Cyanobacteria Scytonema javanicum and Scytonema ocellatum Lipopolysaccharides Elicit Release of Superoxide Anion, Matrix-Metalloproteinase-9, Cytokines and Chemokines by Rat Microglia In Vitro

Cosmopolitan Gram-negative cyanobacteria may affect human and animal health by contaminating terrestrial, marine and freshwater environments with toxins, such as lipopolysaccharide (LPS). The cyanobacterial genus Scytonema (S) produces several toxins, but to our knowledge the bioactivity of genus Scytonema LPS has not been investigated. We recently reported that cyanobacterium Oscillatoria sp. LPS elicited classical and alternative activation of rat microglia in vitro. Thus, we hypothesized that treatment of brain microglia in vitro with either cyanobacteria S. javanicum or S. ocellatum LPS might stimulate classical and alternative activation with concomitant release of superoxide anion (O2−), matrix metalloproteinase-9 (MMP-9), cytokines and chemokines. Microglia were isolated from neonatal rats and treated in vitro with either S. javanicum LPS, S. ocellatum LPS, or E. coli LPS (positive control), in a concentration-dependent manner, for 18 h at 35.9 °C. We observed that treatment of microglia with either E. coli LPS, S. javanicum or S. ocellatum LPS generated statistically significant and concentration-dependent O2−, MMP-9 and pro-inflammatory cytokines IL-6 and TNF-α, pro-inflammatory chemokines MIP-2/CXCL-2, CINC-1/CXCL-1 and MIP-1α/CCL3, and the anti-inflammatory cytokine IL-10. Thus, our results provide experimental support for our working hypothesis because both S. javanicum and S. ocellatum LPS elicited classical and alternative activation of microglia and concomitant release of O2−, MMP-9, cytokines and chemokines in a concentration-dependent manner in vitro. To our knowledge this is the first report on the toxicity of cyanobacteria S. javanicum and S. ocellatum LPS to microglia, an immune cell type involved in neuroinflammation and neurotoxicity in the central nervous system.


Introduction
Cyanobacteria are photoautotrophic Gram-negative bacteria that are found in a wide range of terrestrial, marine and freshwater environments [1]. Overgrowth of cyanobacteria can result in blooms 10,000-and 100,000-fold, respectively, less potent than E. coli LPS in stimulating statistically significant O 2 − production from neonatal rat microglia in vitro.
Toxins 2018, 10, x FOR PEER REVIEW 3 of 15 maximal O2 − release at 1 ng/mL as previously reported [3]. Thus, S. javanicum and S. ocellatum LPS appeared to be 10,000-and 100,000-fold, respectively, less potent than E. coli LPS in stimulating statistically significant O2 − production from neonatal rat microglia in vitro.

Effect of S. javanicum and S. ocellatum LPS on Neonatal Rat Brain Microglia LDH Generation
To determine whether the decrease in O2 − production shown in Figure 1 resulted from concentration-dependent toxicity from E. coli or Scytonema LPS to microglia during the 18 h incubation, LDH release was determined in culture supernatants [19]. LDH release has been used extensively as a marker for cellular toxicity, as is described in the Materials and Methods [3,4].
As shown in Figure 2, LDH release was low for all concentrations of both S. javanicum and S. ocellatum LPS we investigated. In S. javanicum and S. ocellatum-LPS treated microglia, a nonstatistically significant release of LDH was observed at 100,000 ng/mL LPS, reaching 12.1 ± 12.1% and 14.9 ± 10.5% of control, respectively. In contrast, in E. coli LPS-stimulated microglia, a statistically significant LDH release of 35.3 ± 17.7% of control was observed at 100 ng/mL, as previously reported [3]. Thus, the LDH data suggest that both S. javanicum and S. ocellatum LPS elicited classical and alternative activation of microglia and concomitant release of O2 − , MMP-9 and cytokines and chemokines in a concentration-dependent manner while not affecting the microglia cell membrane in vitro at the concentrations tested in these experiments.

Effect of S. javanicum and S. ocellatum LPS on Neonatal Rat Brain Microglia LDH Generation
To determine whether the decrease in O 2 − production shown in Figure 1 resulted from concentration-dependent toxicity from E. coli or Scytonema LPS to microglia during the 18 h incubation, LDH release was determined in culture supernatants [19]. LDH release has been used extensively as a marker for cellular toxicity, as is described in the Materials and Methods [3,4]. As shown in Figure 2, LDH release was low for all concentrations of both S. javanicum and S. ocellatum LPS we investigated. In S. javanicum and S. ocellatum-LPS treated microglia, a non-statistically significant release of LDH was observed at 100,000 ng/mL LPS, reaching 12.1 ± 12.1% and 14.9 ± 10.5% of control, respectively. In contrast, in E. coli LPS-stimulated microglia, a statistically significant LDH release of 35.3 ± 17.7% of control was observed at 100 ng/mL, as previously reported [3]. Thus, the LDH data suggest that both S. javanicum and S. ocellatum LPS elicited classical and alternative activation of microglia and concomitant release of O 2 − , MMP-9 and cytokines and chemokines in a concentration-dependent manner while not affecting the microglia cell membrane in vitro at the concentrations tested in these experiments.

