Neuron-Microglia Interaction Induced Bi-directional Cytotoxicity Associated with Calpain Activation

Ethics Statement

Human fetal tissues were obtained following written approval from adult female patients undergoing therapeutic abortion at 10–14 weeks gestational age at the University of Washington, WA, USA (IRB approval #11449). The use of human fetal tissue was approved by the University of South Carolina, SC, USA (USCeRA#:HSA4636), and is IRB-exempt [45 CFR 46.102(d)]. All primary human fetal neuron cell culture studies were done according to university guidelines in a biocontainment facility approved by the Institutional Biosafety Committee (IBC) of the University of South Carolina.

PBMCs were collected from human subjects enrolled according to protocols approved by the Medical University of South Carolina Institutional Review Board (IRB #9481). Microglial cells were isolated from human adult brain tissues collected from the Medical University of South Carolina operating room during surgical procedures on non-affected/not-damaged brain according to protocols approved by the Medical University of South Carolina, SC, USA (IRB #19380). All experimental protocols were reviewed and approved by the Institutional Review Board of the Medical University of South Carolina, and experiments were undertaken with the understanding and written consent of each subject. The study conforms to the Code of Ethics of the World Medical Association (Declaration of Helsinki), printed in the British Medical Journal (18 July 1964).

Rat Cell Cultures

Rat cortical cultures (neurons and microglia) were established from two-day-old Spraque-Dawley pups from dams (Harlan) as previously reported (Banker & Goslin 1988, Jana et al. 2007). Briefly, cortices were dissected and dissociated by trituration and trypsinization. The dissociated cells were re-suspended in B-27-supplemented neurobasal media (Invitrogen) and plated onto a poly-D-lysine-coated T-flask. Neurons (adherent cells) were treated with 5 μM of the anti-mitotic inhibitor arabinose-C and cultured for 7 days prior to the experiment. Non-adherent cells contained mixed glial cells and were mechanically shaken for 30 min at 200 rpm on a gyratory shaker to isolate microglia (Bhat et al. 1992). The purity of rat cortical neurons and microglia cultures was confirmed by immunochemistry with 90–95% positive staining for neuron-specific nuclear protein (NeuN) and CD11b, respectively. Ventral spinal cord 4.1 (VSC4.1) motoneuron cell line, a generous gift from Dr. S. Appel (Baylor College of Medicine, TX, USA), was grown under previously described conditions (Das et al. 2005).

Primary Human Brain Cell Cultures

Primary human brains at 8 to 12 weeks gestational age obtained from the University of Washington, Seattle, were cultured as described earlier (Chauhan et al. 2003, Mehla & Chauhan 2015). Briefly, meninges and blood vessels were removed from the cortex. The tissue was washed twice with opti-MEM containing 1% penicillin-streptomycin and amphotericin-B (GibcoBRL). Washed tissue was mechanically disrupted by one passage through a 20-ml syringe without a needle. Disrupted cortex tissue was centrifuged at 1,200 rpm for 20 minutes at 4°C. For differentiation to neurons, pellets were resuspended in opti-MEM medium and cultured in Opti-MEM supplemented with 5% FCS, 0.2% N2 supplement (Gibco BRL), and antibiotics. The purity of neuron and astrocyte cultures was verified by immunostaining using antibody against microtubule-associated protein-2 (Sigma) or glial fibrillary acidic protein (GFAP), respectively. The purity of the neuron cultures were 90–95%. Primary human fetal neuron cultures were maintained for at least a month before use in experiments. To reduce the chance of mycoplasma contamination in these cultures, plasmocin, an anti-mycoplasma agent (Invitrogen), was used during the initial culture passages. Human primary neurons were isolated from 12- to 14-week-old fetal brains, as previously described (Chauhan et al. 2003, Mehla & Chauhan 2015). Microglial cells were isolated from human adult brain tissues collected from the Medical University of South Carolina operating room during surgical procedures on non-affected/not-damaged brain according to the previously published method (De Groot et al. 2000).

