Abstract
The cysteinyl leukotrienes (CysLTs) are inflammatory mediators closely associated with neuronal injury after brain ischemia through the activation of their receptors, CysLT1R and CysLT2R. Here we investigated the involvement of both receptors in oxygen-glucose deprivation/recovery (OGD/R)-induced ischemic neuronal injury and the effect of the novel CysLT2R antagonist HAMI 3379 [3-({[(1S,3S)-3- carboxycyclohexyl]amino}carbonyl)-4-(3-{4-[4-(cyclo-hexyloxy)butoxy]phenyl}propoxy)benzoic acid] in comparison with the CysLT1R antagonist montelukast. In primary neurons, neither the nonselective agonist leukotriene D4 (LTD4) nor the CysLT2R agonist N-methyl-leukotriene C4 (NMLTC4) induced neuronal injury, and HAMI 3379 did not affect OGD/R-induced neuronal injury. However, in addition to OGD/R, LTD4 and NMLTC4 induced cell injury and neuronal loss in mixed cultures of cortical cells, and neuronal loss and necrosis in neuron-microglial cocultures. Moreover, they induced phagocytosis and cytokine release (interleukin-1β and tumor necrosis factor-α) from primary microglia, and conditioned medium from the treated microglia induced neuronal necrosis. HAMI 3379 inhibited all of these responses, and its effects were the same as those of CysLT2R interference by CysLT2R short hairpin RNA, indicating CysLT2R dependence. In comparison, montelukast moderately inhibited OGD/R-induced primary neuronal injury and most OGD/R- and LTD4-induced (but not NMLTC4-induced) responses in mixed cultures, cocultures, and microglia. The effects of montelukast were both dependent and independent of CysLT1Rs because interference by CysLT1R small interfering RNA had limited effects on neuronal injury in neuron-microglial cocultures and on cytokine release from microglia. Our findings indicated that HAMI 3379 effectively blocked CysLT2R-mediated microglial activation, thereby indirectly attenuating ischemic neuronal injury. Therefore, CysLT2R antagonists may represent a new type of therapeutic agent in the treatment of ischemic stroke.
Introduction
Cysteinyl leukotrienes (CysLTs), namely leukotriene C4 (LTC4), leukotriene D4 (LTD4), and leukotriene E4, are arachidonic acid–derived lipid mediators (Back et al., 2011). CysLTs act on two G-protein-coupled receptors, cysteinyl leukotriene receptor 1 (CysLT1R) and cysteinyl leukotriene receptor 2 (CysLT2R), and play important regulatory roles (Rovati and Capra, 2007; Singh et al., 2010; Back et al., 2011). In the periphery, CysLT1Rs are involved in various inflammatory diseases such as bronchial asthma and allergic rhinitis (Rovati and Capra, 2007; Singh et al., 2010; Back et al., 2011), and CysLT2Rs increase vascular permeability and aggravate myocardial ischemia/reperfusion injury (Sarau et al., 1999; Heise et al., 2000; Takasaki et al., 2000; Hui et al., 2001; Kamohara et al., 2001; Moos et al., 2008). In the central nervous system, the production of CysLTs increases after ischemic injury in rat brain (Zhou et al., 2006), primary neurons (Ge et al., 2006), and astrocytes (Huang et al., 2008). Increased CysLTs induce central nervous system responses by activating their receptors. It has been reported that the expression of CysLT1Rs and CysLT2Rs is upregulated in the brain after focal cerebral ischemia. These upregulated receptors are distributed in injured neurons in the acute phase (∼24 hours) and in activated microglia and proliferating astrocytes in the late phases (3–28 days) (Fang et al., 2006, 2007; Zhao et al., 2011). These findings suggest that CysLT1Rs and CysLT2Rs mediate acute ischemic neuronal injury and sequential microgliosis and astrocytosis in vivo.
However, whether ischemic neuronal injury in vitro is mediated by activation of CysLT1Rs or CysLT2Rs is poorly understood. In primary neurons, oxygen-glucose deprivation (OGD) induces injury but the agonist LTD4 does not induce ischemia-like neuronal injury (Hu et al., 2007). This suggests that CysLTs may induce ischemic neuronal injury by complex intercellular interactions, especially those between neurons and glia. Astrocytes protect neurons against ischemic or other injury (Barreto et al., 2011b; Terashvili et al., 2012) and are regulated by CysLTs after OGD-induced ischemic injury in different ways. CysLT1Rs mediate astrocyte proliferation after mild ischemia (1-hour OGD), whereas CysLT2Rs mediate astrocyte death after more severe ischemia (4-hour OGD) (Huang et al., 2008), and this may be associated with ischemic neuronal injury. However, whether CysLTRs regulate microglial activation, and thereby mediate ischemic neuronal injury, is not clear. Although neuron-like PC12 cells transfected with CysLT1Rs and CysLT2Rs show distinct sensitivities to ischemic injury (Sheng et al., 2006), it is necessary to investigate whether and how these receptors regulate ischemic neuronal injury.
