Background
Manganese (Mn) is an essential trace element primarily acquired through diet. However, increased exposures in juveniles and adults lead to inflammation and neuronal injury in the cortex and basal ganglia that can cause irreversible neurodegeneration associated with cognitive and motor deficits. Sources of excess Mn exposure include soy-based infant formula [
1], well water [
2], and industrial activities such as mining [
3] and welding [
4]. Activation of astrocytes and microglia in response to Mn neurotoxicity can lead to overproduction of neurotoxic levels of reactive oxygen and nitrogen species (ROS, RNS) [
5], as well as inflammatory cytokines such as TNF [
6]. Additionally, astrocytes concentrate Mn through plasma membrane divalent metal transports, which increase oxidative stress and decrease their capacity for neuronal trophic support by affecting key metabolic coupling pathways such as glutamate uptake and glutathione synthesis [
7,
8]. The combination of enhanced inflammatory gene expression and decreased trophic support may therefore cause a reactive phenotype that promotes neuronal injury.
A number of studies have demonstrated that Mn-induced glial activation is exacerbated by glial-derived pro-inflammatory factors that damage neurons. Data from our laboratory recently demonstrated that NF-κB signaling in microglia plays an essential role in inflammatory responses in Mn toxicity by regulating cytokines and chemokines that amplify the activation of astrocytes [
6]. Although this demonstrates that nuclear factor kappa B (NF-κB) signaling in microglia is essential to inflammatory activation of astrocytes, subsequent effects of astrocytes on microglia and ultimately on neuronal cell death are less well understood. It was reported that Mn exposure induces activation of microglia and signs of dystrophy including increased iron-mediated oxidative stress in the substantia nigra of non-human primates [
9], as well as microglia-induced degeneration of dopaminergic neurons in rats [
10]. Previous studies from our laboratory and others demonstrated that expression of inducible nitric oxide synthase (iNOS/NOS2) and NO production in astrocytes causes injury to surrounding neurons in Mn-exposed mice [
11‐
14]. NOS2 and many other pro-inflammatory factors are highly regulated in glial cells by NF-κB, consistent with data reporting that the NF-κB-mediated pro-inflammatory cytokines CCL2, CCL5, and TNF released by astrocytes are associated with Mn neurotoxicity in in vitro studies of murine glia [
5,
15,
16].
However, only recently has research begun to establish how the communication between both glial cell types mediates Mn-induced neuronal injury [
17]. The astrocyte-specific chemokine, CCL2, is regulated by NF-κB and induces microglial activation in a surgery-induced cognitive dysfunction and neuroinflammatory model, suggesting not only that astrocytes and microglia communicate during stress and injury, but also that NF-κB-mediated factors contribute to glial inflammation. NF-κB is activated in glia in response to Mn, oxidative stress, and other neurotoxic exposures [
15,
16,
18]. We recently reported that NF-κB activation in microglia amplifies the inflammatory response of astrocytes to Mn toxicity, resulting in overproduction of inflammatory cytokines and chemokines such as TNF, IL1, and IL6 [
6]. However, it remains to be determined which NF-κB-regulated inflammatory signaling molecules in astrocytes modulate the inflammatory response of microglia, as well as the effect of this signaling on neuronal injury.
We therefore examined the role of NF-κB in Mn-induced neurotoxicity by exposing pure microglia, astrocytes (from wild-type and astrocyte-specific IKK/NF-κB knockout mice), and mixed glial cultures to varying concentrations of Mn and then treating neurons with the resultant glial-conditioned media (GCM) of each cell type. We hypothesized that Mn would enhance expression of inflammatory cytokines in mixed cultures of astrocytes and microglia to a greater extent than in either cell type alone and would similarly increase neuronal cell death. We measured expression of NF-κB-regulated inflammatory genes by qPCR in glial cells, as well as levels of secreted cytokines in GCM from both WT or IKK KO astrocytes and microglia using spotted array-based ELISA. Mn exposure enhanced expression of multiple inflammatory cytokines and chemokines in mixed glial cultures, which was inhibited by pharmacologic inhibition or genetic inhibition of NF-κB. Additionally, Mn-stimulated GCM increased neuronal apoptosis, which was attenuated by inhibition of NF-κB in astrocytes, suggesting that activation of this signaling pathway in astrocytes is a critical mediator of glial reactivity and neuronal injury from Mn through release of cytokines and chemokines that amplify activation of microglia.
