NEUROFARBA), Section of Pharmacology and Toxicology and 2Department of Health Sciences, Section of Clinical Pharmacology and Oncology, University of Florence, Florence, Italy.

Address correspondence to Elisabetta Bigagli, PhD, Department of NEUROFARBA, University of Florence, Viale Pieraccini 6, 50139 Florence, Italy. Email: elisabetta.bigagli@unifi.it

Abstract

Our purpose was to evaluate long-term neuroglial cocultures as a model for investigating senescence in the nervous system and to assess its similarities with in vivo models. To this aim, we maintained the cultures from 15 days in vitro (mature cultures) up to 27 days in vitro (senescent cultures), measuring senescence-associated, neuronal, dendritic, and astrocytic markers. Whole microRNA expression profiles were compared with those measured in the cortex of 18- and 24-month-old C57Bl/6J aged mice and of transgenic TgCRND8 mice, a model of amyloid-β deposition. Neuroglial cocultures displayed features of cellular senescence (increased senescence-associated-β-galactosidase activity, oxidative stress, γ-H2AX expression, IL-6 production, astrogliosis) that were concentration dependently counteracted by the antiaging compound resveratrol (1–5 µM). Among the 1,080 microRNAs analyzed, 335 were downregulated or absent in 27 compared with 15 days in vitro and resveratrol reversed this effect. A substantial overlapping was found between age-associated changes in microRNA expression profiles in vitro and in TgCRND8 mice but not in physiologically aged mice, indicating that this culture model displays more similarities with pathological than physiological brain aging. Our results demonstrate that neuroglial cocultures aged in vitro can be useful for investigating the cellular and molecular mechanisms of brain aging and for preliminary testing of protective compounds.

Key Words: Aging—miRNA—Microarray—Neuroglial cocultures—Neurodegeneration.

Decision Editor: Rafael de Cabo, PhD

Increasing evidence indicates that cellular senescence is a causal fac-tor of aging in mammals (1). Senescent cells may act as a continuous source of proinflammatory and prooxidant stimuli that affect organ homeostasis and functioning (2). Many morphological and molecular features of cellular senescence, such as major changes in morphol-ogy, activation of lysosomal hydrolase β-galactosidase (senescence-associated-β-galactosidase [SA-β-gal]), increased oxidative damage,

and DNA strand breaks, have been described in vitro in proliferat-ing cells, mainly fibroblasts (3). Single- and double-strand breaks are activators of the DNA damage response, a signal cascade involv-ing the activation of ATM/ATR kinases and the formation of DNA damage foci incorporating the activated histone variant γ-H2AX, resulting in cell cycle arrest, the so-called replicative senescence state (4).

There are few in vitro studies of neuronal senescence and it is not known whether the morphological, functional, and molecular char-acteristics of replicative senescence are relevant for nondividing cells. Studies on long-term primary neuronal cultures have reported that some features of the senescence process are shared by dividing cells and neurons: both cerebellar granular neurons and hippocampal neu-rons show a progressive increase in SA-β-gal activity in culture (5,6). Hippocampal neurons aged in vitro have been reported to remain unchanged in number up to 45 days in vitro and to lose markers of synaptic density and neuritic/dendritic network such as PSD-95 and MAP2 (7) just as in in vivo aged human brains (8). Increased protein oxidative damage has also been reported in aging cultures (9) as well as production of β-amyloid (7). Cortical, hippocampal, and peripheral neurons of aged mice show markers of the senescent phenotype, such as DNA damage foci, activated p38/MAPkinase, increased oxida-tive damage, IL-6 production, heterochromatinization, and SA-β-gal activity; this phenotype is dependent on p21 function (10) similarly to senescent proliferating cells. Thus, it appears that neurons aged in vitro express at least some of the features of replicative senescence and recapitulate to some extent the process of in vivo brain aging.

MicroRNAs (miRNAs) are potential critical players in gene regu-lation during age-associated processes. The unbalanced equilibrium between upregulated miRNAs, suppressing unwanted gene expres-sion, and downregulated miRNAs, whose target genes are important for cellular function, has been noted in the normal aging process in the brain (11) and in neurodegenerative diseases (12). miRNA regulation might play an important role in the shaping of age-related mRNA changes in the human and macaque cerebral cortex (13). Li and colleagues (14) have identified age-associated changes in miRNA expression in the mouse brain partly counteracted by caloric restriction (15). Associations between specific miRNAs and age-associated learning dysfunctions have been reported in animal models and in Alzheimer’s disease (16,17). Some of these miRNAs regulate the expression of genes involved in neuronal plasticity (18) and it is possible that specific miRNAs play a role in the complex array of mechanisms leading to reduced neuronal function in aging.

In vitro research models have the obvious advantage of being less costly and less time consuming, besides sparing animal lives. Most of the studies on in vitro neuronal aging have used primary neu-ronal cultures, that is, homogeneous cellular populations that do not account for the reciprocal interactions between neurons and glial cells, particularly important in brain aging (19).

Thus, we decided to evaluate an integrated in vitro system com-prising both mouse neuronal and glial cells as a model for study-ing brain aging. To this aim, we first defined the temporal evolution of this system, measuring a series of senescence-associated markers. We further investigated the variation of miRNA expression profiles in aging mixed cultures and compared it with that occurring in the cerebral cortex of physiologically aged C57Bl/6J mice (20) and of TgCRND8 transgenic mice, displaying a widespread amyloid deposi-tion in the brain (21). Finally, we tested the effect of the antiaging compound resveratrol (20,22) on senescence-associated markers and miRNA expression in order to provide a first evaluation of the pre-dictive potential of antiaging substances on neuronal–glial cocultures.

Materials and Methods

Cell CulturesThe experimental procedures were in accordance with the standards set forth in the Guide for the Care and Use of Laboratory Animals (National Academy of Science). The uterus was removed from the

gravid mouse under anesthesia. Primary cultures of mixed cortical cells containing both neuronal and glial elements were prepared as previously described (23). Briefly, cerebral cortices were dissected from fetal mice at 17–18 days of gestation, minced by using medium stock (composed of Eagle’s Minimum Essential Medium [with Earle’s salts, glutamine- and NaHCO3-free] supplemented with 38 mmol/L NaHCO3, 22 mmol/L glucose, 100 U/mL penicillin, and 100 µg/mL streptomycin) and incubated for 10 minutes at 37°C in medium stock with 0.25% trypsin and 0.05% DNAse. Enzymatic diges-tion was terminated by incubation (10 minutes at 37°C) in medium stock supplemented with 10% heat-inactivated horse serum and 10% fetal bovine serum, and the cells were mechanically disrupted and counted. After a brief centrifugation, cells were resuspended (approximately 4 × 105 cells/mL) and plated in 15 mm multi-well vessels on a layer of confluent astrocytes in medium stock supple-mented with 10% heat-inactivated horse serum, 10% fetal bovine serum, and 2 mmol/L glutamine. Cultures were kept in an incubator at 37°C, with 100% humidity and 95% air and 5% CO2 atmos-phere. After 4–5 days in vitro, nonneuronal cell division was halted by the application of 3 mmol/L cytosine arabinoside for 24 hours. Cultures were then shifted to a maintenance medium identical to the plating medium but lacking fetal bovine serum, which was then par-tially replaced twice a week. Neurons at 15 days in vitro (DIV) were considered mature and followed over time until cells were viable. At several time points (15, 21, 23, and 27 DIV), cells were harvested to evaluate different senescence parameters.