Pharmacologically, the CysLT1R antagonists pranlukast and montelukast protect against focal and global cerebral ischemia (Zhang et al., 2002; Zhang and Wei, 2003; Yu et al., 2005a,b; Chu et al., 2006; Fang et al., 2006), N-methyl-d-aspartate-induced brain injury (Zhang and Wei, 2005), and brain cold injury (Qian et al., 2006). They attenuate neuronal injury, blood-brain barrier disruption, inflammatory responses, chronic brain injury, and the associated glial scar formation (Zhang et al., 2002; Yu et al., 2005a,b; Chu et al., 2006). However, montelukast has no effect on the OGD-induced reduction in neuronal viability (Hu et al., 2007) and a merely moderate effect on the morphologic neuronal changes after OGD (Wang et al., 2012). On the other hand, because of the previous lack of selective CysLT2R antagonists (other than the nonselective antagonist Bay U9773), their effects on ischemic neuronal injury are unknown. Recently, Bay CysLT2 and HAMI 3379 have been reported to be selective CysLT2R antagonists (Wunder et al., 2010; Ni et al., 2011). Bay CysLT2 protects mice from myocardial infarction and inhibits LTD4-induced Evans Blue leakage in mouse ear vasculature (Ni et al., 2011); it also protects astrocytes from in vitro ischemic injury (Huang et al., 2008). HAMI 3379, 3-({[(1S,3S)-3-carboxycyclohexyl]amino}carbonyl)-4-(3-{4-[4-(cyclo-hexyloxy)butoxy]phenyl}propoxy)benzoic acid, is devoid of CysLT receptor agonism and shows >10,000-fold affinity for CysLT2Rs versus CysLT1Rs (Wunder et al., 2010). We have reported that intracerebroventricular injection of HAMI 3379 protects rats from acute brain injury after focal cerebral ischemia (Shi et al., 2012), but its effects on ischemic neuronal injury in vitro need investigation.
Therefore, here we clarified the roles of CysLT1Rs and CysLT2Rs in ischemic neuronal injury in vitro and determined the effects of HAMI 3379 and the related cellular mechanisms in comparison with the CysLT1R antagonist montelukast. To assess whether oxygen-glucose deprivation/recovery (OGD/R) and CysLTs affect neurons directly or indirectly, the experiments were performed in primary neurons, mixed cortical cells, and neuron-glial cocultures as well as in primary microglia.
Materials and Methods
Primary Cultures of Cortical Neurons, Astrocytes, and Microglia and Mixed Cultures of Cortical Cells.
Primary cortical neurons were isolated from neonatal Sprague-Dawley rats of either sex within 24 hours of birth (Laboratory Animal Center of Zhejiang University, Hangzhou, China) as described previously with modifications (Meloni et al., 2001). In brief, after quick decollation using mechanical shearing, the cerebral cortices were dissected and digested with 0.25% trypsin (Sangon, Shanghai, China) for 10 minutes at 37°C. The dissociated cells then were immediately seeded onto 96-, 24-, or 6-well plates (Falcon, Franklin Lakes, NJ) coated with poly-l-lysine (0.1 mg/ml) (Sigma-Aldrich, St. Louis, MO) at 105, 5 × 105, or 2 × 106 cells per well. Cells were cultured in plating medium [high-glucose Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Temecula, CA) supplemented with 10% fetal bovine serum (FBS), 10% horse serum, 2 mM glutamine,100 units/ml penicillin, and 100 μg/ml streptomycin (Sigma-Aldrich)] for 24 hours. Thereafter, the plating medium was exchanged for feeding medium [high-glucose DMEM supplemented with 5% horse serum, 2 mM glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin, 0.01% N2 (Gibco), and 0.04% B27 (Gibco)]. The day of plating was counted as 0 days in vitro (DIV). On DIV 3, cytosine arabinoside (10 μM) (Sigma-Aldrich) was added for 24 hours to prevent the proliferation of nonneuronal cells. Cultures were maintained at 37°C in a humidified atmosphere (5% CO2 and 95% air), and the feeding medium was renewed every 2–3 days. On DIV 10, ∼95% of the cultured cells were neurons (data not shown) as identified by immunofluorescence staining with rabbit monoclonal antibody against microtubule-associated protein 2 (MAP2) (1:200) (Chemicon, Temecula, CA). All neuronal experiments were performed at DIV 10.
Primary cortical astrocytes were prepared as described previously with modifications (Meloni et al., 2001). In brief, the cerebral cortices were digested with 0.25% trypsin for 15 minutes at 37°C. The dissociated cells then were immediately seeded onto 150-cm2 flasks (one brain per flask) coated with poly-l-lysine (0.1 mg/ml). Cells were cultured in high-glucose DMEM supplemented with 10% FBS, 2 mM glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin. Cultures were maintained at 37°C in a humidified atmosphere (5% CO2 and 95% air), and the medium was renewed every 3 days. On DIV 14, the confluent cultures were shaken on an orbital shaker at 250 rpm for 12–16 hours. The adherent cells then were trypsinized and replated onto poly-l-lysine-coated 24-well plates at 3 × 105 cells/ml. More than 95% of the cultured cells were astrocytes (data not shown), as identified by immunostaining with the astrocyte marker mouse monoclonal anti–glial fibrillary acidic protein (GFAP) antibody (1:600) (Chemicon).
Primary microglial cells were prepared as described previously (Candelario-Jalil et al., 2007; Ni and Aschner, 2010). In brief, the cerebral cortices were digested with 0.25% trypsin for 10 minutes at 37°C. The dissociated cells then were plated in minimum Eagle’s medium containing 10% FBS onto 150-cm2 flasks (one brain per flask) pretreated with poly-l-lysine (0.1 mg/ml). Cultures were grown for 7–9 days at 37°C, and the microglial cells were removed by shaking on an orbital shaker at 250 rpm for 30 minutes. Cells were then centrifuged for 5 minutes at 1500g and replated onto poly-l-lysine-coated 24-well plates at 3 × 105 cells/mm2. The cultures contained >95% microglial cells, as determined by immunostaining with the microglial marker rabbit polyclonal anti–ionized calcium binding adaptor molecule 1 (Iba-1) antibody (1:1000) (Wako, Osaka, Japan).