Materials and methods
Materials
All general chemical reagents including MnCl2 and antibiotics were purchased from Sigma Aldrich (St. Louis, MO). Fluorescent antibodies and dyes were purchased from Life Technologies (Carlsbad, CA). Cell culture medium was acquired from Hyclone (Logan, UT) or Gibco (Life Technologies, Carlsbad, CA). For immunofluorescence studies and flow cytometric experimentation and live-cell imaging of primary neurons, Annexin V-iFluor™ 647 conjugate was purchased from AAT Bioquest (Sunnyvale, CA), and Propidium Iodide and Caspase-3/7 Green were purchased from Life Technologies (Carlsbad, CA).
Primary glial and neuronal isolation
Cortical glia and neurons were isolated from 1-day-old C57Bl/6 or transgenic mouse pups according to procedures described previously [
6,
19], and purity confirmed through immunofluorescence staining using antibodies against GFAP and IBA1 [
20]. Mixed glial cultures were also established from astrocyte-specific IKK2 knockout mice, which were generated in our laboratory by crossing mice expressing a
loxP-targeted (floxed)
Ikk2 allele [
21] with mice expressing the human
Gfap promoter driving expression of cre recombinase [
22]. Progeny were bred to homozygosity for the floxed-
Ikk2 allele, and both male and female littermates from the F4 generation were utilized for cell isolations. Briefly, pups were euthanized by decapitation under isofluorane anesthesia, and cortices (astrocytes) were rapidly dissected out, and meninges removed. Tissue was subject to digestion with Dispase (1.5 U/ml), and a complete media change 24 h after plating to remove non-glial cell types from glial cultures and glia from neuronal cultures. Glial cultures were maintained at 37 °C and 5% CO
2 in minimum essential media supplemented with 10% heat-inactivated fetal bovine serum and a penicillin (0.001 mg/ml), streptomycin (0.002 mg/ml), and neomycin (0.001) antibiotic cocktail (PSN). Neuronal cultures were maintained at 37 °C and 5% CO
2 in neurobasal media supplemented with HEPES, B27, and PSN. Cell media was changed 24 h prior to all treatments. All animal procedures were approved by the Colorado State University Institutional Animal Care and Use Committee and were conducted in accordance with published NIH guidelines. Neuro-2a cells (N2A) were cultured as previously described [
23].
Treatments and glial-conditioned media experiments
Glia were treated with MnCl
2 (solubilized in saline) (0–100 μM) for 8 h similarly to methods and experimentation previously described [
6]. In brief, prior to mRNA assessment or conditioned media experiments, glia were seeded onto six-well tissue culture plates at approximately 3 × 10
5 cells per well, grown to confluence and treated with saline or 100 μM MnCl
2 (in 2 ml total of cell culture medium adequate for growth/health of glia and survival/health of neurons; see the “
Materials and methods” section) for 8 h. To inhibit NF-κB signaling in glia, cells seeded in six-well plates were treated with 5 μM Bay 11-7082 (Bay11; Sigma) or the vehicle dimethyl sulfoxide (DMSO; Sigma) at 0.05% in complete media for 3 h. Exposure to Bay11 or DMSO for 3 h had no effect on cell viability. Media was removed prior to 8 h treatment with Saline or MnCl
2. Conditioned media (denoted GCM for mixed glia-conditioned media, ACM for astrocyte-conditioned media, and MCM for microglia-conditioned media) were pooled per treatment and centrifuged at 800×
g for 10 min to remove detached cells. N2As were seeded in six-well tissue culture plates at 1 × 10
5 cells per well for flow cytometry experiments or 5 × 10
3 cells per well of 96-well tissue culture plate for Presto Blue Viability Assay 24 h prior to 48 h exposure to conditioned media. Primary neurons were seeded in four-well chamber slides for live-cell imaging at 1 × 10
5cells per well and 2 × 10
5cells per well of a 96-well tissue culture plate for Presto Blue Viability Assay 10 days prior to 48 h exposure to conditioned medium. N2As and primary neurons were then assessed for viability and cell death markers.
