A novel long non-coding RNA, Nostrill, regulates iNOS gene transcription and neurotoxicity in microglia.

Background Microglia are resident immunocompetent and phagocytic cells in the CNS. Pro-inammatory microglia, stimulated by environmental microbial signals such as bacterial lipopolysaccharide (LPS), viral RNAs, or inammatory cytokines, are neurotoxic and associated with pathogenesis of several neurodegenerative diseases. Long non-coding RNA (lncRNA) are emerging as important tissue-specic regulators directing cell differentiation and functional states and may help direct proinammatory responses of microglia. Methods Microglial gene expression array analyses and qRT-PCR was used to identify a novel intergenic long-noncoding RNA that was upregulated in LPS-stimulated microglial cell lines, LPS-stimulated primary microglia, and LPS-injected mouse cortical tissue. Silencing and overexpression studies, RNA immunoprecipitation, chromatin immunoprecipitation, chromatin RNA immunoprecipitation assays, and qRT-PCR were used to study the function of this long-noncoding RNA in microglia. In vitro cytotoxicity assays were used to examine the effects of silencing the novel long-noncoding RNA in LPS-stimulated microglia on neurotoxicity. Results We report here that the previously uncharacterized intergenic lncRNA we termed Nostrill is induced by LPS stimulation in both BV2 cells and primary murine microglia, as well as in cortical tissue of LPS-injected mice. Induction of Nostrill is NF-κB dependent and silencing of Nostrill decreased inducible nitric oxide synthase (iNOS) expression and nitric oxide production in BV2 and primary microglial cells. Overexpression of Nostrill increased iNOS expression and nitric oxide production. RNA immunoprecipitation assays demonstrated that Nostrill is physically associated with NF-κB subunit p65 following LPS stimulation. Silencing of Nostrill signicantly reduced NF-κB p65 and RNA polymerase II recruitment to the iNOS promoter and decreased H3K4me3 activating histone modications at iNOS gene loci. In vitro studies demonstrate that silencing of Nostrill in microglia reduced LPS-stimulated microglia neurotoxicity. Conclusions Our data indicate a new regulatory role of NF-κB-induced Nostrill and suggest that Nostrill acts as a co-activator of transcription of iNOS resulting in the production of nitric oxide in microglia through modulation of epigenetic chromatin remodeling. Nostrill may be a target for reducing the neurotoxicity associated with iNOS-mediated inammatory processes in microglia during neurodegeneration. and microglia upregulate lincRNA-Cox 2. Upon bacterial lipopolysaccharide (LPS) stimulation of TLR4, lincRNA-Cox2 interacts with the nucleosome remodeling complex SWI//SNF to modulate pro-inammatory NF-κB signaling. SWI/SNF-associated histone acetylation causes transactivation of late-primary inammatory response genes in LPS- stimulated microglia (31). We have also shown in macrophages and microglia that lincRNA-Tnfaip3 transcript interacts with components of the Hmgb1 complex, and an NF-κB/Hmgb1/lincRNA-Tnfaip3 complex assembles in microglial cells in response to LPS stimulation (32). These preliminary data provide novel and exciting evidence that lincRNAs may be involved in microglia plasticity and polarization in response to environmental cues. Therefore, lincRNAs may participate in pathogenesis of various inammatory and neurodegenerative diseases making them targets for therapeutic interventions.

Results We report here that the previously uncharacterized intergenic lncRNA we termed Nostrill is induced by LPS stimulation in both BV2 cells and primary murine microglia, as well as in cortical tissue of LPS-injected mice. Induction of Nostrill is NF-κB dependent and silencing of Nostrill decreased inducible nitric oxide synthase (iNOS) expression and nitric oxide production in BV2 and primary microglial cells. Overexpression of Nostrill increased iNOS expression and nitric oxide production. RNA immunoprecipitation assays demonstrated that Nostrill is physically associated with NF-κB subunit p65 following LPS stimulation. Silencing of Nostrill signi cantly reduced NF-κB p65 and RNA polymerase II recruitment to the iNOS promoter and decreased H3K4me3 activating histone modi cations at iNOS gene loci. In vitro studies demonstrate that silencing of Nostrill in microglia reduced LPS-stimulated microglia neurotoxicity.
Conclusions Our data indicate a new regulatory role of NF-κB-induced Nostrill and suggest that Nostrill acts as a co-activator of transcription of iNOS resulting in the production of nitric oxide in microglia through modulation of epigenetic chromatin remodeling. Nostrill may be a target for reducing the neurotoxicity associated with iNOS-mediated in ammatory processes in microglia during neurodegeneration.
Background Systemic in ammation due to pathogenic infection is a direct cause of dysregulated neuroimmune responses and is linked to several neurodegenerative pathologies of the central nervous system (CNS) (1)(2)(3)(4)(5)(6)(7). Microglia, the principal neuroimmune cells, participate in the immune processes of pathogen clearance contributing to both neurorecovery and neurotoxicity (5). Microglia exhibit functional plasticity and are able to act as homeostatic surveillance cells (8)(9)(10)(11), anti-in ammatory and neuroprotective cells Transcriptional Regulatory Intergenic LncRNA Locus) is induced by in ammatory mediators and controlled by NF-κB signaling in microglial cell lines and primary microglia following TLR3, and more dramatically, TLR4 stimulation. Silencing or overexpression of Nostrill in microglial cells in uenced iNOS mRNA levels and nitric oxide production. Nostrill is physically associated with NF-κB p65 following LPS stimulation. Knockdown of Nostrill decreased NF-κB p65 and RNA polymerase II recruitment to iNOS promoter region and decreased H3K4me3 activating histone modi cations. Importantly, blocking the expression of Nostrill in microglia reduced proin ammatory toxicity to primary cultured cortical neurons in cellular assays. Identifying pro-in ammatory lincRNAs such as Nostrill that when silenced reduce microglial neurotoxicity may be useful in developing targeted therapeutic strategies that reduce the neurotoxicity associated with immune responses to pathogenic signals and thereby limit neurodegeneration.