The procedures for the mixed cultures of cortical cells were the same as for primary neuron culture except for the addition of cytosine arabinoside (10 μM). On DIV 10, ∼15% of the cultured cells were neurons, ∼70% were astrocytes, and ∼15% were microglia (data not shown), as identified by immunofluorescence staining with rabbit monoclonal anti-MAP2 (1:200), mouse monoclonal anti-GFAP (1:600), and rabbit polyclonal anti-Iba1 (1:1000) antibodies.
All the animal experiments were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The experimental protocols were approved by the Ethics Committee of Laboratory Animal Care and Welfare, School of Medicine, Zhejiang University.
Transwell Cocultures.
The primary neurons (DIV 10) plated on coverslips were transferred onto the insert of a 12-well Transwell system (0.4-μm-pore-size polycarbonate membrane coated with poly-l-lysine) (Corning Incorporated Life Science, Lowell, MA). Harvested primary microglia or astrocytes were plated on coverslips and transferred into the lower compartment of the Transwell. The Transwell cultures then were used in experiments.
RNA Interference of CysLT1R and CysLT2R.
Small interfering RNA (siRNA) duplexes of 21 nucleotides specific for the rat CysLT1R sequence were chemically synthesized, together with a nonsilencing negative control (NC) siRNA. The rat CysLT1R siRNA sense sequence was 5′-CAG CCU UCC AAG UAU ACA UTT-3′; the NC siRNA sense sequence was 5′-CCU ACG CCA CCA AUU UCG UTT-3′ (Genechem Co., Shanghai, China). Transfection of siRNA duplexes was performed according to the instructions of the manufacturer. In brief, the medium was renewed with an appropriate medium with 10% FBS but without antibiotics 24 hours before transfection. Then CysLT1R siRNA or NC siRNA was transiently transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) at a final concentration of 100 nM for 6 hours, and cells were incubated for 48 hours before exposure to OGD, LTD4, and N-methyl-leukotriene C4 (NMLTC4). CysLT1R siRNA reduced CysLT1R mRNA expression by 65.66% and protein expression by 62.77%, as assessed by reverse-transcription polymerase chain reaction and Western blot analysis (Supplemental Fig. 1, A and C).
The rat CysLT2R short hairpin RNA (shRNA) sense sequence was 5′-GAT CCC CCC GTC AAC ATG TAT ACT AGC ATT TTC AAG AGA AAT GCT AGT ATA CAT GTT GAC TTT TTG GA A C-3′; the NC shRNA sense sequence was 5′-GAT CCC CCC TTC TCC GAA CGT GTC ACG TTT CAA GAG ATT CTC CGA ACG TGT CAC GTT TTT TGG AAC-3′. The cDNA of double-stranded shRNA oligo was cloned into pFU-GW-RNAi-GFP lentivirus vector (Genechem Co.) using the HpaI and Xho restriction enzymes. Cells were infected 72 hours before exposure to OGD, LTD4, and NMLTC4. CysLT2R shRNA reduced CysLT2R mRNA expression by 43.84% and protein expression by 50.71%, as assessed by reverse-transcription polymerase chain reaction and Western blot analysis (Supplemental Fig. 1, B and D).
OGD and Pharmacological Treatments.
In vitro ischemia was induced by OGD as described previously (Goldberg and Choi, 1993; Song et al., 2004; Ge et al., 2006, 2007; Hu et al., 2006). In brief, cells were rinsed twice and incubated in Earle’s solution without glucose (concentrations in mM: NaCl, 117; KCl, 5.3; CaCl2, 1.8; NaHCO3, 26; MgSO4, 0.8; NaH2PO4, 1.0). The cells then were moved into an anaerobic chamber filled with 95% N2 and 5% CO2 at 37°C for 1, 2, or 4 hours. This procedure decreased the pO2 in the medium from 151.8 ± 6.9 (mean ± S.D., n = 5, control) to 26.7 ± 4.5 (n = 5, 1-hour OGD), 22.3 ± 6.7 (n = 4, 2-hour OGD), and 23.8 ± 7.4 mm Hg (n = 4, 4-hour OGD). The cells were then fed with high-glucose DMEM supplemented with 1% horse serum and returned to the normal incubator for recovery (24, 48, or 72 hours).
The nonselective CysLT1R/CysLT2R agonist LTD4 (Sigma-Aldrich) and the selective CysLT2R agonist NMLTC4 (Cayman Chemical Co., Ann Arbor, MI) (Yan et al., 2011) were used to activate the receptors. LTD4 and NMLTC4 were added into the culture medium at final concentrations of 0.1–1000 nM for 24, 48, or 72 hours. The selective CysLT1R antagonist montelukast (0.0001–1 μM) (Merck Pharmaceutical Co., Wilmington, DE) and the selective CysLT2R antagonist HAMI 3379 (0.0001–1 μM) (Cayman) were continuously applied from 30 minutes before exposure (OGD, LTD4, and NMLTC4) to the end of experiments. Control cells received the same treatment, except for the exposure to stimuli.
In the primary cultures of microglia, after exposure to OGD/R, LTD4, and NMLTC4 for 48 hours, the medium was collected and centrifuged at 2000 rpm for 10 minutes to remove cells and debris. The medium (conditioned medium) was immediately applied to neuronal cultures to induce necrosis.
Assessments of Cell Viability and Death.
