Inflammation leads to distinct populations of extracellular vesicles from microglia

Inflammation leads to distinct populations of extracellular vesicles from microglia Background: Activated microglia play an essential role in inflammatory responses elicited in the central nervous system (CNS). Microglia-derived extracellular vesicles (EVs) are suggested to be involved in propagation of inflammatory signals and in the modulation of cell-to-cell communication. However, there is a lack of knowledge on the regulation of EVs and how this in turn facilitates the communication between cells in the brain. Here, we characterized microglial EVs under inflammatory conditions and investigated the effects of inflammation on the EV size, quantity, and protein content. Methods: We have utilized western blot, nanoparticle tracking analysis (NTA), and mass spectrometry to characterize EVs and examine the alterations of secreted EVs from a microglial cell line (BV2) following lipopolysaccharide (LPS) and tumor necrosis factor (TNF) inhibitor (etanercept) treatments, or either alone. The inflammatory responses were measured with multiplex cytokine ELISA and western blot. We also subjected TNF knockout mice to experimental stroke (permanent middle cerebral artery occlusion) and validated the effect of TNF inhibition on EV release. Results: Our analysis of EVs originating from activated BV2 microglia revealed a significant increase in the intravesicular levels of TNF and interleukin (IL)-6. We also observed that the number of EVs released was reduced both in vitro and in vivo when inflammation was inhibited via the TNF pathway. Finally, via mass spectrometry, we identified 49 unique proteins in EVs released from LPS-activated microglia compared to control EVs (58 proteins in EVs released from LPS- activated microglia and 37 from control EVs). According to Gene Ontology (GO) analysis, we found a large increase of proteins related to translation and transcription in EVs from LPS. Importantly, we showed a distinct profile of proteins found in EVs released from LPS treated cells compared to control. Conclusions: We demonstrate altered EV production in BV2 microglial cells and altered cytokine levels and protein composition carried by EVs in response to LPS challenge. Our findings provide new insights into the potential roles of EVs that could be related to the pathogenesis in neuroinflammatory diseases. Keywords: Microglia, Extracellular vesicles (EVs), Neuroinflammation, TNF Background [3]. Emerging evidence has shown that microglia are key Microglia are considered the main innate-immune cells of causative players in neuroinflammation, which in turn is the central nervous system (CNS). They continuously sur- believed to play a major role in neurodegenerative vey their microenvironment and have the ability to interact diseases [4]. with neurons to regulate their activity [1]. In the healthy Microglia are highly dynamic cells with the ability to brain, microglia continuously survey their surroundings transform their morphology from ramified to amoeboid with highly dynamic processes [2] and become activated in and alter their phenotypes corresponding to diverse con- response to injury, infection or neurodegenerative processes ditions. Traditionally, macrophages and microglial cells are classified into two different phenotypes, M1 and M2. M1-microglia are proinflammatory, secreting inflamma- * Correspondence: yiyi.yang@med.lu.se; tomas.deierborg@med.lu.se tory cytokines, chemokines, and nitric oxide (NO), Department of Experimental Medical Science, Experimental which is believed to result in neuronal dysfunction and Neuroinflammation Laboratory, Lund University, Lund, Sweden Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 2 of 19 accelerate the progression of neurodegenerative diseases, changed after bacterial infection [17]. Thus, there is a such as Alzheimer’s disease (AD) and Parkinson’s disease critical need for both identification of specific markers (PD). In contrast, M2-microglia are believed to have and particular signaling pathways controlling EV traf- neuroprotective functions, including increased produc- ficking. The mechanism of action of EVs in microglial tion of interleukin (IL)-4 and neurotrophic factors, along communication is poorly understood. In this study, we with an increase in phagocytosis which in turn leads to hypothesized that activation of microglia can secrete a clearance of cell debris and tissue damage [3, 5]. How- distinct population of EV through modulation of specific ever, there is a wide spectrum of microglial activation signaling pathways. We investigated the dynamics of EVs between the two defined phenotypes [6], and microglia from activated microglial (BV2) cells subjected to lipo- might even have specific neuronal functions beyond typ- polysaccharide (LPS) stimulation. We used differential ical pro-/anti-inflammatory responses [7]. A better un- ultracentrifugation to isolate EVs, including microvesi- derstanding of the interactions between microglia and cles and exosomes. EVs were then characterized in terms other cells in the brain is therefore needed in order to of size and concentration by nanoparticle tracking design therapies to ameliorate the detrimental effects of analysis (NTA), while the origin of EVs was indicated by microglial reactions in brain diseases. western blotting using antibodies against CD63, flotillin-1, The main interaction between cells occur through cel- and Alix. Importantly, we also analyzed the levels of in- lular signaling pathways including autocrine, paracrine, flammatory cytokines in EVs. Secretion of EVs was altered and endocrine processes [8] and extracellular vesicles by suppression of inflammation in microglia via inhibition (EVs) can be important to transport signals between of tumor necrosis factor (TNF) signaling in vivo and in cells. In fact, increasing evidence has shown EVs are vitro. Subsequently, qualitative proteomic analysis was considered one of the main participants in cell-to-cell performed to reveal a different protein composition of communication along with having a proposed role in the EVs in response to LPS challenge. Taken together, our spread of pathology in neurodegenerative disease [9, 10]. findings provide new insights into the role of EVs in regu- These vesicles are able to carry pathogen-associated and lating microglial cell communication. damage-associated molecular patterns that act as signals to regulate and propagate the inflammatory response Methods [11–13]. Hence, investigation of EV trafficking under in- Cell culture flammatory conditions may broaden our understanding BV2, an immortalized murine microglial cell line, was of the roles of microglia in neurodegenerative diseases, cultured in growing medium containing Dulbecco’s as well as their potential in therapeutic manipulation. modified Eagle medium (DMEM) (Gibco™GlutaMAX™, The secretion of EVs is a highly conserved process Thermo Fisher Scientific) supplemented with 10% [14]. However, a number of studies using proteomic ana- heat-inactivated fetal bovine serum (FBS) and 1% peni- lysis of EVs released by various cell types, including cillin/streptomycin (Thermo Fisher Scientific) in 5% microglia, have revealed a diverse range of markers and CO in air at 37 °C in a humidified incubator. Cells were alteration of protein composition [15, 16]. The lack of re-cultured every 2 days starting at a concentration of knowledge on EV’s regulation in vitro and in vivo halts a 2×10 cells/ml in T75 flask (Sarstedt). For a large scale clear understanding of EV functions in cell-to-cell com- of EV collection, microglia were plated in T175 flask (Sar- munication. It is likely that different subsets of EVs have stedt). For inflammatory activation, cells were challenged different functional properties, and trafficking of EVs is with 1 μg/ml LPS (Sigma-Aldrich, Clony 0127-B8) for 12 h most likely modulated by specific signaling pathways. and then grown for 12 h in serum-free media prior to col- Although consensus within the field is being reached, lection of EVs. For TNF inhibition experiment, microglia the classification of EVs is not easy. While different sub- were plated in growing medium either with 1 μg/ml LPS, sets of EVs are being described with increasing rate, for 200 ng/ml etanercept, or both for 12 h. EVs were collected simplicity, we will focus on two different classes of EVs from serum-free media 12 h after treatment. of different sizes and origins. EVs shed directly from the plasma membrane are characterized as microvesicles or Animals ectosomes ranging from 100 to 1000 nm. Exosomes are Adult male C57BL/6 mice (between 7 and 8 weeks of generated within the endosomal pathway and terminate age, n = 20) were purchased from Taconic Ltd. (Ry, at the multivesicular endosomal body (MVB), whereby Denmark) and transferred to the Laboratory of Biomedi- they are released upon the MVB fusing with the plasma cine, University of Southern Denmark, where they were membrane. Generally, the size of exosomes is smaller allowed to acclimatize for 7 days prior to surgery. TNF than microvesicles and below 100 nm. knockout (TNF-KO) breeding couples were originally A recent study has shown that the size distribution purchased from The Jackson Laboratory and transferred and protein composition of EVs in macrophages can be to the Laboratory of Biomedicine where they were Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 3 of 19 established as a colony. Animals were housed under di- Measurement of extracellular vesicles size by nanoparticle urnal lighting conditions and given free access to food tracking analysis (NTA) and water [18]. All animal experiments were performed The size and total number of EVs were measured by in accordance with the relevant guidelines and regula- using NanoSight LM10 (Malvern, UK) with the technol- tions approved by the Danish Animal Ethical Committee ogy of Nanoparticle Tracking Analysis (NTA). In liquid (numbers 2011/561-1950 and 2013-15-2934-00924). suspension, particles undergo Brownian motion together with light scattering properties, the size distribution and concentration of EVs samples can be obtained [17]. Sam- Induction of experimental stroke, permanent middle ples were diluted with distilled water to obtain optimal 6 9 cerebral artery occlusion (pMCAO) concentration for detection (10 –10 particles/ml) and The distal part of the left middle cerebral artery was per- injected with a continuous syringe system for 30 s × 5 manently occluded under Hypnorm and Dormicum times at speed 50 μl/min. Data acquisition was undertaken anesthesia (fentanyl citrate (0.315 mg/ml; Jansen-Cilag) at ambient temperature and measured 5 times by NTA. and fluanisone (10 mg/ml; Jansen-Cilag, Birkerød, Data were analyzed with NTA 2.2 software (Malvern, UK) Denmark), and midazolam (5 mg/ml; Hoffmann- La with minimum expected particle size 10 nm. Roche, Hvidovre, Denmark)), respectively. After surgery, mice were injected subcutaneously with 1 ml of 0.9% sa- Western blot analysis line and allowed to recover in a 25 °C controlled envir- Cell pellets and EVs were lysed in RIPA buffer (Sigma-Al- onment. Mice surviving for 5 days were returned to the drich) supplemented with proteinase inhibitors (Thermo conventional animal facility after 24 h. For post-surgical Scientific) and PhosphoStop (Roche Diagnostics GmbH). analgesia, mice were treated with 0.001 mg/20 g bupre- The concentration of cell lysates was determined using norphine hydrochloride (Temgesic, Schering-Plough, bicinchoninic acid assay (BCA) (Thermo Scientific), while Ballerup, Denmark) three times at 8-h intervals, starting concentrations obtained using NanoSight were utilized to immediately prior to surgery. Mice were allowed to sur- ensure even loading of EVs. Samples were loaded onto 4– vive for 1 day (immunofluorescent staining and cytokine 20% Mini-Protean TGX Precast Gels (Bio-Rad) and then measurement) or 5 days (EV analysis) whereafter they transferred to Nitrocellulose membranes (Bio-Rad) using were killed using either an overdose of pentobarbital Trans-Blot Turbo System (Bio-Rad). Membranes were (200 mg/ml) containing lidocaine (20 mg/ml) (Glostrup incubated with following primary antibodies: Alix (Cell Apotek, Glostrup, Denmark) and perfused through the Signaling; 1:1000), flotillin-1 (Cell Signaling; 1:1000), left ventricle using 4% paraformaldehyde (PFA) or killed CD63 (Santa Cruz Biotechnology; 1:1000), inducible nitric by cervical dislocation. The blood and brains were col- oxide synthase (iNOS) (Santa Cruz Biotechnology; lected for further analysis. 1:3000), NLRP3 (Adipogen; 1:1000) and pro-caspase1 (Adipogen; 1:1000). All secondary antibodies were horse-radish protein (HRP) conjugated (Vector; 1:5000 or Extracellular vesicle isolation procedure and transmission 1:10000). Protein bands were detected using Clarity West- electron microscopy (TEM) ern ECL Substrate (Bio-Rad) or Pierce™ ECL Western For isolation of EVs, cells were cultivated in growing Blotting Substrate (ThermoFisher), and imaged on medium DMEM and then deprivation of serum for a Bio-Rad ChemiDoc XRS+. Protein levels were normalized period of 12 h. The media was then collected and sub- to beta-actin (Sigma-Aldrich; 1:15,000). Image lab™ soft- jected to a series of low-speed centrifugation steps ware (Bio-Rad) was used to analyze the results. (500×g for 10 min, 2000×g for 10 min, and 10,000×g for 30 min) at 4 °C in order to remove cells and cellular Multiplex cytokine enzyme-linked immunosorbent assay debris. The supernatant was then collected in centrifuge (ELISA) tubes (Beckman Coulter) and spun at 100,000×g for The concentrations of different cytokines in EVs and in 70 min before the resultant EV pellet was washed in a isolated media as well as serum from mice were mea- large volume of phosphate-buffered saline (PBS) before sured with the MSD Mouse Proinflammatory V-Plex repeating the 100,000×g spin. The pellets containing EVs Plus Kit (Interferonγ (IFNγ), IL-1β, IL-2, IL-4, IL-5, IL-6, were resuspended in 20 μl of PBS and stored at 4 °C or IL-10, IL-12p70, TNF, C-X-C motif chemokine ligand 1 long term at − 20 °C. For electron microscopic analysis, (KC/GRO), Mesoscale) using a QuickPlex SQ120 Plate samples of EVs were fixed with an equal volume of 2% Reader (Mesoscale Discovery, Rockville, USA) according PFA and loaded onto Formvar/carbon-coated electron to the manufacturer’s instructions. The data was ana- microscopic grids. EVs were observed under TEM at lyzed with MSD Discovery Workbench software. The 80 kV. TEM was carried out at Lund University Bioima- levels of cytokines in EVs and media were normalized to ging Center. each samples total protein content in cell lysates. In Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 4 of 19 total, 6 independent EV samples were analyzed; however, 0.1% (v/v) formic acid in LC-mass spectrometry grade those samples that were under the lowest detection limit water (solvent A) and 0.1% (v/v) formic acid in aceto- were removed from the statistical analysis. nitrile (solvent B). Peptides were first loaded with a con- stant pressure mode with a flow rate of solvent A onto Immunohistochemistry the trapping column. Subsequently, peptides were eluted Immunofluorescent double labeling for TNF and CD11b via the analytical column at a constant flow of 300 nl/ was performed on 16-μm thick, cryostat-cut tissue sec- min. During the elution step, the percentage of solvent B tions from C57BL/6 mice with 1-day survival after increased from 5 to 22% in the first 20 min, then in- pMCAO as previously described in detail [18, 19]. creased to 32% in 5 min and finally to 98% in a further 2 min and was keeping it for 8 min. The peptides were Extracellular vesicle fluorescent labeling introduced into the mass spectrometer via a Stainless Following isolation, EVs were labeled with PKH67 Green steel emitter 40 mm (Thermo Fisher) and a spray volt- Fluorescent Cell Linker Midi Kit for General Cell age of 1.9 kV was applied. The capillary temperature was Membrane Labeling (Sigma-Aldrich) according to the set at 275 °C. manufacturer’s instructions. Briefly, EVs were resus- Data acquisition was carried out using a top N-based pended in 1 ml PBS before 1 ml of Diluent C supple- data-dependent method with cycle time of 3 s. The mas- mented 4 μl PKH67 dye. Samples were incubated at ter scan was performed in the Orbitrap in the range of room temperature for 4 min prior to the addition of 350–1500 mass to charge ratio (m/z) at a resolution of 2 ml of 1% bovine serum albumin (BSA) (VWR Inter- 60,000 full width at half-maximum (FWHM). The filling national) to bind excess dye. Samples were then supple- time was set at maximum of 50 ms with limitation of mented with 5 ml PBS and placed in 300 kDa Vivaspin 4×10 ions. In a second stage of tandem mass spec- filters (Sartorius Stedim Biotech GmbH, Goettingen, trometry (MS/MS) ion trap collision-induced dissoci- Germany), prior to centrifugation for 5 min at 4000×g to ation was acquired using parallel mode, filling time remove excess dye. This process was repeated a further maximum 300 ms with limitation of 2 × 10 ions, a pre- two times, followed by a further two washes in a clean cursor ion isolation width of 1.6 m/z and resolution of filter with DMEM (Thermo Fisher Scientific) in place of 15,000 FWHM. Normalized collision energy was set to + + PBS. The same procedure minus EVs was carried out as 35%. Only multiply charged (2 to 5 ) precursor ions control. were selected for MS/MS. The dynamic exclusion list was set to 30 s and relative mass window of 5 ppm. TNF inhibition on dynamics of extracellular vesicle trafficking Bioinformatic analysis PKH67-labeled EVs (2 × 10 particles/ml) were incu- Gene Ontology (GO) classifications and enrichments bated with BV2 cells as indicated previously. After 12 h were performed using FunRich [20]. The identified pro- incubation, cells were washed three times with PBS and teins were compared with web tool Exocarta database one time with 1 M NaCl prior to fixation with 4% PFA and also with the Top100 exosomal proteins from the for 20 min on ice. Cells were then imaged by fluores- database [21]. cence microscope (Olympus IX71) at × 20 magnification and images processed using Cellsens Standard version Data analysis 1.6 software (Olympus). Vesicle uptake was analyzed by MS/MS data were searched with PEAKS (7.5). UniProt measuring fluorescent intensity using ImageJ software Mus musculus (house mouse, including 16,792 se- (National Institutes of Health). quences) was used with non-tryptic specificity allowing up to 3 missed cleavages. A 15 ppm precursor tolerance Mass spectrometry and a 0.1 Da fragment tolerance were used. Oxidation Mass spectrometry was carried out on an Orbitrap (M) and deamidation (NQ) were treated as dynamic Fusion Tribrid MS system (Thermo Scientific) equipped modification and carbamidomethylation (C) as a fixed with a Proxeon Easy-nLC 1000 (Thermo Fisher). modification. Maximum post-translational modification Injected peptides were trapped on an Acclaim PepMap per peptide was 2. Search results were filtered by using C18 column (3-μm particle size, 75-μm inner diameter 1% false discovery rate and 2 unique peptides. × 20 mm length). After trapping, gradient elution of pep- The rest was evaluated using either unpaired t test or tides was performed on an Acclaim PepMap C18 col- one-way ANOVA followed by Tukey’s test