Animals
Animals were housed in AAALAC-accredited facilities, and all experiments were conducted under protocols approved by the Creighton University Institutional Animal Care and Use Committee. C57BL/6J mice were obtained from The Jackson Laboratory. Mice were housed and bred in the animal care facility at Creighton University under a 12/12 h light/dark cycle with ad libitum access to food. For primary microglial and cortical neuronal cell isolation, animals were treated in strict accordance to the approved Institutional Animal Care and Use protocol #0793. For LPS injection In Vivo Model, animals were treated in strict accordance to approved by the Institutional Animal Care and Use Protocol #1086.

LPS Injection In Vivo Model
Male and female C57BL/6J mice of age 6-weeks old were divided into two groups: a vehicle control group receiving intravenous (IV) tail vein injection of Dulbecco's phosphate-buffered saline (50 µl/10 g, DPBS, Thermo Fisher Scienti c, Waltham, MA) or an experimental group receiving an IV injection of LPS at 1 mg/kg (Escherichia coli O111:B4, Sigma-Aldrich, St Louis, MO, USA) in DPBS. At 24 hours, mice were anesthetized with ketamine/xylazine and transcardially perfused with cold phosphate-buffered saline (PBS). Mice were weighed before LPS injection and 24 hours after injection. LPS injected mice usually lose ~ 10% of the body weight, which can be used as an indication of successful tail vein delivery of LPS. Fischer Scienti c, Waltham, MA), 1% l-glutamine, 1% penicillin/streptomycin (ThermoFisher Scienti c, Waltham, MA). Cells were grown in 100-mm tissue culture dishes at 37 ºC in 5% CO 2 and allowed to reach 80% con uency before passage.

Primary Cortical Microglial Cell Culture
Primary microglial cells were isolated from P0-P2 C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME). Use of animals was performed in strict accordance with the Institutional Animal Care and Use committee guidelines as approved by the IACUC committee at Creighton University (protocol #0793). P0-P2 mouse brains were dissected, meninges were removed, and cortices were isolated in ice cold, sterile FBS (Hyclone #SH30072.03, Lot No. AXB30110, Thermo Fischer Scienti c, Waltham, MA) and cells were mechanically triturated. Cells suspension was strained through a sterile 70 µm nylon mesh strainer and plated onto poly-D-lysine coated 75 cm 2 tissue culture asks in DMEM plus 10% FBS and penicillin/streptomycin and allowed to reach con uency over 14 days at 37 ºC in 5% CO 2 . After reaching con uency cells were shook vigorously and on an orbital shaker at 220 rpm to remove microglia.
Microglial were collected and re-seeded at 0.5 × 10 6 cells/ml onto tissue culture plates. After 1 h attachment, oating cells were removed and adherent cells were cultured in DMEM plus 10% FBS and penicillin/streptomycin at 37 ºC in 5% CO 2 . unless rinsed and switched into Neurobasal media for experiments. Microglial purity was determined using immunocytochemical analysis of cortical cell protein expression (described below).

Isolation of Cortical Neurons
Primary cortical cells were isolated from P0-P2 C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME) following methods modi ed from Ahlemeyer et al., 2005 (75). Use of animals was performed in strict accordance with the Institutional Animal Care and Use committee guidelines as approved at Creighton University (Protocol #0793). Brie y, P0-P2 mouse brains were dissected, meninges were removed, and Scienti c, Waltham, MA). The cell suspension was washed three times and resuspended with Neurobasal media supplemented with B-27™ Plus Supplement (GibcoBRL #A35828-01, Thermo Fischer Scienti c, Waltham, MA) and penicillin/streptomycin (#10378016, Thermo sher Scienti c, Waltham, MA) and dissociated with mechanical trituration. Cells suspension was centrifuged for 5 min at 1000 rpm, resuspended in supplemented serum-free Neurobasal media and plated onto poly-D-lysine (#P0899, Sigma, St. Louis, MO) coated tissue culture plates at density of 1.5 × 10 6 cells/well in 6-well plates and 5 × 10 5 cells/well in 24-well plates at 37 ºC in 5% CO 2 for at least one week. Each cortical culture was considered a biological replicate and all experiments were performed in triplicate.
BV2 microglia were placed in suspension above cortical neurons and co-cultured for an additional three days in unsupplemented Neurobasal media at 37 ºC in 5% CO 2 . After co-culture, Transwells® with microglia were removed and cortical neuronal cultures were xed in culture media plus 3.7% formaldehyde at 37 o C in 5% CO 2 . Cortical neuronal cultures were assessed using immunocytochemistry (described below).