TreatmentsCultures were treated with resveratrol 1 and 5 µM (Sigma–Aldrich Chemicals Co., St Louis, MO) continuously from the maturity stage of 15 DIV (young/adult cultures) to senescence at 27 DIV. The medium was partially changed every 2  days to maintain the con-centrations relatively constant over time (24). Concentrations were selected on the basis of the existing literature showing effects of this substance in vitro (24,25).

Lactate Dehydrogenase ReleaseCell damage was evaluated by measuring the amount of lactate dehy-drogenase released from injured cells into the extracellular fluid by using the Citotoxicity Detection Kit (lactate dehydrogenase; Roche Applied Science, Penzberg, Upper Bavaria, Germany).

Senescence-Associated-β-Galactosidase ActivitySA-β-gal was performed following the protocol of the senescence detection kit (β-galactosidase Staining Kit, Cell Signaling, Danvers, MA). Cells positive for SA-β-gal blue staining were observed under a microscope and counted by two independent investigators with a ×20 objective; 10 fields were counted and averaged for each slide.

Western BlottingHarvested cells were lysed in radioimmunoprecipitation assay buffer (25 mM Tris–HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate) with 1% protease and phosphatase inhibitor Cocktail (Sigma–Aldrich) and disrupted by sonication (Microson XL-2000; Misonix, Farmingdale, NY). Lysates were clarified by centrifugation and supernatants collected at −20°C. Protein content was estimated by using the Bio-Rad DC protein assay kit (Bio-Rad, Segrate, Milan, Italy). Thirty micrograms of protein were subjected to 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis separation (NuPAGE, Novex; Invitrogen, Carlsbad,

2 Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00

CA) and transferred to polyvinylidene fluoride membranes (Millipore, Billerica, MA). Immunostaining was performed with specific primary antibodies: NeuN monoclonal antibody, clone A60 (Millipore), 1:1,000; MAP2 rabbit antibody (Cell Signaling), 1:1,000; glial fibril-lary acidic protein rabbit antibody (Dako, Glostrup, Denmark), 1:3,000; Phospho-Histone H2A.X (Ser 139)  rabbit antibody (Cell Signaling), 1:500; Dicer mouse monoclonal antibody (Santa Cruz Biotechnology, Dallas, TX), 1:1,000; glyceraldehyde 3-phosphate dehydrogenase (14C10) rabbit antibody (Cell Signaling), 1:3,000 and peroxidase-conjugated secondary antibodies: anti-rabbit IgG antibody (Cell Signaling), 1:4,000 and anti-mouse IgG (Chemicon, Temecula, CA), 1:5,000, for 2 hours at room temperature (RT). Proteins were visualized using the enhanced chemiluminescence pro-cedure with Immobilon Horseradish Peroxidase Substrate (Millipore) and immune-reactive bands were acquired through the Image Quant 350 (GE Healthcare Life Sciences, Buckinghamshire, UK) and quan-tified by densitometric analysis using the Quantity-One software (Bio-Rad Laboratories, Hercules, CA). Each density measure was normalized by using the corresponding glyceraldehyde 3-phosphate dehydrogenase level as internal control.

Dot BlottingEqual aliquots (30 μg) of protein were applied to a nitrocellulose membrane (Millipore) and allowed to dry for 30 minutes at RT. After blocking with 6% nonfat dry milk for 1 hour at RT, the membranes were incubated overnight at RT with the anti-hydrox-ynonenal-Michael adducts, reduced rabbit pAb Calbiochem (Millipore), 1:500, followed by incubation with anti-rabbit IgG horseradish peroxidase-linked antibody (Cell Signaling), 1:4,000 for 1 hour at RT. Chemiluminescence was developed by Immobilon Horseradish Peroxidase Substrate (Millipore) and immunoreac-tive spots were quantified using Quantity-One software (Bio-Rad Laboratories).

ImmunocytochemistryThe following primary antibodies were used: NeuN mouse monoclo-nal antibody, clone A60 (Millipore), 1:100; MAP2 rabbit antibody (Cell Signaling), 1:100; glial fibrillary acidic protein rabbit antibody (Dako), 1:500; Phospho-Histone H2A.X (Ser 139) rabbit antibody (Cell Signaling), 1:50; nitrotyrosine rabbit antibody (Millipore), 1:100, followed by the suitable fluorescent secondary antibodies: AlexaFluor 488 goat anti-mouse (Invitrogen), 1:200 or AlexaFluor 594 goat anti-rabbit (Invitrogen), 1:200.

Image Acquisition and AnalysisMicroscopic analysis was performed with a fluorescence microscope (Labophot-2, Nikon, Tokyo, Japan) connected to a CCD camera. Densitometric analyses were performed with the freely available ImageJ NIH software. Results were expressed in arbitrary units.

Carbonyl Residues AssayCarbonyl residues were spectrophotometrically determined at 370 nm by the method of Correa-Salde and Albesa (26).

Detection of Advanced Glycation End-ProductsAdvanced glycation end-products determination was carried out in the culture media according to a previously described method (27). The fluorescence intensity was measured at 440 nm after excita-tion at 370 nm, using a fluorescence spectrophotometer (Multilabel

Counter 1240 Victor 3, Perking Elmer, Waltham, MA) at RT. Results were expressed as fluorescence intensity/mg of proteins.

Cytokine DeterminationNeuroglial coculture supernatants were collected at the differ-ent time points and stored at −80°C for cytokine determination by enzyme-linked immunosorbent assay (MCYTOMAG-70K, MILLIPLEX MAP Mouse Cytokine/Chemokine Magnetic Bead Panel—Immunology Multiplex Assay, Merck-Millipore, Darmstadt, Germany).

AnimalsAll procedures were carried out in agreement with the European Union Regulations on the Care and Use of Laboratory Animals (OJ of ECL 358/1, December 18, 1986), according to the Italian Ministry of Public Health regulations. C57Bl/6J mice were sacrificed by cervi-cal dislocation; the cerebral cortex from C57Bl/6J young and aged mice (3-, 18-, and 24-month-old, respectively, n = 4 per group) was dissected, immediately frozen in liquid nitrogen and stored at −80°C for further molecular analyses. Transgenic hemizygous TgCRND8 mice (4–6 months of age, corresponding to middle stage Aβ deposi-tion, n = 4) were sacrificed by cervical dislocation and frozen samples of their cerebral cortex were also immediately sectioned, frozen with liquid nitrogen, and stored at −80°C. The group of young C57Bl/6J served also as wild-type control, as the transgenic mice are on a C57BL/6 background.