for multiple umn (100 Å 3 μm, 150 mm, 75 μm). The outlet of the comparisons. All statistical analysis was done using the analytical column was coupled directly to the mass spec- GraphPad Prism 7.0 software for Macintosh (GraphPad trometer using a Proxeon nanospray source. The mobile Software, San Diego, CA, USA). Data are presented as phases for liquid chromatography (LC) separation were means ± SD. A confidence interval of 95% was set as Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 5 of 19 significant. The exact P values are given in the figure revealed that LPS-stimulated microglia release larger EV legends. Figures were organized using Adobe Illustrator. populations compared to control (Fig. 1g). EV samples were blotted for different EV markers including a marker Results for plasma membrane (flotillin-1) and an endosomal Proinflammatory responses from LPS-stimulated BV2 marker (Alix) as well as the EV marker CD63 for MVB microglial cells (Fig. 1e) to elucidate the subcellular origin of the EVs [25]. First, we examined the activation of BV2 cells after 12 h We observed that in EVs released from microglia after culture in the presence of 1 μg/ml LPS followed by LPS activation had a higher ratio of flotillin-1 and CD63 deprivation of serum for 12 h to elicit a strong inflam- when compared to those from non-activated cells when matory reaction. The inflammatory enzyme iNOS is loading equal amounts of EVs in each lane, suggesting al- expressed by activated microglia [22] and as expected, tered EV biogenesis and release after an inflammatory the level of iNOS was significantly increased in cells stimulus. upon LPS stimulation (Fig. 1a). Moreover, the protein levels of other important inflammatory mediators, Increased production of TNF and IL-6 in EVs upon LPS NLRP3 (Nod-like receptor protein 3) and pro-caspase1 activation (involved in the maturation, production, and release of Next, we studied the cytokine release from LPS-activated IL-1β and IL-18 [23]) were found to be elevated BV2 microglia to evaluate the free concentration of re- considerably (Fig. 1b, c). According to our previous stud- leased cytokines and the cytokine concentration in EVs. ies, the viability of BV2 cells is not affected by LPS acti- Culture medium from activated and non-activated micro- vation [22, 24]. These results suggest that an activated glia was collected, and EVs were isolated from equal pro-inflammatory status of microglia remained over the amounts of medium. EVs, along with the EV-depleted 12 h EV collection period following LPS treatment. media, were then subjected to analysis by multiplex ELISA. Out of ten inflammatory cytokines analyzed, the Changes in EV size distribution in response to LPS levels of two pro-inflammatory cytokines, TNF and IL-6, activation were found to be significantly increased in EVs from Transmission electron microscopy (TEM) was per- activated microglia (Fig. 2a, b). TNF and IL-6 are two formed to visualize microglial-derived EVs (Fig. 1d). Im- representative pro-inflammatory cytokines produced by ages from control condition and LPS treatment revealed microglia related to neurodegenerative diseases [3]. heterogeneous populations of EVs from microglia in the Notably, there was also a significant upregulation in the range between 100 and 1000 nm in diameter. LPS treat- concentration of these two cytokines in medium ment seemed to induce microglial release of larger EVs, (Additional file 1). However, other pro-inflammatory around 200–300 nm in diameter. cytokines such as IL-5 and IL-1β were found to be To further characterize EVs released under increased only in the medium, but not in EVs pro-inflammatory condition by LPS treatment, we com- (Additional files 1 and 2). Importantly, the level of pared size distribution of EVs derived from non-activated TNF was much higher increased, 22-fold, compared and LPS-activated microglia using Nanoparticle Tracking to 5-fold in IL-6. Thus, we further investigated the ef- Analysis (NTA). The range of EVs detected with NTA was fect of TNF in regulation of EV release in the follow- from 50 to 700 nm. Different size subpopulations of EVs ing study. were observed in EVs released from activated and non-activated microglial cells (Fig. 1f). According to the Reduction of EVs by inhibition of inflammation via TNF diameter of EVs, we can classify them into two subpopula- pathway tions: one ranging from 50 to 100 nm can be considered In view of the specific increase of TNF in the EVs after as exosomes and another subpopulation with size LPS-stimulation, we set out to further characterize the exceeding 100 nm can be regarded as microvesicles role of TNF in the microglial release of EV. To investi- (MVs). As can be seen in Fig. 1f, the population of MVs gate the mechanistic basis for EV regulation on LPS ranging from 300 to 400 nm has a higher frequency in the signaling in microglial cells, we quantified the number of LPS-activated EVs. We found EVs released from EVs released from microglia after TNF inhibition with LPS-activated cells to be significantly larger (178 ± etanercept upon LPS activation. Notably, activated 5.66 nm) compared to the size of EVs from control cul- microglia secreted a 30-fold increase in the number of tures (159 ± 4.95 nm) (Fig. 1f; p < 0.001). D90 measure- EVs under LPS stimulation (Fig. 3a). In contrast, the ment shows the upper limit of 90% measured particles, effect of LPS on EV release was completely attenuated and in control condition, the D90 value for EVs was 205 ± down to control levels using etanercept (200 ng/ml), a 3.61 nm, whereas the value was 254 ± 11.06 nm in TNF inhibitor that blocks both soluble and transmem- LPS-activated samples (Fig. 1f; p < 0.001). These results brane forms of TNF (Fig. 3a). In the presence of the Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 6 of 19 Fig. 1 Characterization of microglia-derived extracellular vesicles. Microglia (BV2) were activated by treatment with LPS for 12 h before extracellular vesicles were isolated from the media of both LPS-treated (LPS) or control (CTRL) cells. a Western blot analysis of protein expression levels of iNOS in cell lysates from control and activated microglia with LPS stimulation (Mean ± SD, n = 5). b, c Components of the inflammasome, NLRP3 and pro-Caspase 1, were measured by western blot in cell lysates with representative pictures of blots (Mean ± SD, n = 5). d Representative TEM imaging of extracellular vesicle populations from CTRL and LPS-derived microglia. The imaging illustrates heterogeneity and sphere structure of extracellular vesicles. Typical microvesicles are pointed with red arrows. Scale bars: 500 nm. e Western blots showed alterations of expression levels of vesicle markers indicated different origins of extracellular vesicles from CTRL and LPS. Three biological independent samples were blotted in each condition. f The size of extracellular vesicles was determined in diameter from CTRL and LPS-treated microglia. The mean size shows the average diameter of extracellular vesicles in samples (n = 12). D90 demonstrates the upper limit of extracellular vesicles size in 90% of the population (n = 12). g Representative histograms of extracellular vesicles size distributions collected from CTRL and LPS conditions. Sample from LPS condition was diluted 25 times more than control to obtain similar concentration of EVs to demonstrate size distribution. Concentrations (× 10 particles/ml) by size (nm) of recorded extracellular vesicles are showed. The major subpopulations in EV samples are indicated with digitals showing the mean diameter of extracellular vesicles. Histograms were generated from five independent measurements by NTA 2.2 software (Unpaired t test, *P < 0.05; ***P < 0.001; ****P < 0.0001) TNF inhibitor, the amount of EVs released from cells indicating the reduction of EVs was due to TNF treat- was also reduced compared with control condition ment on EV release not by affecting EV uptake/turnover (Fig. 3a). Next, we wanted to understand whether the (Additional file 3). reduction of EV is caused by decreased EV uptake Next, we evaluated the degree of inflammatory status through inhibition of TNF. Thus, we assessed the cap- in microglia in relation to the amount of EVs secreted. ability of EV uptake under different conditions as indi- To that aim, the level of iNOS was analyzed by western cated above. We found no significant difference between blot on cells previously treated in different conditions. the conditions on internalization of PHK76-labeled EVs, We found two-fold reduction in iNOS levels following Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 7 of 19 were statistically elevated in WT at day 1 and unchanged in TNF-KO mice (Additional file 4). IL-5 and IL-12p70 were also found statistically elevated in WT mice 1 day after manipulation, but not in TNF-KO (Fig. 4a, c). While in the case of IL-1β, IL-6, and KC/GRO, expres- sion levels were considerably increased 1 day after pMCAO in both types of mice compared with unmanip- ulated mice (Fig. 4b, d, f). Notably, such increases in- duced by pMCAO were significantly attenuated by deficiency of TNF in mice. Levels of IL-10 were remark- ably lower in TNF-KO mice compared with WT mice after pMCAO, but not before (Fig. 4e). Levels of IL2, IL-4, and IFNγ were remained at baseline levels at day 1 in both types of mice subjected to pMCAO (Additional file 4). In conclusion, we found clear evidence of signifi- cant upregulation of proinflammatory cytokines at day 1 Fig. 2 Increased levels of proinflammatroy cytokines in microglia- after pMCAO in both types of mice. However, in derived extracellular vesicles upon LPS activation. The levels of TNF-KO mice, such inflammation induced by pMCAO cytokines were analyzed by multiplex ELISA plate. a Bar graph shows was remarkably attenuated at day 1 after stroke. a significant upregulation of TNF (n = 3). b Bar graph shows a significant upregulation of IL-6 (n = 6)(Mean ± SD, Unpaired t test, Decreased EVs in TNF knockout mice after focal cerebral *P < 0.05; ***P < 0.001) ischemia Given the impact of pMCAO on systemic inflammation TNF inhibition. Interestingly, the reduction of EVs re- in WT and TNF-KO mice, we examined microglia acti- leased upon TNF inhibition was reduced 16-fold, sug- vation in the brain. We observed that TNF co-localized gesting that TNF signaling is particularly important with the microglial marker, CD11b, in the peri-infarct when it comes to reducing the number of EV released in area 1 day after pMCAO (Fig. 3d), which shows that proinflammatory activation of microglia (Fig. 3a). To- focal cerebral ischemia could induce an initial phase of gether, these results implicate that the secretion of EVs microglia activation involving TNF signaling and tissue is dramatically impeded by TNF inhibition in injury, as we have shown before [30]. Moreover, the vol- LPS-activated microglia, which is only partly associated ume of infarct was also assessed in WT mice and to an overall reduction in the inflammatory status. TNF-KO mice 5 days after pMCAO in the previous study, which has shown that the injury was significantly Evaluation of systemic inflammation in mice after focal larger in TNF-KO mice than WT mice [18]. Previously, cerebral ischemia we have shown that the infarct volume correlates with As the level of TNF has shown to be upregulated in EVs the number of EVs in plasma after this stroke model under LPS activation in vitro, we next studied whether [31]. Thus, we analyzed the number of EVs in the the secretion of EVs in vivo was affected by complete ab- plasma of TNF-KO mice after pMCAO. The production lation of TNF signaling in a strong neuroinflammatory of EVs increased from both genotypes 5 days after situation. To this purpose, we chose an experimental pMCAO indicating inflammation occurred, which in line stroke model, permanent middle cerebral artery occlu- with our previous findings in vitro (Fig. 3a, c). Import- sion (pMCAO), as an in vivo inflammatory model. Ex- antly, we found here that a complete ablation of TNF perimental stroke in rat and mouse is known to induce successfully reduced the counts of EVs in the plasma of neuroinflammation in the brain [26, 27] as well as alter TNF-KO mice 5 days after induction of permanent focal the inflammatory response in the periphery [28, 29]. Our cerebral ischemia (Fig. 3c), but not at day 1 (Add- earlier study has revealed significant increase of TNF re- itional file 5). These results are consistent with our in ceptors, toll-like receptor (TLR) 2 and IL-1β at mRNA vitro data and indicate the number of EVs is related to levels in wild type (WT) and TNF-KO mice 1 day after an inflammatory event and that TNF signaling is import- pMCAO, indicating inflammation occurred in the brain ant in the mechanism regulating EV release. [18]. We also wanted to assess systemic inflammatory response in mice after pMCAO. Therefore, we measured Identification of microglial EVs proteins levels of different cytokines in serum using Multiplex To further elucidate cellular communication by EVs ELISA from WT and TNF-KO mice without manipula- under inflammation, we set out to identify proteins in tion and 1 day after pMCAO. TNF expression levels association to EVs released by BV2 cells. To this aim, Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 8 of 19 Fig. 3 Reduction in the number of microglia-derived extracellular vesicles by inhibition of TNF signaling pathway. Extracellular vesicles were visualized by TEM. Concentrations of extracellular vesicles in conditional media and plasma from mice were measure by NTA. a Comparison of extracellular vesicles concentrations in different conditional media. BV2 microglia cell line was treated with either LPS (1 μg/ml) or etanercept (200 ng/ml) or in the presence of both. Control (CTRL) was cells without any treatment (Mean ± SD, one-way ANOVA, ***P <0.001, n =3). b Expression levels of iNOS in BV2 cells previously treated with different conditions were analyzed by western blot. iNOS bands were not able to be visualized in the conditions of control and etanercept due to out of the detection limit (Mean ± SD, Unpaired t-test, *P < 0.05, n =4). c Bar graph shows a comparison of the amount of extracellular vesicles in plasma from mice (WT, n = 6; TNF-KO, n = 6) without surgery and mice (WT, n =9; TNF-KO, n = 8) 5 days after a stroke model, permanent middle cerebral artery occlusion (pMCAO) (Mean ± SD, one-way ANOVA followed by Tukey’s test for multiple comparisons, *P < 0.05, ***P < 0.001). d The brain section in area of infarct from C57BL/6 mouse subjected to pMCAO stained with anti-TNF (green) and anti-CD11b (red), nuclei stained with DAPI. TNF+ cells are also stained for CD11b indicated with arrow; scale bar: 20 μm proteomic analysis was performed using mass spectrom- Gene Ontology annotations related to membrane and etry on EV samples from LPS-activated and control extracellular exosome (Fig. 5c, d). The identified proteins microglia. Biological independent duplicates of EVs were were also compared with the ExoCarta database, which pooled together and analyzed. In total, 86 proteins were has exosomal proteins identified from previous publica- identified with two peptides confirmed and high confi- tions [21]. In our samples, only one protein has not been dence, as -10lgP is set for 20 as threshold, using PEAKS reported in ExoCarta database (Fig. 5a). Compared with analysis program (Tables 1 and 2)[32]. In total, 37 pro- top 100-ranked proteins presented in the database, we teins from control and 58 proteins from LPS-stimulated identified 17 of them in our study (Fig. 5a). Notably, 89.5% microglia were successfully mapped with UniProt data- proteins from this study have not been reported before base. Among them, we found 9 proteins in common and first identified in microglia (Table 4), suggesting that (Table 3) and 49 specific proteins present in LPS condi- these cells are releasing specific EVs. tion. The analysis showed enzymes, chaperones, ribosomal structure proteins, and membrane receptors previously re- Functional profiles of the quantified microglia EVs proteins ported in other immune cells in our samples (Fig. 5 d, e). Next, we wanted to know the functions of the quantified Of these, the majority of the identified proteins were asso- proteins differentially secreted in EVs under LPS activa- ciated with RNA binding, and more than half of them had tion. The 86 quantified proteins in the microglial-derived Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 9 of 19 IL-1 ab IL-5 2.5 9 ** **** **** 2.0 1.5 0.6 1.0 0.3 0.5 0.0 0.0 WT TNF-KO WT TNF-KO WT TNF-KO WT TNF-KO Control 1 D after pMCAO Control 1 D after pMCAO IL-6 IL-12p70 c d ** * *** * *** 0 0 WT TNF-KO WT TNF-KO WT TNF-KO WT TNF-KO Control Control 1 D after pMCAO 1 D after pMCAO KC/GRO IL-10 e f ** ** ** 2.0 ** 1.5 1.0 0.5 20 0.0 0 WT TNF-KO WT TNF-KO WT TNF-KO WT TNF-KO Control Control 1 D after pMCAO 1 D after pMCAO Fig. 4 Systemic inflammation in mice after focal cerebral ischemia. The levels of cytokines in serum were analyzed by multiplex ELISA from WT and TNF-KO mice without manipulation and 1 day after focal cerebral ischemia. a and c Bar graphs show significant upregulation of IL-5 and IL- 12p70 in WT mice after pMCAO. b, d, f Bar graphs show significant upregulation of IL-1β, IL-6, and KC/GRO in both types of mice subjected to pMCAO. e Bar graph shows remarkable reduction of IL-10 in TNF-KO mice compared with WT mice 1 day after pMCAO. (Mean ± SD, one-way ANOVA followed by Tukey’s test for multiple comparisons, n =3–6, *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001) EVs were analyzed using FunRich software [21] to validate flotillin-1 and CD63 indicating more MVs released. In the our data referred to the Vesiclepedia Exosome Database categories of molecular function, the majority of the pro- [33]. The analysis was based on their cellular compart- teins were annotated to RNA binding, 38.9% in control ments to which they belong, their molecular function and and 72.4% in LPS (Fig. 5c). Importantly, the proteins con- the biological process in which they are involved. In the tributed to ribosome function raised from 2.8 to 56.9% analysis, some of the proteins were annotated in more after LPS challenge (Fig. 5c). The profile of proteins in than one cellular component, molecular function, and bio- molecular function was dramatically altered after inflam- logical process. matory stimulation. It is likely that microglia released a Firstly, enrichment in pathways from membrane and distinct population of EVs related to transduction and extracellular exosome was observed for GO analysis on translation after activation of LPS. Intriguingly, the profile cellular component (Fig. 5d). There were large increases of EVs was peculiarly different after activation according in proteins from LPS-stimulated EVs from ribosome, focal to classification of biological process (Fig. 5e). EVs were adhesion, extracellular matrix, and membrane (Fig. 5d). In found involved in regulation of neuron projection devel- contrast, a reduction of proteins associated with extracel- opment at “rest” status (control), however, with activation lular exosome was found in LPS-stimulated EVs, which is the proteins identified in EVs shifted to ribosomal assem- also confirmed by western blot with increased ratio of bly and translation. The role of microglia was changed pg/mL pg/mL pg/mL pg/mL