Immunocytochemistry
Cortical primary microglial or cortical neuronal cultures were xed with 3.7% formaldehyde in cell culture media, rinsed in PBS. Cells were permeabilized with 0.2% Triton X-100 in PBS for 10 minutes, washed, and blocked for 1 hour in PBS, 0.2% BSA, and 0.2% Triton X-100. Primary antibodies were applied and incubated overnight at 4 ºC in PBS, 0.2% BSA, 0.2% Triton X-100. Cells were incubated with anti-Iba-1 (microglial marker, Abcam, rabbit anti-GFAP (1:400, Millipore Cat# AB5541, RRID:AB_177521), and mouse anti-beta tubulin III/TUJI (1:200, Millipore Cat# MAB1637, RRID:AB_2210524). Secondary antibodies were applied for 1 h at a concentration of 1:500 for goat anti-rabbit IgG (H + L) Alexa Fluor 488 conjugate and goat anti-mouse IgG (H + L) Alexa Fluor 594 488 conjugate (Pierce, Rockford, IL). Nuclei were visualized using a DAPI stain (300 mmol, MP Biomedicals, Santa Ana, CA). Qualitative and quantitative analysis of immunocytochemistry was performed by acquiring images with a Leica DMI4000B inverted microscope with a cooled CCD camera (Q Imaging, Surrey, BC) and uorescent capabilities. For quanti cation of the percent of cells expressing cell-speci c proteins was determined by counting the number of immunopositive cells for each marker and dividing that number by the total number of cells counted in the eld. For quanti cation of relative uorescence intensity units associated with the immunocytochemistry experiments localizing protein expression in cortical cell cultures. In all experiments, images were analyzed with Volocity (PerkinElmer,USA) and ImageQuant (GE Healthcare, USA) software were used for image analysis and presentation. For image data, 3 eld views of at least 100 cells from 3 separate experiments were analyzed for each condition.

Measurement of Cell Viability -propidium iodine incorporation
Cell viability was measured using propidium iodine incorporation methods as described by the manufacturer (Invitrogen, #P1304MP, Thermo Fischer Scienti c, Waltham, MA). Propidium iodide will permeate dead cells and is used to detect cell death/viability. Brie y, following co-culture with microglia cortical neuronal cultures were RNase-Treated by equilibrating for 5 min in 2X SCC buffer (0.3M NaCl, 0.03M sodium citrate, pH 7.0) and then incubated in 100 µg/ml RNase-free RNase in 2X SCC for 20 min at 37 o C. Cells were rinsed three times in 2X SCC and counterstained with 500 nM PI in 2X SCC for 5 min. Cells were rinsed three times in 2X SCC, excess buffer was removed, placed in 1X PBS and imaged immediately. Neuronal cultures were viewed for propidium iodide (PI) red-uorescent nuclear and chromosome counterstaining. Images were acquired via the EVOS M5000 cell imaging system (Excitation 535 nm/Emission 617 nm) and images saved for later analysis using Firmware, EVOS FLoid Software (Thermo Fischer Scienti c, Waltham, MA). In all experiments, acquired images were analyzed with Volocity (PerkinElmer,USA) and ImageQuant (GE Healthcare, USA) software were used for image analysis and presentation. Experiments were performed in triplicate.

Small interfering RNAs and transfection
For gene silencing, the small interfering RNA (siRNA) duplexes for mouse Nostrill were synthesized using Integrated DNA Technologies. The siRNA sequences targeting Nostrill were as follows: sense, 5′-CGAGAUAGGCUGAGGACUU − 3′; antisense, 5′-AAGUCCUCAGCCUAUCUCG − 3′. The nonspeci c scrambled siRNA sequence UUCUCCGAACGUGUCACGUUU was used for the control. Cells were treated with siRNAs ( nal concentration, 60 nM) using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. For Nostrill overexpression, Nostrill cDNA was ampli ed through PCR, inserted into the PTarget (Promega, Madison, WI) expression vector to generate PTarget-Nostrill, and subsequently sequenced. According to the manufacturer's protocol, cells were transfected with plasmid DNA using Lipofectamine 2000. Quantitative RT-PCR was used to determine the signi cant alteration of each target gene.

RT-PCR Analysis
For real-time PCR analysis of cytokines, total RNA was isolated from cells with Trizol reagent (Applied Biosystems). An amount of 200 ng total RNA was reverse-transcribed using the iScript Reverse Transcription Supermix (Bio-Rad, Hercules, CA). Comparative real-time PCR was performed using the Invitrogen™ SYBR GreenER™ qPCR SuperMix Universal (Thermo Fisher Scienti c, Waltham, MA) on the Bio-Rad CFX96 Touch™ Real-Time PCR Detection System. The sequences for all the primers are listed in Supplementary Table 1. Normalization was performed using Gapdh. Relative expression was calculated using the comparative Ct (ΔΔCt) method.