Real-time RT-PCR and qPCR array analysis
Confluent mixed glia, purified astrocytes, or purified microglia were treated with MnCl
2 (0–100 μM) for 8 h prior to RNA isolation. RNA was isolated using the RNEasy Mini kit (Qiagen, Valencia, CA), and purity and concentration were determined using a Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Following purification, RNA (250–1000 ng) was used as template for reverse transcriptase (RT) reactions using the iScript RT kit (BioRad, Hercules CA). The resulting cDNA was immediately profiled for mRNA expression according to the 2-ΔΔCT method [
24]. Primer sequences for all genes profiled are provided in Table
1.
Table 1
Primer table. Primer sequences of measured genes in qPCR experiments
Nos2
| NM_010927.3 | For: TCA CGC TTG GGT CTT GTT | 149 |
Rev: CAG GTC ACT TTG GTA GGA TTT |
Tnfα
| NM_013693.3 | For: CTT GCC TGA TTC TTG CTT CTG | 140 |
Rev: GCC ACC ACT TGC TCC TAC |
Il1-β
| NM_008361.3 | For: GCA GCA GCA CAT CAA CAA G | 90 |
Rev: CAC GGG AAA GAC ACA GGT AG |
Ccl2
| NM_011331.2 | For: TTAAAAACCTGGATCGGAACCAA | 121 |
Rev: GCATTAGCTTCAGATTTACGGGT |
Ccl5
| NM_013653.3 | For: GCT GCT TTG CCT ACC TCT CC | 104 |
Rev: TCG AGT GAC AAA CAC GAC TGC |
Il-6
| NM_031168.1 | For:CTG CAA GAG ACT TCC ATC CAG | 131 |
Rev:AGT GGT ATA GAC AGG TCT GTT GG |
β-actin
| NM_007393.3 | For: GCT GTG CTA TGT TGC TCT AG | 117 |
Rev: CGC TCG TTG CCA ATA GTG |
Hprt
| NM_013556.2 | For: TCA GTC AAC GGG GGA CAT AAA | 142 |
Rev: GGG GCT GTA CTG CTT AAC CAG |
C3
| NM_009778.3 | For: GAG CGA AGA GAC CAT CGT ACT | 83 |
Rev: TCT TTA GGA AGT CTT GCA CAG TG |
Ccr2
| XM_011243064.2 | For: ATC CAC GGC ATA CTA TCA ACA TC | 89 |
Rev: TCG TAG TCA TAC GGT GTG GTG |
Presto Blue viability assay
N2A cells (5000 cells per well) or primary neurons (2 × 105cells/well) were grown or plated on 96-well plates for 24 h or 10 days, respectively, before treatment with GCM. After 48 h, cells were imaged using the PrestoBlue Cell viability reagent (Life Technologies, Carlsbad, CA) per the manufacturer’s protocol.
Spotted protein array ELISA assays
Measurement of cytokines in glia-conditioned media (resultant from treatments for mRNA expression profiling of glia) was sampled from glia prior to application on neurons and stored at − 80 °C. Stored media was thawed, and cytokines were measured using a custom mouse 7-plex ELISA (Q-Plex™ Mouse Cytokine Arrays, Quansys Biosciences, Logan, UT) according to manufacturer instructions and imaged on a ChemiDoc XRS (Life Science Research, Hercules, CA) to capture images. Levels of cytokines and chemokines were calculated from standard curves using Q-View imaging software (Quansys Biosciences, Logan, UT).