pg/mL pg/mL Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 10 of 19 Table 1 List of proteins identified in control EV samples Protein ID Gene name Protein name Score P08226 APOE Apolipoprotein E 506 P21956 MFGE8 Lactadherin 136 P21460 CST3 Cystatin-C 123 Q9WU78 PDCD6IP Programmed cell death 6-interacting protein 121 P11152 LPL Lipoprotein lipase 107 Q8VDN2 ATP1A1 Sodium/potassium-transporting ATPase subunit alpha-1 105 P62737 ACTA2 Actin, aortic smooth muscle 84 Q61753 PHGDH D-3-phosphoglycerate dehydrogenase 80 P07901 HSP90AA1 Heat shock protein HSP 90-alpha 80 P10605 CTSB Cathepsin B 75 P01942 HBA Hemoglobin subunit alpha 73 P63017 HSPA8 Heat shock protein 8 71 Q62419 SH3GL1 Endophilin-A2 70 Q8R366 IGSF8 Immunoglobulin superfamily member 8 69 Q68FD5 CLTC Clathrin heavy chain 1 67 P10923 SPP1 Osteopontin 64 Q61207 PSAP Prosaposin 64 P06869 PLAU Urokinase-type plasminogen activator 62 Q8CGP5 HIST1H2AF Histone H2A type 1-F 61 P52480 PKM Pyruvate kinase 59 P17182 ENO1 Alpha-enolase 56 P04104 KRT1 Keratin, type II cytoskeletal 1 50 P68369 TUBA1A Tubulin alpha-1A chain 47 P01887 B2M Beta-2-microglobulin 45 P09405 NCL Nucleolin 44 P09055 ITGB1 Integrin beta-1 42 P01901 H2-K1 H-2 class I histocompatibility antigen, K-B alpha chain 41 P01899 H2-D1 H-2 class I histocompatibility antigen, D-B alpha chain 36 P29341 PABPC1 Polyadenylate-binding protein 1 41 P06797 CTSL Cathepsin L1 38 P10852 SLC3A2 4F2 cell-surface antigen heavy chain 37 P08905 LYZ2 Lysozyme C-2 34 P10126 EEF1A1 Elongation factor 1-alpha 1 28 Q61937 NPM1 Nucleophosmin 27 Q9DBJ1 PGAM1 Phosphoglycerate mutase 1 25 P28798 GRN Granulins 24 P14206 RPSA 40S ribosomal protein SA 24 Proteins retrieved in control extracellular vesicle samples from mass spectrometry. Protein ID and gene name are according to Uniprot Knowledgebase. Score values were obtained using MASCOT. The Score shows how well the observed protein matches to the stated protein in the database. Only protein identifications supported by at least two high confident peptides (confidence > 95%) were considered from “rest” state to “activated” state by not only produc- In larger proportion, 49 proteins were exclusively tion of cytokine and chemokine, but also with altered se- present in EVs samples from LPS activation, including cretion of EVs, which could be directly linked to some cytoskeleton proteins and ribosomal proteins. Im- detrimental inflammatory responses related to neurode- portantly, three of them were associated with inflamma- generative diseases. tory and neuropathological pathways: tubulin beta 4B Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 11 of 19 Table 2 List of proteins identified in LPS EV samples Protein ID Gene name Protein name Score P68372 TUBB4B Tubulin beta-4B chain 258 P62242 RPS8 40S ribosomal protein S8 201 P47963 RPL13 60S ribosomal protein L13 181 Q9CZX8 RPS19 40S ribosomal protein S19 178 P15864 HIST1H1C Histone H1.2 172 Q8VEK3 HNRNPU Heterogeneous nuclear ribonucleoprotein U 148 Q9CR57 RPL14 60S ribosomal protein L14 146 P63276 RPS17 40S ribosomal protein S17 138 P11152 LPL Lipoprotein lipase 130 P16858 GAPDH Glyceraldehyde-3-phosphate dehydrogenase 128 P68369 TUBA1A Tubulin alpha-1A chain 123 P62082 RPS7 40S ribosomal protein S7 106 Q9D8E6 RPL4 60S ribosomal protein L4 101 P35979 RPL12 60S ribosomal protein L12 101 P14131 RPS16 40S ribosomal protein S16 100 P70696 HIST1H2BA Histone H2B type 1-A 100 P12970 RPL7A 60S ribosomal protein L7a 98 P60710 ACTB Actin, cytoplasmic 1 94 Q9CXW4 RPL11 60S ribosomal protein L11 85 Q9CZM2 RPL15 60S ribosomal protein L15 84 P02301 H3F3C Histone H3.3C 83 P47911 RPL6 60S ribosomal protein L6 83 P43276 HIST1H1B Histone H1.5 83 Q8CGP5 HIST1H2AF Histone H2A type 1-F 82 P14115 RPL27A 60S ribosomal protein L27a 74 P27659 RPL3 60S ribosomal protein L3 72 O55142 RPL35A 60S ribosomal protein L35a 70 P25206 MCM3 DNA replication licensing factor MCM3 68 P62702 RPS4X 40S ribosomal protein S4, X isoform 63 P10126 EEF1A1 Elongation factor 1-alpha 1 63 Q8BP67 RPL24 60S ribosomal protein L24 62 P14206 RPSA 40S ribosomal protein SA 62 P11499 HSP90AB1 Heat shock protein HSP 90-beta 60 P08226 APOE Apolipoprotein E 60 P35980 RPL18 60S ribosomal protein L18 59 P63017 HSPA8 Heat shock protein 8 59 P80318 CCT3 T-complex protein 1 subunit gamma 58 P62918 RPL8 60S ribosomal protein L8 57 O08585 CLTA Clathrin light chain A 55 Q9Z1Q9 VARS Valine-tRNA ligase 54 P01942 HBA Hemoglobin subunit alpha 51 P97351 RPS3A 40S ribosomal protein S3a 50 P62806 HIST1H4A Histone H4 48 P62270 RPS18 40S ribosomal protein S18 44 Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 12 of 19 Table 2 List of proteins identified in LPS EV samples (Continued) Protein ID Gene name Protein name Score P62911 RPL32 60S ribosomal protein L32 43 P14869 RPLP0 60S acidic ribosomal protein P0 41 Q9JIK5 DDX21 Nucleolar RNA helicase 2 41 P62281 RPS11 40S ribosomal protein S11 41 Q68FD5 CLTC Clathrin heavy chain 1 40 P62717 RPL18A 60S ribosomal protein L18a 39 P86048 RPL10L 60S ribosomal protein L10-like 39 P62267 RPS23 40S ribosomal protein S23 37 P25444 RPS2 40S ribosomal protein S2 36 P14148 RPL7 60S ribosomal protein L7 31 P53026 RPL10A 60S ribosomal protein L10a 28 P62264 RPS14 40S ribosomal protein S14 28 P62908 RPS3 40S ribosomal protein S3 23 P04918 SAA3 Serum amyloid A-3 protein 23 Proteins retrieved in LPS extracellular vesicles samples from mass spectrometry. Protein ID and gene name are according to Uniprot Knowledgebase. Score values were obtained using MASCOT. The score shows how well the observed protein matches to the stated protein in the database. Only protein identifications supported by at least two high confident peptides (confidence > 95%) were considered (TUBB4B), heterogeneous nuclear ribonucleoprotein U roles in neurodegenerative diseases [36–38]. Microglia (HNRNPU) and serum amyloid A 3 (SAA3). Nine pro- are an essential component in innate immunity in the teins were shared in samples from control and LPS con- CNS. Activation of microglia is followed by the initiation ditions (Fig. 5b). Proteins related to immune system of intracellular machinery that leads to production of process and stimuli response were detected in both sam- cytotoxic and proinflammatory cytokines and chemo- ples: lipoprotein lipase (LPL), apolipoprotein E (APOE), kines, which promote progression of inflammation and and heat shock protein 8 (HSPA8). Notably, APOE was affect neighboring cells through different mechanisms. particular with high score in the EVs before and after The secretion of proinflammatory signals can be con- LPS activation. Variants of APOE (APOE4) is known as ducted in classic secretion manner or non-canonical the strongest risk factor for late onset Alzheimer’s dis- manner via vesicles. Previous studies have shown the im- ease [34] and has also been suggested to have a proin- portant role of EVs in regulation of cytokine production flammatory effect on microglia [35]. on recipient cells and propagation of pathogenic pro- teins [11, 39]. Discussion The functions of microglia in the brain are diverse Compelling evidence has suggested that the activation of depending on stimulus and different brain regions. In innate immunity and neuroinflammation play crucial response to different inflammatory/homeostatic condi- tions, they can be either beneficial or detrimental. Table 3 The common proteins shared in CTRL and LPS EV Increasing evidence has suggested that there is a spectrum samples of activation in microglia based on the profile of secreted Protein ID Gene name Protein name molecules [3]. However, the importance of EV released by P08226 APOE Apolipoprotein E microglia has not been well characterized. Thus, in the P11152 LPL Lipoprotein lipase present study, we directly evaluated effects of LPS-stimulation on EVs derived from BV2 microglia, in P01942 HBA Hemoglobin subunit alpha terms of physical and biological properties. Our results P63017 HSPA8 Heat shock protein 8 demonstrate that (i) upon LPS-activation, the size dis- Q68FD5 CLTC Clathrin heavy chain 1 tribution of EVs released from microglia increases in Q8CGP5 HIST1H2AF Histone H2A type 1-F size, indicating larger vesicles are released under in- P68369 TUBA1A Tubulin alpha-1A chain flammatory conditions; (ii) IL-6 and in particular TNF P10126 EEF1A1 Elongation factor 1-alpha 1 are increasingly secreted in microglia-derived EVs after LPS challenge; (iii) inhibiting the TNF signaling path- P14206 RPSA 40S ribosomal protein SA way resulted in a robust reduction in the number of Proteins retrieved in extracellular vesicle samples from mass spectrometry. Protein ID and gene name are according to Uniprot Knowledgebase vesicles released from LPS-treated microglia and in Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 13 of 19 Fig. 5 Bioinformatic analysis of the identified proteins in BV2 microglia-derived extracellular vesicles. Gene ontology (GO) analysis of the identified proteins was performed using Exocarta based software FunRich. The comparison was carried on proteins expressed exclusively in each condition. a Comparison of 86 identified proteins from microglial extracellular vesicles with online database Exocarta and the top 100 proteins commonly reported in the same database. b Comparison of the number of proteins identified and quantified in control microglia and LPS-activated microglia derived extracellular vesicles samples. c Bioinformatic analysis from FunRich shows comparison of GO analysis on Molecular Function on proteins from CTRL and LPS-activated microglia released extracellular vesicles (GO terms are with P < 0.01). d Bioinformatic analysis from FunRich shows comparison of GO analysis on Cellular Component on proteins from CTRL and LPS-activated microglia released extracellular vesicles (GO terms are with P < 0.01). e Bioinformatic analysis from FunRich shows comparison of GO analysis on Biological Process on proteins from CTRL and LPS-activated microglia released extracellular vesicles (GO terms are with P < 0.01) Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 14 of 19 Table 4 New proteins found in present microglia-derived EV samples Protein ID Gene name Protein name P08226 APOE Apolipoprotein E P21460 CST3 Cystatin-C Q8VDN2 ATP1A1 Sodium/potassium-transporting ATPase subunit alpha-1 P62737 ACTA2 Actin, aortic smooth muscle Q61753 PHGDH D-3-phosphoglycerate dehydrogenase P10605 CTSB Cathepsin B P01942 HBA Hemoglobin subunit alpha Q62419 SH3GL1 Endophilin-A2 Q8R366 IGSF8 Immunoglobulin superfamily member 8 Q68FD5 CLTC Clathrin heavy chain 1 P10923 SPP1 Osteopontin Q61207 PSAP Prosaposin P06869 PLAU Urokinase-type plasminogen activator Q8CGP5 HIST1H2AF Histone H2A type 1-F P04104 KRT1 Keratin, type II cytoskeletal 1 P68369 TUBA1A Tubulin alpha-1A chain P01887 B2M Beta-2-microglobulin P09405 NCL Nucleolin P09055 ITGB1 Integrin beta-1 P01901 H2-K1 H-2 class I histocompatibility antigen, K-B alpha chain P01899 H2-D1 H-2 class I histocompatibility antigen, D-B alpha chain P29341 PABPC1 Polyadenylate-binding protein 1 P06797 CTSL Cathepsin L1 P10852 SLC3A2 4F2 cell-surface antigen heavy chain P08905 LYZ2 Lysozyme C-2 P10126 EEF1A1 Elongation factor 1-alpha 1 Q61937 NPM1 Nucleophosmin P28798 GRN Granulins P14206 RPSA 40S ribosomal protein SA P68372 TUBB4B Tubulin beta-4B chain P62242 RPS8 40S ribosomal protein S8 P47963 RPL13 60S ribosomal protein L13 Q9CZX8 RPS19 40S ribosomal protein S19 P15864 HIST1H1C Histone H1.2 Q8VEK3 HNRNPU Heterogeneous nuclear ribonucleoprotein U Q9CR57 RPL14 60S ribosomal protein L14 P63276 RPS17 40S ribosomal protein S17 P62082 RPS7 40S ribosomal protein S7 Q9D8E6 RPL4 60S ribosomal protein L4 P35979 RPL12 60S ribosomal protein L12 P14131 RPS16 40S ribosomal protein S16 P70696 HIST1H2BA Histone H2B type 1-A P12970 RPL7A 60S ribosomal protein L7a P60710 ACTB Actin, cytoplasmic 1 Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 15 of 19 Table 4 New proteins found in present microglia-derived EV samples (Continued) Protein ID Gene name Protein name Q9CXW4 RPL11 60S ribosomal protein L11 Q9CZM2 RPL15 60S ribosomal protein L15 P02301 H3F3C Histone H3.3C P47911 RPL6 60S ribosomal protein L6 P43276 HIST1H1B Histone H1.5 Q8CGP5 HIST1H2AF Histone H2A type 1-F P14115 RPL27A 60S ribosomal protein L27a P27659 RPL3 60S ribosomal protein L3 O55142 RPL35A 60S ribosomal protein L35a P25206 MCM3 DNA replication licensing factor MCM3 P62702 RPS4X 40S ribosomal protein S4, X isoform P10126 EEF1A1 Elongation factor 1-alpha 1 Q8BP67 RPL24 60S ribosomal protein L24 P14206 RPSA 40S ribosomal protein SA P11499 HSP90AB1 Heat shock protein HSP 90-beta P08226 APOE Apolipoprotein E P35980 RPL18 60S ribosomal protein L18 P63017 HSPA8 Heat shock protein 8 P80318 CCT3 T-complex protein 1 subunit gamma P62918 RPL8 60S ribosomal protein L8 P62911 RPL32 60S ribosomal protein L32 P14869 RPLP0 60S acidic ribosomal protein P0 Q9JIK5 DDX21 Nucleolar RNA helicase 2 P62281 RPS11 40S ribosomal protein S11 P62717 RPL18A 60S ribosomal protein L18a P86048 RPL10L 60S ribosomal protein L10-like P62267 RPS23 40S ribosomal protein S23 P25444 RPS2 40S ribosomal protein S2 P14148 RPL7 60S ribosomal protein L7 P53026 RPL10A 60S ribosomal protein L10a P62264 RPS14 40S ribosomal protein S14 P62908 RPS3 40S ribosomal protein S3 P04918 SAA3 Serum amyloid A-3 protein Proteins identified in present study were compared with proteins uploaded in Exocarta database from microglial origin. Protein ID and gene name are according to Uniprot Knowledgebase mice subjected to pMCAO; and (iv) in response to LPS, and translation in activated state and include a particular BV2 microglia release EVs with a distinct proteomic population of MVs budding from the plasma membrane. profile related to transcription and translation. The mechanism underlying release of EV is still not EVs can either enhance or suppress inflammation and clear. TNF can induce neurotoxicity by modulating glu- act as main factors to regulate inflammation and im- tamate production that results in excitotoxic neuronal munity [40, 41]. There is evidence showing microglia death [43]. One study conducted by Wang et al. has can release IL-1β, upon exposure to ATP derived from shown that TNF promotes the release of EVs from astro- astrocytes, by shedding MVs which contain the entire cytes through increased expression of glutaminase, machinery important for processing of it including the which convert glutamine to glutamate [44]. TNF can P2X7 receptors [42]. This is consistent with our findings also induce extensively production of glutamate from that EVs contain more molecules related to transcription microglia in an autocrine manner to cause excitotoxicity Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 16 of 19 and contribute to neuronal damage [45]. Thereby, these changes in the levels of either plasma membrane or endo- findings together with our results suggest that microglial somal markers to elucidate where these vesicles had origi- EV release could be potentially modulated by TNF nated. Indeed, we observed an increase in signal from through specifically regulated mechanisms. plasma membrane markers when compared to endosomal Our data suggests that inhibition of TNF signaling in markers upon LPS activation. When combined with the microglia may impact on inflammation, with an ob- increase in overall size observed with NanoSight, this sup- served reduction both in the release of different cyto- ports our theory that more MVs are released under in- kines and in the total number of EVs released, implying flammatory conditions. that the release of EVs in activated microglia seems to Furthermore, our proteomic analysis indicated that the be regulated. This idea is further supported by data from two populations of EVs were dramatically different in our experimental stroke model that showed altered EV categories of molecular function and biological process counts in plasma from TNF-KO mice compared to WT. under GO analysis. From the qualitative proteomic ana- These results contradict previous work that showed an lysis, 86 proteins were identified with high accuracy. increase in the amount of MVs in mice treated with Compared to Exocarta database, most of the proteins TNF inhibitor for 5 days after pMCAO compared to have been reported in previous studies in exosomes from saline-treated mice [31]. It is therefore reasonable to as- other cell types, where 89.5% of the quantified proteins sume that the difference we observed is due to the dif- were firstly identified in microglial EVs. The small over- ferent experimental set-up. A systemic administration of lap with previous studies is most likely due to a lack of TNF inhibitor in the mice 30 min after surgery could studies utilizing microglia-derived EVs, with a sole study merely have a transient effect on the TNF signaling path- performed on EVs from the N9 microglial cell line re- way, whereas using the TNF-KO model is expected to be sponsible for all microglial proteins present in ExoCarta more stable to investigate and characterize long-term ef- [15]. It was also clear from the FunRich analysis [20] fects of EVs release. that the proteins detected in EVs from LPS-treated cells From our previous study, pMCAO could initiate had different functions to those from control cells, with microglia activation and inflammation in the brain [18, proteins involved in RNA binding and structural compo- 28]. Therefore, we also evaluated systemic inflammation nents of the ribosome more prevalent in LPS-derived in mice subjected to pMCAO. Although several inflam- EVs. While the analysis on the cellular origin of the EV matory factors, such as IL-1β and IL-6, were significantly proteins revealed that the extracellular exosome and increased in both types of mice at day 1 after manipula- membrane were dominant from both conditions, it is of tion, such induction was attenuated in TNF-KO mice. interest to note that EVs isolated from inflammatory Another study has shown that EVs derived from macro- condition were detected with more membrane and less phages stimulated with bacterial infection are able to in- exosome proteins, consistent with a shift towards MV crease secretion of proinflammatory cytokines in release rather than exosome. recipient cells, including TNF [17]. Taken together, we For the first time, our data indicates that microglia can speculate that inflammatory propagation can be mit- change its EVs releasing machinery after LPS activation igated by a reduction in the number of microglia-derived with an increase in the overall number of EVs, and more EVs, thus halting inflammation via TNF signaling. How- EVs budding from the membrane. It is tempting to ever, this is complicated by the fact that there are two speculate that non-activated microglia have an expedient forms of TNF, soluble TNF (solTNF), which is related to and controlled release of EVs, whereas in an inflamma- neurotoxicity and inflammation, while the transmem- tory condition microglia release a large variation of EVs brane TNF (tmTNF) is involved in functional recovery that will perturb the normal homeostatic function of and neuroprotection [18, 46]. Hence, both the cellular microglia. We also think that, upon LPS-stimulation, contribution to TNF signaling and specifically the form microglia respond to release populations of extracellular of TNF carried in microglial EVs is important to evaluate cargoes with a unique proteomic profile related to RNA final outcome of experiment, neuroprotective or neuro- transcription