Total RNA ExtractionTotal RNA including miRNAs was isolated from neuroglial cocul-tures and mice brain cortex by using the Absolutely RNA miRNA Kit (Agilent Technologies, Santa Clara, CA) according to the manufac-turer’s protocol. RNA concentration and purity was determined by using a NanoPhotometer spectrophotometer (IMPLEN, München, Germany). RNA integrity was checked with a 2100 Bioanalyzer and RNA 6000 Nano LabChip kit (Agilent).

miRNA Microarray ProtocolmiRNA profiling was performed on 15, 27, and 27 DIV resveratrol-treated (1 and 5 µM) neuroglial coculture cells and on samples of cerebral cortex harvested from 3-, 18-, and 24-month-old C57Bl/6J mice and from TgCRND8 mice. Total RNA samples (100 ng) con-taining miRNAs were labeled with cyanine 3-pGp (Cy3) using the Agilent microRNA Complete Labeling and Hyb Kit. To assess the labeling and hybridization efficiencies, the RNA samples were first spiked with synthesized oligonucleotides by using a MicroRNA Spike-In Kit (Agilent). Each sample was placed on an Agilent mice miRNA array (Release 18.0, 8x60K) and covered with an Agilent 8-plex Gasket chamber. In this array, each miRNA is printed 20 times with replicate probes. Slides were then hybridized for 20 hours at 55°C using an Agilent hybridization system. After hybridization, the microarray slides were washed by using Agilent Gene Expression Wash Buffers.

Bioinformatic MethodsFluorescent signal intensities were detected by using Agilent Scan Control 7.0 Software on an Agilent DNA Microarray Scanner, at a resolution of 2  μm. Data were acquired using Agilent Feature Extraction 9.5.3.1 software for miRNA microarray, generating GeneView files containing summarized signal intensities for each

Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00 3

miRNA, by combining intensities of replicate probes and back-ground subtraction.

DIANA-microT and Target scan databases were used to pre-dict the miRNA target genes. DIANA-miRPath v2.0 (http://www.microrna.gr/miRPathv2.), a web-based application that performs an enrichment analysis of predicted target genes in all available KEGG pathways, was used to identify biological pathways potentially altered by the expression of the miRNAs simultaneously modu-lated in neuroglial cocultures and in in vivo models of aging and neurodegeneration.

miRNA Validation: Quantitative Real-Time Reverse Transcription–Polymerase Chain ReactionTo confirm the results of microarray miRNAs expression profiling, we performed quantitative real-time reverse transcription–polymer-ase chain reaction (qRT–PCR) in selected affected (miR-17, miR-124, let-7a, mir-29) and nonaffected miRNAs (mir-34), using SYBR green qRT–PCR assay. Reverse transcription of RNA was performed using miScript Reverse Transcription Kit (Qiagen, Dusseldorf, Germany). qRT–PCR assays were carried out in 7900HT Fast Real-Time PCR System (Applied Biosystems, California, CA) using miScript SYBR Green PCR Kit and miScript Primer Assay (Qiagen). The amplifica-tion profile was denaturated at 95°C for 15 minutes, followed by 40 cycles of 94°C for 15 seconds, 55°C for 30 seconds, and 70°C for 30 seconds. The expression levels of miRNAs were normalized to RNU6B and calculated as 2−ΔΔct.

StatisticsStatistical analysis was carried out using GRAPHPAD Prism 5 soft-ware. As the number of observations did not allow the application of a test to evaluate normal distribution, we performed comparisons among groups using the nonparametric Mann–Whitney U test (p < .05 was considered statistically significant). Data are means ± SE.

Results

Cell Culture Morphology and CytotoxicityMeasuring lactate dehydrogenase release in the medium, we found a low and constant level of cell death (about 20%) in neuroglial cocul-tures both at 15 and 27 DIV, which was not modified by resvera-trol (Figure 1, panel A). Demonstrative microscopy images indicate that 15 DIV represent a stage of maturation of neurons in which they form clusters and an extensive network of neurites. At 27 DIV, no visible sign of neuronal deterioration was detected (Figure  1, panel B).

Senescence-Related MarkersCytochemical determination of SA-β-galThe appearance of senescence was evaluated starting at 15 DIV (the maturity stage) until 27 DIV. Neurons positive for SA-β-gal increased as a function of time, reaching a maximum between 23 and 27 DIV (p < .05, 20 and 23 DIV vs 15 DIV; p < .01, 27 DIV vs 15 DIV). Beyond 27 DIV, neuroglial cocultures were no longer viable and we considered this stage as full senescence (Figure  1, panel C). Microscopy images indicate that at 15 DIV, almost no neu-ronal cells were positive for SA-β-gal, whereas at 27 DIV, the neu-roglial coculture showed widespread blue staining in neurons inside the clusters and an even more intense staining outside the clusters (Figure 1, panel D). At 27 DIV, a light blue and diffuse staining for SA-β-gal also appeared in the underlying astrocytic layer (Figure 1,

panel D). Resveratrol (1–5  µM) administered from 15 to 27 DIV reduced SA-β-gal activity in 27 DIV neuroglial cocultures at levels below those of 15 DIV (p < .01 vs 27 DIV; Figure 1, panels C and D).

Western blotting and immunocytochemistry for γ-H2AXWe observed a time-dependent increase of γH2AX, a marker of senescence-related DNA damage in neuroglial cocultures at 15, 20, 23, and 27 DIV (p < .05, 27 DIV vs 15 DIV). Resveratrol concen-tration dependently reduced the senescence-associated increase in γH2AX (p < .05, 27 DIV Resv 5 vs 27 DIV; Figure  2, panel A). Representative immunofluorescence images for γH2AX (in red) and the neuronal marker NeuN (in green) show the colocalization of γH2AX with NeuN positive neurons (in yellow; Figure 2, panel B).

Western Blotting and Immunocytochemistry for Neuronal (NeuN), Dendritic (MAP2), and Astrocytic (GFAP) markersThe neuronal marker NeuN decreased from 15 to 20 DIV, remained stable from 20 to 23 DIV, and significantly decreased at 27 DIV (p < .05, 27 DIV vs 15 DIV; Figure 2, panel C). Moreover, immu-nofluorescent images showed the progressive emptying of neuronal clusters at 27 DIV (Figure 2, panel D). Immunoreactivity for MAP2, a marker of dendritic spines, indicated a reduced presence of pro-cesses in 27 DIV cocultures (Figure 2, panel F). The relative protein quantifications further supported this result; in fact, the trend of MAP2 protein expression levels was quite similar to that observed for NeuN, displaying a constant loss of MAP2 protein expression from 15 DIV to 20, 23, and 27 DIV (p < .05 27 DIV vs 15 DIV; Figure  2, panel E). Resveratrol did not modify the senescence-dependent loss either in NeuN or in MAP2 protein expression (Figure 2, panels C and E).