RNA interference
RNA interference (siRNA, small interfering RNA) oligonucleotides were purchased from Integrated DNA Technologies (IDT DNA, Coralville, IA). RNAi duplexes were designed against splice common variants of the target gene and were validated using a dose-response assay with increasing concentrations of the suspended oligo (900–1200 ng/ml) using a standard scrambled dicer substrate RNA (DsiRNA) as control. RNAi oligonucleotides were transfected using the TransIT-X2 delivery system (Mirus Bio, Madison, WI) 48 h before 100 μM MnCl2 treatment. Separate siRNA systems were used to ensure specific knockdown of Ccl2 and C3 mRNA. The Ccl2 dsiRNA duplex sequences are (5′➔3′) UGAAGCUAAUGCAUCCACUACCUTT; UAAACAAUACCUUGGAAUCUCAAACAC (IDT DsiRNA; denoted siCcl2). The C3 dsiRNA duplex sequences are (5′➔3′) UAAUAAAGCUUCAGUUGUAUUUCAA; UUGAAAUACAACUGAAGCUUUAUUAGA (IDT DsiRNA; denoted siC3).
Flow cytometry
The percent of Annexin-V positive (+) and Propidium Iodide positive (+) in neuroblastoma (N2A) cultures before and after treatment with or without conditioned media or MnCl
2 for 48 h followed by flow cytometric analysis as described [
6]. Briefly, cells were labeled using Annexin-V and Propidium Iodide (PI) at room temperature for 1 h. After labeling, the cells were washed twice in incubation buffer and resuspended at a final volume of 500 μL of PBS and stored at 37 °C until analysis. Flow cytometry was performed on a Beckman Coulter CyAn ADP flow cytometer operated with Summit Software for data collection at Colorado State University’s Flow Cytometry Core Facility. All further data analysis was done utilizing FlowJo software (version 10.1; FlowJo, Ashland, OR).
Statistical analysis
Experiments were performed in triplicate, with replicates consisting of independent cultures using a minimum of (n = 4) four plates or cover slips per replicate study. Comparison of two means was performed by Student’s t test, while comparison of three or more means was performed using one-way ANOVA while those consisting of comparison of three or more among the genetic variations of wild-type and knock-out treatment groups followed the Tukey-Kramer multiple comparison post hoc test using Prism software (v6.0 h, Graphpad Software, Inc., San Diego, CA). For all experiments, data was reported as standard error mean (± S.E.M) and P < 0.05 was considered significant, although the level of significance was often much greater (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001).
Discussion
Mn-induced neurotoxicity in humans and in animal models is accompanied by reactive gliosis and inflammation that is damaging to surrounding neurons, although mechanisms regulating this neuroinflammatory phenotype remain poorly understood. Multiple studies have reported that microglia and astrocytes respond to Mn exposure with elevated levels of NF-κB-regulated inflammatory cytokines and other inflammatory mediators [
11,
17,
25]. Furthermore, other studies have evaluated glial responses as a result of Mn-induced neuronal cell death [
13]; however, research assessing the role that glial-glial and glial-neuronal communication plays in neurodegeneration in a model of manganism is lacking.
To identify glial-derived factors involved in neuronal injury as a result of Mn exposure, we first assessed Mn-induced inflammatory gene expression in three glial populations. We treated astrocytes and microglia, either together or separately, for 8 h with 0, 30, or 100 μM MnCl
2. Overall, in mixed glial cell cultures, there was a dose-dependent increase in several NF-κB-regulated inflammatory genes including
Nos2,
Il6, and
Il1β (Fig.
1a) that was much greater than in pure astrocytes (Fig.
1b) or pure microglia alone (Fig.