At the end of experiments, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (Sigma-Aldrich) was added to each well to reach a final concentration of 0.5 mg/ml for 4 hours at 37°C. The medium then was removed, and 100 μl dimethyl sulfoxide (Sangon) was added. After 5-minute incubation, the product formazan was measured at 570 nm (OD570) in a Microplate Recorder (Elx800; Bio-TEK Instruments Inc., Winooski, VT). To detect cell death, the lactate dehydrogenase (LDH) activity in the medium was determined according to the protocols of the LDH kit (Jiancheng Bioengineering Institute, Nanjing, China) (Hui et al., 2004). In brief, an aliquot of the medium was mixed with NAD and lactate solution, and the LDH product was measured at 450 nm (OD450) in the microplate recorder. One unit represents an increase of 1 μmol of pyruvate catalyzed by LDH in 100 ml of recovery medium at 37°C for 15 minutes. Results are reported as percentages of control.
To detect apoptosis and necrosis, cells grown on coverslips were stained with 10 μg/ml Hoechst dye 33342 (Sigma-Aldrich) and 10 μg/ml propidium iodide (PI) (Sigma-Aldrich) for 10 minutes at 37°C. After repeated washing, the cells were fixed in cold methanol (−20°C). The cells then were photographed under a fluorescence microscope (Olympus BX51; Olympus, Tokyo, Japan). The apoptotic cells were determined as condensed or fragmented nuclei with strong bright Hoechst dye 33342 staining, and the necrotic cells as condensed nuclei with red PI staining. Apoptotic or necrotic cells were counted by a researcher who was blind to the treatments and reported as percentages of total cells.
Immunocytochemical Examination.
Neurons were immunostained with rabbit monoclonal anti-MAP2 antibody (1:200) (Chemicon), astrocytes were immunostained with mouse monoclonal anti-GFAP antibody (1:600) (Chemicon), and microglia were stained with rabbit polyclonal anti-Iba-1 antibody (1:1000) (Wako). Goat anti-rabbit Cy3 (1:200) (Chemicon) and goat anti-mouse fluorescein isothiocyanate (1:200) (Chemicon) antibodies were used as secondary antibodies. Cells cultured on coverslips were fixed in ice-cold methanol (−20°C) for 5 minutes and incubated in phosphate-buffered saline (PBS) containing 10% normal goat serum for 2 hours to block nonspecific binding of IgG. The cells then were incubated at 4°C overnight with the primary antibodies; after washing three times, the cells were incubated with the secondary antibodies for 2 hours at room temperature. Finally, the stained cells were observed by fluorescence microscopy (Olympus BX51). Control slips were treated with normal goat serum instead of the primary antibody, and no positive immunostaining was detected (data not shown).
Microglial Phagocytosis.
To determine the phagocytic activity of microglia, fluorescent microspheres (1-μm-diameter fluorescent carboxylate-modified microspheres, Millipore F8819; Invitrogen) were added for 1 hour at the end of experiments. Then microglia were harvested and washed with PBS, resuspended in 0.5 ml PBS, and transferred into cytometric tubes. The fluorescence was analyzed on the FL-3 channel of a flow cytometer (BD Biosciences, San Jose, CA).
Cytokine Measurement.
The medium was collected from microglial cultures after 6-hour recovery following OGD or 6-hour exposure to LTD4 and NMLTC4 and centrifuged at 2000 rpm for 10 minutes to remove cells and debris. The cytokines interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α) in the medium were measured using enzyme-linked immunosorbent assay kits for IL-1β (Wuhan Boster Biologic Technology Co., Ltd., Wuhan, China) and TNF-α (R&D Systems, Minneapolis, MN) according to the manufacturers’ instructions.
Statistical Analysis.
Data are reported as mean ± S.E.M. The significance of differences was analyzed by one-way analysis of variance (ANOVA) followed by Dunnett’s multiple comparison test (SPSS 10.0 for Windows, 1999; SPSS Inc., Chicago, IL). A value of P < 0.05 was considered statistically significant.
Results
OGD/R, but Not CysLT Receptor Agonists, Induces Injury in Primary Neurons.
We first determined whether the agonists of CysLT receptors have a directly injurious effect on neurons, similar to OGD/R-induced ischemia-like injury. As confirmation of ischemic neuronal injury, we found that OGD for 1, 2, and 4 hours decreased neuronal viability by 18, 30, and 50% and increased LDH release by 50, 90, and 169%, respectively (Fig. 1A). These time-dependent changes were gradually aggravated after 24, 48, and 72 hours of recovery (Fig. 1A). An optimal OGD/R condition was confirmed as 1-hour OGD followed by 24-hour recovery, in which neuron viability decreased by ∼20% and LDH release increased by ∼70%. Thus, the next experiments were performed under this condition. Because the production of CysLTs (LTC4, LTD4, and leukotriene E4) is elevated after OGD/R in both cultured neurons and astrocytes (Ge et al., 2006, 2007; Huang et al., 2008), we next determined whether CysLTs induce OGD/R-like neuronal injury by activating their receptors. The results showed that neither the nonselective CysLT1R/CysLT2R agonist LTD4 nor the selective CysLT2R agonist NMLTC4 (0.1–1000 nM) affected neuronal viability and LDH release (Fig. 1, B and C).
To elucidate the involvement of CysLTRs in ischemic neuronal injury, we further determined whether the antagonists and RNA interference of CysLT1R and CysLT2R affect OGD/R-induced neuronal injury. The results showed that the CysLT1R antagonist montelukast (0.001–1 μM) attenuated the reduction of OGD/R-induced neuronal viability and the increased LDH release (Fig. 2A), but CysLT1R siRNA, the CysLT2R antagonist HAMI 3379, and CysLT2R shRNA had no such effect (Fig. 2, A and B). OGD/R-induced reduction in the number of MAP2-positive neurons (Fig. 2C) and the increase in necrotic neurons (Fig. 2D) were partly reversed by montelukast (0.01 μM) but not by CysLT1R siRNA, HAMI 3379 (0.01 μM), or CysLT2R shRNA. OGD/R induced little neuronal apoptosis (Hoechst dye 33342 staining, <10%), and none of the treatments affected apoptosis (data not shown).