Griess Analysis
Media collected from microglial cultures were evaluated using a Nitric Oxide Assay Kit to determine nitric oxide composition through measurement of nitrate (NO 3 ) and nitrite (NO 2 ) levels according to manufacturer's instructions (#EMSNO, Thermo Fischer Scienti c, Waltham, MA). Brie y, 1X reagent diluent, NADH, and nitrate reductase were prepared as recommended in the kit instructions. Samples were diluted 1:2 with 1X reagent diluent and ltered through a 10,000 MWCO lter. NADPH was oxidized with 10 µL of lactate dehydrogenase (1500 U/ml in 30 mM sodium pyruvate) after incubation with nitrate reductase and incubated at 37 o C for 10 min. Nitrate standards were prepared by serial dilution following manufacturer's instructions. Griess reagent I and II were added to standard, control, and sample wells. Plates were tapped to mix and incubated at room temperature for 10 min. Plates were read using an optical density at 540 ± 20 nm on Synergy HTX multi-mode reader (BioTek US, Winooski, VT). Technical triplicate readings were averaged and experiments were run in biological triplicates.

RNA Immunoprecipitation Assay
The formaldehyde crosslinking RIP was performed as described (26). Brie y, precleaning lysates with 20 µl of PBS washed Magna ChIP Protein A + G Magnetic Beads (Millipore, Massachusetts). The precleaned lysate (250 µl) was then diluted with the whole cell extract buffer (250 µl), mixed with the speci c antibody-coated beads, and incubated with rotation at 4 °C for 4 h, followed by 4 times washing with the whole cell extract buffer containing protease and RNase inhibitors. The collected immunoprecipitated RNP complexes and input were digested in RNA PK Buffer pH 7.0 (100 mM NaCl, 10 mM TrisCl pH 7.0,1 mM EDTA, 0.5% SDS) with addition of 10 µg Proteinase K and incubated at 50 °C for 45 min with end-to-end shaking at 400 rpm. Formaldehyde cross-links were reversed by incubation at 65 °C with rotation for 4 h. RNA was extracted from these samples using Trizol according to the manufacturer's protocol (Invitrogen) and treated with DNA-free DNase Treatment & Removal I kit according to the manufacturer's protocol (Ambion, Austin, TX). The presence of RNA was measured by quantitative RT-PCR using the CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad). Gene-speci c PCR primer pairs are listed in Supplementary Table 1. The following antibodies were used for RIP analysis: anti-NF-κB p65 (Santa Cruz), normal mouse IgG (Santa Cruz).

Chromatin Immunoprecipitation Assay
Chromatin immunoprecipitation (ChIP) assays were performed as described previously (10). Brie y, cells were xed with 1% formaldehyde for 10 minutes, collected in ice-cold PBS, and resuspended in an SDS lysis buffer. Genomic DNA was then sheared to lengths ranging from 200 to 1000 bp by sonication. One percent of the cell extracts was taken as input, and the rest of the extracts was incubated with either anti-NF-κB p65 (Santa Cruz), anti-H3K4me3 (Cell Signaling), anti-RNA Polymerase 2 (Millipore), or normal mouse IgG (Santa Cruz) overnight at 4 °C, followed by precipitation with protein G-agarose beads. The immunoprecipitates were sequentially washed once with a low-salt buffer, once with a high-salt buffer, once with an LiCl buffer, and twice with a Tris buffer. The DNA-protein complex was eluted, and proteins were then digested with proteinase K for 1 h at 45 °C. The DNA was detected by real-time quantitative PCR analysis. Gene-speci c PCR primer pairs are listed in Supplementary Table 1.
Chromatin Isolation by RNA Puri cation ChIRP analysis was performed as previously reported (36). Brie y, a pool of tiling oligonucleotide probes with a nity speci c to the Nostrill sequence was used and glutaraldehyde cross-linked for chromatin isolation. The sequences for each probe are listed in Supplementary Table 1; probe 1, 3, 5, and 7 are mixed as the probe pool Odd and probe 2, 4, 6, and 8 as the probe pool Even. The DNA sequences of the chromatin immunoprecipitates were con rmed and quanti ed by real-time PCR using the same primer sets covering the gene promoter regions of interest as for ChIP analysis. A pool of oligo probes for LacZ were served as controls. The percent input method was used to normalize the ChIRP data.