In order to evaluate the appearance of senescent characteristics in the underlying astrocytic layer, we also examined astrocyte cell morphology and the astrocytic protein GFAP. In 27 DIV neuroglial cocultures, astrocytes became hypertrophic, exhibited intense immu-nostaining for GFAP, and displayed an extensive network of large processes (Figure 3, panel B). Results of GFAP protein expression levels were consistent with those of immunofluorescence, show-ing an increase in GFAP protein expression from 15 DIV to 20, 23, and 27 DIV (Figure 3, panel A; p < .05, 23 and 27 DIV vs 15 DIV). Resveratrol concentration dependently reduced GFAP protein expression to levels below those of 15 DIV neuroglial cocultures (Figure 3, panel A).

Oxidative Stress and Inflammation MarkersProtein oxidation, measured as carbonyl residues, progressively increased from 15 DIV to 20, 23, and 27 DIV. At 27 DIV, we detected the highest levels of protein oxidation compared with 15 DIV neuroglial cocultures; this effect was reduced by 1 µM resvera-trol (p < .05; Figure  3, panel C). Dot blotting for 4-hydroxynon-enal showed increased lipid peroxidation in neuroglial cocultures at 27 DIV and reduction of this effect by 1 µM resveratrol (Figure 3, panel D). Densitometric analysis of nitrotyrosine did not show sig-nificant differences between 15 and 27 DIV neuroglial cocultures (75.10 ± 6.82 in 15 DIV vs 74.58 ± 2.81 in 27 DIV). Resveratrol did not exert any significant effect (data not shown). No statistically sig-nificant differences were found in culture media levels of advanced glycation end-products between 15 and 27 DIV neuroglial cocul-tures (613.67 ± 12.85 vs 627.71 ± 25.36 fluorescence intensity/mg of protein).

4 Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00

The production of the proinflammatory cytokine IL-6 was strongly increased in 27 DIV compared with 15 DIV and resveratrol 1–5 µM counteracted this effect (p < .01, 15 DIV vs 27 DIV; p < .01, 27 DIV Resv 5 vs 27 DIV; p < .01, 27 DIV Resv 1 vs 27 DIV; Figure 3, panel E).

miRNA Expression Profiles in Neuroglial CoculturesmiRNA profiling was performed on neuroglial cocultures at 15 and 27 DIV and in 27 DIV treated with 1 and 5 µM resveratrol. Three hundred thirty-five of the 1,080 miRNAs analyzed were differen-tially expressed (downregulated or completely absent) between 15 and 27 DIV.

To rule out the possibility of effects due to changes in miRNA precursors processing, we analyzed the expression of the enzyme Dicer by Western blot and found no difference in the levels of the protein in young and senescent cultures (data not shown).

Resveratrol reversed the extensive downregulation observed in 27 DIV neuroglial cocultures and this effect was concentration dependent for 202 miRNAs. An unsupervised hierarchical cluster (Figure 4, panel A) shows the relationships among the experimen-tal groups as a tree with branch lengths reflecting the similarity between experimental groups. miRNA expression profiles of senes-cent neuroglial cocultures treated with resveratrol were similar to those of young cells (Figure 4, panel A). Among the miRNAs downregulated in senescent cocultures and restored after resvera-trol treatment, we found several members of the mir-466-467-669 cluster, mir-7, mir-24-3p, mir-124-3p, mir-328-3p, miR-17-92 clus-ter, and let-7a. Putative gene targets of some miRNAs downregu-lated in senescent cultures were identified by DIANA-microT and are shown in Table 1. miR-715, miR-199a-5p, and miR-199a-3p were the only three found to be upregulated in 27 DIV cells, with a >2 fold-change (FC).

Figure 1. Panel A: lactate dehydrogenase (LDH) release in the neuroglial coculture medium at 15 days in vitro (DIV), 27 DIV and in 27 DIV cells treated with 1 and 5 µM resveratrol. Results are expressed as percentage of Triton-X-induced LDH release (mean ± SE of three experiments). Panel B: representative microscopic images of neuroglial coculture at 15 DIV, 27 DIV, and of 27 DIV cells treated with 1 and 5 µM resveratrol. Panel C: number of cells positive for senescence-associated-β-galactosidase (SA-β-gal) blue staining/microscopic field in neuroglial cocultures at 15, 20, 23, 27 DIV and in 27 DIV cells treated with 1 and 5 µM resveratrol (mean ± SE of three experiments). *p < .05 vs 15 DIV; **p < .01 vs 15 DIV; ##p < .01 vs 27 DIV. Panel D: representative microscopic images of SA-β-gal blue-staining in neuroglial cocultures at 15, 27 DIV, and of 27 DIV cells treated with 1 and 5 µM resveratrol.

Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00 5

Figure 2. Panel A: γ-H2AX protein expression in neuroglial coculture at 15, 20, 23, 27 days in vitro (DIV) and in 27 DIV cells treated with 1 and 5 µM resveratrol (mean ± SE of three experiments). *p < .05 vs 15 DIV; #p < .05 vs 27 DIV. Panel B: representative immunofluorescence images of γ-H2AX (red) and of the neuronal marker NeuN (green) in neuroglial cocultures at 15, 27 DIV and in 27 DIV cells treated with 1 and 5 µM resveratrol. Yellow color indicates co-localization. Panel C: NeuN protein expression in neuroglial cocultures at 15, 20, 23, 27 DIV and in 27 DIV cells treated with 1 and 5 µM resveratrol (mean ± SE of three experiments). *p < .05 vs 15 DIV. Panel D: representative immunofluorescence images of neuronal clusters (NeuN positive cells in green) at 15, 27 DIV, and at 27 DIV treated with 1 and 5 µM resveratrol. Panel E: MAP2 protein expression in neuroglial cocultures at 15, 20, 23, 27 DIV and in 27 DIV cells treated with 1 and 5 µM resveratrol (mean ± SE of three experiments). *p < .05 vs 15 DIV. Panel F: representative immunofluorescence images of MAP2 (green) in neuroglial cocultures at 15, 27 DIV and in 27 DIV cells treated with 1 and 5 µM resveratrol.

6 Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00

Comparison Between In Vitro and In Vivo miRNA Expression ProfilingmiRNA profiling was used to compare in vitro and in vivo mod-els. The age-associated changes in miRNA profiles of neuroglial cultures were compared with those occurred in the cortex of aged C57Bl/6J mice (18 and 24 months of age) and TgCRND8 mice aged 4–6 months (corresponding to a middle stage of Aβ depo-sition). A total of 85 miRNAs resulted in being downregulated (with an FC < 0.5) in 27 DIV cultures and in TgCRND8 mice, compared with their respective young counterparts (15 DIV and wild-type mice). Instead, only eight miRNAs were downregu-lated both in 27 DIV and in 24-month-old C57Bl/6J mice, and 17 miRNAs in 27 DIV and in 18-month-old C57Bl/6J mice, each compared with their respective young counterparts (15 DIV and young mice). These data and the degree of overlap among the different groups are represented in the Venn diagram in Figure 4, panel B.

Interestingly, three miRNAs downregulated in 27 DIV were simul-taneously downregulated in 18- and 24-month-old C57Bl/6J mice and in TgCRND8 mice (mir-7b-3p, mir-452-5p, mir-290-5p). We found mir-29a and mir-29b, mir-466, mir-467, and mir-669 among the miR-NAs exclusively downregulated in 27 DIV and the TgCRND8 mouse cortex. Putative gene targets for some of these downregulated miR-NAs were identified by DIANA-microT and are shown in Table 2.