These results suggested that neither CysLT1Rs nor CysLT2Rs might induce neuronal injury directly nor mediate ischemic neuronal injury. HAMI 3379 did not directly protect neurons from ischemic injury. Montelukast might inhibit the ischemic injury through receptor-independent mechanisms because CysLT1R siRNA had no inhibitory effect.
OGD/R- and Agonist-Induced Neuronal Injury in Mixed Cultures of Cortical Cells.
To confirm the roles of CysLT1Rs and CysLT2Rs in animal experiments in vivo, we investigated neuronal injury in mixed cultures of cortical cells, a cellular environment mimicking the intact brain. We found that 1-, 2-, and 4-hour OGD increased LDH release by 40, 70, and 80%, respectively; this increase was aggravated after 24, 48, and 72 hours of recovery (Fig. 3A, upper panel). On the other hand, the agonists LTD4 (0.1–1000 nM) (Fig. 3B, upper panel) and NMLTC4 (0.1–1000 nM) (Fig. 3C, upper panel) concentration- and time-dependently reduced cell viability and increased LDH release. The moderate exposures were used as the optimal conditions in the next experiments, i.e., 1-hour OGD and 24-hour recovery and 48-hour exposure to 100 nM LTD4 or NMLTC4. In addition, the time- and concentration-dependent pattern of cell viability reduction was the same as that of the increase in LDH release (data not shown). Montelukast (0.01–1 μM), but not CysLT1R siRNA, attenuated the OGD/R-increased LDH release (Fig. 3A, middle panel), whereas both of them slightly inhibited the LTD4-increased LDH release (Fig. 3B, middle panel) but not the NMLTC4-increased release (Fig. 3C, middle panel). HAMI 3379 (0.0001–1 μM, except 0.0001 μM for OGD/R) and CysLT2R shRNA inhibited the increased LDH release by all three stimuli (Fig. 3, lower panels). The effects of these treatments on cell viability reduction were the same as those on increased LDH release (data not shown).
To characterize the cellular damage, we examined the cell morphology by immunostaining the mixed cultures. After exposure to OGD/R, LTD4 (100 nM), and NMLTC4 (100 nM), the number of MAP2-positive neurons was reduced (Fig. 4A). The number of GFAP-positive astrocytes was slightly increased (but reduced by NMLTC4), and the microglia had an activated appearance (from ramified to round and ameboid) (Supplemental Fig. 2). The reduction in neuron number induced by OGD/R and LTD4 was slightly inhibited by montelukast (0.01–1 μM for OGD/R; 0.001–1 μM for LTD4), HAMI 3379 (0.001–1 μM for OGD/R; 0.01–1 μM for LTD4), and CysLT2R shRNA but not by CysLT1R siRNA (Fig. 4, B and C). The NMLTC4-induced reduction was remarkably inhibited by HAMI 3379 (0.01–1 μM) and CysLT2R shRNA but not by montelukast and CysLT1R siRNA (Fig. 4C).
These results suggested that CysLT2R might mediate ischemic injury in the mixed cultures, especially the neuronal injury, and this injury might result from interactions between neurons and glia. HAMI 3379 showed a remarkable protective effect on the ischemic injury. Montelukast might exert the inhibitory effect on neuronal loss and necrosis via receptor-independent mechanisms, because of the lack of support from the CysLT1R knockdown results.
OGD/R- and Agonist-Induced Neuronal Injury in Neuron-Glial Transwell Cocultures.
To clarify the interactions between neurons and glia, we performed neuron-microglial and neuron-astrocyte Transwell cocultures. In neuron-microglial cocultures, we found that OGD/R, LTD4 (100 nM), and NMLTC4 (100 nM) reduced the numbers of neurons (Fig. 5A) from 860 ± 32 (control) to 700 ± 35 cells/mm2 (18.6%, n = 10, P < 0.05, OGD/R), 918 ± 29 to 745 ± 31 cells/mm2 (18.8%, n = 10, P < 0.05, LTD4), and 889 ± 33 to 657 ± 45 cells/mm2 (26.1%, n = 10, P < 0.05, NMLTC4). Montelukast (0.1 and 1 μM) and CysLT1R siRNA inhibited the neuron number reduction induced by LTD4 (but not NMLTC4); montelukast also showed a nonsignificant trend to inhibit the OGD/R-induced reduction (Fig. 5B). However, HAMI 3379 (0.01–1 μM) and CysLT2R shRNA significantly inhibited the neuron number reduction induced by all three stimuli (Fig. 5C).
PI/Hoechst dye 33342 staining showed that OGD/R, LTD4 (100 nM), and NMLTC4 (100 nM) induced little apoptosis (data not shown) but remarkable necrosis in the cocultured neurons (Fig. 6A). Necrotic neurons increased from 4.78 ± 0.89% (control) to 17.82 ± 0.78% (n = 10, P < 0.001, OGD/R), 5.10 ± 0.45% to 16.52 ± 0.98% (n = 10, P < 0.001, LTD4), and 5.60 ± 0.35% to 20.53 ± 0.42% (n = 10, P < 0.001, NMLTC4). Montelukast (0.1 and 1 μM) inhibited the OGD/R- and LTD4-induced neuronal necrosis, and CysLT1R siRNA inhibited the LTD4-induced necrosis, but neither affected the NMLTC4-induced necrosis (Fig. 6B). HAMI 3379 (0.01–1 μM) and CysLT2R shRNA significantly inhibited the necrosis induced by all three stimuli (Fig. 6C).