Statistical Analysis
Data are expressed as mean values and error bars represent standard error of the mean (SEM). Student T test with Bonferroni's correction or one-way ANOVA followed by Tukey-Kramer post hoc tests were performed where appropriate. For determination of signi cant differences between percents and for multiple comparisons between culture conditions, two-way ANOVA followed by Tukey-Kramer multiple analyses post hoc tests were used. Values of p < 0.05 were considered to be signi cant.
GM14005 and AK15331 were not upregulated or not expressed at detectable levels in the cortical tissues of this in vivo LPS-injected mouse model system (Fig. 1C). While not upregulated in culture cells in response to LPS, NR_029444 was upregulated 4.51 ± 1.24 folds in vivo (Fig. 1C). Since Nostrill was the most highly upregulated lincRNA following LPS stimulation in both BV2 and primary microglial cells and was signi cantly upregulated in the in vivo LPS-injection mouse model system, Nostrill was chosen for further investigation.
Nostrill upregulation following LPS stimulation increased in a dose-dependent manner up to ~ 12 folds that of unstimulated control levels when BV2 microglia were incubated with LPS at 10 µg/ml (Fig. 1D).
Real-time PCR analysis was used to determine the time course of Nostrill expression after LPS exposure. Temporal expression Nostrill in response to LPS stimulation increased to ~ 8 folds above control levels at 2 and 4 h of TLR4 stimulation and peaked at 6 hr to 22.5 ± 2.08 folds. Nostrill levels returned to baseline by 24 h (Fig. 1E). Nostrill expression in BV2 microglia also increased in response to other known proin ammatory mediators, including tumor necrosis factor (TNF-α), the TLR3 ligand, olyinosinic:polycytidylic acid (Poly (I:C)), and Interferon gamma (IFN-γ) (Fig. 1F). Interestingly, Nostrill expression is not in uenced by stimulation with the anti-in ammatory cytokine, Interleukin-4 (IL-4) (Fig. 1F).
Knockdown or overexpression of Nostrill attenuates the upregulation of in ammatory genes trigged by LPS Since activity of the NF-κB pathway is involved in Nostrill upregulation following TLR4 stimulation, we sought to determine whether knockdown or overexpression of Nostrill affected the expression of NF-κBresponsive genes. Short interfering RNA (siRNA) targeting Nostrill was used to knockdown Nostrill expression. SiRNA-Nostrill signi cantly reduced basal levels of Nostrill to 0.25 ± 0.1 folds compared to unstimulated control levels (Fig. 3A). Silencing Nostrill in BV2 cells and then stimulating with LPS reduced Nostrill upregulation to that of basal levels seen in unstimulated cells (Fig. 3A). Scrambled siRNA (siRNA-control) did not block LPS-induced upregulation of Nostrill in BV2 cells (Fig. 3A). The overexpression construct (using the PTarget mammalian expression vector) was used to enhance Nostrill expression (PTarget-Nostrill). Real-time PCR showed that BV2 microglia transfected with PTarget-Nostrill increased Nostrill over 1300 folds compared to PTarget control construct (PTarget-Empty) that served as the controls (Fig. 3B). Quantitative real-time PCR was used to examine mRNA expression of several downstream target genes of NF-κB signaling when Nostrill was silenced or overexpressed ( Fig. 3C-H). The anti-in ammatory cytokine Arginase 1 (Arg1) mRNA levels decreased following LPS stimulation and silencing of Nostrill did not signi cantly in uence Arg1 mRNA levels compared to siRNA-control transfected cells treated with LPS (Fig. 3C). Silencing of Nostrill reduced Arg1 mRNA levels signi cantly compared to siRNA-control in the absence of LPS stimulation (Fig. 3C); however the Arg1 mRNA levels in this condition were not signi cantly different from siRNA-control or siRNA-Nostrill plus LPS stimulation (Fig. 3C). Following silencing of Nostrill and stimulation with LPS, mRNA levels of proin ammatory cytokines interleukin 6 (IL-6), interleukin 1 beta (IL-1β), and tumor necrosis factor alpha (TNF-α) were not signi cantly different from siRNA-control levels ( Fig. 3D-F). Monocyte chemoattractant protein 1 (MCP1/Ccl2) and inducible nitric oxide synthase (iNOS) mRNA levels were signi cantly reduced in siRNA-Nostrill-transfected BV2 microglia (Fig. 3G-H). SiRNA-Nostrill transfection alone in the absence of LPS stimulation signi cantly decreased Ccl2, IL-1β, TNF-α, and iNOS. Overexpression of Nostrill (PTarget-Nostrill) signi cantly increased IL-6, Ccl2, IL-1β, TNF-α, and iNOS in the absence of LPS stimulation as compared to unstimulated PTarget controls (Fig. 3D-F). Interestingly, overexpression of Nostrill followed by LPS stimulation signi cantly increased Ccl2 ~ 28 folds more than in unstimulated PTarget-Nostrill cells (Fig. 3G) and iNOS expression ~ 60 folds more than in unstimulated PTarget-Nostrill cells (Fig. 3H).
Loss-or gain-of-function of Nostrill regulates the production of nitric oxide by LPS-stimulated BV2 and primary microglia In BV2 microglia, iNOS gene transcription increased signi cantly with doses of LPS from 0-10 µg/ml (Fig. 5A). Real-time PCR analyses showed that at 10 µg/ml, iNOS mRNA levels reached 188.2 ± 20.3 folds that of control, unstimulated levels (Fig. 5B). The concentration of 10 µg/ml LPS was used to stimulate siRNA-control and siRNA-Nostrill transfected BV2 microglia as well as PTarget-Empty and PTarget-Nostrill transfected microglia to assess NO 2 production using Griess assays (Fig. 5B). In the absence of LPS stimulation, silencing of Nostrill signi cantly reduced NO 2 production in BV2 microglia compared to untreated, control transfected cells (Fig. 5B). Treatment with LPS, signi cantly increased NO 2 production was detected in control transfected cells as compared to untreated, control transfected cells (Fig. 5B). Silencing of Nostrill signi cantly decreased NO 2 production following LPS treatment as compared to LPS treated, siRNA-control transfected cells (p ≤ 0.001, Fig. 5B). The level of NO 2 production following silencing of Nostrill and LPS treatment was signi cantly different from untreated siRNA-control and siRNA-Nostrill transfected cells that were not treated with LPS (Fig. 5B). Overexpression of Nostrill (PTarget-Nostrill) in BV2 microglia signi cantly increased NO 2 production as compared to untreated and LPS treated PTarget-Nostrill transfected control cells (p ≤ 0.001, Fig. 5B). LPS treatment of PTarget-Nostrill cells did not signi cantly increase NO 2 production in BV2 microglial as compared to untreated PTarget-Nostrill transfected cells (Fig. 5B). In primary microglial cells, silencing of Nostrill signi cantly decreased NO 2 production in unstimulated siRNA-Nostrill transfected primary microglia as compared to unstimulated control cells (Fig. 5C). Following LPS stimulation, silencing of Nostrill signi cantly reduced NO 2 production in siRNA-Nostrill transfected cells as compared to siRNA-control transfected cells (Fig. 5C).