DIANA-miRPath v2.0 analysis identified several pathways as significantly enriched (p < .01) for the 85 miRNAs downregulated in 27 DIV cultures and in TgCRND8 mice. As it can be seen in Supplementary Table 1, many of the identified pathways are catego-rized under core biological processes in the brain, such as neurotro-phin signaling, axon guidance pathway, dopaminergic, GABAergic, and glutamatergic transmission as well as long-term depression and potentiation. Moreover, mTOR and the PI3K-Akt signaling pathway were also significantly modulated. The list of miRNAs belonging to these pathways is available in Supplementary Table 2.

Figure 3. Panel A: GFAP protein expression in neuroglial coculture at 15, 20, 23, 27 days in vitro (DIV) and in 27 DIV cells treated with 1 and 5 µM resveratrol (mean ± SE of three experiments). *p < .05 vs 15 DIV; #p < .05 vs 27 DIV. Panel B: representative immunofluorescence images of GFAP (red) in neuroglial cocultures at 15, 27 DIV and in 27 DIV cells treated with 1 and 5 µM resveratrol. Panel C: carbonyl residues in neuroglial cocultures at 15, 20, 23, 27 DIV and in 27 DIV cells treated with 1 and 5 µM resveratrol (mean ± SE of three experiments). *p < .05 vs 15 DIV; **p < .01 vs 15 DIV; #p < .05 vs 27 DIV. Panel D: representative dot blot images of hydroxynonenal (marker of lipid peroxidation) in neuroglial cocultures at 15, 27 DIV, and 27 DIV treated with 1 and 5 µM resveratrol. Panel E: IL-6 levels in the culture media of neuroglial cocultures at 15, 27 DIV, and 27 DIV treated with 1 and 5 µM resveratrol. **p < .01 vs 15 DIV; ##p < .01 vs 27 DIV.

Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00 7

miRNA Expression by qRT–PCR AnalysisqRT–PCR confirmed the downregulation (indicated as FC) of miR-17 (FC 27 DIV/15 DIV: 0.09), miR-124 (FC 27 DIV/15 DIV: 0.115), let-7a (FC 27 DIV/15 DIV: 0.27), mir-29b (FC 27 DIV/15 DIV: 0.35). Given that mir-29b was among the miRNAs downregu-lated in 27 DIV cultures and in TgCRND8 mice, its downregulation was also confirmed in the cortex of TgCRND8 mice with an FC TgCRND8/young of 0.27. mir-34a was also chosen as one of the nonaffected miRNAs and also in this case, qRT–PCR confirmed the microarray results demonstrating that this miRNA is expressed only at very low levels and that it is not modified in senescent cultures compared with the young ones. With the same approach, we also confirmed the effect of resveratrol in reverting the downregulation observed in 27 DIV for the above-mentioned miRNAs: mir-17: FC 27 DIV Resv 1 µM/15 DIV: 1.29 and FC 27 DIV Resv 5 µM/15 DIV: 1.52; miR-124: FC 27 DIV Resv 1 µM/15 DIV: 3.2 and FC 27 DIV Resv 5 µM/15 DIV: 3.7; let-7a: FC 27 DIV Resv 1 µM/15 DIV: 4.2 and FC 27 DIV Resv 5 µM/15 DIV: 5.2; mir-29: FC 27 DIV Resv 1 µM/15 DIV: 1.8 and FC 27 DIV Resv 5 µM/15 DIV: 4.4.

Discussion

Neurons are particularly vulnerable to age-related changes that affect their function and contribute to the onset of age-related neurodegen-erative pathologies. Neuroglial cocultures are an integrated model accounting for the reciprocal and profound interactions between neurons and glia, currently used only for short-term experiments. Here, we studied the temporal evolution and senescence of this integrated system. We observed a significant variability in obtaining long-term cell survival in vitro, possibly due to variations in cell iso-lation procedures or to the frequent medium changes. However, once the system is functioning, it can provide useful information. In fact, we were able to demonstrate that a relatively long period in culture determines the appearance of a senescent phenotype, offering the possibility of investigating age-related changes. Interestingly, neuro-glial cocultures show some of the senescence features observed in

Table 2. List of miRNAs Downregulated Both in 27 DIV, Compared With 15 DIV Cells, and in TgCRND8 Mice Compared With Wild-Type Mice, With an FC < 0.5, Their Predicted Target Gene, and Biological Significance

miRNA Predicted Gene Target

Biological Significance

mir-7b-3p histone H3.3, PAI-1 Regulation of chromatin, Senescence

mir-452-3p GLUD1 Loss of neurons, dendritic spines

mir-290-5p Histone demetilase UTX

Aging

mir-29/mir-328 BACE-1 Amiloid precursor protein (APP) processing

mir-669h-3p Contactin 4 Axon guidancemir-466m-5p APP Amyloid plaquesmir-467 HIF-1, E2f7 Hypoxia, p53-dependent

cell cycle arrest/ senescence

mir-669c-3p Beclin 1 Autophagic processmir-181b/d-5p HMGB-1, IL-6 Stress response in

astrocytes, inflammation

Notes: DIV = days in vitro; FC = fold-change; miRNA = microRNA.

Table 1. List of Downregulated miRNAs in 27 DIV Compared With 15 DIV Cells, With an FC < 0.5, Their Predicted Target Gene, and Bio-logical Significance

miRNA Predicted Gene Target

Biological Significance

miR-17-92 p21 Cell cycle regulation, senescence, apoptosis

mir-669c-3p GLB1 Marker of senescencemir-466 ATM DNA damage response

cascademir-24-3p/mir-124-3p H2AX Senescence, DNA

damagemir-7a-5p PARP-1 Senescence,

DNA damage, neurodegeneration

miR-149-3p GFAP Astrocyte activationmiR-188-5p Nrp-2 Dendritic spine

reductionlet-7 IL6 InflammationmiR-23b Apaf-1 Neuronal deathmir-26b-5p E2F7 Senescence

Notes: DIV = days in vitro; FC = fold-change; miRNA = microRNA.

Figure  4. Panel A: Unsupervised hierarchical clustering of microRNA (miRNA) expression profiles in neuroglial cocultures at 15, 27 days in vitro (DIV), and 27 DIV treated with 1 and 5 µM resveratrol. Each row represents a different miRNA; the color code indicates unexpressed miRNAs (gray) or expressed miRNAs (blue). Panel B: Venn diagram showing the number of miRNAs downregulated by aging (fold change [FC] < 0.5) and the degree of overlapping between the different models. Top left and right circles: miRNAs downregulated both in senescent neuroglial cocultures and in the cerebral cortex of aging C57Bl/6J mice (18- and 24-month-old), each compared with the respective young controls; bottom circle: miRNAs downregulated both in senescent neuroglial cocultures and in TgCRND8 mice cortex, each compared with the respective young/wild-type controls.

8 Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00

proliferating peripheral cells such as the progressive accumulation of SA-β-gal, as demonstrated by others in pure neuronal cells (5,6,28).