and translation. The existence of various toxic. In our study, we measured total TNF in EVs in- RNA molecules in EVs is well-established including cluding tmTNF and solTNF. However, the specific form mRNA and microRNA. In fact, microRNAs can function of TNF is likely to be important for the inflammatory as ligands for TLRs and induce immune responses or in- and cytoprotective outcome. We believe such studies in- hibit activation by suppressing TLR signaling [13, 47]. vestigating the specific form of TNF carried in EVs are These RNAs are selectively sorted to EVs under different important in future studies. mechanisms [47]. However, the mechanisms responsible We also studied the protein composition to further for this packaging are not clear. Thus, our study implies characterize the EVs released under control or inflamma- proteins potentially involved in such mechanisms that tory conditions. Using western blot analysis, we looked for are increased when inflammation takes place. According Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 17 of 19 to our proteomic data, we speculate that EVs could actu- Additional files ally carry the whole machinery not only RNAs to recipi- Additional file 1: Supplementary figures for cytokines in conditioned ent cells and surrounding tissue. medium from microglia show significant upregulations of TNF (n = 7), IL- Finally, by comparing the protein cargoes from 1β (n = 7), IL5 (n = 7), and IL-6 (n = 3). Measured by multiplex ELISA LPS-activated and non-activated BV2 microglia, we were (Unpaired t test, *P < 0.05; ***P < 0.001). (PDF 201 kb) able to identify potential candidates likely to be involved Additional file 2: Supplementary figures for cytokines in microglia- derived extracellular vesicles not altered after LPS treatment. Measured by in the communication of microglial cells with other ef- multiplex ELISA (Unpaired t test, *P < 0.05; ***P < 0.001). (PDF 180 kb) fector cells under inflammation. We found 49 proteins Additional file 3: Supplementary figures for the effect of TNF inhibition exclusively present in EVs in the presence of LPS. on dynamics of EV trafficking. Images were taken and then measured for Among them, TUBB4B, HNRNPU, and SAA3 are pro- fluorescent intensity. A) Representative images of BV2 cells cultured with PHK76-labeled EVs 12 h after different treatments, including pre- teins related to inflammation and neuropathology. In- treatment of cells with either LPS (1 μg/ml) or etanercept (200 ng/ml) or deed, TUBB4B is a member of tubulin family and known in presence of both. Control (CTRL) was cells without any treatment. Cells as a component of cytoskeleton [48]. Liu X et al, sug- without EV were regarded as baseline. Merged images of the indicated areas show PHK76 internalized cells (Scale bar, 50 μm). B) Comparison of gested that TUBB4B may be a part of the same disease total fluorescent intensity (IntDen) in BV2 cells after incubation of dye- pathway as leucine-rich repeat kinase 2 (LRRK2), which labeled EVs. No significant differences were found between the is a crucial factor to understand the etiology of Parkin- conditions (one-way ANOVA, n = 3). (PDF 9799 kb) son’s disease (PD) [49]. HNRNPU acts as a key factor to Additional file 4: Supplementary figures for cytokines in serum from WT and TNF-KO mice before and 1 day after pMCAO. Measured by multi- maintain 3D structure of chromatin and has been re- plex ELISA (one-way ANOVA followed by Tukey’s test for multiple ported as a posttranscriptional regulator for NF-κB in- comparisons, n =3–6, **P < 0.01). (PDF 30 kb) flammatory pathway [50]. In addition, SAA3 is a Additional file 5: Supplementary figures for quantification of extracellular member of serum amyloid A (SAA) and acute phase vesicles in plasma from WT and TNF-KO mice subjected to pMCAO (Unpaired t test, n = 3). (PDF 18 kb) protein accompanying with other inflammatory cyto- kines and chemokines [51]. It has been implied to play a role in the inflammatory processes occurring in Alzhei- Abbreviations AD: Alzheimer’s disease; APOE: Apolipoprotein E; BCA: Bicinchoninic acid assay; mer’s disease (AD) and multiple sclerosis (MS) [51]. In BSA: Bovine serum albumin; CNS: Central nervous system; DMEM: Dulbecco’s addition to those proteins found exclusively in LPS EVs, modified Eagle medium; ELISA: Enzyme-linked immunosorbent assay; we identified APOE in EVs from both LPS and control EVs: Extracellular vesicles; FBS: Fetal bovine serum; FWHM: Full-width at half- maximum; GO: Gene Ontology; HNRNPU: Heterogeneous nuclear conditions. This protein is commonly present in mem- ribonucleoprotein U; HRP: Horse-radish protein; HSP8: Heat shock protein 8; branes and is considered one of the most important li- IFNγ:Interferonγ; IL: Interleukin; iNOS: Inducible nitric oxide synthase; KC/GRO: C-X- poproteins involved in cholesterol shuttling between C motif chemokine ligand 1; LC: Liquid chromatography; LPL: Lipoprotein lipase; LPS: Lipopolysaccharide; LRRK2: Leucine-rich repeat kinase 2; m/z:Masstocharge astrocytes and neurons along with being involved in re- ratio; MS: Multiple sclerosis; MS/MS: Tandem mass spectrometry; modeling and reorganization of neuronal networks after MVB: Multivesicular endosomal compartments; MVs: Microvesicles; NLRP3: Nod- injury [34]. It is also one of the major genetic risk factors like receptor protein 3; NO: Nitric oxide; NTA: Nanoparticle tracking analysis; PBS: Phosphate-buffered saline; PD: Parkinson’s disease; PFA: Paraformaldehyde; for late onset sporadic AD and can function as a ligand pMCAO: Permanent middle cerebral artery occlusion; SAA3: Amyloid A-3 protein; in receptor-mediated endocytosis with extracellular solTNF: Soluble form of tumor necrosis factor; TEM: Transmission electron β-amyloid [34, 52]. microscopy; TLR: Toll-like receptor; tmTNF: Transmembrane form of tumor necrosis factor; TNF: Tumor necrosis factor; TNF-KO: Tumor necrosis factor knockout; TUBB4B: Tubulin beta 4B Conclusions The present data show that upon activation by LPS, BV2 Acknowledgements microglia release EVs with a distinct proteomic profile We acknowledge technical support for mass spectrometry from the national infrastructure BioMS, Lund University, by Carol Nilsson, Sven Kjellström, and compared to control. Our data suggests that under these Yan Hong. And we would like to thank Lina Gefors for assistance with TEM inflammatory conditions, MVs are the predominate EV, imaging of extracellular vesicles at Bioimaging Center, Lund University. containing increased levels of TNF in particular, and to a lesser degree IL-6. We further provide evidence in vitro Funding We gratefully acknowledge funding support from the Strategic Research and in vivo that TNF signaling is important in quantita- Area MultiPark at Lund University, Lund, Sweden; the Swedish Research tively controlling EV release. Furthermore, through Council (No. 2012-2229), the Basal Ganglia Disorders Linnaeus Consortium proteomic analysis, we are able to provide lists of pro- (BAGADILICO); the Swedish Alzheimer Foundation; A.E. Berger Foundation; Swedish Brain Foundation; Crafoord Foundation; Swedish Dementia Associ- teins with the potential to modulate EV trafficking in ation; G&J Kock Foundation; Swedish National Stroke Foundation; Swedish microglia, in particular a change in EV proteins related Parkinson Foundation; Stohnes Foundation; the Royal Physiographic Soci- to neuronal maintenance and protein translation after ety; Olle Engkvist Byggmästare Foundation; Sparbanken Färs & Frosta Foun- dation, and the Danish Medical Research Council (DFF-4181-00033). LPS activation. We believe that EV regulation in micro- glia and its specific role in neuroinflammation will be Availability of data and materials important to fully understand the inflammatory patho- The datasets used and analyzed during the current study are available from genesis in neurodegenerative diseases. the corresponding author on reasonable request. Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 18 of 19 Authors’ contributions 16. Kowal J, Arras G, Colombo M, Jouve M, Morath JP, Primdal-Bengtson B, et al. YY and TD designed the studies and participated in the data analysis, Proteomic comparison defines novel markers to characterize heterogeneous interpretation, and writing of the manuscript. BHC contributed to cytokine populations of extracellular vesicle subtypes. Proc Natl Acad Sci U S A. 2016; analysis from mice. KLL contributed to all surgical and histological aspects of 113:E968–77. the study including the analysis of tissue specimens. YY performed most 17. Reales-Calderón JA, Vaz C, Monteoliva L, Molero G, Gil C. Candida albicans experiments and analyzed and interpreted data. ABS helped to interpret data modifies the protein composition and size distribution of THP-1 and participate in the design of the study. CD assisted with extracellular macrophage-derived extracellular vesicles. J Proteome Res. 2017;16:87–105. vesicle experiment and provided useful input to the drafting of the paper. 18. Lambertsen KL, Clausen BH, Babcock AA, Gregersen R, Fenger C, Nielsen HH, All authors have read and approved the final manuscript. et al. Microglia protect neurons against ischemia by synthesis of tumor necrosis factor. J Neurosci. 2009;29:1319–30. Ethics approval 19. Clausen BH, Lambertsen KL, Babcock AA, Holm TH, Dagnaes-Hansen F, All animal experiments were performed in accordance with the relevant Finsen B. Interleukin-1beta and tumor necrosis factor-alpha are expressed guidelines and regulations approved by the Danish Animal Ethical Committee by different subsets of microglia and macrophages after ischemic stroke in (numbers 2011/561-1950 and 2013-15-2934-00924). mice. J Neuroinflammation. 2008;5:46–18. 20. Benito Martin A, Peinado H. FunRich proteomics software analysis, let the Competing interests fun begin! Proteomics. 2015;15:2555–6. The authors declare that they have no competing interests. 21. Keerthikumar S, Chisanga D, Ariyaratne D, Saffar Al H, Anand S, Zhao K, et al. ExoCarta: a web-based compendium of Exosomal cargo. J Mol Biol. 2016; 428:688–92. Publisher’sNote 22. Burguillos MA, Svensson M, Schulte T, Boza-Serrano A, Garcia-Quintanilla A, Springer Nature remains neutral with regard to jurisdictional claims in published Kavanagh E, et al. Microglia-secreted galectin-3 acts as a toll-like receptor 4 maps and institutional affiliations. ligand and contributes to microglial activation. Cell Rep. 2015;10:1626–38. 23. Schroder K, Tschopp J. The inflammasomes. Cell. 2010;140:821–32. Author details 24. Burguillos MA, Magnusson C, Nordin M, Lenshof A, Augustsson P, Hansson Department of Experimental Medical Science, Experimental MJ, et al. Microchannel acoustophoresis does not impact survival or Neuroinflammation Laboratory, Lund University, Lund, Sweden. Department function of microglia, leukocytes or tumor cells. PLoS One. 2013;8:e64233. of Biochemistry and Structural Biology, Lund University, Lund, Sweden. 25. Yáñez-Mó M, Siljander PRM, Andreu Z, Zavec AB, Borràs FE, Buzás EI, et al. Department of Neurobiology Research, Institute of Molecular Medicine, Biological properties of extracellular vesicles and their physiological University of Southern Denmark, Odense, Denmark. BRIGDE—Brain functions. J Extracell Vesicles. 2015;4:27066. Research–Inter-Disciplinary Guided Excellence, Department of Clinical 5 26. Vendrame M, Gemma C, De Mesquita D, Collier L, Bickford PC, Sanberg CD, Research, University of Southern Denmark, Odense, Denmark. Department et al. Anti-inflammatory effects of human cord blood cells in a rat model of of Neurology, Odense University Hospital, Odense, Denmark. stroke. Stem Cells Dev. 2005;14:595–604. 27. Inácio AR, Ruscher K, Leng L, Bucala R, Deierborg T. Macrophage migration Received: 1 March 2018 Accepted: 15 May 2018 inhibitory factor promotes cell death and aggravates neurologic deficits after experimental stroke. J Cereb Blood Flow Metab. 2011;31:1093–106. 28. Inácio AR, Liu Y, Clausen BH, Svensson M, Kucharz K, Yang Y, et al. Endogenous References IFN-β signaling exerts anti-inflammatory actions in experimentally induced 1. Béchade C, Cantaut-Belarif Y, Bessis A. Microglial control of neuronal activity. focal cerebral ischemia. J Neuroinflammation. 2015;12:211. Front Cell Neurosci. 2013;7:32. 29. Clausen BH, Lambertsen KL, Dagnaes-Hansen F, Babcock AA, Linstow von 2. Tremblay M-È, Stevens B, Sierra A, Wake H, Bessis A, Nimmerjahn A. The role CU, Meldgaard M, et al. Cell therapy centered on IL-1Ra is neuroprotective of microglia in the healthy brain. J Neurosci Soc Neurosci. 2011;31:16064–9. in experimental stroke. Acta Neuropathologica. Springer. 2016;131:775–91. 3. Heneka MT, Kummer MP, Latz E. Innate immune activation in 30. Lambertsen KL, Meldgaard M, Ladeby R, Finsen B. A quantitative study neurodegenerative disease. Nat Rev Immunol. 2014;14:463–77. of microglial-macrophage synthesis of tumor necrosis factor during 4. Labzin LI, Heneka MT, Latz E. Innate immunity and neurodegeneration. acute and late focal cerebral ischemia in mice. J Cereb Blood Flow Annu Rev Med. 2018;69:437–49. Metab. 2005;25:119–35. 5. Wyss-Coray T, Rogers J. Inflammation in Alzheimer disease-a brief review of 31. Clausen BH, Degn M, Martin NA, Couch Y, Karimi L, Ormhøj M, et al. the basic science and clinical literature. Cold Spring Harb Perspect Med. Systemically administered anti-TNF therapy ameliorates functional outcomes 2012;2:a006346. after focal cerebral ischemia. J Neuroinflammation. 2014;11:203. 6. Ransohoff RM. A polarizing question: do M1 and M2 microglia exist? Nat 32. Ma B, Zhang K, Hendrie C, Liang C, Li M, Doherty-Kirby A, et al. PEAKS: Neurosci. 2016;19:987–91. powerful software for peptide de novo sequencing by tandem mass 7. Masgrau R, Guaza C, Ransohoff RM, Galea E. Should we stop saying ‘glia’ spectrometry. Rapid Commun Mass Spectrom. 2003;17:2337–42. and ‘neuroinflammation’? Trends Mol Med. 2017;23:486–500. 33. Kalra H, Simpson RJ, Ji H, Aikawa E, Altevogt P, Askenase P, et al. 8. Turola E, Furlan R, Bianco F, Matteoli M, Verderio C. Microglial microvesicle Vesiclepedia: a compendium for extracellular vesicles with continuous secretion and intercellular signaling. Front Physiol. 2012;3:149. community annotation. PLoS Biol. 2012;10:e1001450. 9. Gupta A, Pulliam L. Exosomes as mediators of neuroinflammation. J 34. Kim J, Basak JM, Holtzman DM. The role of apolipoprotein E in Alzheimer's Neuroinflammation. 2014;11:1–10. disease. Neuron. 2009;63:287–303. 10. Quek C, Hill AF. The role of extracellular vesicles in neurodegenerative 35. Chen S, Averett NT, Manelli A, LaDu MJ, May W, Ard MD. Isoform-specific diseases. Biochem Biophys Res Commun. 2017;483:1178–86. effects of apolipoprotein E on secretion of inflammatory mediators in adult 11. Szabo GT, Tarr B, Paloczi K, Eder K, Lajko E, Kittel A, et al. Critical role of rat microglia. J Alzheimers Dis. 2005;7:25–35. extracellular vesicles in modulating the cellular effects of cytokines. Cell Mol 36. Sjögren M, Folkesson S, Blennow K, Tarkowski E. Increased intrathecal Life Sci. 2014;71:4055–67. inflammatory activity in frontotemporal dementia: pathophysiological 12. Kumar A, Stoica BA, Loane DJ, Yang M, Abulwerdi G, Khan N, et al. implications. J Neurol Neurosurg Psychiatry. 2004;75:1107–11. Microglial-derived microparticles mediate neuroinflammation after traumatic 37. Zhang B, Gaiteri C, Bodea L-G, Wang Z, McElwee J, Podtelezhnikov AA, et al. brain injury. J Neuroinflammation. 2017;14:47. Integrated systems approach identifies genetic nodes and networks in late- 13. Fleshner M, Crane CR. Exosomes, DAMPs and miRNA: features of stress onset Alzheimer’s disease. Cell. 2013;153:707–20. physiology and immune homeostasis. Trends Immunol. 2017;38:768–76. 14. Raposo G, Stoorvogel W. Extracellular vesicles: exosomes, microvesicles, and 38. Gerhard A, Pavese N, Hotton G, Turkheimer F, Es M, Hammers A, et al. In friends. J Cell Biol. 2013;200:373–83. vivo imaging of microglial activation with [11C](R)-PK11195 PET in 15. Potolicchio I, Carven GJ, Xu X, Stipp C, Riese RJ, Stern LJ, et al. Proteomic idiopathic Parkinson's disease. Neurobiol Dis. 2006;21:404–12. analysis of microglia-derived exosomes: metabolic role of the 39. Fevrier B, Vilette D, Archer F, Loew D, Faigle W, Vidal M, et al. Cells release aminopeptidase CD13 in neuropeptide catabolism. J Immunol. 2005;175: prions in association with exosomes. Proc. Natl. Acad. Sci. U.S.A. National 2237–43. Acad. Sciences. 2004;101:9683–8. Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 19 of 19 40. Yoon YJ, Kim OY, Gho YS. Extracellular vesicles as emerging intercellular communicasomes. BMB Rep. 2014;47:531–9. 41. Clayton A, Mitchell JP, Court J, Mason MD, Tabi Z. Human tumor-derived exosomes selectively impair lymphocyte responses to interleukin-2. Cancer Res. 2007;67:7458–66. 42. Bianco F, Pravettoni E, Colombo A, Schenk U, Möller T, Matteoli M, et al. Astrocyte-derived ATP induces vesicle shedding and IL-1 beta release from microglia. J Immunol. 2005;174:7268–77. 43. Ye L, Huang Y, Zhao L, Li Y, Sun L, Zhou Y, et al. IL-1β and TNF-α induce neurotoxicity through glutamate production: a potential role for neuronal glutaminase. J Neurochem. 2013;125:897–908. 44. Wang K, Ye L, Lu H, Chen H, Zhang Y, Huang Y, et al. TNF-α promotes extracellular vesicle release in mouse astrocytes through glutaminase. J Neuroinflammation. 2017;14:87. 45. Takeuchi H, Jin S, Wang J, Zhang G, Kawanokuchi J, Kuno R, et al. Tumor necrosis factor-α induces neurotoxicity via glutamate release from hemichannels of activated microglia in an autocrine manner. J Biol Chem. 2006;281:21362–8. 46. Madsen PM, Clausen BH, Degn M, Thyssen S, Kristensen LK, Svensson M, et al. Genetic ablation of soluble tumor necrosisfactorwithpreservationof membrane tumor necrosis factor is associated with neuroprotection after focal cerebral ischemia. J Cereb Blood Flow Metab. 2016;36:1553–69. 47. Jiang L, Vader P, Schiffelers RM. Extracellular vesicles for nucleic acid delivery: progress and prospects for safe RNA-based gene therapy. Gene Ther. 2017;24:157–66. 48. Hammond JW, Cai D, Verhey KJ. Tubulin modifications and their cellular functions. Curr Opin Cell Biol. 2008;20:71–6. 49. Liu X, Cheng R, Ye X, Verbitsky M, Kisselev S, Mejia Santana H, et al. Increased rate of sporadic and recurrent rare genic copy number variants in Parkinson’s disease among Ashkenazi Jews. Mol Genet Genomic Med. 2013; 1:142–54. 50. Lu Y, Liu X, Xie M, Liu M, Ye M, Li M, et al. The NF-κB-responsive long noncoding RNA FIRRE regulates posttranscriptional regulation of inflammatory gene expression through interacting with hnRNPU. J Immunol. 2017;199:3571–82. 51. Barbierato M, Borri M, Facci L, Zusso M, Skaper SD, Expression GP. Differential responsiveness of central nervous system glial cell populations to the acute phase protein serum amyloid a. Sci Rep. 2017;7:12158. 52. Winblad B, Amouyel P, Andrieu S, Ballard C, Brayne C, Brodaty H, et al. Defeating Alzheimer’s disease and other dementias: a priority for European science and society. Lancet Neurol. 2016;15:455–532. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Neuroinflammation Springer Journals

Inflammation leads to distinct populations of extracellular vesicles from microglia

Free
19 pages

Loading next page...
 
/lp/springer_journal/inflammation-leads-to-distinct-populations-of-extracellular-vesicles-pbRfuR0T4U
Publisher
BioMed Central
Copyright
Copyright © 2018 by The Author(s).