In neuron-astrocyte cocultures, we found that OGD/R, LTD4 (100 nM), and NMLTC4 (100 nM) reduced the number of neurons from 920 ± 23 (control) to 760 ± 25 cells/mm2 (17.4%, n = 10, P < 0.05, OGD/R), 874 ± 24 to 764 ± 31 cells/mm2 (12.6%, n = 10, P < 0.05, LTD4), and 859 ± 31 to 723 ± 27 cells/mm2 (15.8%, n = 10, P < 0.05, NMLTC4). These reductions were not inhibited by montelukast or HAMI 3379 (0.0001–1 μM) or by CysLT1R siRNA and CysLT2R shRNA (data not shown). OGD/R, LTD4, and NMLTC4 also induced neuronal necrosis, which was not significantly inhibited by montelukast, HAMI 3379, CysLT1R siRNA, or CysLT2R shRNA (Supplemental Fig. 3).
These results suggested that CysLT2R might be a determinant in ischemic neuronal injury in mixed cultures and cocultures, and this injury can be inhibited by HAMI 3379. CysLT1R and montelukast also somewhat affected the neuronal injury. Microglia might be the target cells of CysLT2R-regulated ischemic neuronal injury, whereas the roles of astrocytes in CysLT1R- and CysLT2R-mediated neuronal injury remain unclear.
OGD/R- and Agonist-Induced Phagocytosis and Cytokine Release in Primary Microglia, and Effect of Conditioned Microglial Medium on Neuronal Necrosis.
To elucidate the regulation of microglial activation by CysLT1R and CysLT2R, we assessed phagocytosis and proinflammatory cytokine release as indicators of microglial activation. On microscopic examination, we found that OGD/R, LTD4, and NMLTC4 significantly induced microglial phagocytosis (Supplemental Fig. 4A). OGD/R increased the phagocytosis by 125%, and this increase was significantly reduced by montelukast (0.1 and 1 μM), HAMI 3379 (0.01–1 μM), and CysLT2R shRNA but not by CysLT1R siRNA (Supplemental Fig. 4, B and C, left panels). LTD4 (100 nM) increased the phagocytosis by 73% and NMLTC4 (100 nM) by 93%. These increases were significantly reduced by HAMI 3379 (0.01–1 μM for LTD4 and 0.0001–1 μM for NMLTC4) and CysLT2R shRNA but not by montelukast or CysLT1R siRNA (Supplemental Fig. 4, B and C, middle and right panels). Flow cytometry showed that OGD/R, LTD4 (100 nM), and NMLTC4 (100 nM) enhanced microglial phagocytosis (Fig. 7A). Montelukast (0.01 μM) inhibited the OGD/R-induced phagocytosis, and HAMI 3379 and CysLT2R shRNA inhibited the phagocytosis induced by all three stimuli, but CysLT1R siRNA had no effect (Fig. 7B).
Moreover, OGD/R, LTD4, and NMLTC4 increased the release of TNF-α by 379, 353, and 355% and that of IL-1β by 239, 208, and 236%, respectively. The OGD/R- and LTD4-induced TNF-α and IL-1β increases were significantly inhibited by montelukast (0.01 μM), HAMI 3379 (0.01 μM), CysLT1R siRNA, and CysLT2R shRNA (Fig. 8, A and B, upper and middle panels). The NMLTC4-induced TNF-α and IL-1β increases were significantly inhibited by HAMI 3379 and CysLT2R shRNA but not by CysLT1R siRNA; only the NMLTC4-increased TNF-α release was slightly inhibited by montelukast (Fig. 8, A and B, lower panels).
Finally, we measured the necrosis in primary neurons induced by the conditioned medium from microglia that had been pretreated with OGD/R, LTD4, and NMLTC4. The results showed that, after these pretreatments, the conditioned medium induced neuronal necrosis, which was significantly inhibited by HAMI 3379 and CysLT2R shRNA (Fig. 8C). However, montelukast showed an inhibitory trend but no significant effect on OGD/R- and LTD4-induced necrosis, and CysLT1R siRNA had no effect (Fig. 8C).
These results indicated that HAMI 3379 inhibits CysLT2R-mediated microglial phagocytosis and cytokine release, which might cause neuronal injury. CysLT1Rs partially mediate the cytokine release, and montelukast might inhibit microglial activation through CysLT1R-independent and -dependent mechanisms.
Discussion
In the present study, we found that the selective CysLT2R antagonist HAMI 3379 attenuated OGD/R-induced ischemic neuronal injury, which was indirectly mediated by inhibiting microglial activation. Our findings also revealed the regulatory roles of CysLT2Rs in microglial activation and the resultant ischemic neuronal injury for the first time. In comparison, the CysLT1R-selective antagonist montelukast attenuated OGD/R-induced ischemic neuronal injury, possibly mediated by both receptor-independent and -dependent mechanisms.
As the target of HAMI 3379, the CysLT2R is important and complex in its mediation of ischemic neuronal injury. It seems not to be a direct determining factor for primary neuronal injury, although CysLT2R-transfected PC12 cells exhibit more severe ischemic injury after OGD (Sheng et al., 2006). We showed that neither the nonselective CysLT1R/CysLT2R agonist LTD4 nor the selective CysLT2R agonist NMLTC4 injured primary neurons. Moreover, OGD/R-induced ischemic neuronal injury was not attenuated by HAMI 3379 and CysLT1R or CysLT2R RNA interference. Therefore, the CysLT1Rs and CysLT2Rs may not directly mediate neuronal injury or ischemic neuronal injury. Rather, the CysLT2R indirectly regulates ischemic neuronal injury through intercellular interactions. The interactions are defined by the following findings. First, in mixed cultures of cortical cells, the agonists LTD4 and NMLTC4, similar to OGD/R, actually induced neuronal injury. Second, in neuron-microglial and neuron-astrocyte cocultures, neuronal injury was induced by LTD4 and NMLTC4. CysLT2R inhibition or blockade remarkably attenuated the neuronal injury in neuron-microglial cocultures (not in neuron-astrocyte cocultures), but CysLT1R siRNA did not have this effect. These findings suggest that CysLT2Rs may regulate neuronal injury via microglial activation. Third, neuronal injury was induced by the conditioned medium from microglia pretreated with OGD/R, LTD4, and NMLTC4, and this injury was inhibited by HAMI 3379 and CysLT2R shRNA. This finding demonstrated the role of microglial CysLT2Rs in the induction of neuronal injury.