Nostrill promotes iNOS transcription through chromatin modi cations associated with the p65 protein of the NF-κB subunits
Given that Nostrill expression is associated with NF-κB signaling and NF-κB-mediated gene transcription of iNOS in particular, we sought to determine whether Nostrill can directly interact with NF-κB subunit proteins. Formaldehyde cross-linking RNA immunoprecipitation (RIP) analysis was performed on LPS stimulated and unstimulated BV2 microglia (Fig. 6A). Immunoprecipitation of p65 demonstrated a 11.2 ± 0.1 folds of Nostrill enrichment in LPS stimulated cells as compared to control untreated cells (Fig. 6A).
Immunoprecipitation using a non-speci c IgG antibody did not demonstrate enrichment of Nostrill in LPS stimulated or control cells (Fig. 6A). No signi cant interaction between actin and NF-κB p65 was observed in LPS stimulated or control cells (Fig. 6A). We used Chromatin Immunoprecipitation (ChIP) assays and silencing of Nostrill expression to evaluate Nostrill involvement in docking NF-κB p65 at iNOS promoter region. Primer sets positioned at several different locations within the iNOS promoter region (iNOS 1-7) were used to determine whether silencing of Nostrill in uenced the association of p65 to iNOS promoter region (Fig. 6B). Enrichment of iNOS promoter region at 4 and 7 locations was signi cant in siRNA-control transfected BV2 cells following LPS stimulation compared to controls (Fig. 6C).
Silencing of Nostrill signi cantly reduced ampli cation of iNOS transcriptional region by primer sets 4 and 7 following cross-linking to p65 in unstimulated BV2 cells to 2.7 ± 1.4 folds and 0.77 ± 0.37 folds, respectively (Fig. 6C). Importantly, silencing of Nostrill in LPS stimulated BV2 cells signi cantly reduced enrichment of the iNOS transcriptional region ampli ed by primer sets 4 and 7 to 0.19 ± 0.05 folds and 1.14 ± 0.0 folds, respectively (Fig. 6C). In unstimulated and LPS stimulated BV2 cells, siRNA-control and siRNA-Nostrill treatments did not show enrichment for association with genomic actin transcription site (Fig. 6C). ChIP analyses also showed that following LPS stimulation, crosslinking of RNA Polymerase II (Pol II) to the chromatin of siRNA-control transfected BV2 cells showed signi cant enrichment of iNOS transcriptional region ampli ed by primer sets 4 and 7 to 49.9 ± 9.6 folds and 21.4 ± 6.2 folds, respectively (Fig. 6D). Silencing of Nostrill signi cantly reduced enrichment of iNOS transcriptional region by primer sets 4 and 7 to 11.36 ± 2.0 folds and 1.14 ± 0.53 folds, respectively (Fig. 6D).
Modi cation of histone proteins such as H3 is necessary for gene transcription. Commonly trimethylation of H3K4 (H3K4me3) is associated with and is necessary for enhanced transcription of nearby genes. ChIP analysis using H3K4me3 antibodies demonstrated enhanced association of H3K4me3 with iNOS transcriptional region ampli ed by primer sets 4 and 7 as well as other locations within the iNOS transcriptional region by primer sets 2, 5 and 6 (Fig. 6E). Primer sets 5 and 6 show enrichment following LPS stimulation in siRNA-control transfected BV2 cells and inhibition of enrichment in siRNA-Nostrill transfected LPS-stimulated BV2 cells (Fig. 6E). Primer set iNOS 2 demonstrated enrichment following LPS stimulation in siRNA-control and siRNA-Nostrill-transfected BV2 cells (Fig. 6E).
To further examine whether Nostrill is directly associated with the chromatin modi cations necessary for iNOS transcription, chromatin isolation by RNA puri cation (ChIRP) was performed (Fig. 6F). For increased sensitivity and speci city, probes to Nostrill were split between even and odd pools to test whether Nostrill was directly recruited to the iNOS promoter region. Following pull-down of Nostrill biotinylated probes, quantitative PCR demonstrated increased interactions of Nostrill and iNOS promoter region at the same set 4 and 7 locations as for p65 docking and Pol II recruitment and H3K4me3 enrichment in BV2 cells following LPS stimulation (Fig. 6F). Increased interactions were detected in both even and odd probe pools (Fig. 6F). Quantitative PCR using genomic actin primers did not show interactions between Nostrill and serves as negative control (Fig. 6F).