Effect of S. javanicum and S. ocellatum LPS on Neonatal Rat Brain Microglia Proinflammatory TXB 2 Generation
Eicosanoids, such as TXB 2 , have been implicated in neurodegenerative disease by contributing to neuroinflammation [30]. We have reported that cyanobacteria Microcystis aeruginosa and Oscillatoria sp. LPS stimulated rat microglia to release TXB 2 in vitro [3,4,19]. As shown in Supplementary Table S1, both S. javanicum, and S. ocellatum LPS-treated microglia showed a non-statistically significant TXB 2 release as compared to untreated microglia.

Effect of S. javanicum and S. ocellatum LPS on Neonatal Rat Brain Microglia Pro-Inflammatory MMP-9 Generation
MMP-9 and other matrix metalloproteinases produced during neuroinflammation may affect the blood brain barrier causing disruption and resulting neuropathology [31]. Our laboratory has previously reported that rat microglia release MMP-9 upon stimulation with cyanobacteria Microcystis aeruginosa and/or Oscillatoria sp. LPS [4,19]. MMP-9 release in supernatants was measured via ELISA. As shown in Figure 3, E. coli LPS-treated microglia released statistically significant levels of MMP-9 from 1-100 ng/mL LPS. S. javanicum LPS-treated microglia also released statistically significant levels of MMP-9 but at 10,000-100,000 ng/mL LPS. In contrast, S. ocellatum LPS did not induce statistically significant release of MMP-9 from treated microglia. Thus, S. javanicum LPS appeared to be 10,000-fold, less potent than E. coli LPS in stimulating statistically significant MMP-9 production from neonatal rat microglia in vitro.

Effect of S. javanicum and S. ocellatum LPS on Neonatal Rat Brain Microglia Proinflammatory TXB2 Generation
Eicosanoids, such as TXB2, have been implicated in neurodegenerative disease by contributing to neuroinflammation [30]. We have reported that cyanobacteria Microcystis aeruginosa and Oscillatoria sp. LPS stimulated rat microglia to release TXB2 in vitro [3,4,19]. As shown in Supplementary Table S1, both S. javanicum, and S. ocellatum LPS-treated microglia showed a nonstatistically significant TXB2 release as compared to untreated microglia.

Effect of S. javanicum and S. ocellatum LPS on Neonatal Rat Brain Microglia Pro-Inflammatory MMP-9 Generation
MMP-9 and other matrix metalloproteinases produced during neuroinflammation may affect the blood brain barrier causing disruption and resulting neuropathology [31]. Our laboratory has previously reported that rat microglia release MMP-9 upon stimulation with cyanobacteria Microcystis aeruginosa and/or Oscillatoria sp. LPS [4,19]. MMP-9 release in supernatants was measured via ELISA. As shown in Figure 3, E. coli LPS-treated microglia released statistically significant levels of MMP-9 from 1-100 ng/mL LPS. S. javanicum LPS-treated microglia also released statistically significant levels of MMP-9 but at 10,000-100,000 ng/mL LPS. In contrast, S. ocellatum LPS did not induce statistically significant release of MMP-9 from treated microglia. Thus, S. javanicum LPS appeared to be 10,000-fold, less potent than E. coli LPS in stimulating statistically significant MMP-9 production from neonatal rat microglia in vitro.  was substracted from all data. Data expressed as pg/mL is the mean ± SEM from three independent experiments (n), each experiment with triplicate determinations. * p < 0.05, ** p < 0.01, *** p < 0.001 LPS versus untreated control (0).

Discussion
Microglia activated by stimuli such as infections [50] and neurodegenerative diseases [51] display either the pro-inflammatory M1 or the anti-inflammatory M2 phenotypes that participate in the initiation and resolution of inflammation [43]. One activator of microglia is LPS which activates microglia via its lipid A moiety resulting in the concomitant generation of inflammatory mediators including matrix metalloproteases, arachidonic acid metabolites, cytokines, chemokines, and free