Isolation and Stimulation of PBMCs from MS Patients

PBMCs from MS patients were isolated as described previously (Smith et al. 2011). Cells were stimulated with 10 μg/ml of anti-CD3 and 5 μg/ml of anti-CD28 (BD, Franklin Lakes, NJ) following pre-treatment with the calpain inhibitor calpeptin (CP, 5 μM) in DMSO for 72 h. After treatment, supernatants were collected and stored at −80°C until use.

Treatment of Rat Primary Cortical Neurons

Primary rat neurons (RN) from cortex were stimulated with IFN-γ (100 units/ml) in the absence or presence of 1 μM CP. After 24 h of incubation, conditioned media (CM) as well as cell pellets were collected (Supplemental Fig. 1A) and analyzed (Fig. 1). CM-RN were collected, speed-vacuum concentrated (10×), and used for treatment of microglia as well as VSC4.1 motoneurons.

To determine if damaged neurons could activate microglia, primary rat microglia (RM) from cortex were pre-treated for one hour with CM-RN±CP followed by collection of CM and cells (Supplemental Fig. 1B) for subsequent analysis (Fig. 2). CM-RM were concentrated and used for further treatment.

Treatment of VSC4.1 Cells

VSC4.1 motoneurons were exposed to CM-RM, CM-RM±CP (1 h pre-treatment), CM-RM±SNJ-1945 (1 h pre-treatment), or CM-RM + siRNA for calpain 1 or CM-RM + siRNA for calpain 2 (1 h pre-treatment). After 24 hr of treatment, CM as well as cells were collected (Supplemental Fig. 1C) and analyzed (Fig. 3). The delivery of siRNA for calpain 1 (μcalpain) and calpain 2 (mcalpain) into VSC4.1 motoneurons was performed using Lipofectamine RNAiMAX (Invitrogen) following the manufacturer’s instructions, as described earlier (Das et al. 2013b). Cell transfection was carried out 4–6 h, and the medium was then replaced with normal growth medium. Control cells were mock-transfected, and all experimental results were confirmed using scrambled siRNA. After 24 h cells were exposed to CM-RM.

To determine whether conditioned medium (CM) from damaged neurons (CM-RN) could affect fresh naïve neuron viability and whether calpain inhibitor could prevent such damage, VSC4.1 motoneurons were treated with CM-RN±CP (1 h pre-treatment). Two types of negative controls were included: inactivation of IFN-γ by heating at 55°C for 30 min before primary rat culture stimulation and neutralization of IFN-γ by anti-IFN-γ antibody. VSC4.1 cells were either challenged with heat-inactivated IFN-γ (HI–IFN-γ) or with neutralizing antibody to IFN-γ in the absence or presence of 1 μM CP. CM and cells (Supplemental Fig. 1D) were collected and subjected to appropriate analysis (Fig. 4).

Treatment of Human Primary Cells

Fetal human neurons (HN) were exposed to 50 μl of PBMCs, PBMCs_Act, or PBMCs_Act plus 1 μM or 5 μM CP (PBMCs_Act+CP; 1 h pre-treatment) for 24 hr. After incubation, CM as well as cell pellets were collected (Supplemental Fig. 2A) and subjected to appropriate analysis (Fig. 5). PBMCs_Act were concentrated and used for human microglia (HM) treatment.

Primary HM isolated from undamaged cortex from unaffected brains of patients were treated with CM-HN_PBMCsAct±CP. The CM from incubated microglia as well as the cell pellet were collected (Supplemental Fig. 2B) a (Fig. 6). CM-HM were concentrated and used for fresh or naïve HN treatment.

To determine the effect of calpain upon microglia-mediated damage, fresh naive fetal HN were treated with CM-HM in the presence or absence of CP. CM and cells were collected after treatments (Supplemental Fig. 2C) and subjected to further analysis (Fig. 7). Following treatments, the mixture of attached and detached cell populations was used for estimation of the cell viability by the Trypan blue dye exclusion test (Das et al. 2004).

Calpain Activity Assay

Calpain activity in total cell lysates was determined based on the reaction with a calpain substrate, fluorogenic peptide (Ac-LLY-AFC), using a calpain activity assay kit (Abcam, Cambridge, MA). Briefly, cells were lysed then incubated with the substrate and reaction buffer for 1 h at 37°C in the dark. Proteolysis of the fluorescent probe (AFC) was monitored using a fluorimeter with filter settings of 400 nm for excitation and 505 nm for emission.