Effect of Nostrill expression on the neurotoxicity of LPS-stimulated microglia
To investigate whether silencing Nostrill reduced microglial neurotoxicity, we designed co-culture experiments where BV2 microglia were transfected with silencing or overexpression constructs, stimulated with LPS (10 µg/ml), washed to remove LPS, and then co-cultured with cortical neurons for three days in vitro (Fig. 7A). At the end of co-culturing, neurons were xed and immunostained for β tubulin III, a neuronal speci c cytoskeletal associated protein that is localized to neuronal processes. Neuronal cell bodies were detected by DAPI nuclear staining (Fig. 7A, B). Immunocytochemistry of neurons for β tubulin III following co-culture show extensive neuronal differentiation and neuronal process outgrowth following co-culture with siRNA-Control and siRNA-Nostrill transfected microglia (Fig. 7B). LPS stimulation of siRNA-Control transfected cells resulted in signi cant loss of immunoreactivity to β tubulin III seen in healthy neuronal processes (Fig. 7B). Co-culture of neurons with LPS stimulated BV2 cells following siRNA-Nostrill transfection noticeably reduced the dramatic loss of β tubulin III expression and loss of neuronal processes seen in LPS stimulated, siRNA-Control transfected BV2 co-cultures (Fig. 7B). Relative uorescence of β tubulin III immunoreactivity was quanti ed and normalized to the unstimulated siRNA-Control transfection condition (Fig. 7C). Silencing of Nostrill in LPS-stimulated BV2 microglia signi cantly improved β tubulin III immunoreactivity by nearly 0.37 ± 0.09 folds or ~ 37% (p < 0.05, Fig. 7C). Overexpression of Nostrill in microglia resulted in co-cultured neurons with little β tubulin III immunoreactivity (Fig. 7D). LPS-stimulation of BV2 microglia overexpressing Nostrill also resulted in co-culture conditions that did not support neuronal expression of β tubulin III (Fig. 7D). Unstimulated BV2 cells transfected with PTarget control did not affect β tubulin III expression and neurite outgrowth (Fig. 7D). Quanti cation of β tubulin III immunoreactivity in cortical neurons cocultured with LPS-stimulated PTarget transfected BV2s was signi cantly decreased to 0.24 ± 0.09 folds that of PTarget transfected controls (Fig. 7E, p < 0.001). An ~ 85% reduction in β tubulin III immunoreactivity was observed in cortical neurons co-cultured with BV2 cells overexpressing Nostrill (PTarget-Nostrill) as compared to PTarget-Empty transfected controls (Fig. 7E). LPS stimulation of BV2 cells overexpressing Nostrill did not demonstrate a signi cant additive effect on β tubulin III immunoreactivity as compared to overexpression of Nostrill alone (Fig. 7E, p > 0.05).
Propidium iodide (PI) immunoreactivity is commonly used as a reliable indicator of cell death in vitro as it is excluded from living cells and taken up by dying or dead cells (39). Using the co-culture experimental design (Fig. 7A) neurons were stained with propidium iodide after co-culture with LPS-stimulated BV2 microglia following silencing and overexpression of Nostrill. Unstimulated and control transfected BV2 cells (siRNA-Control and PTarget-Empty) served as controls (Fig. 8). LPS-stimulation of siRNA-Control transfected microglia signi cantly increased PI immunoreactivity to 6.5 ± 0.20 folds that of unstimulated controls (Fig. 8A, B). Silencing of Nostrill in unstimulated controls (siRNA-Nostrill) did not signi cantly reduce PI uptake by neurons as compared to unstimulated siRNA-control conditions but signi cantly reduced the increased PI update following LPS-stimulation (siRNA-Nostrill + LPS) to near control levels at 1.37 ± 0.33 folds (Fig. 8A-B). Overexpression of Nostrill in unstimulated BV2 cells resulted in increased PI uptake in co-cultured neurons to 7.22 ± 0.99 folds that of neurons co-cultured with unstimulated PTarget-Empty control BV2 cells (Fig. 8C-D). This increase in PI uptake in neurons was similar to neurons cocultured with LPS-stimulated PTarget -Empty control BV2 cells (Fig. 8C-D). Co-culture of neurons with BV2 cells overexpressing Nostrill and then stimulated with LPS signi cantly increased PI uptake in neurons to 10.58 ± 3.28 folds that of control co-culture conditions (Fig. 8C-D) but were not signi cantly different from PI uptake by neurons in LPS-stimulated PTarget-Nostrill or unstimulated PTarget-Nostrill co-culture conditions (Fig. 8D).