Discussion
Microglia activated by stimuli such as infections [50] and neurodegenerative diseases [51] display either the pro-inflammatory M1 or the anti-inflammatory M2 phenotypes that participate in the initiation and resolution of inflammation [43]. One activator of microglia is LPS which activates microglia via its lipid A moiety resulting in the concomitant generation of inflammatory mediators including matrix metalloproteases, arachidonic acid metabolites, cytokines, chemokines, and free radicals both in vivo and in vitro [19].
Our working hypothesis was that cyanobacteria S. javanicum and S. ocellatum LPS would induce an M1 or classical activation phenotype in primary neonatal rat microglia in vitro and O 2 − release.
In fact, both S. javanicum and S. ocellatum LPS stimulated microglia released statistically significant O 2 − in a concentration-dependent manner similar to E. coli LPS, which was used as a positive control. In addition to O 2 − , S. javanicum and S. ocellatum LPS-treated microglia generated pro-inflammatory cytokines and chemokines in a concentration-dependent manner: MIP-2/CXCL-2 > IL-6 > CINC-1/CXCL-1 > MIP-1α/CCL3 > TNF-α. Although S. javanicum LPS was less potent than E. coli LPS in inducing the M1 phenotype, and less efficacious in stimulating release of four cytokines and chemokines, the release of CINC-1/CXCL-1 was enhanced compared to E. coli LPS.
In contrast, S. ocellatum LPS, with the sole exception of MIP-1α/CCL3, was both less potent and less efficacious in activating an M1 microglia phenotype with concomitant release of MIP-2/CXCL-2, IL-6, CINC-1/CXCL-1, and TNF-α. Two recent studies characterizing microglial activation by cyanobacteria M. aeruginosa and Oscillatoria sp. LPS [3,4] allow for an interesting comparison of cyanobacterial LPS efficacy and potency in the concomitant generation of pro-inflammatory cytokines and chemokines. Of the four cyanobacterial LPS our laboratory has studied so far, M. aeruginosa appears to be the most efficacious in stimulating secretion of MIP-2/CXCL-2, IL-6, MIP-1α/CCL3, and TNF-α, while S. ocellatum LPS was the least efficacious. As compared to Oscillatoria sp. LPS [3], S. javanicum LPS appeared to be more efficacious in stimulating secretion of MIP-2/CXCL-2, IL-6, and CINC-1/CXCL-1 from rat microglia, but resulted in similar concentrations of MIP-1α/CCL3 and TNF-α. Thus, our study provides experimental support for our working hypothesis, namely that cyanobacteria S. javanicum and S. ocellatum LPS (0.1-100,000 ng/mL) activated an M1 or classical activation phenotype in primary rat microglia, with no significant toxicity to microglia in vitro.
The second component of our working hypothesis was to investigate whether S. javanicum and S. ocellatum LPS-treated rat microglia activated a M2 or alternative activation phenotype with concomitant release of the anti-inflammatory mediator IL-10. The M2 microglia phenotype and anti-inflammatory mediators have been associated with tissue repair processes [52]. Both S. javanicum and S. ocellatum LPS-treated rat microglia demonstrated statistically significant concentration-dependent release of the anti-inflammatory cytokine IL-10. Although both S. javanicum and S. ocellatum LPS were less potent than E. coli LPS in stimulating release of IL-10, S. javanicum LPS had similar efficacy as E. coli LPS. Thus, our present results complement our recently published study on the effects of cyanobacterium Oscillatoria sp. LPS on alternative activation of rat microglia and concomitant IL-10 release [3]. In terms of potency, both Scytonema species LPS were 10-fold less potent as they did not stimulate maximal IL-10 release until 100,000 ng/mL LPS, whereas Oscillatoria sp. LPS resulted in peak IL-10 release at 10,000 ng/mL LPS [3]. We currently hypothesize that the observed differences in potency and efficacy amongst the cyanobacterial LPS could be due to differing lipid A structures [25]. The structures of most cyanobacterial LPS is currently unknown, so determination of LPS structure appears necessary for further characterization of their in vitro and in vivo effects on microglial activation states [25,26].
Taken together, our data lend support to our working hypothesis by demonstrating that in vitro treatment of primary neonatal rat microglia with cyanobacteria S. javanicum and S. ocellatum LPS will result in both classical or M1 and alternative or M2 activation in a concentration-dependent manner. As our current study was conducted in vitro, and because it has been reported that E. coli and Salmonella tiphyimurium LPS activate microglia upon systemic administration [53][54][55], future studies are required to determine whether systemic cyanobacterial S. javanicum and S. ocellatum LPS will activate microglia in the CNS, as well as concomitant pro-inflammatory and anti-inflammatory mediator release.

Conclusions
In conclusion, the present investigation on the toxicology of both S. javanicum and S. ocellatum LPS to microglia in vitro extends our previous studies with cyanobacteria Microcystis aeruginosa and Oscillatoria sp. LPS, and contributes to our understanding of the potential toxicity of cyanobacterial LPS in general, and the genus Scytonema in particular, to the brain immune system.

Isolation of Primary Rat Neonatal Microglia
Adherence to the National Institutes of Health guidelines on the use of experimental animals and protocols approved by Midwestern University's Research and Animal Care Committee were followed in all experiments (Midwestern University Protocol File # 941.1 titled " Neuroinflammation, microglia and marine natural products"was approved on 5 January 2015). The cerebral cortices of 1-2 day-old Sprague-Dawley rats (Charles Rivers, Hartford, CT, USA), were surgically removed, placed in cold DMEM containing 120 U/mL P and 12 µg/mL S, the meninges removed, and brain tissue minced and dissociated with trypsin-EDTA at 35.9 • C for 3-5 min. The mixed glial cell suspension was plated in 225-cm 2 vented cell culture flasks with DMEM medium supplemented with 10% FBS containing 120 U/mL P and 12 µg/mL S and grown in a humidified 5% CO 2 incubator at 35.9 • C for 12-14 days. Upon confluence of the astrocyte layer (day 14) and every week thereafter, microglia were detached using an orbital shaker (150 rpm, 0.5 h, 35.9 • C, 5% CO 2 ), centrifuged (400× g, 25 min, 4 • C), and cell number and viability assessed by trypan blue dye exclusion. Rat neonatal microglia (2 × 10 5 cells/well) averaging greater than 95% viability were plated in non-pyrogenic polystyrene 24-well flat-bottom culture clusters with DMEM supplemented with 10% FBS containing 120 U/mL P and 12 µg/mL S, and then transferred to a humidified incubator at 35.9 • C and 5% CO 2 , 24 h prior to the experiments. The purity of the isolated rat brain neonatal microglia was routinely determined using a mouse monoclonal anti-rat CD11b antibody (Cat # MCA275R, AbD SeroTec, Raleigh, NC, USA), and was greater than 98%.

Activation of Microglia with LPS (Experimental Protocol)
To determine the effect of S. javanicum and S. ocellatum LPS on neonatal rat microglia classical and alternative activation and concomitant mediator release (O 2 − , thromboxane B 2 , cytokines, and chemokines), 1.8-2.0 × 10 5 neonatal microglia in DMEM + 10% FBS + 0.1% P + S were plated into each well of a 24-well flat-bottom culture cluster, and then stimulated with either 0. was measured in the presence of FCC (50 µM) and HBSS, with or without SOD (700 Units), which inhibited >95% of FCC reduction during a 70 min incubation. All experimental treatments were run in duplicate and in a final volume of 1 mL. Changes in FCC absorbance were measured at 550 nm using a DU-800 spectrophotometer (Beckman Coulter, Inc., Indianapolis, Indiana, USA). Differences in the amount of reduced FCC, in the presence and absence of SOD, were used to determine microglia O 2 − generation using the molecular extinction coefficient of 21.0 × 10 3 M −1 cm −1 and data expressed in nmol.

Lactate Dehydrogenase Assay
To assess cell viability following preincubation of microglia with either S. javanicum, S. ocellatum or E. coli LPS as described in our experimental protocol, the conditioned medium was harvested and LDH release was determined spectrophotometrically [19,60]. Microglia LDH release was expressed as a percent of total LDH released into the conditioned media. Total LDH release resulted from 0.1% Triton X-100-lysed microglia (intracellular LDH) plus LDH present in the extracellular media, because the fetal bovine serum contained LDH (data not shown). Unless LDH release from LPS-treated microglia was significantly greater than that observed from the vehicle-treated group, shown as 0 or control in the corresponding figures, the 18 h LPS treatment was considered to have had no effect on microglia viability.

Assay for Microglia TXB 2 Generation
After preincubation of neonatal rat microglia with either E. coli, S. javanicum, or S. ocellatum LPS for 18 h, TXB 2 production was determined by immunoassay (Cayman Chemical, Ann Arbor, MI, USA) according to the manufacturer's protocol. Results were expressed as pg/mL and the minimum detectable concentration was 7.8 pg/mL TXB 2 .

Assay for Microglia MMP-9 Generation
After 18 h treatment of neonatal rat microglia with E. coli, S. javanicum, or S. ocellatum LPS, MMP-9 generation was determined by ELISA (Cat# DY8174-05, R&D Systems, Minneapolis, MN, USA) according to manufacturer's protocol. Results were expressed as pg/mL and the minimum detectable concentration was 78.10 pg/mL MMP-9.

Statistical Analysis of the Data
Data was expressed as means ± SEM of triplicate determinations of 3 similar experiments. Data was analyzed with Prism software package version 7 from GraphPad (San Diego, CA, USA). Appropriate multiway analysis of variance was performed on all sets of data. Where significant interactions were encountered, simple effects were tested with a one-way analysis of variance followed by a Dunnett multiple comparisons test. Differences were considered statistically significant at p < 0.05 [3].