Subject
Biomedicine; Neurosciences; Neurology; Neurobiology; Immunology
eISSN
1742-2094
D.O.I.
10.1186/s12974-018-1204-7
Publisher site
See Article on Publisher Site

Abstract

Background: Activated microglia play an essential role in inflammatory responses elicited in the central nervous system (CNS). Microglia-derived extracellular vesicles (EVs) are suggested to be involved in propagation of inflammatory signals and in the modulation of cell-to-cell communication. However, there is a lack of knowledge on the regulation of EVs and how this in turn facilitates the communication between cells in the brain. Here, we characterized microglial EVs under inflammatory conditions and investigated the effects of inflammation on the EV size, quantity, and protein content. Methods: We have utilized western blot, nanoparticle tracking analysis (NTA), and mass spectrometry to characterize EVs and examine the alterations of secreted EVs from a microglial cell line (BV2) following lipopolysaccharide (LPS) and tumor necrosis factor (TNF) inhibitor (etanercept) treatments, or either alone. The inflammatory responses were measured with multiplex cytokine ELISA and western blot. We also subjected TNF knockout mice to experimental stroke (permanent middle cerebral artery occlusion) and validated the effect of TNF inhibition on EV release. Results: Our analysis of EVs originating from activated BV2 microglia revealed a significant increase in the intravesicular levels of TNF and interleukin (IL)-6. We also observed that the number of EVs released was reduced both in vitro and in vivo when inflammation was inhibited via the TNF pathway. Finally, via mass spectrometry, we identified 49 unique proteins in EVs released from LPS-activated microglia compared to control EVs (58 proteins in EVs released from LPS- activated microglia and 37 from control EVs). According to Gene Ontology (GO) analysis, we found a large increase of proteins related to translation and transcription in EVs from LPS. Importantly, we showed a distinct profile of proteins found in EVs released from LPS treated cells compared to control. Conclusions: We demonstrate altered EV production in BV2 microglial cells and altered cytokine levels and protein composition carried by EVs in response to LPS challenge. Our findings provide new insights into the potential roles of EVs that could be related to the pathogenesis in neuroinflammatory diseases. Keywords: Microglia, Extracellular vesicles (EVs), Neuroinflammation, TNF Background [3]. Emerging evidence has shown that microglia are key Microglia are considered the main innate-immune cells of causative players in neuroinflammation, which in turn is the central nervous system (CNS). They continuously sur- believed to play a major role in neurodegenerative vey their microenvironment and have the ability to interact diseases [4]. with neurons to regulate their activity [1]. In the healthy Microglia are highly dynamic cells with the ability to brain, microglia continuously survey their surroundings transform their morphology from ramified to amoeboid with highly dynamic processes [2] and become activated in and alter their phenotypes corresponding to diverse con- response to injury, infection or neurodegenerative processes ditions. Traditionally, macrophages and microglial cells are classified into two different phenotypes, M1 and M2. M1-microglia are proinflammatory, secreting inflamma- * Correspondence: yiyi.yang@med.lu.se; tomas.deierborg@med.lu.se tory cytokines, chemokines, and nitric oxide (NO), Department of Experimental Medical Science, Experimental which is believed to result in neuronal dysfunction and Neuroinflammation Laboratory, Lund University, Lund, Sweden Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 2 of 19 accelerate the progression of neurodegenerative diseases, changed after bacterial infection [17]. Thus, there is a such as Alzheimer’s disease (AD) and Parkinson’s disease critical need for both identification of specific markers (PD). In contrast, M2-microglia are believed to have and particular signaling pathways controlling EV traf- neuroprotective functions, including increased produc- ficking. The mechanism of action of EVs in microglial tion of interleukin (IL)-4 and neurotrophic factors, along communication is poorly understood. In this study, we with an increase in phagocytosis which in turn leads to hypothesized that activation of microglia can secrete a clearance of cell debris and tissue damage [3, 5]. How- distinct population of EV through modulation of specific ever, there is a wide spectrum of microglial activation signaling pathways. We investigated the dynamics of EVs between the two defined phenotypes [6], and microglia from activated microglial (BV2) cells subjected to lipo- might even have specific neuronal functions beyond typ- polysaccharide (LPS) stimulation. We used differential ical pro-/anti-inflammatory responses [7]. A better un- ultracentrifugation to isolate EVs, including microvesi- derstanding of the interactions between microglia and cles and exosomes. EVs were then characterized in terms other cells in the brain is therefore needed in order to of size and concentration by nanoparticle tracking design therapies to ameliorate the detrimental effects of analysis (NTA), while the origin of EVs was indicated by microglial reactions in brain diseases. western blotting using antibodies against CD63, flotillin-1, The main interaction between cells occur through cel- and Alix. Importantly, we also analyzed the levels of in- lular signaling pathways including autocrine, paracrine, flammatory cytokines in EVs. Secretion of EVs was altered and endocrine processes [8] and extracellular vesicles by suppression of inflammation in microglia via inhibition (EVs) can be important to transport signals between of tumor necrosis factor (TNF) signaling in vivo and in cells. In fact, increasing evidence has shown EVs are vitro. Subsequently, qualitative proteomic analysis was considered one of the main participants in cell-to-cell performed to reveal a different protein composition of communication along with having a proposed role in the EVs in response to LPS challenge. Taken together, our spread of pathology in neurodegenerative disease [9, 10]. findings provide new insights into the role of EVs in regu- These vesicles are able to carry pathogen-associated and lating microglial cell communication. damage-associated molecular patterns that act as signals to regulate and propagate the inflammatory response Methods [11–13]. Hence, investigation of EV trafficking under in- Cell culture flammatory conditions may broaden our understanding BV2, an immortalized murine microglial cell line, was of the roles of microglia in neurodegenerative diseases, cultured in growing medium containing Dulbecco’s as well as their potential in therapeutic manipulation. modified Eagle medium (DMEM) (Gibco™GlutaMAX™, The secretion of EVs is a highly conserved process Thermo Fisher Scientific) supplemented with 10% [14]. However, a number of studies using proteomic ana- heat-inactivated fetal bovine serum (FBS) and 1% peni- lysis of EVs released by various cell types, including cillin/streptomycin (Thermo Fisher Scientific) in 5% microglia, have revealed a diverse range of markers and CO in air at 37 °C in a humidified incubator. Cells were alteration of protein composition [15, 16]. The lack of re-cultured every 2 days starting at a concentration of knowledge on EV’s regulation in vitro and in vivo halts a 2×10 cells/ml in T75 flask (Sarstedt). For a large scale clear understanding of EV functions in cell-to-cell com- of EV collection, microglia were plated in T175 flask (Sar- munication. It is likely that different subsets of EVs have stedt). For inflammatory activation, cells were challenged different functional properties, and trafficking of EVs is with 1 μg/ml LPS (Sigma-Aldrich, Clony 0127-B8) for 12 h most likely modulated by specific signaling pathways. and then grown for 12 h in serum-free media prior to col- Although consensus within the field is being reached, lection of EVs. For TNF inhibition experiment, microglia the classification of EVs is not easy. While different sub- were plated in growing medium either with 1 μg/ml LPS, sets of EVs are being described with increasing rate, for 200 ng/ml etanercept, or both for 12 h. EVs were collected simplicity, we will focus on two different classes of EVs from serum-free media 12 h after treatment. of different sizes and origins. EVs shed directly from the plasma membrane are characterized as microvesicles or Animals ectosomes ranging from 100 to 1000 nm. Exosomes are Adult male C57BL/6 mice (between 7 and 8 weeks of generated within the endosomal pathway and terminate age, n = 20) were purchased from Taconic Ltd. (Ry, at the multivesicular endosomal body (MVB), whereby Denmark) and transferred to the Laboratory of Biomedi- they are released upon the MVB fusing with the plasma cine, University of Southern Denmark, where they were membrane. Generally, the size of exosomes is smaller allowed to acclimatize for 7 days prior to surgery. TNF than microvesicles and below 100 nm. knockout (TNF-KO) breeding couples were originally A recent study has shown that the size distribution purchased from The Jackson Laboratory and transferred and protein composition of EVs in macrophages can be to the Laboratory of Biomedicine where they were Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 3 of 19 established as a colony. Animals were housed under di- Measurement of extracellular vesicles size by nanoparticle urnal lighting conditions and given free access to food tracking analysis (NTA) and water [18]. All animal experiments were performed The size and total number of EVs were measured by in accordance with the relevant guidelines and regula- using NanoSight LM10 (Malvern, UK) with the technol- tions approved by the Danish Animal Ethical Committee ogy of Nanoparticle Tracking Analysis (NTA). In liquid (numbers 2011/561-1950 and 2013-15-2934-00924). suspension, particles undergo Brownian motion together with light scattering properties, the size distribution and concentration of EVs samples can be obtained [17]. Sam- Induction of experimental stroke, permanent middle ples were diluted with distilled water to obtain optimal 6 9 cerebral artery occlusion (pMCAO) concentration for detection (10 –10 particles/ml) and The distal part of the left middle cerebral artery was per- injected with a continuous syringe system for 30 s × 5 manently occluded under Hypnorm and Dormicum times at speed 50 μl/min. Data acquisition was undertaken anesthesia (fentanyl citrate (0.315 mg/ml; Jansen-Cilag) at ambient temperature and measured 5 times by NTA. and fluanisone (10 mg/ml; Jansen-Cilag, Birkerød, Data were analyzed with NTA 2.2 software (Malvern, UK) Denmark), and midazolam (5 mg/ml; Hoffmann- La with minimum expected particle size 10 nm. Roche, Hvidovre, Denmark)), respectively. After surgery, mice were injected subcutaneously with 1 ml of 0.9% sa- Western blot analysis line and allowed to recover in a 25 °C controlled envir- Cell pellets and EVs were lysed in RIPA buffer (Sigma-Al- onment. Mice surviving for 5 days were returned to the drich) supplemented with proteinase inhibitors (Thermo conventional animal facility after 24 h. For post-surgical Scientific) and PhosphoStop (Roche Diagnostics GmbH). analgesia, mice were treated with 0.001 mg/20 g bupre- The concentration of cell lysates was determined using norphine hydrochloride (Temgesic, Schering-Plough, bicinchoninic acid assay (BCA) (Thermo Scientific), while Ballerup, Denmark) three times at 8-h intervals, starting concentrations obtained using NanoSight were utilized to immediately prior to surgery. Mice were allowed to sur- ensure even loading of EVs. Samples were loaded onto 4– vive for 1 day (immunofluorescent staining and cytokine 20% Mini-Protean TGX Precast Gels (Bio-Rad) and then measurement) or 5 days (EV analysis) whereafter they transferred to Nitrocellulose membranes (Bio-Rad) using were killed using either an overdose of pentobarbital Trans-Blot Turbo System (Bio-Rad). Membranes were (200 mg/ml) containing lidocaine (20 mg/ml) (Glostrup incubated with following primary antibodies: Alix (Cell Apotek, Glostrup, Denmark) and perfused through the Signaling; 1:1000), flotillin-1 (Cell Signaling; 1:1000), left ventricle using 4% paraformaldehyde (PFA) or killed CD63 (Santa Cruz Biotechnology; 1:1000), inducible nitric by cervical dislocation. The blood and brains were col- oxide synthase (iNOS) (Santa Cruz Biotechnology; lected for further analysis. 1:3000), NLRP3 (Adipogen; 1:1000) and pro-caspase1 (Adipogen; 1:1000). All secondary antibodies were horse-radish protein (HRP) conjugated (Vector; 1:5000 or Extracellular vesicle isolation procedure and transmission 1:10000). Protein bands were detected using Clarity West- electron microscopy (TEM) ern ECL Substrate (Bio-Rad) or Pierce™ ECL Western For isolation of EVs, cells were cultivated in growing Blotting Substrate (ThermoFisher), and imaged on medium DMEM and then deprivation of serum for a Bio-Rad ChemiDoc XRS+. Protein levels were normalized period of 12 h. The media was then collected and sub- to beta-actin (Sigma-Aldrich; 1:15,000). Image lab™ soft- jected to a series of low-speed centrifugation steps ware (Bio-Rad) was used to analyze the results. (500×g for 10 min, 2000×g for 10 min, and 10,000×g for 30 min) at 4 °C in order to remove cells and cellular Multiplex cytokine enzyme-linked immunosorbent assay debris. The supernatant was then collected in centrifuge (ELISA) tubes (Beckman Coulter) and spun at 100,000×g for The concentrations of different cytokines in EVs and in 70 min before the resultant EV pellet was washed in a isolated media as well as serum from mice were mea- large volume of phosphate-buffered saline (PBS) before sured with the MSD Mouse Proinflammatory V-Plex repeating the 100,000×g spin. The pellets containing EVs Plus Kit (Interferonγ (IFNγ), IL-1β, IL-2, IL-4, IL-5, IL-6, were resuspended in 20 μl of PBS and stored at 4 °C or IL-10, IL-12p70, TNF, C-X-C motif chemokine ligand 1 long term at − 20 °C. For electron microscopic analysis, (KC/GRO), Mesoscale) using a QuickPlex SQ120 Plate samples of EVs were fixed with an equal volume of 2% Reader (Mesoscale Discovery, Rockville, USA) according PFA and loaded onto Formvar/carbon-coated electron to the manufacturer’s instructions. The data was ana- microscopic grids. EVs were observed under TEM at lyzed with MSD Discovery Workbench software. The 80 kV. TEM was carried out at Lund University Bioima- levels of cytokines in EVs and media were normalized to ging Center. each samples total protein content in cell lysates. In Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 4 of 19 total, 6 independent EV samples were analyzed; however, 0.1% (v/v) formic acid in LC-mass spectrometry grade those samples that were under the lowest detection limit water (solvent A) and 0.1% (v/v) formic acid in aceto- were removed from the statistical analysis. nitrile (solvent B). Peptides were first loaded with a con- stant pressure mode with a flow rate of solvent A onto Immunohistochemistry the trapping column. Subsequently, peptides were eluted Immunofluorescent double labeling for TNF and CD11b via the analytical column at a constant flow of 300 nl/ was performed on 16-μm thick, cryostat-cut tissue sec- min. During the elution step, the percentage of solvent B tions from C57BL/6 mice with 1-day survival after increased from 5 to 22% in the first 20 min, then in- pMCAO as previously described in detail [18, 19]. creased to 32% in 5 min and finally to 98% in a further 2 min and was keeping it for 8 min. The peptides were Extracellular vesicle fluorescent labeling introduced into the mass spectrometer via a Stainless Following isolation, EVs were labeled with PKH67 Green steel emitter 40 mm (Thermo Fisher) and a spray volt- Fluorescent Cell Linker Midi Kit for General Cell age of 1.9 kV was applied. The capillary temperature was Membrane Labeling (Sigma-Aldrich) according to the set at 275 °C. manufacturer’s instructions. Briefly, EVs were resus- Data acquisition was carried out using a top N-based pended in 1 ml PBS before 1 ml of Diluent C supple- data-dependent method with cycle time of 3 s. The mas- mented 4 μl PKH67 dye. Samples were incubated at ter scan was performed in the Orbitrap in the range of room temperature for 4 min prior to the addition of 350–1500 mass to charge ratio (m/z) at a resolution of 2 ml of 1% bovine serum albumin (BSA) (VWR Inter- 60,000 full width at half-maximum (FWHM). The filling national) to bind excess dye. Samples were then supple- time was set at maximum of 50 ms with limitation of mented with 5 ml PBS and placed in 300 kDa Vivaspin 4×10 ions. In a second stage of tandem mass spec- filters (Sartorius Stedim Biotech GmbH, Goettingen, trometry (MS/MS) ion trap collision-induced dissoci- Germany), prior to centrifugation for 5 min at 4000×g to ation was acquired using parallel mode, filling time remove excess dye. This process was repeated a further maximum 300 ms with limitation of 2 × 10 ions, a pre- two times, followed by a further two washes in a clean cursor ion isolation width of 1.6 m/z and resolution of filter with DMEM (Thermo Fisher Scientific) in place of 15,000 FWHM. Normalized collision energy was set to + + PBS. The same procedure minus EVs was carried out as 35%. Only multiply charged (2 to 5 ) precursor ions control. were selected for MS/MS. The dynamic exclusion list was set to 30 s and relative mass window of 5 ppm. TNF inhibition on dynamics of extracellular vesicle trafficking Bioinformatic analysis PKH67-labeled EVs (2 × 10 particles/ml) were incu- Gene Ontology (GO) classifications and enrichments bated with BV2 cells as indicated previously. After 12 h were performed using FunRich [20]. The identified pro- incubation, cells were washed three times with PBS and teins were compared with web tool Exocarta database one time with 1 M NaCl prior to fixation with 4% PFA and also with the Top100 exosomal proteins from the for 20 min on ice. Cells were then imaged by fluores- database [21]. cence microscope (Olympus IX71) at × 20 magnification and images processed using Cellsens Standard version Data analysis 1.6 software (Olympus). Vesicle uptake was analyzed by MS/MS data were searched with PEAKS (7.5). UniProt measuring fluorescent intensity using ImageJ software Mus musculus (house mouse, including 16,792 se- (National Institutes of Health). quences) was used with non-tryptic specificity allowing up to 3 missed cleavages. A 15 ppm precursor tolerance Mass spectrometry and a 0.1 Da fragment tolerance were used. Oxidation Mass spectrometry was carried out on an Orbitrap (M) and deamidation (NQ) were treated as dynamic Fusion Tribrid MS system (Thermo Scientific) equipped modification and carbamidomethylation (C) as a fixed with a Proxeon Easy-nLC 1000 (Thermo Fisher). modification. Maximum post-translational modification Injected peptides were trapped on an Acclaim PepMap per peptide was 2. Search results were filtered by using C18 column (3-μm particle size, 75-μm inner diameter 1% false discovery rate and 2 unique peptides. × 20 mm length). After trapping, gradient elution of pep- The rest was evaluated using either unpaired t test or tides was performed on