Measurement of Nitrite and ROS

The production of cytoactive molecules was measured as reported previously (Das et al. 2007, Davidge et al. 1995). Briefly, NO production was calculated from the amount of nitrite detected by Griess reaction. Absorbance was determined at 550 nm using a microplate reader. Net intracellular accumulation of ROS was detected using fluorescent probe 2′,7′-dichlorofluorescin diacetate (DCF-DA), and the fluorescence was measured at 530 nm after excitation at 480 nm.

Assay for Cyclooxygenase-2 Activity

COX-2 activity was determined using a COX-2 activity assay kit (Cayman Chemical, Ann Arbor, MI) according to the manufacturer’s protocol, as previously described (Sribnick et al. 2010). Briefly, cells (1×10⁶) were homogenized in cold buffer 0.1 M Tris-HCl, pH 7.8 containing 1 mM EDTA, mixed with assay buffer, heme, and COX-1 inhibitor. Colorimetric substrate and arachidonic acid were added sequentially and incubated at 37°C for 10 min, and the absorbance was read at 590 nm.

Colorimetric Assays for the Measurement of Caspase-8, Caspase-9, and Caspase-3 Activities

Measurements of caspase activities in cells were performed using commercially available caspase-8, caspase-9, and caspase-3 assay kits (Sigma), as previously reported (Das et al. 2006a). Briefly, activities were assayed based on hydrolysis of the Ac-IETD p-nitroaniline (pNA) by caspase-8, Ac-LEHD pNA by caspase-9, and Ac-DEVD pNA by caspase-3. The absorbance read at 405 nm. Caspase activities in μmol of pNA released from the substrate per min per ml of cell lysate were calculated based on the formula: OD × d/ε^mM × t × v, where d was dilution factor, t was reaction time, v was volume of sample and the constant value ε^mM=10.5. Data were expressed as % of change compared to control ± SEM, where control was set as 100%.

Statistical Analysis

Results of at least 3 different experiments were expressed as the percentage of control where control values were set to 100%. Significant differences between groups were determined by one-way analysis of variance ANOVA followed by Tukey’s multiple comparison test using GraphPad PRISM 5.0. Significant differences were indicated by *p<0.05, **p<0.01, or ***p<0.001.

IFN-γ Induced Cell Death in Primary Rat Neurons

Primary RN culture was used as an in vitro model to investigate the cytotoxic action of the Th1 pro-inflammatory cytokine IFN-γ, a potent autoimmune inducer known to be upregulated in MS. As expected, neurons were affected by IFN-γ treatment, as indicated by a 2.6 fold reduction of viability after 24 h of treatment (Fig. 1A) and 3-fold increased (p<0.001) calpain activity (Fig. 1B). The increased vulnerability of neurons to IFN-γ injury was almost completely abolished by the calpain inhibitor CP.

Next, the mechanisms involved in cell death were investigated, and whether calpain inhibition could block neuron injury was determined. Neurons exposed to IFN-γ demonstrated a significant increase in ROS production (3.6 fold, p<0.01) at 24 h post-exposure (Fig. 1C). Activation of caspase-3, considered a central regulator of apoptosis, consistently increased by 3.1 fold (p<0.01), confirming that cells were already committed to apoptosis (Fig. 1D). Treatment with CP attenuated both ROS generation and caspase-3 activity following Th1 cytokine insult (Fig. 1, C–D). This suggests that calpain acts to increase the vulnerability of neurons to IFN-γ by affecting both ROS generation and caspase-3 activity.

Injured Primary Rat Neurons Secrete Factors that Activated Rat Microglia In Vitro

Deleterious microglial activation may be triggered by mediators released from damaged neurons causing a self-propelling cycle, a mechanism for chronic neuronal loss proposed in diverse neurodegenerative diseases (Block & Hong 2005, Gao et al. 2003, Levesque et al. 2010). In line with this, CM-RN stimulated with IFN-γ was used to activate primary RM. Microglial cells were vulnerable to the toxic mediators that were released into the CM after neuronal damage, as indicated by reduced cell viability (2.9 fold, p<0.05, Fig. 2A). The significant increases were observed in a variety of pro-inflammatory and apoptotic mediators such as calpain activity (3.3 fold, p<0.01, Fig. 2B), ROS generation (2.3 fold, p<0.05, Fig. 2C), NO production (2.1 fold, p<0.05, Fig. 2D), COX-2 activity (3.2 fold, p<0.001, Fig. 2E), and caspase-3 activity (4.2 fold, p<0.001, Fig. 2F). Calpain inhibition suppressed the effects of these mediators with caspase 3 activity being considered the most significant (Fig. 2F). This suggests that secretory factors, especially calpain, from damaged RNs activated microglial cells in vitro.

Calpain Inhibition/Knockdown Prevented VSC 4.1 Cell Death Stimulated by Secreted Factors Released From Damaged Primary Neurons

In parallel studies using VSC4.1 motoneurons, treatment with CM-RM was shown to decrease cell viability (Fig. 3A) while calpain activity (Fig. 3B), ROS production (Fig. 3C), and caspase-3 activity (Fig. 3D) were significantly increased (p<0.001). Calpain inhibition by cell-permeable inhibitor CP or the novel water-soluble calpain inhibitor SNJ-1945 significantly suppressed these effects (p<0.001; Fig. 3, B–D).

A similar effect was observed when a specific siRNA-oligonucleotide for calpain 1 or calpain 2 was transfected into VSC4.1 cells as revealed by decreased calpain activity (Fig. 3B). The microglia-induced damage indicated by ROS production (Fig. 3C), and caspase 3-activity (Fig. 3D) was significantly reduced (p<0.001) whereas non-specific siRNA had no inhibitory effect (data not shown). This is indicative that inhibition of calpain secreted from damaged primary neurons may prevent cell death.

Secreted Factors Released From IFN-γ Damaged Primary Neurons Induced VSC 4.1 Motoneuron Cell Death

To investigate whether damaged neurons could affect fresh naïve neuron viability, VSC4.1 cells were treated with CM-RN±CP (1 h pre-treatment). CM-RN were incubated CP and SNJ for 1 h prior to incubation with VSC4.1 cells. In a parallel experiment, primary microglia were also challenged with heat inactivated IFN-γ or neutralizing antibody treated CM-RN±CP. Treatment of CM-RN with CP was done 1 h prior to incubation with VSC4.1. As expected, VSC 4.1 cells showed increased sensitivity to CM-RN (Fig. 4A). No significant change in viability was seen after treatment with CM-RN+CP or comparable sets of cells treated with heat inactivated and neutralizing antibody treated CM-RN±CP (Fig. 4A). As demonstrated, the CM-RN treatment not only decreased cell viability but also increased calpain activity (Fig. 4, B–C) and elevated oxidative stress response, as revealed by ROS production (Fig. 4D). CP treatment inhibited calpain activity (Fig. 4, panels B–C) as well as production of ROS (Fig. 4D), and this inhibitory effect was closely linked to IFN-γ activity being dramatically decreased after heat inactivation of IFN-γ (Fig. 4, B–D). Calpain activity and production of ROS were also evaluated after treatment of CM-RN with anti-IFN-γ neutralizing antibody (Fig. 4, B–D), and they were significantly decreased. Therefore, secretory factors from damaged primary neuron both increased the activity of cell death markers and decreased cell viability.

Activated PBMCs from MS Patients Secreted Factors that Induced Damage to Human Neurons in vitro and Calpain Inhibition Prevented Damage

The second model used activated PBMCs as insulting stimulus in human primary cultures. Previously, activated PBMCs were shown to be a source of auto-reactive T cells releasing Th1/Th17 pro-inflammatory cytokines (Imam et al. 2007, Smith et al. 2011). This model was chosen based on the general consensus supporting a role of T cells subsets and their contribution to the disease initiation and perpetuation in EAE and MS. First, the major player and the mechanism that contributed to IFN-γ–initiated activation of neuron-microglia associated with calpain-mediated bi-directional cytotoxicity in rat cultures were determined. Secondly, whether the same or a similar mechanism contributing to bi-directional cytotoxicity is elicited by activated PBMCs from MS patients was explored.

Neurons treated with PBMCs_Act showed reduced viability that was reversed by CP treatment (Fig. 5A). Calpain activity of neurons treated with PBMCs vs. PBMCs_Act was also monitored. As shown in Fig. 5B, calpain activity was increased by 86.77 ± 15.26% and 183.1 ± 27.89% in neurons treated with PBMCs and PBMCs_Act, respectively, suggesting that PBMCs were already secreting calpain even without CD3/CD28 in vitro activation. Calpain activity was decreased when HN were treated with PBMCs_Act +CP (2.1 fold, p<0.01, Fig. 5B).

Our results on HN indicate that treatment with PBMCs_Act significantly increased ROS production (2.8 fold, p<0.001) while CP treatment significantly decreased ROS generation (1.9 fold, p<0.001, Fig. 5C). As expected, COX-2 activity was greatly increased (6.3 fold, p<0.001), and CP treatment reduced COX-2 activity (3 fold, p<0.001) in exposed neurons (Fig. 5D).

Whether PBMCs_Act could stimulate caspase-8, an apoptotic initiator in the extrinsic pathway, was also determined. While a marked increase in caspase-8 activity occurred after 24 h of stimulation (5.7 fold, p<0.001), this effect was abolished by 1.6 fold following CP treatment (Fig. 5E). Our data demonstrate that the loss of HN viability was associated with the consequent activation of the intrinsic pathway, caspase-9 (6.2 fold, p<0.001, Fig. 5F), and the final executioner, caspase-3. Caspase-3 was increased by 86.8% and 136.3% in HN treated with PBMCs and PBMCs_Act, respectively (Fig. 5G). In contrast, treatment of HN with PBMCs_Act+CP decreased caspase-3 activity by 1.9 fold (p<0.001), as shown in Fig. 5G. Thus, treatment with activated PBMCs increased calpain activation and HN cell death via the intrinsic pathway in vitro.

Calpain Inhibition Prevented Human Microglia Damage in vitro

Whether toxic mediators released from damaged HN could cause microglia injury was also investigated. Our results indicate that treatment with CM-HN reduced microglia viability (p<0.001, Fig. 6A) and demonstrated increases in calpain activity (2.9 fold, p<0.01, Fig. 6B), ROS production (p<0.01, Fig. 6C), COX-2 activity (p<0.01, Fig. 6D), and caspase-3 activity (p<0.01, Fig. 6D). The increase in ROS generation was accompanied by a significant increase in NO production (1.8 fold, p<0.001, Fig. 6F). Furthermore, calpain inhibition suppressed the effects of these toxic mediators and restored microglial viability (Fig. 6, A–F), which suggests that calpain secreted by HNs induced microglial damage.

Activated Microglia Released Factors that Induced Apoptosis in Naïve Human Neurons: the Effect Reversed by Calpain Inhibition

To explore the role of activated human microglia on cell viability, the effect of CM-HM on neurons was examined. Treatment of HN with CM-HM produced significant injury, causing decreased cell viability to 61.5% (p<0.01, Fig. 7A). The neuronal toxicity of CM-HM was attenuated by pretreatment with CP (p<0.01, Fig. 7A). Our study using human primary cultures confirmed the role of calpain as a key signal molecule released from activated neurons, as increased extracellular calpain activity was observed upon exposure of naïve HN to CM-HM (3.1 fold, p<0.01, Fig. 7B). To further reinforce the role of calpain, released by activated microglial cells into CM, parallel neuronal cultures were co-incubated with CP. As shown in Fig. 7A, CP reduced percentage of apoptotic cells and neuronal viability was almost completely restored.

Decreased cell viability was also associated with a concomitant increase in ROS production (4.2 fold, p<0.01, Fig. 7C), suggesting the involvement of oxidative damage in this process. Consistent with this result, activity of caspase-3, a critical player in apoptosis, was also increased (3.1 fold, p<0.01, Fig. 7D). In contrast to these apoptotic features, pretreatment of HN with CP followed by treatment with CM-HM suppressed the effects of these neurotoxic mediators (Fig. 7, panels B–D). Therefore, calpain secretion by microglia induces apoptosis in HN and the viability of HN could be restored following treatment with CP.