1c), suggesting that microglia-astrocyte communication amplifies the inflammatory response to Mn. Interestingly, the complement component
C3, which is uniquely expressed by activated astrocytes in the CNS and correlates closely with a neurotoxic inflammatory phenotype (A1 astrocytes) [
26], was more highly induced in mixed glia (~ 4-fold more in 100 μM compared to 0 μM) than in pure astrocytes (only ~ 2-fold increase), whereas pure microglia showed a dose-dependent decrease upon Mn-exposure, consistent with findings that show
C3 is not significantly expressed by microglia [
26]. This demonstrates that
C3 is induced in astrocytes following Mn exposure and also that microglia potentiate expression of inflammatory gene expression in astrocytes, similar to in vivo studies reporting that release of the inflammatory cytokines C1q, IL1a, and TNF by microglia induces reactive astrocytosis in vivo [
26]. Additionally, the chemokine,
Ccl2, was induced following Mn exposure in mixed glia and in pure astrocytes but not in pure microglial cultures, consistent with studies demonstrating that
Ccl2 is astrocyte-derived and increases microglial activation and neuroinflammation [
27,
28]. This suggests that cell-cell communication between microglia and astrocytes promotes an increase in astrocytic-specific inflammatory gene expression compared to astrocytes cultured in the absence of microglia, possibly through increased astrocyte expression of
Ccl2. Other studies have also implicated CCL2-CCR2 signaling in astrocyte-mediated microglial activation in central nervous system (CNS) inflammation [
27,
28], suggesting that the increase in both
Ccl2 and
Ccr2 in mixed glia is what contributes to an increased inflammatory gene response due to Mn treatment.
To determine whether glia release neurotoxic factors as a result of Mn exposure, we incubated N2A cells with conditioned media from cultures of mixed glia (GCM), astrocytes (ACM), or microglia (MCM) and found that 48 h of exposure to 100 μM Mn-GCM caused a greater decrease in neuronal viability compared to either ACM or MCM. Although purified astrocytes and microglia elicited some loss of neuronal viability in response to Mn, this was magnified with both cell types present, consistent with other studies demonstrating that Mn increases production in inflammatory cytokines and chemokines [
5,
15,
17]. These data also suggest that cell-cell signaling between astrocytes and microglia amplifies the overall neuroinflammatory response to Mn, similar to recent studies from our lab demonstrating that microglia-derived inflammatory cytokines were greater in the presence of astrocytes than in pure microglial cultures [
6]. Flow cytometric analysis also indicated that N2A cell death was enhanced by exposure to Mn-GCM, as shown by staining for Annexin V and Propidium Iodide for apoptotic and necrotic cells, respectively (Fig.
2c, d). This suggests that mixed populations of astrocytes and microglia produce more damaging levels of inflammatory mediators than either cell type alone, likely due to glial-glial communication that intensifies a reactive phenotype.
Because many inflammatory cytokines and chemokines associated with innate immune response in Mn-exposed glial cells are regulated by NF-κB [
16,
25,
29], we tested the function of this signaling pathway in regulating astrocyte cross-communication with microglia. Treating purified astrocytes or mixed glia with the NF-κB/IKK2 inhibitor, Bay 11-7082, prior to treatment with 100 μM Mn resulted in significant reductions in glial activation, based on expression of inflammatory genes. The inflammatory genes upregulated (
Nos2,
C3,
Il6,
Il1β,
Ccl5,
Ccl2,
Ccr2, and
Tnfα) by Mn exposure were significantly decreased in both mixed glia and pure astrocytes, suggesting that Mn-induced inflammatory gene expression in mixed glia and astrocytes requires activation of NF-κB.
The functional effect of Bay-11 in Mn-treated glia was determined in neuronal viability studies using GCM or ACM derived from Bay-11-pretreated glial cells (Fig.
4). Bay-11 pretreatment was significantly neuroprotective (Fig.
4a) in GCM-treated N2A cells, whereas ACM showed a similar but less potent trend (Fig.
4b). Direct treatment of N2A cells with Mn demonstrated that the LD50 was 30 μM and that Mn also directly increased the number of apoptotic (Annexin V-positive) and necrotic (PI-positive) neuronal cells, although to a lesser extent than Mn-treated GCM or ACM (Fig.