an Acclaim PepMap C18 col- one-way ANOVA followed by Tukey’s test for multiple umn (100 Å 3 μm, 150 mm, 75 μm). The outlet of the comparisons. All statistical analysis was done using the analytical column was coupled directly to the mass spec- GraphPad Prism 7.0 software for Macintosh (GraphPad trometer using a Proxeon nanospray source. The mobile Software, San Diego, CA, USA). Data are presented as phases for liquid chromatography (LC) separation were means ± SD. A confidence interval of 95% was set as Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 5 of 19 significant. The exact P values are given in the figure revealed that LPS-stimulated microglia release larger EV legends. Figures were organized using Adobe Illustrator. populations compared to control (Fig. 1g). EV samples were blotted for different EV markers including a marker Results for plasma membrane (flotillin-1) and an endosomal Proinflammatory responses from LPS-stimulated BV2 marker (Alix) as well as the EV marker CD63 for MVB microglial cells (Fig. 1e) to elucidate the subcellular origin of the EVs [25]. First, we examined the activation of BV2 cells after 12 h We observed that in EVs released from microglia after culture in the presence of 1 μg/ml LPS followed by LPS activation had a higher ratio of flotillin-1 and CD63 deprivation of serum for 12 h to elicit a strong inflam- when compared to those from non-activated cells when matory reaction. The inflammatory enzyme iNOS is loading equal amounts of EVs in each lane, suggesting al- expressed by activated microglia [22] and as expected, tered EV biogenesis and release after an inflammatory the level of iNOS was significantly increased in cells stimulus. upon LPS stimulation (Fig. 1a). Moreover, the protein levels of other important inflammatory mediators, Increased production of TNF and IL-6 in EVs upon LPS NLRP3 (Nod-like receptor protein 3) and pro-caspase1 activation (involved in the maturation, production, and release of Next, we studied the cytokine release from LPS-activated IL-1β and IL-18 [23]) were found to be elevated BV2 microglia to evaluate the free concentration of re- considerably (Fig. 1b, c). According to our previous stud- leased cytokines and the cytokine concentration in EVs. ies, the viability of BV2 cells is not affected by LPS acti- Culture medium from activated and non-activated micro- vation [22, 24]. These results suggest that an activated glia was collected, and EVs were isolated from equal pro-inflammatory status of microglia remained over the amounts of medium. EVs, along with the EV-depleted 12 h EV collection period following LPS treatment. media, were then subjected to analysis by multiplex ELISA. Out of ten inflammatory cytokines analyzed, the Changes in EV size distribution in response to LPS levels of two pro-inflammatory cytokines, TNF and IL-6, activation were found to be significantly increased in EVs from Transmission electron microscopy (TEM) was per- activated microglia (Fig. 2a, b). TNF and IL-6 are two formed to visualize microglial-derived EVs (Fig. 1d). Im- representative pro-inflammatory cytokines produced by ages from control condition and LPS treatment revealed microglia related to neurodegenerative diseases [3]. heterogeneous populations of EVs from microglia in the Notably, there was also a significant upregulation in the range between 100 and 1000 nm in diameter. LPS treat- concentration of these two cytokines in medium ment seemed to induce microglial release of larger EVs, (Additional file 1). However, other pro-inflammatory around 200–300 nm in diameter. cytokines such as IL-5 and IL-1β were found to be To further characterize EVs released under increased only in the medium, but not in EVs pro-inflammatory condition by LPS treatment, we com- (Additional files 1 and 2). Importantly, the level of pared size distribution of EVs derived from non-activated TNF was much higher increased, 22-fold, compared and LPS-activated microglia using Nanoparticle Tracking to 5-fold in IL-6. Thus, we further investigated the ef- Analysis (NTA). The range of EVs detected with NTA was fect of TNF in regulation of EV release in the follow- from 50 to 700 nm. Different size subpopulations of EVs ing study. were observed in EVs released from activated and non-activated microglial cells (Fig. 1f). According to the Reduction of EVs by inhibition of inflammation via TNF diameter of EVs, we can classify them into two subpopula- pathway tions: one ranging from 50 to 100 nm can be considered In view of the specific increase of TNF in the EVs after as exosomes and another subpopulation with size LPS-stimulation, we set out to further characterize the exceeding 100 nm can be regarded as microvesicles role of TNF in the microglial release of EV. To investi- (MVs). As can be seen in Fig. 1f, the population of MVs gate the mechanistic basis for EV regulation on LPS ranging from 300 to 400 nm has a higher frequency in the signaling in microglial cells, we quantified the number of LPS-activated EVs. We found EVs released from EVs released from microglia after TNF inhibition with LPS-activated cells to be significantly larger (178 ± etanercept upon LPS activation. Notably, activated 5.66 nm) compared to the size of EVs from control cul- microglia secreted a 30-fold increase in the number of tures (159 ± 4.95 nm) (Fig. 1f; p < 0.001). D90 measure- EVs under LPS stimulation (Fig. 3a). In contrast, the ment shows the upper limit of 90% measured particles, effect of LPS on EV release was completely attenuated and in control condition, the D90 value for EVs was 205 ± down to control levels using etanercept (200 ng/ml), a 3.61 nm, whereas the value was 254 ± 11.06 nm in TNF inhibitor that blocks both soluble and transmem- LPS-activated samples (Fig. 1f; p < 0.001). These results brane forms of TNF (Fig. 3a). In the presence of the Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 6 of 19 Fig. 1 Characterization of microglia-derived extracellular vesicles. Microglia (BV2) were activated by treatment with LPS for 12 h before extracellular vesicles were isolated from the media of both LPS-treated (LPS) or control (CTRL) cells. a Western blot analysis of protein expression levels of iNOS in cell lysates from control and activated microglia with LPS stimulation (Mean ± SD, n = 5). b, c Components of the inflammasome, NLRP3 and pro-Caspase 1, were measured by western blot in cell lysates with representative pictures of blots (Mean ± SD, n = 5). d Representative TEM imaging of extracellular vesicle populations from CTRL and LPS-derived microglia. The imaging illustrates heterogeneity and sphere structure of extracellular vesicles. Typical microvesicles are pointed with red arrows. Scale bars: 500 nm. e Western blots showed alterations of expression levels of vesicle markers indicated different origins of extracellular vesicles from CTRL and LPS. Three biological independent samples were blotted in each condition. f The size of extracellular vesicles was determined in diameter from CTRL and LPS-treated microglia. The mean size shows the average diameter of extracellular vesicles in samples (n = 12). D90 demonstrates the upper limit of extracellular vesicles size in 90% of the population (n = 12). g Representative histograms of extracellular vesicles size distributions collected from CTRL and LPS conditions. Sample from LPS condition was diluted 25 times more than control to obtain similar concentration of EVs to demonstrate size distribution. Concentrations (× 10 particles/ml) by size (nm) of recorded extracellular vesicles are showed. The major subpopulations in EV samples are indicated with digitals showing the mean diameter of extracellular vesicles. Histograms were generated from five independent measurements by NTA 2.2 software (Unpaired t test, *P < 0.05; ***P < 0.001; ****P < 0.0001) TNF inhibitor, the amount of EVs released from cells indicating the reduction of EVs was due to TNF treat- was also reduced compared with control condition ment on EV release not by affecting EV uptake/turnover (Fig. 3a). Next, we wanted to understand whether the (Additional file 3). reduction of EV is caused by decreased EV uptake Next, we evaluated the degree of inflammatory status through inhibition of TNF. Thus, we assessed the cap- in microglia in relation to the amount of EVs secreted. ability of EV uptake under different conditions as indi- To that aim, the level of iNOS was analyzed by western cated above. We found no significant difference between blot on cells previously treated in different conditions. the conditions on internalization of PHK76-labeled EVs, We found two-fold reduction in iNOS levels following Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 7 of 19 were statistically elevated in WT at day 1 and unchanged in TNF-KO mice (Additional file 4). IL-5 and IL-12p70 were also found statistically elevated in WT mice 1 day after manipulation, but not in TNF-KO (Fig. 4a, c). While in the case of IL-1β, IL-6, and KC/GRO, expres- sion levels were considerably increased 1 day after pMCAO in both types of mice compared with unmanip- ulated mice (Fig. 4b, d, f). Notably, such increases in- duced by pMCAO were significantly attenuated by deficiency of TNF in mice. Levels of IL-10 were remark- ably lower in TNF-KO mice compared with WT mice after pMCAO, but not before (Fig. 4e). Levels of IL2, IL-4, and IFNγ were remained at baseline levels at day 1 in both types of mice subjected to pMCAO (Additional file 4). In conclusion, we found clear evidence of signifi- cant upregulation of proinflammatory cytokines at day 1 Fig. 2 Increased levels of proinflammatroy cytokines in microglia- after pMCAO in both types of mice. However, in derived extracellular vesicles upon LPS activation. The levels of TNF-KO mice, such inflammation induced by pMCAO cytokines were analyzed by multiplex ELISA plate. a Bar graph shows was remarkably attenuated at day 1 after stroke. a significant upregulation of TNF (n = 3). b Bar graph shows a significant upregulation of IL-6 (n = 6)(Mean ± SD, Unpaired t test, Decreased EVs in TNF knockout mice after focal cerebral *P < 0.05; ***P < 0.001) ischemia Given the impact of pMCAO on systemic inflammation TNF inhibition. Interestingly, the reduction of EVs re- in WT and TNF-KO mice, we examined microglia acti- leased upon TNF inhibition was reduced 16-fold, sug- vation in the brain. We observed that TNF co-localized gesting that TNF signaling is particularly important with the microglial marker, CD11b, in the peri-infarct when it comes to reducing the number of EV released in area 1 day after pMCAO (Fig. 3d), which shows that proinflammatory activation of microglia (Fig. 3a). To- focal cerebral ischemia could induce an initial phase of gether, these results implicate that the secretion of EVs microglia activation involving TNF signaling and tissue is dramatically impeded by TNF inhibition in injury, as we have shown before [30]. Moreover, the vol- LPS-activated microglia, which is only partly associated ume of infarct was also assessed in WT mice and to an overall reduction in the inflammatory status. TNF-KO mice 5 days after pMCAO in the previous study, which has shown that the injury was significantly Evaluation of systemic inflammation in mice after focal larger in TNF-KO mice than WT mice [18]. Previously, cerebral ischemia we have shown that the infarct volume correlates with As the level of TNF has shown to be upregulated in EVs the number of EVs in plasma after this stroke model under LPS activation in vitro, we next studied whether [31]. Thus, we analyzed the number of EVs in the the secretion of EVs in vivo was affected by complete ab- plasma of TNF-KO mice after pMCAO. The production lation of TNF signaling in a strong neuroinflammatory of EVs increased from both genotypes 5 days after situation. To this purpose, we chose an experimental pMCAO indicating inflammation occurred, which in line stroke model, permanent middle cerebral artery occlu- with our previous findings in vitro (Fig. 3a, c). Import- sion (pMCAO), as an in vivo inflammatory model. Ex- antly, we found here that a complete ablation of TNF perimental stroke in rat and mouse is known to induce successfully reduced the counts of EVs in the plasma of neuroinflammation in the brain [26, 27] as well as alter TNF-KO mice 5 days after induction of permanent focal the inflammatory response in the periphery [28, 29]. Our cerebral ischemia (Fig. 3c), but not at day 1 (Add- earlier study has revealed significant increase of TNF re- itional file 5). These results are consistent with our in ceptors, toll-like receptor (TLR) 2 and IL-1β at mRNA vitro data and indicate the number of EVs is related to levels in wild type (WT) and TNF-KO mice 1 day after an inflammatory event and that TNF signaling is import- pMCAO, indicating inflammation occurred in the brain ant in the mechanism regulating EV release. [18]. We also wanted to assess systemic inflammatory response in mice after pMCAO. Therefore, we measured Identification of microglial EVs proteins levels of different cytokines in serum using Multiplex To further elucidate cellular communication by EVs ELISA from WT and TNF-KO mice without manipula- under inflammation, we set out to identify proteins in tion and 1 day after pMCAO. TNF expression levels association to EVs released by BV2 cells. To this aim, Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 8 of 19 Fig. 3 Reduction in the number of microglia-derived extracellular vesicles by inhibition of TNF signaling pathway. Extracellular vesicles were visualized by TEM. Concentrations of extracellular vesicles in conditional media and plasma from mice were measure by NTA. a Comparison of extracellular vesicles concentrations in different conditional media. BV2 microglia cell line was treated with either LPS (1 μg/ml) or etanercept (200 ng/ml) or in the presence of both. Control (CTRL) was cells without any treatment (Mean ± SD, one-way ANOVA, ***P <0.001, n =3). b Expression levels of iNOS in BV2 cells previously treated with different conditions were analyzed by western blot. iNOS bands were not able to be visualized in the conditions of control and etanercept due to out of the detection limit (Mean ± SD, Unpaired t-test, *P < 0.05, n =4). c Bar graph shows a comparison of the amount of extracellular vesicles in plasma from mice (WT, n = 6; TNF-KO, n = 6) without surgery and mice (WT, n =9; TNF-KO, n = 8) 5 days after a stroke model, permanent middle cerebral artery occlusion (pMCAO) (Mean ± SD, one-way ANOVA followed by Tukey’s test for multiple comparisons, *P < 0.05, ***P < 0.001). d The brain section in area of infarct from C57BL/6 mouse subjected to pMCAO stained with anti-TNF (green) and anti-CD11b (red), nuclei stained with DAPI. TNF+ cells are also stained for CD11b indicated with arrow; scale bar: 20 μm proteomic analysis was performed using mass spectrom- Gene Ontology annotations related to membrane and etry on EV samples from LPS-activated and control extracellular exosome (Fig. 5c, d). The identified proteins microglia. Biological independent duplicates of EVs were were also compared with the ExoCarta database, which pooled together and analyzed. In total, 86 proteins were has exosomal proteins identified from previous publica- identified with two peptides confirmed and high confi- tions [21]. In our samples, only one protein has not been dence, as -10lgP is set for 20 as threshold, using PEAKS reported in ExoCarta database (Fig. 5a). Compared with analysis program (Tables 1 and 2)[32]. In total, 37 pro- top 100-ranked proteins presented in the database, we teins from control and 58 proteins from LPS-stimulated identified 17 of them in our study (Fig. 5a). Notably, 89.5% microglia were successfully mapped with UniProt data- proteins from this study have not been reported before base. Among them, we found 9 proteins in common and first identified in microglia (Table 4), suggesting that (Table 3) and 49 specific proteins present in LPS condi- these cells are releasing specific EVs. tion. The analysis showed enzymes, chaperones, ribosomal structure proteins, and membrane receptors previously re- Functional profiles of the quantified microglia EVs proteins ported in other immune cells in our samples (Fig. 5 d, e). Next, we wanted to know the functions of the quantified Of these, the majority of the identified proteins were asso- proteins differentially secreted in EVs under LPS activa- ciated with RNA binding, and more than half of them had tion. The 86 quantified proteins in the microglial-derived Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 9 of 19 IL-1 ab IL-5 2.5 9 ** **** **** 2.0 1.5 0.6 1.0 0.3 0.5 0.0 0.0 WT TNF-KO WT TNF-KO WT TNF-KO WT TNF-KO Control 1 D after pMCAO Control 1 D after pMCAO IL-6 IL-12p70 c d ** * *** * *** 0 0 WT TNF-KO WT TNF-KO WT TNF-KO WT TNF-KO Control Control 1 D after pMCAO 1 D after pMCAO KC/GRO IL-10 e f ** ** ** 2.0 ** 1.5 1.0 0.5 20 0.0 0 WT TNF-KO WT TNF-KO WT TNF-KO WT TNF-KO Control Control 1 D after pMCAO 1 D after pMCAO Fig. 4 Systemic inflammation in mice after focal cerebral ischemia. The levels of cytokines in serum were analyzed by multiplex ELISA from WT and TNF-KO mice without manipulation and 1 day after focal cerebral ischemia. a and c Bar graphs show significant upregulation of IL-5 and IL- 12p70 in WT mice after pMCAO. b, d, f Bar graphs show significant upregulation of IL-1β, IL-6, and KC/GRO in both types of mice subjected to pMCAO. e Bar graph shows remarkable reduction of IL-10 in TNF-KO mice compared with WT mice 1 day after pMCAO. (Mean ± SD, one-way ANOVA followed by Tukey’s test for multiple comparisons, n =3–6, *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001) EVs were analyzed using FunRich software [21] to validate flotillin-1 and CD63 indicating more MVs released. In the our data referred to the Vesiclepedia Exosome Database categories of molecular function, the majority of the pro- [33]. The analysis was based on their cellular compart- teins were annotated to RNA binding, 38.9% in control ments to which they belong, their molecular function and and 72.4% in LPS (Fig. 5c). Importantly, the proteins con- the biological process in which they are involved. In the tributed to ribosome function raised from 2.8 to 56.9% analysis, some of the proteins were annotated in more after LPS challenge (Fig. 5c). The profile of proteins in than one cellular component, molecular function, and bio- molecular function was dramatically altered after inflam- logical process. matory stimulation. It is likely that microglia released a Firstly, enrichment in pathways from membrane and distinct population of EVs related to transduction and extracellular exosome was observed for GO analysis on translation after activation of LPS. Intriguingly, the profile cellular component (Fig. 5d). There were large increases of EVs was peculiarly different after activation according in proteins from LPS-stimulated EVs from ribosome, focal to classification of biological process (Fig. 5e). EVs were adhesion, extracellular matrix, and membrane (Fig. 5d). In found involved in regulation of neuron projection devel- contrast, a reduction of proteins associated with extracel- opment at “rest” status (control), however, with activation lular exosome was found in LPS-stimulated EVs, which is the proteins identified in EVs shifted to ribosomal assem- also confirmed