Discussion
Microglial proin ammatory states elicited by systemic immune responses to bacterial or viral pathogens are associated with a variety of neurodegenerative and autoimmune diseases (40)(41)(42)(43)(44). In order to determine whether inhibition of speci c molecular mechanisms regulating microglial-mediated neuroin ammation will be effective for prevention or treatment of these diseases, it is necessary to increase our understanding of the molecular processes regulating microglial proin ammatory states.
Interestingly, the expression of several of these lncRNAs, such as lincRNA-Cox2 is regulated by the proin ammatory transcription factor NF-κB (31,38,65). NF-κB-mediated transcription of lincRNAs likely coincides with the new protein synthesis driven by early response genes (66) and may allow for lincRNA and protein interactions necessary for secondary and late gene transcription. Similar to lincRNA-Cox2, one of the major ndings of this study is that LPS-induced upregulation of Nostrill appears to be dependent upon NF-κB signaling since two inhibitors of the NF-κB signaling pathway (SC-514, an IKK-2 inhibitor, and JSH-23, an NF-κB p65 inhibitor) signi cantly attenuated Nostrill expression in LPSstimulated BV2 cells (Fig. 2). Silencing of Nostrill in LPS-stimulated BV2 and primary microglia reduced iNOS gene transcription (Fig. 3-5) and nitric oxide production (Fig. 6). Overexpression of Nostrill in BV2 cells increased inducible NOS (iNOS) mRNA and nitric oxide production (Fig. 6) suggesting that Nostrill acts to drive iNOS gene transcription and NO synthesis in LPS-stimulated microglia. Since iNOS is a secondary response gene the regulatory mechanisms of iNOS gene transcription may require protein synthesis and chromatin remodeling (67). RIP analyses showed Nostrill may regulate iNOS gene transcription by interacting with NF-κB p65 and then associating with regions within iNOS promoter sites as indicated by ChIP. Silencing of Nostrill revealed further that the assembly of RNA polymerase II and modi ed histone H3K4me3 at iNOS promoter region following LPS-stimulation was in uenced by Nostrill. Recruitment of Nostrill to the gene locus of the secondary response gene iNOS was further con rmed by ChIRP. Interestingly, Nostrill was recently identi ed as a CAP-associated lincRNA but Nostrill's functional role was not studied (68). As one of two recently identi ed CAP-associated lincRNAs, Nostrill could readily act as a scaffold in long-range chromatin and protein interactions to assist with iNOS transcription (68). Several reports have speculated that lincRNAs may function as scaffold molecules to affect gene expression (25,47,63). LncRNAs may function as scaffold molecules because they are able to interact with RNA-binding proteins such as polycomb repressive complex 1 (PRC1) or MyBBP1A (24,69,70).
Speci cally, previous work has shown that lincRNA-Cox2 directly interacts with MyBBP1A and may be necessary for MyBBP1A assembly into the SWI/SNF complex (31). These data suggest Nostrill may also function to scaffold the transcriptionally active p65 protein of NF-κB, H3K4me3, and RNA polymerase II at iNOS promoter region. The assembly of other RNA-binding proteins with Nostrill is undetermined and is under investigation.
Proin ammatory activation of microglia causing the overproduction and/or sustained production of nitric oxide (NO) contributes to neurotoxicity (57,71). Mechanisms of neurotoxicity involving lincRNA function in microglia may underly the development and persistence of neurodegenerative and autoimmune disease processes (6,17,25,40,53). For example, the antimicrobial immune response of microglia to release NO in response to bacterial infection is known to contribute to the destruction of myelin and death of CNS neurons during proin ammatory phases of multiple sclerosis (13,40,72). Targeted inhibition of proin ammatory pathways in microglia may reduce neurotoxicity and help mitigate or treat such neuroin ammatory disorders (6). A few studies have shown that siRNA delivery to the CNS can exacerbate or reduce aspects of lincRNA-regulated microglial proin ammatory responses in vivo (35,64,73,74) indicating their utility when the lincRNA function is fully understood. Our proof-of-concept, in vitro, co-culture experiments showed that silencing of Nostrill in microglia inhibits LPS-stimulated neurotoxicity while overexpression of Nostrill leads to neurotoxicity ( Fig. 7-8). Both immunocytochemistry and PIuptake data in this in vitro system provide support for the hypothesis that blocking proin ammatory, lincRNA-mediated gene transcription can reduce neurotoxicity. These in vitro studies are the rst step to investigating the neurobiological relevance of targeting Nostrill in microglia following activation by pathogenic signals such as bacterial LPS. Future studies are designed to further investigate the potential therapeutic effects of Nostrill silencing in primary microglia and in vivo model systems.

Conclusions
Our data indicate a new regulatory role of NF-κB-induced Nostrill and suggest that Nostrill acts as a coactivator of transcription of iNOS resulting in the production of nitric oxide in microglia through modulation of epigenetic chromatin remodeling. Nostrill may be a target for reducing the neurotoxicity associated with iNOS-mediated in ammatory processes in microglia during neurodegeneration. Continued lincRNA studies in microglia will further expand our understanding of the utility RNA drug targets for neurodegenerative and autoimmune diseases.