Resveratrol induces a strong, concentration-dependent decrease in the number of cells positive for SA-β-gal when administered to neuroglial cocultures from 15 to 27 DIV. We reported similar effects in human fibroblasts undergoing replicative senescence (24). Increased SA-β-gal activity, together with an increased expression of γ-H2AX, has also been observed in the brain of old mice (10). The cerebral cortex is a major site of γ-H2AX expression in both neurons and astrocytes in senescent mice (29). Interestingly, we observed a progressive increase in γ-H2AX protein expression in our neuroglial cocultures, suggesting that DNA damage response activation takes place in neurons also in vitro. Resveratrol reduced γ-H2AX protein levels as in proliferating cell lines and in human lymphocytes (30).

Neuroglial cocultures show a time-related increase in protein oxidative damage, described also in aging neurons (9); the increase in carbonyl residues is considered one of the contributors to age-related neurodegenerative diseases (31). Instead, no difference in nitrotyrosine levels was found upon senescence in vitro, in agree-ment with previous work in a similar model (32), showing that some proteins accumulate nitration, whereas others show a reduced nitra-tion profile in cortical neurons. In vivo, nitrotyrosine was found to be both decreased (33) and increased (34) in aged rats.

Senescent cultures are also characterized by a strong increase in the production of the proinflammatory cytokine, IL-6, which is completely counteracted by resveratrol. Several reports indicate that senescent cells express a senescence-associated secretory phenotype consisting of an increased secretion of growth factors, inflamma-tory cytokines, and proteolytic enzymes and involved in microen-vironment remodeling (35). As a whole, neuroglial cocultures show features typical of in vivo aging and a well-established antiaging molecule, such as resveratrol, reverses these changes.

Besides recapitulating some characteristics of cellular senes-cence, neuroglial cocultures also display age-associated degenera-tive changes such as progressive neuronal loss and a decline in the dendritic network. A  reduction in synaptic and dendritic markers has been described in hippocampal neurons aged in vitro (7). Dong and colleagues (28) reported that the number of neurons decreases after DIV 25 in vitro with a concomitant increase in SA-β-gal. In the present experiments, we demonstrate that changes in neurons are accompanied by activation of glial cells, resulting in hypertrophy of astrocytes and age-associated gliosis.

The activation of astrocytes produces a state of low-level inflam-mation, playing an important role both in aging and in age-related neurological diseases (36). Aged astrocytes do not support neuronal survival in mixed neuronal–glial cultures as efficiently as young cells (37). In vivo data also indicate that aged astrocytes show features of the senescence-associated secretory phenotype, as described in fibro-blast and endothelial cells (19). We can thus conclude that in neuro-glial cocultures in vitro, some neurons undergo senescence, whereas others are progressively lost, as indicated by the time-dependent decrease in markers of neuronal nuclei and synaptic contacts. In this context, it is feasible that the activated astrocytes lose their function of support for neurons.

The antiaging effect of resveratrol has been widely explored in both cellular and animal models (22,24). Resveratrol has been reported as neuroprotective in rodent models of neurodegeneration and central nervous system injury (38). Resveratrol also prevents age-related cognitive decline in rodent models (39) and humans (40). In our model, resveratrol failed to counteract the senescence-dependent loss of neuronal cell bodies and connections but exerted

a concentration-dependent effect in reducing astrogliosis. An analo-gous effect was reported by Genade and Lang (41) who found that resveratrol prevented the age-associated dysregulation of GFAP expression without avoiding the loss of neurons in the midbrain of the fast-aging Nothobranchius guentheri fish.

Alterations of senescent cells are accompanied by changes in miRNA expression patterns. miRNAs act as master negative regu-lators of gene expression and a single miRNA can target many mRNAs in parallel, thus having the potential to simultaneously modify multiple cell pathways such as proliferation, differentiation, senescence, and death.

We observed a substantial miRNA downregulation in neuroglial cocultures aged in vitro compared with their younger counterparts. A  general decline in miRNA expression was previously reported during the aging process in peripheral proliferating cells such as fibroblasts (42) and in peripheral blood mononuclear cells from elderly individuals (43,44). These changes were not accompanied by modifications of Dicer expression in senescent cultures, suggest-ing that miRNA downregulation represents a general mechanism in senescent cultures rather than a technical artifact. Several miR-NAs downregulated in 27 DIV neuroglial coculture such as miR-17, miR-19b, miR-20a, and miR-106b were previously found to be less abundant in cells from elderly humans. The miR-17-92 cluster has been reported to target several proteins involved in cell cycle regula-tion (such as p21) and is downregulated in human replicative (45) and stress-induced senescence (46). The overall effect of resveratrol was to shift the miRNA expression pattern of senescent neuroglial cocultures toward that observed in mature neuroglial cocultures (15 DIV). This observation is of particular interest and may illustrate new mechanisms by which resveratrol exerts its antiaging effects.

Focusing on those miRNAs that target genes encoding for pro-teins whose expression was actually measured in long-term neu-roglial cocultures, we were able to confirm most of the predicted miRNA–mRNA interactions listed in Table 1. In fact, the upregula-tion of the genes encoding for GFAP, IL-6, and H2AX, predicted on the basis of the downregulation of the respective targeting miRNAs 149-3p, let-7a/b/c/g-5p, and 124-3p/24-3p, has been actually con-firmed by the Western blot data, showing increased levels of these three proteins. Similarly, the predicted downregulation of Nrp-2 and Apaf-1 genes, encoding for proteins involved in the dendritic spine reduction and neuronal death, is supported by the finding that MAP2 and NeuN proteins are downregulated in the senescent cultures.

The gene encoding for β-galactosidase, GLB, was also predicted by the DIANA-microT to be upregulated, based on the downregu-lation of mir-669c-3p in aged cocultures, and this is in agreement with the increased β-galactosidase activity found at senescence. Interestingly, the expression of this family of miRNAs was restored by resveratrol, which also drastically reduced β-galactosidase stain-ing in aged cultures.

Thus, these data disclose previously unknown putative miRNA–mRNA relationships and suggest a functional role for these interac-tions in the frame of the senescence process.

By comparing the age-associated changes in the expression pro-files of miRNAs in the in vitro model with those measured in the cortex of physiologically aged 18- and 24-month-old mice, we found only a few similarities such as the downregulation of mir-7b-3p, mir-452-5p, and mir-290-5p. These were not exclusively related to aging, being observed also in the cortex of TgCRND8 mice. Interestingly, DIANA-miRPath analysis indicates that miR-7b-3p and miR-452-5p are involved in mTOR and HIF-1 signaling pathways, key modula-tors of aging, and age-related disease (47,48).

Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00 9

The expression profiles of miRNAs in the cellular model appear to resemble more closely those of the TgCRND8 mice cortex, an in vivo model of pathological aging and neurodegeneration. Among the 67 miRNAs downregulated by aging both in long-term neuroglial cocultures and in TgCRND8 mice cortex, the miR-29 family mem-bers were previously reported to have target sites on the gene encod-ing for β-amyloid cleavage enzyme 1 (BACE-1), required for the generation of toxic Aβ species (49). Overall, the DIANA-miRPath pathway enrichment analysis indicated that the most significant set of miRNAs deregulated both in senescent neuroglial cocultures and in TgCRND8 mice cortex is involved in pathways related to the con-trol of many aspects of survival, development, function, and senes-cence of neurons.