by western blot with increased ratio of bly and translation. The role of microglia was changed pg/mL pg/mL pg/mL pg/mL pg/mL pg/mL Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 10 of 19 Table 1 List of proteins identified in control EV samples Protein ID Gene name Protein name Score P08226 APOE Apolipoprotein E 506 P21956 MFGE8 Lactadherin 136 P21460 CST3 Cystatin-C 123 Q9WU78 PDCD6IP Programmed cell death 6-interacting protein 121 P11152 LPL Lipoprotein lipase 107 Q8VDN2 ATP1A1 Sodium/potassium-transporting ATPase subunit alpha-1 105 P62737 ACTA2 Actin, aortic smooth muscle 84 Q61753 PHGDH D-3-phosphoglycerate dehydrogenase 80 P07901 HSP90AA1 Heat shock protein HSP 90-alpha 80 P10605 CTSB Cathepsin B 75 P01942 HBA Hemoglobin subunit alpha 73 P63017 HSPA8 Heat shock protein 8 71 Q62419 SH3GL1 Endophilin-A2 70 Q8R366 IGSF8 Immunoglobulin superfamily member 8 69 Q68FD5 CLTC Clathrin heavy chain 1 67 P10923 SPP1 Osteopontin 64 Q61207 PSAP Prosaposin 64 P06869 PLAU Urokinase-type plasminogen activator 62 Q8CGP5 HIST1H2AF Histone H2A type 1-F 61 P52480 PKM Pyruvate kinase 59 P17182 ENO1 Alpha-enolase 56 P04104 KRT1 Keratin, type II cytoskeletal 1 50 P68369 TUBA1A Tubulin alpha-1A chain 47 P01887 B2M Beta-2-microglobulin 45 P09405 NCL Nucleolin 44 P09055 ITGB1 Integrin beta-1 42 P01901 H2-K1 H-2 class I histocompatibility antigen, K-B alpha chain 41 P01899 H2-D1 H-2 class I histocompatibility antigen, D-B alpha chain 36 P29341 PABPC1 Polyadenylate-binding protein 1 41 P06797 CTSL Cathepsin L1 38 P10852 SLC3A2 4F2 cell-surface antigen heavy chain 37 P08905 LYZ2 Lysozyme C-2 34 P10126 EEF1A1 Elongation factor 1-alpha 1 28 Q61937 NPM1 Nucleophosmin 27 Q9DBJ1 PGAM1 Phosphoglycerate mutase 1 25 P28798 GRN Granulins 24 P14206 RPSA 40S ribosomal protein SA 24 Proteins retrieved in control extracellular vesicle samples from mass spectrometry. Protein ID and gene name are according to Uniprot Knowledgebase. Score values were obtained using MASCOT. The Score shows how well the observed protein matches to the stated protein in the database. Only protein identifications supported by at least two high confident peptides (confidence > 95%) were considered from “rest” state to “activated” state by not only produc- In larger proportion, 49 proteins were exclusively tion of cytokine and chemokine, but also with altered se- present in EVs samples from LPS activation, including cretion of EVs, which could be directly linked to some cytoskeleton proteins and ribosomal proteins. Im- detrimental inflammatory responses related to neurode- portantly, three of them were associated with inflamma- generative diseases. tory and neuropathological pathways: tubulin beta 4B Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 11 of 19 Table 2 List of proteins identified in LPS EV samples Protein ID Gene name Protein name Score P68372 TUBB4B Tubulin beta-4B chain 258 P62242 RPS8 40S ribosomal protein S8 201 P47963 RPL13 60S ribosomal protein L13 181 Q9CZX8 RPS19 40S ribosomal protein S19 178 P15864 HIST1H1C Histone H1.2 172 Q8VEK3 HNRNPU Heterogeneous nuclear ribonucleoprotein U 148 Q9CR57 RPL14 60S ribosomal protein L14 146 P63276 RPS17 40S ribosomal protein S17 138 P11152 LPL Lipoprotein lipase 130 P16858 GAPDH Glyceraldehyde-3-phosphate dehydrogenase 128 P68369 TUBA1A Tubulin alpha-1A chain 123 P62082 RPS7 40S ribosomal protein S7 106 Q9D8E6 RPL4 60S ribosomal protein L4 101 P35979 RPL12 60S ribosomal protein L12 101 P14131 RPS16 40S ribosomal protein S16 100 P70696 HIST1H2BA Histone H2B type 1-A 100 P12970 RPL7A 60S ribosomal protein L7a 98 P60710 ACTB Actin, cytoplasmic 1 94 Q9CXW4 RPL11 60S ribosomal protein L11 85 Q9CZM2 RPL15 60S ribosomal protein L15 84 P02301 H3F3C Histone H3.3C 83 P47911 RPL6 60S ribosomal protein L6 83 P43276 HIST1H1B Histone H1.5 83 Q8CGP5 HIST1H2AF Histone H2A type 1-F 82 P14115 RPL27A 60S ribosomal protein L27a 74 P27659 RPL3 60S ribosomal protein L3 72 O55142 RPL35A 60S ribosomal protein L35a 70 P25206 MCM3 DNA replication licensing factor MCM3 68 P62702 RPS4X 40S ribosomal protein S4, X isoform 63 P10126 EEF1A1 Elongation factor 1-alpha 1 63 Q8BP67 RPL24 60S ribosomal protein L24 62 P14206 RPSA 40S ribosomal protein SA 62 P11499 HSP90AB1 Heat shock protein HSP 90-beta 60 P08226 APOE Apolipoprotein E 60 P35980 RPL18 60S ribosomal protein L18 59 P63017 HSPA8 Heat shock protein 8 59 P80318 CCT3 T-complex protein 1 subunit gamma 58 P62918 RPL8 60S ribosomal protein L8 57 O08585 CLTA Clathrin light chain A 55 Q9Z1Q9 VARS Valine-tRNA ligase 54 P01942 HBA Hemoglobin subunit alpha 51 P97351 RPS3A 40S ribosomal protein S3a 50 P62806 HIST1H4A Histone H4 48 P62270 RPS18 40S ribosomal protein S18 44 Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 12 of 19 Table 2 List of proteins identified in LPS EV samples (Continued) Protein ID Gene name Protein name Score P62911 RPL32 60S ribosomal protein L32 43 P14869 RPLP0 60S acidic ribosomal protein P0 41 Q9JIK5 DDX21 Nucleolar RNA helicase 2 41 P62281 RPS11 40S ribosomal protein S11 41 Q68FD5 CLTC Clathrin heavy chain 1 40 P62717 RPL18A 60S ribosomal protein L18a 39 P86048 RPL10L 60S ribosomal protein L10-like 39 P62267 RPS23 40S ribosomal protein S23 37 P25444 RPS2 40S ribosomal protein S2 36 P14148 RPL7 60S ribosomal protein L7 31 P53026 RPL10A 60S ribosomal protein L10a 28 P62264 RPS14 40S ribosomal protein S14 28 P62908 RPS3 40S ribosomal protein S3 23 P04918 SAA3 Serum amyloid A-3 protein 23 Proteins retrieved in LPS extracellular vesicles samples from mass spectrometry. Protein ID and gene name are according to Uniprot Knowledgebase. Score values were obtained using MASCOT. The score shows how well the observed protein matches to the stated protein in the database. Only protein identifications supported by at least two high confident peptides (confidence > 95%) were considered (TUBB4B), heterogeneous nuclear ribonucleoprotein U roles in neurodegenerative diseases [36–38]. Microglia (HNRNPU) and serum amyloid A 3 (SAA3). Nine pro- are an essential component in innate immunity in the teins were shared in samples from control and LPS con- CNS. Activation of microglia is followed by the initiation ditions (Fig. 5b). Proteins related to immune system of intracellular machinery that leads to production of process and stimuli response were detected in both sam- cytotoxic and proinflammatory cytokines and chemo- ples: lipoprotein lipase (LPL), apolipoprotein E (APOE), kines, which promote progression of inflammation and and heat shock protein 8 (HSPA8). Notably, APOE was affect neighboring cells through different mechanisms. particular with high score in the EVs before and after The secretion of proinflammatory signals can be con- LPS activation. Variants of APOE (APOE4) is known as ducted in classic secretion manner or non-canonical the strongest risk factor for late onset Alzheimer’s dis- manner via vesicles. Previous studies have shown the im- ease [34] and has also been suggested to have a proin- portant role of EVs in regulation of cytokine production flammatory effect on microglia [35]. on recipient cells and propagation of pathogenic pro- teins [11, 39]. Discussion The functions of microglia in the brain are diverse Compelling evidence has suggested that the activation of depending on stimulus and different brain regions. In innate immunity and neuroinflammation play crucial response to different inflammatory/homeostatic condi- tions, they can be either beneficial or detrimental. Table 3 The common proteins shared in CTRL and LPS EV Increasing evidence has suggested that there is a spectrum samples of activation in microglia based on the profile of secreted Protein ID Gene name Protein name molecules [3]. However, the importance of EV released by P08226 APOE Apolipoprotein E microglia has not been well characterized. Thus, in the P11152 LPL Lipoprotein lipase present study, we directly evaluated effects of LPS-stimulation on EVs derived from BV2 microglia, in P01942 HBA Hemoglobin subunit alpha terms of physical and biological properties. Our results P63017 HSPA8 Heat shock protein 8 demonstrate that (i) upon LPS-activation, the size dis- Q68FD5 CLTC Clathrin heavy chain 1 tribution of EVs released from microglia increases in Q8CGP5 HIST1H2AF Histone H2A type 1-F size, indicating larger vesicles are released under in- P68369 TUBA1A Tubulin alpha-1A chain flammatory conditions; (ii) IL-6 and in particular TNF P10126 EEF1A1 Elongation factor 1-alpha 1 are increasingly secreted in microglia-derived EVs after LPS challenge; (iii) inhibiting the TNF signaling path- P14206 RPSA 40S ribosomal protein SA way resulted in a robust reduction in the number of Proteins retrieved in extracellular vesicle samples from mass spectrometry. Protein ID and gene name are according to Uniprot Knowledgebase vesicles released from LPS-treated microglia and in Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 13 of 19 Fig. 5 Bioinformatic analysis of the identified proteins in BV2 microglia-derived extracellular vesicles. Gene ontology (GO) analysis of the identified proteins was performed using Exocarta based software FunRich. The comparison was carried on proteins expressed exclusively in each condition. a Comparison of 86 identified proteins from microglial extracellular vesicles with online database Exocarta and the top 100 proteins commonly reported in the same database. b Comparison of the number of proteins identified and quantified in control microglia and LPS-activated microglia derived extracellular vesicles samples. c Bioinformatic analysis from FunRich shows comparison of GO analysis on Molecular Function on proteins from CTRL and LPS-activated microglia released extracellular vesicles (GO terms are with P < 0.01). d Bioinformatic analysis from FunRich shows comparison of GO analysis on Cellular Component on proteins from CTRL and LPS-activated microglia released extracellular vesicles (GO terms are with P < 0.01). e Bioinformatic analysis from FunRich shows comparison of GO analysis on Biological Process on proteins from CTRL and LPS-activated microglia released extracellular vesicles (GO terms are with P < 0.01) Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 14 of 19 Table 4 New proteins found in present microglia-derived EV samples Protein ID Gene name Protein name P08226 APOE Apolipoprotein E P21460 CST3 Cystatin-C Q8VDN2 ATP1A1 Sodium/potassium-transporting ATPase subunit alpha-1 P62737 ACTA2 Actin, aortic smooth muscle Q61753 PHGDH D-3-phosphoglycerate dehydrogenase P10605 CTSB Cathepsin B P01942 HBA Hemoglobin subunit alpha Q62419 SH3GL1 Endophilin-A2 Q8R366 IGSF8 Immunoglobulin superfamily member 8 Q68FD5 CLTC Clathrin heavy chain 1 P10923 SPP1 Osteopontin Q61207 PSAP Prosaposin P06869 PLAU Urokinase-type plasminogen activator Q8CGP5 HIST1H2AF Histone H2A type 1-F P04104 KRT1 Keratin, type II cytoskeletal 1 P68369 TUBA1A Tubulin alpha-1A chain P01887 B2M Beta-2-microglobulin P09405 NCL Nucleolin P09055 ITGB1 Integrin beta-1 P01901 H2-K1 H-2 class I histocompatibility antigen, K-B alpha chain P01899 H2-D1 H-2 class I histocompatibility antigen, D-B alpha chain P29341 PABPC1 Polyadenylate-binding protein 1 P06797 CTSL Cathepsin L1 P10852 SLC3A2 4F2 cell-surface antigen heavy chain P08905 LYZ2 Lysozyme C-2 P10126 EEF1A1 Elongation factor 1-alpha 1 Q61937 NPM1 Nucleophosmin P28798 GRN Granulins P14206 RPSA 40S ribosomal protein SA P68372 TUBB4B Tubulin beta-4B chain P62242 RPS8 40S ribosomal protein S8 P47963 RPL13 60S ribosomal protein L13 Q9CZX8 RPS19 40S ribosomal protein S19 P15864 HIST1H1C Histone H1.2 Q8VEK3 HNRNPU Heterogeneous nuclear ribonucleoprotein U Q9CR57 RPL14 60S ribosomal protein L14 P63276 RPS17 40S ribosomal protein S17 P62082 RPS7 40S ribosomal protein S7 Q9D8E6 RPL4 60S ribosomal protein L4 P35979 RPL12 60S ribosomal protein L12 P14131 RPS16 40S ribosomal protein S16 P70696 HIST1H2BA Histone H2B type 1-A P12970 RPL7A 60S ribosomal protein L7a P60710 ACTB Actin, cytoplasmic 1 Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 15 of 19 Table 4 New proteins found in present microglia-derived EV samples (Continued) Protein ID Gene name Protein name Q9CXW4 RPL11 60S ribosomal protein L11 Q9CZM2 RPL15 60S ribosomal protein L15 P02301 H3F3C Histone H3.3C P47911 RPL6 60S ribosomal protein L6 P43276 HIST1H1B Histone H1.5 Q8CGP5 HIST1H2AF Histone H2A type 1-F P14115 RPL27A 60S ribosomal protein L27a P27659 RPL3 60S ribosomal protein L3 O55142 RPL35A 60S ribosomal protein L35a P25206 MCM3 DNA replication licensing factor MCM3 P62702 RPS4X 40S ribosomal protein S4, X isoform P10126 EEF1A1 Elongation factor 1-alpha 1 Q8BP67 RPL24 60S ribosomal protein L24 P14206 RPSA 40S ribosomal protein SA P11499 HSP90AB1 Heat shock protein HSP 90-beta P08226 APOE Apolipoprotein E P35980 RPL18 60S ribosomal protein L18 P63017 HSPA8 Heat shock protein 8 P80318 CCT3 T-complex protein 1 subunit gamma P62918 RPL8 60S ribosomal protein L8 P62911 RPL32 60S ribosomal protein L32 P14869 RPLP0 60S acidic ribosomal protein P0 Q9JIK5 DDX21 Nucleolar RNA helicase 2 P62281 RPS11 40S ribosomal protein S11 P62717 RPL18A 60S ribosomal protein L18a P86048 RPL10L 60S ribosomal protein L10-like P62267 RPS23 40S ribosomal protein S23 P25444 RPS2 40S ribosomal protein S2 P14148 RPL7 60S ribosomal protein L7 P53026 RPL10A 60S ribosomal protein L10a P62264 RPS14 40S ribosomal protein S14 P62908 RPS3 40S ribosomal protein S3 P04918 SAA3 Serum amyloid A-3 protein Proteins identified in present study were compared with proteins uploaded in Exocarta database from microglial origin. Protein ID and gene name are according to Uniprot Knowledgebase mice subjected to pMCAO; and (iv) in response to LPS, and translation in activated state and include a particular BV2 microglia release EVs with a distinct proteomic population of MVs budding from the plasma membrane. profile related to transcription and translation. The mechanism underlying release of EV is still not EVs can either enhance or suppress inflammation and clear. TNF can induce neurotoxicity by modulating glu- act as main factors to regulate inflammation and im- tamate production that results in excitotoxic neuronal munity [40, 41]. There is evidence showing microglia death [43]. One study conducted by Wang et al. has can release IL-1β, upon exposure to ATP derived from shown that TNF promotes the release of EVs from astro- astrocytes, by shedding MVs which contain the entire cytes through increased expression of glutaminase, machinery important for processing of it including the which convert glutamine to glutamate [44]. TNF can P2X7 receptors [42]. This is consistent with our findings also induce extensively production of glutamate from that EVs contain more molecules related to transcription microglia in an autocrine manner to cause excitotoxicity Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 16 of 19 and contribute to neuronal damage [45]. Thereby, these changes in the levels of either plasma membrane or endo- findings together with our results suggest that microglial somal markers to elucidate where these vesicles had origi- EV release could be potentially modulated by TNF nated. Indeed, we observed an increase in signal from through specifically regulated mechanisms. plasma membrane markers when compared to endosomal Our data suggests that inhibition of TNF signaling in markers upon LPS activation. When combined with the microglia may impact on inflammation, with an ob- increase in overall size observed with NanoSight, this sup- served reduction both in the release of different cyto- ports our theory that more MVs are released under in- kines and in the total number of EVs released, implying flammatory conditions. that the release of EVs in activated microglia seems to Furthermore, our proteomic analysis indicated that the be regulated. This idea is further supported by data from two populations of EVs were dramatically different in our experimental stroke model that showed altered EV categories of molecular function and biological process counts in plasma from TNF-KO mice compared to WT. under GO analysis. From the qualitative proteomic ana- These results contradict previous work that showed an lysis, 86 proteins were identified with high accuracy. increase in the amount of MVs in mice treated with Compared to Exocarta database, most of the proteins TNF inhibitor for 5 days after pMCAO compared to have been reported in previous studies in exosomes from saline-treated mice [31]. It is therefore reasonable to as- other cell types, where 89.5% of the quantified proteins sume that the difference we observed is due to the dif- were firstly identified in microglial EVs. The small over- ferent experimental set-up. A systemic administration of lap with previous studies is most likely due to a lack of TNF inhibitor in the mice 30 min after surgery could studies utilizing microglia-derived EVs, with a sole study merely have a transient effect on the TNF signaling path- performed on EVs from the N9 microglial cell line re- way, whereas using the TNF-KO model is expected to be sponsible for all microglial proteins present in ExoCarta more stable to investigate and characterize long-term ef- [15]. It was also clear from the FunRich analysis [20] fects of EVs release. that the proteins detected in EVs from LPS-treated cells From our previous study, pMCAO could initiate had different functions to those from control cells, with microglia activation and inflammation in the brain [18, proteins involved in RNA binding and structural compo- 28]. Therefore, we also evaluated systemic inflammation nents of the ribosome more prevalent in LPS-derived in mice subjected to pMCAO. Although several inflam- EVs. While the analysis on the cellular origin of the EV matory factors, such as IL-1β and IL-6, were significantly proteins revealed that the extracellular exosome and increased in both types of mice at day 1 after manipula- membrane were dominant from both conditions, it is of tion, such induction was attenuated in TNF-KO mice. interest to note that EVs isolated from inflammatory Another study has shown that EVs derived from macro- condition were detected with more membrane and less phages stimulated with bacterial infection are able to in- exosome proteins, consistent with a shift towards MV crease secretion of proinflammatory cytokines in release rather than exosome. recipient cells, including TNF [17]. Taken together, we For the first time, our data indicates that microglia can speculate that inflammatory propagation can be mit- change its EVs releasing machinery after LPS activation igated by a reduction in the number of microglia-derived with an increase in the overall number of EVs, and more EVs, thus halting inflammation via TNF signaling. How- EVs budding from the membrane. It is tempting to ever, this is complicated by the fact that there are two speculate that non-activated microglia have an expedient forms of TNF, soluble TNF (solTNF), which is related to and controlled release of EVs, whereas in an inflamma- neurotoxicity and inflammation, while the transmem- tory condition microglia release a large variation of EVs brane TNF (tmTNF) is involved in functional recovery that will perturb the normal homeostatic function of and neuroprotection [18, 46]. Hence, both the cellular microglia. We also think that, upon LPS-stimulation, contribution to TNF signaling and specifically the form microglia respond to release populations of extracellular