4c–e). This finding supports that glial-derived factors contribute to N2A cell death. We previously reported that glial uptake of Mn is ~ 70%, leaving behind ~ 30% of Mn in the conditioned medium [
6]; thus, uptake of 70% of 100 μM Mn treatment would result in ~ 30 μM in the resultant conditioned media, which caused greater cell death than direct treatment of N2A cells with 30 μM Mn. This finding suggests that N2A cell death from conditioned media exposure is not solely due to residual Mn in the media but is due in part to glial-released inflammatory factors, most likely regulated by NF-κB. This is further supported by the flow cytometric analysis of GCM and ACM from Bay-11 experiments in which there is increased neuroprotection as demonstrated by the significant decrease in apoptotic (Annexin V) and necrotic (PI) positive staining in N2A cells exposed to Bay-ll-pretreated GCM (Fig.
4f, h) or ACM (Fig.
4g, i). Taken together, these findings support the involvement of NF-κB activation in glial cells in the neurotoxicity of Mn by demonstrating that pharmacologic inhibition of NF-κB in glia is anti-inflammatory and neuroprotective.
To more precisely establish the involvement of NF-κB in Mn-induced glial inflammation and neuronal cell death, we isolated primary mixed glia from astrocyte-specific IKK knockout (KO) mice and exposed the mixed glia (containing IKK KO astrocytes and wild-type microglia) to 100 μM Mn for 8 h. We showed that there was a significant decrease in inflammatory gene expression based on levels of both mRNA and protein for multiple NF-κB-regulated inflammatory genes (Fig.
5), suggesting not only that NF-κB is directly involved in Mn-induced inflammatory gene expression, but also that genetically inhibiting NF-κB in astrocytes in a mixed glial culture mitigates overall expression of inflammatory genes in both astrocytes as well as microglia following exposure to Mn.
Exposing N2A cells to GCM from astrocyte-specific IKK2 knockout mixed glial cultures resulted in almost complete neuroprotection in GCM-treated N2A cells (Fig.
6a), with significantly fewer apoptotic (Annexin V) and necrotic (PI) positive staining (Fig.
6b, c). Additionally, incubation of primary cortical neurons with GCM from astrocyte-specific IKK2 knockout mixed glial cultures resulted in similar neuroprotection in live cell imaging experiments (Fig.
7), with decreased apoptotic (Annexin V), necrotic (PI), and active caspase 3/7-positive cells. This suggests that IKK2-dependent activation of NF-κB in astrocytes in response to Mn is critical for inflammatory injury to neurons, likely through the release of numerous neurotoxic inflammatory mediators that both directly injure neurons and amplify the inflammatory response of microglia.
To begin to identify astrocyte-derived factors that could result in amplification of glial inflammatory signaling and subsequent neuronal injury, we knocked down (KD) two NF-κB-mediated genes specific to astrocytes,
C3 and
Ccl2 [
26,
27], and determined the effect on inflammatory gene expression and neuronal injury (Fig.
8). Knockdown of either
C3 or
Ccl2 did not diminish inflammatory gene expression in mixed glial cultures, indicating that these factors are not directly involved in glial cross-communication leading to amplification of inflammatory signaling. This is not surprising, given previous studies from our laboratory identifying TNF as a key regulator of glial reactivity in response to Mn [
6]. However, TNF-dependent activation of NF-kB and subsequent production of C3 or CCL2 could still be directly toxic to associated neurons following exposure to Mn. To test this hypothesis, we incubated N2A cells with GCM from mixed glial cultures in which RNAi knocked down astrocytic expression of C3 or Ccl2. Treatment of N2A cells with GCM from C3 KD mixed glial cultures did not prevent Mn-dependent injury, whereas CCL2 KD GCM significantly decreased N2A cell death (Fig.
8c), demonstrating that CCL2 expression in astrocytes is necessary for glial-mediated neuronal cell death in response to Mn treatment. These data not only suggest that NF-κB activation in astrocytes is specifically involved in Mn-induced neuronal injury but also that astrocyte-derived CCL2 is a key contributor to neuronal cell death in a model of Mn neurotoxicity. The factor(s) regulated by CCL2 in microglia and astrocytes that cause direct injury to neurons following Mn exposure remain to be further elucidated.