In conclusion, we demonstrate that neuroglial cocultures show many features of cellular senescence and respond to the exposure to a known antiaging/neuroprotective substance such as resveratrol. The comparison of global miRNA expression in vitro and in vivo suggests that neuroglial cocultures resemble pathological more than normal brain aging. These findings may provide insights into the epi-genetic regulatory mechanisms of age-induced brain alterations and help identify miRNAs involved in these processes. Thus, this model can be useful for investigating the cellular and molecular mecha-nisms involved in neuronal aging and for the screening of antiaging/neuroprotective compounds.

Supplementary Material

Supplementary Material can be found at: http://biomedgerontology.oxfordjournals.org/

Funding

Supported by the University of Florence (519ATEN389), and by the Ente Cassa di Risparmio di Firenze (519CRF057).

AcknowledgmentsThe authors thank Dr. L.  Cinci for counseling on layout and formatting of the results, Dr. C.  Grossi for TgCRND8 mice care, Dr. K.  Tortora for collaboration in performing the experiments, Dr. E.  Bonechi for cytokine determination. The authors sincerely thank Mrs. Mary Forrest for linguistic revision of the manuscript. L.G., P.D., and D.E.P.-G. conceived and designed the study; E.B.  performed the experiments and wrote the manuscript; C.L. performed the microarray experiments and analyzed the data; T.S. iso-lated and treated neuroglial cocultures; F.C. provided the TgCRND8 mice samples; all the authors revised and approved the manuscript; all authors agreed to be responsible for any aspect regarding the accuracy and integrity of the entire work.

Conflict of Interest

The authors declare that they have no conflicting interests.

References 1. López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. The

hallmarks of aging. Cell. 2013;153:1194–1217. doi:10.1016/j.cell. 2013.05.039

2. Campisi J. Cellular senescence: putting the paradoxes in perspective. Curr Opin Genet Dev. 2011;21:107–112. doi:10.1016/j.gde.2010.10.005

3. Mocali A, Giovannelli L, Dolara P, Paoletti F. The comet assay approach to senescent human diploid fibroblasts identifies different phenotypes and

clarifies relationships among nuclear size, DNA content, and DNA dam-age. J Gerontol A Biol Sci Med Sci. 2005;60:695–701. PMID:15983170

4. d’Adda dF. Living on a break: cellular senescence as a DNA-damage response. Nat Rev Cancer. 2008;8:512–522. doi:10.1038/nrc2440

5. Bhanu MU, Mandraju RK, Bhaskar C, Kondapi AK. Cultured cerebel-lar granule neurons as an in vitro aging model: topoisomerase IIβ as an additional biomarker in DNA repair and aging. Toxicol In Vitro. 2010;24:1935–1945. doi:10.1016/j.tiv.2010.08.003

6. Geng YQ, Guan JT, Xu XH, Fu YC. Senescence-associated beta-galactosi-dase activity expression in aging hippocampal neurons. Biochem Biophys Res Commun. 2010;396:866–869. doi:10.1016/j.bbrc.2010.05.011

7. Bertrand SJ, Aksenova MV, Aksenov MY, Mactutus CF, Booze RM. Endogenous amyloidogenesis in long-term rat hippocampal cell cultures. BMC Neurosci. 2011;12:38. doi:10.1186/1471-2202-12-38

8. Masliah E, Crews L, Hansen L. Synaptic remodeling during aging and in Alzheimer’s disease. J Alzheimers Dis. 2006;9(3 suppl):91–99. PMID:16914848

9. Aksenova MV, Aksenov MY, Markesbery WR, Butterfield DA. Aging in a dish: age-dependent changes of neuronal survival, protein oxidation, and creatine kinase BB expression in long-term hippocampal cell culture. J Neurosci Res. 1999;58:308–317. PMID:10502287

10. Jurk D, Wang C, Miwa S, et al. Postmitotic neurons develop a p21-depend-ent senescence-like phenotype driven by a DNA damage response. Aging Cell. 2012;11:996–1004. doi:10.1111/j.1474-9726

11. Liang R, Bates DJ, Wang E. Epigenetic control of micro-RNA expression and aging. Curr Genomics. 2009;10:184–193. doi:10.2174/138920209788185225

12. Gascon E, Gao FB. Cause or effect: misregulation of microRNA path-ways in neurodegeneration. Front Neurosci. 2012;6:48. doi:10.3389/fnins.2012.00048

13. Somel M, Guo S, Fu N, et al. MicroRNA, mRNA, and protein expression link development and aging in human and macaque brain. Genome Res. 2010;20:1207–1218. doi:10.1101/gr.106849.110

14. Li N, Bates DJ, An J, Terry DA, Wang E. Up-regulation of key microRNAs, and inverse down-regulation of their predicted oxidative phosphorylation target genes, during aging in mouse brain. Neurobiol Aging. 2011;32:944–955. doi:10.1016/j.neurobiolaging.2009.04.020

15. Khanna A, Muthusamy S, Liang R, Sarojini H, Wang E. Gain of survival signaling by down-regulation of three key miRNAs in brain of calorie-restricted mice. Aging (Albany NY). 2011;3:223–236. PMID:21415464

16. Liu QY, Chang MN, Lei JX, et al. Identification of microRNAs involved in Alzheimer’s progression using a rabbit model of the disease. Am J Neuro-degener Dis. 2014;3:33–44. PMID:24754001

17. Zovoilis A, Agbemenyah HY, Agis-Balboa RC, et  al. microRNA-34c is a novel target to treat dementias. EMBO J. 2011;30:4299–4308. doi:10.1038/emboj.2011.327

18. Lippi G, Steinert JR, Marczylo EL, et  al. Targeting of the Arpc3 actin nucleation factor by miR-29a/b regulates dendritic spine morphology. J Cell Biol. 2011;194:889–904. doi:10.1083/jcb.201103006

19. Salminen A, Ojala J, Kaarniranta K, Haapasalo A, Hiltunen M, Soininen H. Astrocytes in the aging brain express characteristics of senescence-asso-ciated secretory phenotype. Eur J Neurosci. 2011;34:3–11. doi:10.1111/j.1460-9568.2011.07738.x

20. Pitozzi V, Jacomelli M, Catelan D, et  al. Long-term dietary extra-virgin olive oil rich in polyphenols reverses age-related dysfunctions in motor coordination and contextual memory in mice: role of oxidative stress. Rejuvenation Res. 2012;15:601–612. doi:10.1089/rej.2012.1346

21. Bellucci A, Luccarini I, Scali C, et al. Cholinergic dysfunction, neuronal damage and axonal loss in TgCRND8 mice. Neurobiol Dis. 2006;23:260–272. PMID:16766197

22. Baur JA, Pearson KJ, Price NL, et  al. Resveratrol improves health and survival of mice on a high-calorie diet. Nature. 2006;444:337–342. PMID:17086191

23. Pellegrini-Giampietro DE, Peruginelli F, Meli E, et  al. Protection with metabotropic glutamate 1 receptor antagonists in models of ischemic neuronal death: time-course and mechanisms. Neuropharmacology. 1999;38:1607–1619. PMID:10530822

10 Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00

24. Giovannelli L, Pitozzi V, Jacomelli M, et al. Protective effects of resvera-trol against senescence-associated changes in cultured human fibroblasts. J Gerontol A Biol Sci Med Sci. 2011;66:9–18. doi:10.1093/gerona/glq161

25. Pitozzi V, Mocali A, Laurenzana A, et al. Chronic resveratrol treatment ame-liorates cell adhesion and mitigates the inflammatory phenotype in senes-cent human fibroblasts. J Gerontol A Biol Sci Med Sci. 2013;68:371–381. doi:10.1093/gerona/gls183

26. Correa-Salde V, Albesa I. Reactive oxidant species and oxidation of protein and haemoglobin as biomarkers of susceptibility to stress caused by chlo-ramphenicol. Biomed Pharmacother. 2009;63:100–104. doi:10.1016/j.biopha.2008.05.001

27. Musolino C, Allegra A, Saija A, et  al. Changes in advanced oxidation protein products, advanced glycation end products, and s-nitrosylated proteins, in patients affected by polycythemia vera and essential throm-bocythemia. Clin Biochem. 2012;45:1439–1443. doi:10.1016/j.clinbio-chem.2012.07.100

28. Dong W, Cheng S, Huang F, et  al. Mitochondrial dysfunction in long-term neuronal cultures mimics changes with aging. Med Sci Monit. 2011;17:BR91–BR96. PMID:21455101

29. Barral S, Beltramo R, Salio C, Aimar P, Lossi L, Merighi A. Phosphoryla-tion of histone H2AX in the mouse brain from development to senescence. Int J Mol Sci. 2014;15:1554–1573. doi:10.3390/ijms15011554

30. Halicka HD, Zhao H, Li J, et al. Potential anti-aging agents suppress the level of constitutive mTOR- and DNA damage- signaling. Aging (Albany NY). 2012;4:952–965. PMID:23363784

31. Rahmadi A, Steiner N, Munch G. Advanced glycation endproducts as ger-ontotoxins and biomarkers for carbonyl-based degenerative processes in Alzheimer’s disease. Clin Chem Lab Med. 2011;49:385–391. doi:10.1515/CCLM.2011.079

32. Lesuisse C, Martin LJ. Long-term culture of mouse cortical neurons as a model for neuronal development, aging, and death. J Neurobiol. 2002;51:9–23. PMID:11920724

33. Cakatay U, Telci A, Kayalì R, Tekeli F, Akçay T, Sivas A. Relation of oxida-tive protein damage and nitrotyrosine levels in the aging rat brain. Exp Gerontol. 2001;36:221–229. PMID:11226738

34. Shin CM, Chung YH, Kim MJ, Lee EY, Kim EG, Cha CI. Age-related changes in the distribution of nitrotyrosine in the cerebral cortex and hip-pocampus of rats. Brain Res. 2002;931:194–199. PMID:11897106

35. Coppé JP, Desprez PY, Krtolica A, Campisi J. The senescence-associated secretory phenotype: the dark side of tumor suppression. Annu Rev Pathol. 2010;5:99–118. doi:10.1146/annurev-pathol-121808-102144

36. Middeldorp J, Hol EM. GFAP in health and disease. Prog Neurobiol. 2011;93:421–443. doi:10.1016/j.pneurobio.2011.01.005

37. Pertusa M, Garcia-Matas S, Rodriguez-Farre E, Sanfeliu C, Cristofol R. Astrocytes aged in vitro show a decreased neuroprotective capacity. J Neu-rochem. 2007;101:794–805. PMID:17250685

38. Huber K, Superti-Furga G. After the grape rush: sirtuins as epigenetic drug targets in neurodegenerative disorders. Bioorg Med Chem. 2011;19:3616–3624. doi:10.1016/j.bmc.2011.01.018

39. Oomen CA, Farkas E, Roman V, van der Beek EM, Luiten PG, Meerlo P. Resveratrol preserves cerebrovascular density and cognitive func-tion in aging mice. Front Aging Neurosci. 2009;1:4. doi:10.3389/neuro.24.004.2009

40. Kennedy DO, Wightman EL, Reay JL, et  al. Effects of resveratrol on cerebral blood flow variables and cognitive performance in humans: a double-blind, placebo-controlled, crossover investigation. Am J Clin Nutr. 2010;91:1590–1597. doi:10.3945/ajcn.2009.28641

41. Genade T, Lang DM. Resveratrol extends lifespan and preserves glia but not neurons of the Nothobranchius guentheri optic tectum. Exp Gerontol. 2013;48:202–212. doi:10.1016/j.exger.2012.11.013

42. Brosh R, Shalgi R, Liran A, et  al. p53-Repressed miRNAs are involved with E2F in a feed-forward loop promoting proliferation. Mol Syst Biol. 2008;4:229. doi:10.1038/msb.2008.65

43. Noren Hooten N, Abdelmohsen K, Gorospe M, Ejiogu N, Zonderman AB, Evans MK. microRNA expression patterns reveal differential expression of target genes with age. PLoS One. 2010;5:e10724. doi:10.1371/journal.pone.0010724

44. ElSharawy A, Keller A, Flachsbart F, et al. Genome-wide miRNA signa-tures of human longevity. Aging Cell. 2012;11:607–616. doi:10.1111/j.1474-9726.2012.00824.x

45. Hackl M, Brunner S, Fortschegger K, et  al. miR-17, miR-19b, miR-20a, and miR-106a are down-regulated in human aging. Aging Cell. 2010;9:291–296. doi:10.1111/j.1474-9726.2010.00549.x

46. Li G, Luna C, Qiu J, Epstein DL, Gonzalez P. Alterations in microRNA expression in stress-induced cellular senescence. Mech Ageing Dev. 2009;130:731–741. doi:10.1016/j.mad.2009.09.002

47. Johnson SC, Rabinovitch PS, Kaeberlein M. mTOR is a key modulator of ageing and age-related disease. Nature. 2013;493:338–345. doi:10.1038/nature11861

48. Ogunshola OO, Antoniou X. Contribution of hypoxia to Alzheimer’s dis-ease: is HIF-1alpha a mediator of neurodegeneration? Cell Mol Life Sci. 2009;66:3555–3563. doi:10.1007/s00018-009-0141-0

49. Hébert SS, Horré K, Nicolaï L, et  al. Loss of microRNA clus-ter miR-29a/b-1 in sporadic Alzheimer’s disease correlates with increased BACE1/beta-secretase expression. Proc Natl Acad Sci USA. 2008;105:6415–6420. doi:10.1073/pnas.0710263105

Journals of Gerontology: BIOLOGICAL SCIENCES, 2015, Vol. 00, No. 00 11