of TNF carried in microglial EVs is important to evaluate cargoes with a unique proteomic profile related to RNA final outcome of experiment, neuroprotective or neuro- transcription and translation. The existence of various toxic. In our study, we measured total TNF in EVs in- RNA molecules in EVs is well-established including cluding tmTNF and solTNF. However, the specific form mRNA and microRNA. In fact, microRNAs can function of TNF is likely to be important for the inflammatory as ligands for TLRs and induce immune responses or in- and cytoprotective outcome. We believe such studies in- hibit activation by suppressing TLR signaling [13, 47]. vestigating the specific form of TNF carried in EVs are These RNAs are selectively sorted to EVs under different important in future studies. mechanisms [47]. However, the mechanisms responsible We also studied the protein composition to further for this packaging are not clear. Thus, our study implies characterize the EVs released under control or inflamma- proteins potentially involved in such mechanisms that tory conditions. Using western blot analysis, we looked for are increased when inflammation takes place. According Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 17 of 19 to our proteomic data, we speculate that EVs could actu- Additional files ally carry the whole machinery not only RNAs to recipi- Additional file 1: Supplementary figures for cytokines in conditioned ent cells and surrounding tissue. medium from microglia show significant upregulations of TNF (n = 7), IL- Finally, by comparing the protein cargoes from 1β (n = 7), IL5 (n = 7), and IL-6 (n = 3). Measured by multiplex ELISA LPS-activated and non-activated BV2 microglia, we were (Unpaired t test, *P < 0.05; ***P < 0.001). (PDF 201 kb) able to identify potential candidates likely to be involved Additional file 2: Supplementary figures for cytokines in microglia- derived extracellular vesicles not altered after LPS treatment. Measured by in the communication of microglial cells with other ef- multiplex ELISA (Unpaired t test, *P < 0.05; ***P < 0.001). (PDF 180 kb) fector cells under inflammation. We found 49 proteins Additional file 3: Supplementary figures for the effect of TNF inhibition exclusively present in EVs in the presence of LPS. on dynamics of EV trafficking. Images were taken and then measured for Among them, TUBB4B, HNRNPU, and SAA3 are pro- fluorescent intensity. A) Representative images of BV2 cells cultured with PHK76-labeled EVs 12 h after different treatments, including pre- teins related to inflammation and neuropathology. In- treatment of cells with either LPS (1 μg/ml) or etanercept (200 ng/ml) or deed, TUBB4B is a member of tubulin family and known in presence of both. Control (CTRL) was cells without any treatment. Cells as a component of cytoskeleton [48]. Liu X et al, sug- without EV were regarded as baseline. Merged images of the indicated areas show PHK76 internalized cells (Scale bar, 50 μm). B) Comparison of gested that TUBB4B may be a part of the same disease total fluorescent intensity (IntDen) in BV2 cells after incubation of dye- pathway as leucine-rich repeat kinase 2 (LRRK2), which labeled EVs. No significant differences were found between the is a crucial factor to understand the etiology of Parkin- conditions (one-way ANOVA, n = 3). (PDF 9799 kb) son’s disease (PD) [49]. HNRNPU acts as a key factor to Additional file 4: Supplementary figures for cytokines in serum from WT and TNF-KO mice before and 1 day after pMCAO. Measured by multi- maintain 3D structure of chromatin and has been re- plex ELISA (one-way ANOVA followed by Tukey’s test for multiple ported as a posttranscriptional regulator for NF-κB in- comparisons, n =3–6, **P < 0.01). (PDF 30 kb) flammatory pathway [50]. In addition, SAA3 is a Additional file 5: Supplementary figures for quantification of extracellular member of serum amyloid A (SAA) and acute phase vesicles in plasma from WT and TNF-KO mice subjected to pMCAO (Unpaired t test, n = 3). (PDF 18 kb) protein accompanying with other inflammatory cyto- kines and chemokines [51]. It has been implied to play a role in the inflammatory processes occurring in Alzhei- Abbreviations AD: Alzheimer’s disease; APOE: Apolipoprotein E; BCA: Bicinchoninic acid assay; mer’s disease (AD) and multiple sclerosis (MS) [51]. In BSA: Bovine serum albumin; CNS: Central nervous system; DMEM: Dulbecco’s addition to those proteins found exclusively in LPS EVs, modified Eagle medium; ELISA: Enzyme-linked immunosorbent assay; we identified APOE in EVs from both LPS and control EVs: Extracellular vesicles; FBS: Fetal bovine serum; FWHM: Full-width at half- maximum; GO: Gene Ontology; HNRNPU: Heterogeneous nuclear conditions. This protein is commonly present in mem- ribonucleoprotein U; HRP: Horse-radish protein; HSP8: Heat shock protein 8; branes and is considered one of the most important li- IFNγ:Interferonγ; IL: Interleukin; iNOS: Inducible nitric oxide synthase; KC/GRO: C-X- poproteins involved in cholesterol shuttling between C motif chemokine ligand 1; LC: Liquid chromatography; LPL: Lipoprotein lipase; LPS: Lipopolysaccharide; LRRK2: Leucine-rich repeat kinase 2; m/z:Masstocharge astrocytes and neurons along with being involved in re- ratio; MS: Multiple sclerosis; MS/MS: Tandem mass spectrometry; modeling and reorganization of neuronal networks after MVB: Multivesicular endosomal compartments; MVs: Microvesicles; NLRP3: Nod- injury [34]. It is also one of the major genetic risk factors like receptor protein 3; NO: Nitric oxide; NTA: Nanoparticle tracking analysis; PBS: Phosphate-buffered saline; PD: Parkinson’s disease; PFA: Paraformaldehyde; for late onset sporadic AD and can function as a ligand pMCAO: Permanent middle cerebral artery occlusion; SAA3: Amyloid A-3 protein; in receptor-mediated endocytosis with extracellular solTNF: Soluble form of tumor necrosis factor; TEM: Transmission electron β-amyloid [34, 52]. microscopy; TLR: Toll-like receptor; tmTNF: Transmembrane form of tumor necrosis factor; TNF: Tumor necrosis factor; TNF-KO: Tumor necrosis factor knockout; TUBB4B: Tubulin beta 4B Conclusions The present data show that upon activation by LPS, BV2 Acknowledgements microglia release EVs with a distinct proteomic profile We acknowledge technical support for mass spectrometry from the national infrastructure BioMS, Lund University, by Carol Nilsson, Sven Kjellström, and compared to control. Our data suggests that under these Yan Hong. And we would like to thank Lina Gefors for assistance with TEM inflammatory conditions, MVs are the predominate EV, imaging of extracellular vesicles at Bioimaging Center, Lund University. containing increased levels of TNF in particular, and to a lesser degree IL-6. We further provide evidence in vitro Funding We gratefully acknowledge funding support from the Strategic Research and in vivo that TNF signaling is important in quantita- Area MultiPark at Lund University, Lund, Sweden; the Swedish Research tively controlling EV release. Furthermore, through Council (No. 2012-2229), the Basal Ganglia Disorders Linnaeus Consortium proteomic analysis, we are able to provide lists of pro- (BAGADILICO); the Swedish Alzheimer Foundation; A.E. Berger Foundation; Swedish Brain Foundation; Crafoord Foundation; Swedish Dementia Associ- teins with the potential to modulate EV trafficking in ation; G&J Kock Foundation; Swedish National Stroke Foundation; Swedish microglia, in particular a change in EV proteins related Parkinson Foundation; Stohnes Foundation; the Royal Physiographic Soci- to neuronal maintenance and protein translation after ety; Olle Engkvist Byggmästare Foundation; Sparbanken Färs & Frosta Foun- dation, and the Danish Medical Research Council (DFF-4181-00033). LPS activation. We believe that EV regulation in micro- glia and its specific role in neuroinflammation will be Availability of data and materials important to fully understand the inflammatory patho- The datasets used and analyzed during the current study are available from genesis in neurodegenerative diseases. the corresponding author on reasonable request. Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 18 of 19 Authors’ contributions 16. Kowal J, Arras G, Colombo M, Jouve M, Morath JP, Primdal-Bengtson B, et al. YY and TD designed the studies and participated in the data analysis, Proteomic comparison defines novel markers to characterize heterogeneous interpretation, and writing of the manuscript. BHC contributed to cytokine populations of extracellular vesicle subtypes. Proc Natl Acad Sci U S A. 2016; analysis from mice. KLL contributed to all surgical and histological aspects of 113:E968–77. the study including the analysis of tissue specimens. YY performed most 17. Reales-Calderón JA, Vaz C, Monteoliva L, Molero G, Gil C. Candida albicans experiments and analyzed and interpreted data. ABS helped to interpret data modifies the protein composition and size distribution of THP-1 and participate in the design of the study. CD assisted with extracellular macrophage-derived extracellular vesicles. J Proteome Res. 2017;16:87–105. vesicle experiment and provided useful input to the drafting of the paper. 18. Lambertsen KL, Clausen BH, Babcock AA, Gregersen R, Fenger C, Nielsen HH, All authors have read and approved the final manuscript. et al. Microglia protect neurons against ischemia by synthesis of tumor necrosis factor. J Neurosci. 2009;29:1319–30. Ethics approval 19. Clausen BH, Lambertsen KL, Babcock AA, Holm TH, Dagnaes-Hansen F, All animal experiments were performed in accordance with the relevant Finsen B. Interleukin-1beta and tumor necrosis factor-alpha are expressed guidelines and regulations approved by the Danish Animal Ethical Committee by different subsets of microglia and macrophages after ischemic stroke in (numbers 2011/561-1950 and 2013-15-2934-00924). mice. J Neuroinflammation. 2008;5:46–18. 20. Benito Martin A, Peinado H. FunRich proteomics software analysis, let the Competing interests fun begin! Proteomics. 2015;15:2555–6. The authors declare that they have no competing interests. 21. Keerthikumar S, Chisanga D, Ariyaratne D, Saffar Al H, Anand S, Zhao K, et al. ExoCarta: a web-based compendium of Exosomal cargo. J Mol Biol. 2016; 428:688–92. Publisher’sNote 22. Burguillos MA, Svensson M, Schulte T, Boza-Serrano A, Garcia-Quintanilla A, Springer Nature remains neutral with regard to jurisdictional claims in published Kavanagh E, et al. Microglia-secreted galectin-3 acts as a toll-like receptor 4 maps and institutional affiliations. ligand and contributes to microglial activation. Cell Rep. 2015;10:1626–38. 23. Schroder K, Tschopp J. The inflammasomes. Cell. 2010;140:821–32. Author details 24. Burguillos MA, Magnusson C, Nordin M, Lenshof A, Augustsson P, Hansson Department of Experimental Medical Science, Experimental MJ, et al. Microchannel acoustophoresis does not impact survival or Neuroinflammation Laboratory, Lund University, Lund, Sweden. Department function of microglia, leukocytes or tumor cells. PLoS One. 2013;8:e64233. of Biochemistry and Structural Biology, Lund University, Lund, Sweden. 25. Yáñez-Mó M, Siljander PRM, Andreu Z, Zavec AB, Borràs FE, Buzás EI, et al. Department of Neurobiology Research, Institute of Molecular Medicine, Biological properties of extracellular vesicles and their physiological University of Southern Denmark, Odense, Denmark. BRIGDE—Brain functions. J Extracell Vesicles. 2015;4:27066. Research–Inter-Disciplinary Guided Excellence, Department of Clinical 5 26. Vendrame M, Gemma C, De Mesquita D, Collier L, Bickford PC, Sanberg CD, Research, University of Southern Denmark, Odense, Denmark. Department et al. Anti-inflammatory effects of human cord blood cells in a rat model of of Neurology, Odense University Hospital, Odense, Denmark. stroke. Stem Cells Dev. 2005;14:595–604. 27. Inácio AR, Ruscher K, Leng L, Bucala R, Deierborg T. Macrophage migration Received: 1 March 2018 Accepted: 15 May 2018 inhibitory factor promotes cell death and aggravates neurologic deficits after experimental stroke. J Cereb Blood Flow Metab. 2011;31:1093–106. 28. Inácio AR, Liu Y, Clausen BH, Svensson M, Kucharz K, Yang Y, et al. Endogenous References IFN-β signaling exerts anti-inflammatory actions in experimentally induced 1. Béchade C, Cantaut-Belarif Y, Bessis A. Microglial control of neuronal activity. focal cerebral ischemia. J Neuroinflammation. 2015;12:211. Front Cell Neurosci. 2013;7:32. 29. Clausen BH, Lambertsen KL, Dagnaes-Hansen F, Babcock AA, Linstow von 2. Tremblay M-È, Stevens B, Sierra A, Wake H, Bessis A, Nimmerjahn A. The role CU, Meldgaard M, et al. Cell therapy centered on IL-1Ra is neuroprotective of microglia in the healthy brain. J Neurosci Soc Neurosci. 2011;31:16064–9. in experimental stroke. Acta Neuropathologica. Springer. 2016;131:775–91. 3. Heneka MT, Kummer MP, Latz E. Innate immune activation in 30. Lambertsen KL, Meldgaard M, Ladeby R, Finsen B. A quantitative study neurodegenerative disease. Nat Rev Immunol. 2014;14:463–77. of microglial-macrophage synthesis of tumor necrosis factor during 4. Labzin LI, Heneka MT, Latz E. Innate immunity and neurodegeneration. acute and late focal cerebral ischemia in mice. J Cereb Blood Flow Annu Rev Med. 2018;69:437–49. Metab. 2005;25:119–35. 5. Wyss-Coray T, Rogers J. Inflammation in Alzheimer disease-a brief review of 31. Clausen BH, Degn M, Martin NA, Couch Y, Karimi L, Ormhøj M, et al. the basic science and clinical literature. Cold Spring Harb Perspect Med. Systemically administered anti-TNF therapy ameliorates functional outcomes 2012;2:a006346. after focal cerebral ischemia. J Neuroinflammation. 2014;11:203. 6. Ransohoff RM. A polarizing question: do M1 and M2 microglia exist? Nat 32. Ma B, Zhang K, Hendrie C, Liang C, Li M, Doherty-Kirby A, et al. PEAKS: Neurosci. 2016;19:987–91. powerful software for peptide de novo sequencing by tandem mass 7. Masgrau R, Guaza C, Ransohoff RM, Galea E. Should we stop saying ‘glia’ spectrometry. Rapid Commun Mass Spectrom. 2003;17:2337–42. and ‘neuroinflammation’? Trends Mol Med. 2017;23:486–500. 33. Kalra H, Simpson RJ, Ji H, Aikawa E, Altevogt P, Askenase P, et al. 8. Turola E, Furlan R, Bianco F, Matteoli M, Verderio C. Microglial microvesicle Vesiclepedia: a compendium for extracellular vesicles with continuous secretion and intercellular signaling. Front Physiol. 2012;3:149. community annotation. PLoS Biol. 2012;10:e1001450. 9. Gupta A, Pulliam L. Exosomes as mediators of neuroinflammation. J 34. Kim J, Basak JM, Holtzman DM. The role of apolipoprotein E in Alzheimer's Neuroinflammation. 2014;11:1–10. disease. Neuron. 2009;63:287–303. 10. Quek C, Hill AF. The role of extracellular vesicles in neurodegenerative 35. Chen S, Averett NT, Manelli A, LaDu MJ, May W, Ard MD. Isoform-specific diseases. Biochem Biophys Res Commun. 2017;483:1178–86. effects of apolipoprotein E on secretion of inflammatory mediators in adult 11. Szabo GT, Tarr B, Paloczi K, Eder K, Lajko E, Kittel A, et al. Critical role of rat microglia. J Alzheimers Dis. 2005;7:25–35. extracellular vesicles in modulating the cellular effects of cytokines. Cell Mol 36. Sjögren M, Folkesson S, Blennow K, Tarkowski E. Increased intrathecal Life Sci. 2014;71:4055–67. inflammatory activity in frontotemporal dementia: pathophysiological 12. Kumar A, Stoica BA, Loane DJ, Yang M, Abulwerdi G, Khan N, et al. implications. J Neurol Neurosurg Psychiatry. 2004;75:1107–11. Microglial-derived microparticles mediate neuroinflammation after traumatic 37. Zhang B, Gaiteri C, Bodea L-G, Wang Z, McElwee J, Podtelezhnikov AA, et al. brain injury. J Neuroinflammation. 2017;14:47. Integrated systems approach identifies genetic nodes and networks in late- 13. Fleshner M, Crane CR. Exosomes, DAMPs and miRNA: features of stress onset Alzheimer’s disease. Cell. 2013;153:707–20. physiology and immune homeostasis. Trends Immunol. 2017;38:768–76. 14. Raposo G, Stoorvogel W. Extracellular vesicles: exosomes, microvesicles, and 38. Gerhard A, Pavese N, Hotton G, Turkheimer F, Es M, Hammers A, et al. In friends. J Cell Biol. 2013;200:373–83. vivo imaging of microglial activation with [11C](R)-PK11195 PET in 15. Potolicchio I, Carven GJ, Xu X, Stipp C, Riese RJ, Stern LJ, et al. Proteomic idiopathic Parkinson's disease. Neurobiol Dis. 2006;21:404–12. analysis of microglia-derived exosomes: metabolic role of the 39. Fevrier B, Vilette D, Archer F, Loew D, Faigle W, Vidal M, et al. Cells release aminopeptidase CD13 in neuropeptide catabolism. J Immunol. 2005;175: prions in association with exosomes. Proc. Natl. Acad. Sci. U.S.A. National 2237–43. Acad. Sciences. 2004;101:9683–8. Yang et al. Journal of Neuroinflammation (2018) 15:168 Page 19 of 19 40. Yoon YJ, Kim OY, Gho YS. Extracellular vesicles as emerging intercellular communicasomes. BMB Rep. 2014;47:531–9. 41. Clayton A, Mitchell JP, Court J, Mason MD, Tabi Z. Human tumor-derived exosomes selectively impair lymphocyte responses to interleukin-2. Cancer Res. 2007;67:7458–66. 42. Bianco F, Pravettoni E, Colombo A, Schenk U, Möller T, Matteoli M, et al. Astrocyte-derived ATP induces vesicle shedding and IL-1 beta release from microglia. J Immunol. 2005;174:7268–77. 43. Ye L, Huang Y, Zhao L, Li Y, Sun L, Zhou Y, et al. IL-1β and TNF-α induce neurotoxicity through glutamate production: a potential role for neuronal glutaminase. J Neurochem. 2013;125:897–908. 44. Wang K, Ye L, Lu H, Chen H, Zhang Y, Huang Y, et al. TNF-α promotes extracellular vesicle release in mouse astrocytes through glutaminase. J Neuroinflammation. 2017;14:87. 45. Takeuchi H, Jin S, Wang J, Zhang G, Kawanokuchi J, Kuno R, et al. Tumor necrosis factor-α induces neurotoxicity via glutamate release from hemichannels of activated microglia in an autocrine manner. J Biol Chem. 2006;281:21362–8. 46. Madsen PM, Clausen BH, Degn M, Thyssen S, Kristensen LK, Svensson M, et al. Genetic ablation of soluble tumor necrosisfactorwithpreservationof membrane tumor necrosis factor is associated with neuroprotection after focal cerebral ischemia. J Cereb Blood Flow Metab. 2016;36:1553–69. 47. Jiang L, Vader P, Schiffelers RM. Extracellular vesicles for nucleic acid delivery: progress and prospects for safe RNA-based gene therapy. Gene Ther. 2017;24:157–66. 48. Hammond JW, Cai D, Verhey KJ. Tubulin modifications and their cellular functions. Curr Opin Cell Biol. 2008;20:71–6. 49. Liu X, Cheng R, Ye X, Verbitsky M, Kisselev S, Mejia Santana H, et al. Increased rate of sporadic and recurrent rare genic copy number variants in Parkinson’s disease among Ashkenazi Jews. Mol Genet Genomic Med. 2013; 1:142–54. 50. Lu Y, Liu X, Xie M, Liu M, Ye M, Li M, et al. The NF-κB-responsive long noncoding RNA FIRRE regulates posttranscriptional regulation of inflammatory gene expression through interacting with hnRNPU. J Immunol. 2017;199:3571–82. 51. Barbierato M, Borri M, Facci L, Zusso M, Skaper SD, Expression GP. Differential responsiveness of central nervous system glial cell populations to the acute phase protein serum amyloid a. Sci Rep. 2017;7:12158. 52. Winblad B, Amouyel P, Andrieu S, Ballard C, Brayne C, Brodaty H, et al. Defeating Alzheimer’s disease and other dementias: a priority for European science and society. Lancet Neurol. 2016;15:455–532.

Journal

Journal of NeuroinflammationSpringer Journals

Published: May 28, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off