Furthermore, in primary microglial cultures, we found that CysLT2R agonists induced microglial phagocytosis and cytokine release, and its antagonist HAMI 3379 and shRNA inhibited the responses of microglia to the agonists and OGD/R. In contrast, CysLT1Rs only regulated microglial cytokine release. Therefore, the role of CysLT2Rs in ischemic neuronal injury could be described as follows: the endogenously released CysLTs during ischemic injury, or exogenously administered agonists, activate microglial phagocytosis and cytokine release, thereby causing neuronal injury (Fig. 9).
Currently, the pharmacological effects of HAMI 3379 are poorly understood. It has been reported that HAMI 3379 concentration-dependently inhibits and reverses the increase in perfusion pressure and decrease in contractility induced by LTC4 in isolated Langendorff-perfused guinea pig heart (Wunder et al., 2010). Recently, we reported that intracerebroventricular injection of HAMI 3379 protects against acute brain injury after focal cerebral ischemia in rats (Shi et al., 2012). In the present study, we found that HAMI 3379, unlike montelukast, had no effect on OGD/R-induced neuronal injury in primary neurons. However, it inhibited OGD/R- or LTD4-induced injury similarly to, or 10- to 100-fold more potently than, montelukast in mixed cultures of cortical cells and inhibited all of the responses to NMLTC4. HAMI 3379 also inhibited neuronal injury in neuron-microglial cocultures more effectively than montelukast but had no effect in neuron-astrocyte cocultures, indicating that it mainly acts on microglial activation. In primarily cultured microglia, HAMI 3379 inhibited the phagocytotic activity and cytokine release induced by OGD/R, LTD4, and NMLTC4, and conditioned medium from the treated microglia-induced neuronal necrosis, whereas montelukast only inhibited the release of cytokines. These results strongly suggest that HAMI 3379 has more potent protective effects than montelukast on ischemic neuronal injury in the relatively intact cellular environment. Because HAMI 3379 had the same effects as CysLT2R knockdown by shRNA, its effects might be CysLT2R-dependent. However, HAMI 3379 could not completely reverse some of injurious responses, even those induced by NMLTC4, to control levels. This might result from insensitive responses mediated by non-CysLT receptors or unknown metabolites (see description below).
However, montelukast and CysLT1Rs are also partially involved in ischemic neuronal injury. CysLT1R knockdown by siRNA only attenuated a few responses, i.e., LTD4-induced injury in mixed cultures, LTD4-induced neuronal loss and necrosis in neuron-microglial cocultures, and OGD/R- and LTD4-induced cytokine release from microglia. Thus, CysLT1Rs may partially mediate microglial activation, possibly through regulating cytokine release, and thereby be involved in neuronal injury. However, montelukast had protective effects on multiple responses to ischemia-like injury and agonist stimulation. It inhibited OGD/R-induced injury in primary neurons, attenuated most of the OGD/R- and LTD4-induced changes in mixed cultures and neuron-glial cocultures, but had no effect on NMLTC4-induced changes (except the increased TNF-α release). Montelukast seems to have both CysLT1R-independent and -dependent effects; it blocked CysLT1R responses the same as CysLT1R siRNA and also inhibited the other responses to OGD/R and LTD4. Especially, it attenuated OGD/R injury in primary neurons, which differed from HAMI 3379.
It has been reported that montelukast has a broader spectrum of pharmacological effects than originally thought, including the primary effects of CysLT1R antagonism and the secondary effects, for example, of inhibition of cyclic nucleotide phosphodiesterases (PDEs) and nucleotide P2Y receptor signaling (Tintinger et al., 2010). Inhibition of PDEs by montelukast may be beneficial to ischemic neuronal injury, because the resultant accumulation of cAMP protects neurons from ischemic brain injury (Tsukada et al., 2004; Lin et al., 2009) and inhibitors of PDE3 (cilostazol) and PDE4 (rolipram) have protective effects on neurons (Tanaka et al., 2010; Schaal et al., 2012). In addition, its inhibitory effects on P2Y receptors (Mamedova et al., 2005; Pugliese et al., 2009; Lau et al., 2011) may be protective, because downregulation of the novel P2Y-like receptor GPR17 protects from ischemic neuronal injury after focal cerebral ischemia in rats (Ciana et al., 2006; Zhao et al., 2012). Moreover, montelukast has antioxidative effects in peripheral tissues (Muthuraman and Sood, 2010; Coskun et al., 2011; Mohamadin et al., 2011); in a preliminary study, we also found its moderate CysLT1R-independent antioxidative activity in primary neurons (unpublished data). This may be another beneficial effect of montelukast on ischemic neuronal injury because oxidative stress is a major cause of the ischemic injury (Perez Velazquez et al., 1997; Gan et al., 2012; Zhou et al., 2012).
Microglia are immunologically active residents in the brain and play dual roles under physiologic and pathologic conditions (Yenari et al., 2010; Saijo and Glass, 2011; Gomes-Leal, 2012). Activated microglia protect neurons against ischemic injury by phagocytosis of debris or dead cells and by releasing anti-inflammatory cytokines and neurotrophic factors (Denes et al., 2007; Imai et al., 2007; Lambertsen et al., 2009). However, overactivated microglia contribute to neuronal damage by releasing harmful substances, including inflammatory cytokines, reactive oxygen species, and proteinases (Dheen et al., 2007; Lv et al., 2011; Zhang et al., 2012). Our results revealed that CysLT-activated microglia mediated ischemic neuronal injury in our experimental conditions, and this might have resulted from releasing harmful cytokines, such as IL-1β and TNF-α (Amantea et al., 2010; Lambertsen et al., 2012). This response was mediated mainly by CysLT2Rs and somewhat by CysLT1Rs through regulation of cytokine release as well. However, the neuronal injury induced by conditioned medium from pretreated microglia was not significantly inhibited by montelukast and CysLT1R siRNA; this could be explained by the possibility that they also inhibit the CysLT1R-mediated release of protective substances (Denes et al., 2007; Imai et al., 2007; Lambertsen et al., 2009). Because microglia-mediated inflammation occurs in several brain diseases (Yenari et al., 2010; Saijo and Glass, 2011; Gomes-Leal, 2012), CysLT2R antagonists such as HAMI 3379 may represent a potential therapeutic agent for these diseases as well as ischemic stroke.
Unexpectedly, it is not yet clear whether CysLT1Rs or CysLT2Rs on astrocytes play regulatory roles in ischemic neuronal injury. Astrocytes in the mixed cultures only showed slight changes in number after exposure to the three stimuli. In neuron-astrocyte cocultures, OGD/R induced relatively severe neuronal necrosis, which might result from both direct ischemic injury and astrocyte-mediated injury. In general, astrocytes play protective roles in the ischemic injury of the neurons by releasing protective molecules like neurotrophic factors and removing toxic substances (Takano et al., 2009; Barreto et al., 2011a). On the other hand, proliferating or activated astrocytes may also be associated with neuronal injury (Katayama et al., 2010; Sullivan et al., 2010; Qu et al., 2011), which may explain our OGD/R results. In addition, LTD4 and NMLTC4 also induced milder neuronal necrosis (∼3%) in the neuron-astrocyte cocultures. Because neither agonist had effects on neurons, this effect might result from astrocytes. Astrocytes might mediate neuron injury via acting on their CysLT2Rs, because NMLTC4 is a selective CysLT2R agonist (Yan et al., 2011) and LTD4 at 100 nM mainly acts on CysLT2Rs in astrocytes (Huang et al., 2008). In spite of this, the mild neuronal necrosis induced by both agonists was not affected by the antagonists or RNA interference of CysLT2Rs as well as those of CysLT1Rs.
Our results show that the CysLT2R antagonist HAMI 3379 protects neurons from ischemic injury by inhibiting microglial activation, and the CysLT1R antagonist montelukast also has a broad but relatively weaker protective effect that may be both dependent on and independent of CysLT1Rs. These findings demonstrate that CysLT2Rs play an important role in microglial activation and the resultant neuronal injury, and CysLT1Rs are also involved in microglia-mediated responses through regulation of microglial cytokine release (Fig. 9). Therefore, CysLT2R antagonists may represent a new type of therapeutic agent in the treatment of diseases with neuronal injury such as ischemic stroke and may synergize the effects of CysLT1R antagonists. However, the mechanisms underlying the regulation of microglial activation by CysLT2Rs need further investigation.
Acknowledgments
The authors thank Dr. I.C. Bruce for critically reading and revising this manuscript.
Authorship Contributions
Participated in research design: Wei, X.-Y. Zhang.
Conducted experiments: X.-Y. Zhang, Wang, Yu, Xu, L.-H. Zhang, Chen, Fang, Lu, W.-P. Zhang.
Performed data analysis: X.-Y. Zhang, Wang, Xu, Yu, Shi.
Wrote or contributed to the writing of the manuscript: X.-Y. Zhang, Wei.
Footnotes
- Received January 31, 2013.
- Accepted June 6, 2013.
This study was supported by the National Natural Science Foundation of China (81273491, 81072618, and 81173041) and the Zhejiang Provincial Natural Science Foundation (LY12H31010).
↵This article has supplemental material available at jpet.aspetjournals.org.
Abbreviations
- CysLT
- cysteinyl leukotriene
- CysLT1R
- cysteinyl leukotriene receptor 1
- CysLT2R
- cysteinyl leukotriene receptor 2
- DIV
- days in vitro
- DMEM
- Dulbecco’s modified Eagle’s medium
- FBS
- fetal bovine serum
- GFAP
- glial fibrillary acidic protein
- HAMI 3379
- 3-({[(1S,3S)-3- carboxycyclohexyl]amino}carbonyl)-4-(3-{4-[4-(cyclo-hexyloxy)butoxy]phenyl}propoxy)benzoic acid
- Iba-1
- ionized calcium binding adaptor molecule 1
- IL-1β
- interleukin-1β
- LDH
- lactate dehydrogenase
- LTC4
- leukotriene C4
- LTD4
- leukotriene D4
- MAP2
- microtubule-associated protein 2
- NC
- negative control
- NMLTC4
- N-methyl-leukotriene C4
- OD
- optical density
- OGD
- oxygen-glucose deprivation
- OGD/R
- oxygen-glucose deprivation/recovery
- PBS
- phosphate-buffered saline
- PDE
- phosphodiesterase
- PI
- propidium iodide
- shRNA
- short hairpin RNA
- siRNA
- small interfering RNA
- TNF-α
- tumor necrosis factor-α
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics