Endothelial Japanese encephalitis virus infection enhances migration and adhesion of leukocytes to brain microvascular endothelia via MEK-dependent expression of ICAM1 and the CINC and RANTES chemokines

Authors

  • Ching-Yi Lai,

    1. Department of Life Sciences, National Chung Hsing University, Taichung, Taiwan
    2. Department of Education and Research, Taichung Veterans General Hospital, Taichung, Taiwan
    Search for more papers by this author
  • Yen-Chuan Ou,

    1. Division of Urology, Taichung Veterans General Hospital, Taichung, Taiwan
    Search for more papers by this author
  • Cheng-Yi Chang,

    1. Department of Surgery, Fong-Yuan Hospital, Taichung, Taiwan
    Search for more papers by this author
  • Hung-Chuan Pan,

    1. Department of Neurosurgery, Taichung Veterans General Hospital, Taichung, Taiwan
    Search for more papers by this author
  • Chen-Jung Chang,

    1. Department of Medical Imaging and Radiological Sciences, Central Taiwan University of Sciences and Technology, Taichung, Taiwan
    Search for more papers by this author
  • Su-Lan Liao,

    1. Department of Education and Research, Taichung Veterans General Hospital, Taichung, Taiwan
    Search for more papers by this author
  • Hong-Lin Su,

    Corresponding author
    • Department of Life Sciences, National Chung Hsing University, Taichung, Taiwan
    Search for more papers by this author
  • Chun-Jung Chen

    Corresponding author
    1. Department of Education and Research, Taichung Veterans General Hospital, Taichung, Taiwan
    2. Center for General Education, Tunghai University, Taichung, Taiwan
    3. Institute of Biomedical Sciences, National Chung Hsing University, Taichung, Taiwan
    4. Graduate School of Nursing, Hung-Kuang University, Taichung, Taiwan
    • Department of Life Sciences, National Chung Hsing University, Taichung, Taiwan
    Search for more papers by this author

Address correspondence and reprint requests to Hong-Lin Su, Department of Life Sciences, National Chung-Hsing University, No. 250, Kuo-Kuang Rd., Taichung 402, Taiwan. E-mail: suhonglin@ gmail.com and Chun-Jung Chen, Department of Education and Research, Taichung Veterans General Hospital, No. 160, Sec. 3, Taichung-Kang Rd., Taichung 407, Taiwan. E-mail: cjchen@vghtc.gov.tw

Abstract

Currently, the underlying mechanisms and the specific cell types associated with Japanese encephalitis-associated leukocyte trafficking are not understood. Brain microvascular endothelial cells represent a functional barrier and could play key roles in leukocyte central nervous system trafficking. We found that cultured brain microvascular endothelial cells were susceptible to Japanese encephalitis virus (JEV) infection with limited amplification. This type of JEV infection had negligible effects on cell viability and barrier integrity. Instead, JEV-infected endothelial cells attracted more leukocytes adhesion onto surfaces and the supernatants promoted chemotaxis of leukocytes. Infection with JEV was found to elicit the elevated production of intercellular adhesion molecule-1, cytokine-induced neutrophil chemoattractant-1, and regulated-upon-activation normal T-cell expressed and secreted, contributing to the aforementioned leukocyte adhesion and chemotaxis. We further demonstrated that extracellular signal-regulated kinase was a key upstream regulator which stimulated extensive endothelial gene induction by up-regulating cytosolic phospholipase A2, NF-κB, and cAMP response element-binding protein via signals involving phosphorylation. These data suggest that JEV infection could activate brain microvascular endothelial cells and modify their characteristics without compromising the barrier integrity, making them favorable for the recruitment and adhesion of circulating leukocytes, thereby together with other unidentified barrier-disrupting mechanisms contributing to Japanese encephalitis and associated neuroinflammation.

Abbreviations used
BBB

blood–brain barrier

CINC-1

cytokine-induced neutrophil chemoattractant 1

CREB

cAMP response element-binding protein

DMEM

Dulbecco's modified Eagle medium

ERK

extracellular signal-regulated kinase

ICAM-1

intercellular adhesion molecule 1

JEV

Japanese encephalitis virus

MAFP

methyl arachidonyl fluorophonate

PBMC

peripheral blood mononuclear cells

PKA

protein kinase A

RANTES

regulated-upon-activation normal T-cell expressed and secreted

TNF

tumor necrosis factor

The blood–brain barrier (BBB) is composed of brain microvascular endothelial cells, astrocytes, pericytes, neurons, and basement membrane, complexed by a network of tight junctions, adherens junctions, and gap junctions. These highly specialized complexes form a paracellular diffusion barrier at the vascular interface with circulating blood. Under physiological conditions, the BBB maintains CNS homeostasis and restricts immune cell migration and diffusion of soluble molecules from the systemic compartments of the body into the CNS. Among the involved cell types, the structural and functional integrity of the BBB is critically dependent on the tight junctions between brain microvascular endothelial cells (Hawkins and Davis 2005; Persidsky et al. 2006).

Neurotropic virus-associated neuropathy is characterized by the presence of infectious virus particles, immune cells, inflammatory mediators, and eventual neuronal dysfunction/destruction in the parenchymal tissues of the CNS. Generally, BBB integrity is compromised during infection and this BBB disruption dictates the aforementioned alterations and brain injury in several neurotropic viruses (Soilu-Hänninen et al. 1994; Afonso et al. 2008; Schäfer et al. 2011). Although most studies demonstrated the detrimental consequences of BBB breakdown during neurotropic virus infection, the opening of the BBB prevents certain lethal virus CNS infections, such as silver-haired bat rabies virus (Roy and Hooper 2007). Breakdown of the BBB also occurs during flavivirus infection and the compromise of the BBB is a critical event in the resultant pathology (Chaturvedi et al. 1991; Verma et al. 2009).

Japanese encephalitis virus (JEV), an enveloped, single-stranded, positive-sense, neurotropic flavivirus, is an important human pathogen transmitted by the mosquito (Chambers et al. 1990; Solomon et al. 2000). Neurological complications such as inflammation and neuronal death contribute to the mortality and morbidity associated with JEV-induced encephalitis (German et al. 2006; Ghoshal et al. 2007). During JEV infection, the neuronal death and the mortality rate increase in patients with elevated levels of inflammatory mediators in the serum and cerebrospinal fluids (Ravi et al. 1997; Winter et al. 2004). The increased production of inflammatory mediators is also associated with high virus titers in the brain and increased mortality in Japanese encephalitis animal models (German et al. 2006; Ghoshal et al. 2007; Das et al. 2008; Saxena et al. 2008). Although the exact mechanisms of neurotropic virus-associated CNS invasion and encephalitis are yet to be clearly defined, increasing evidence suggests the crucial role of BBB in controlling viral entry and immune cell infiltration into the nervous tissues. Several clinical and experimental studies demonstrated the dysfunction and/or disruption of the BBB in Japanese encephalitis subjects and these alterations were positively correlated with the severity of encephalitis (Mathur et al. 1992; Liou and Hsu 1998; German et al. 2006; Mishra et al. 2009).

Brain microvascular endothelial cells are the major structural and functional elements of the BBB. Infection of BBB endothelial cells and changes in their properties as a result of virus infection has been observed in several viruses (Soilu-Hänninen et al. 1994; Verma et al. 2009; Xu et al. 2012). Although the viruses can be detected in BBB endothelial cells after systemic infection (Liou and Hsu 1998), our understanding of cellular mechanisms associated with JEV-induced BBB disruption, specifically the contribution of BBB endothelial cells is limited. Therefore, the aim of this study was to examine the effects of JEV infection on the blood–brain barrier and the cellular properties of cultured brain microvascular endothelial cells.

Materials and methods

Virus

JEV NT113 was propagated in C6/36 cells utilizing Dulbecco's modified Eagle medium (DMEM) containing 5% fetal bovine serum (Chen et al. 2012). For virus inactivation, JEV stocks were modified by UV exposure (254 nm exposure for 30 min) or boiling (94°C incubation for 15 min). Baby hamster kidney cells (BHK21, BCRC-60041, Bioresource Collection and Research Center, Hsinchu, Taiwan) were used to determine viral titers (Chen et al. 2012).

Brain microvascular endothelial cells

The protocol for this animal study was approved by the Animal Experimental Committee of Taichung Veterans General Hospital. Brain microvascular endothelial cells were isolated from adult female Sprague–Dawley rats (BioLASCO Co., Ltd, Taipei, Taiwan) and cultured according to previously reported methods with some modifications (András et al. 2003). Briefly, the gray matters were minced and digested for 2 h at 37°C with 1 mg/mL collagenase in DMEM. The cell pellets were separated by centrifugation for 20 min at 1000 g in 20% bovine serum albumin in DMEM. The microvessels obtained in the pellets were digested further with 1 mg/mL collagenase–dispase in DMEM for 1.5 h at 37°C. Microvessel endothelial cells were then separated on a 33% continuous Percoll gradient, collected, and seeded onto collagen-coated dishes. Cells were cultured in DMEM containing 20% horse serum, 40 μg/mL of endothelial cell growth supplements, and 4 μg/mL of puromycin. Two days after the initial plating, cells were fed with culture medium without puromycin, and fed every 2 days afterward.

Viability assessment

[3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] (MTS; Promega, Madison, WI, USA) assay was performed to measure cell viability in a 96-well plate according to the manufacturer's instructions.

Reverse transcriptase-polymerase chain reaction (RT-PCR)

For determination of mRNA expression, total RNA extraction and RT-PCR were performed as previously reported (Chen et al. 2012). Oligonucleotides used in this study were as follows: 5′-CTGGAGAGCACAAACAGCAGAG3′ and 5′-AAGGCCGCAGAGCAAAAGAAGC-3′ for ICAM-1; 5′-ATGGTCTCAGCCA CCCGCTCG-3′ and 5′-TTACTTGGGGACACCCTTTAGCAT-3′ for CINC-1; 5′-GGGTCATCCTGGATGGAGGT-3′ and 5′-CACCGTCATCCTCGTTGC-3′ for RANTES; and 5′-CACTTGGCG GTTCTTTCG-3′ and 5′-GGAGCAATGATCTTGATCTTC-3′ for β-actin. Quantitative real-time PCR was performed on ABI StepOne (Applied Biosystems, Foster City, CA, USA) to determine the level of JEV replication. Relative gene expression was determined by the ΔΔCT method. Primers used for amplifications were as follows: JEV NS3, 5′-AGAGCACCAAGGGAATGAAATAGT and 5′-AATAGGTTGTAGTTGGGCACTCTG and GAPDH, 5′-CCCCCAATGTATCCGTTGTG and 5′-TAGCCCAGGAT GCCCTTTAGT.

Western blot analysis

Protein extracts were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and electrophoretically transferred to polyvinylidene difluoride membranes. After blocking, the membranes were incubated with antibodies against the following: phosphorylated extracellular signal-regulated kinase (ERK), phosphorylated cytosolic phospholipase A2 (cPLA2), phosphorylated IκB kinase (IKK-α/β), phosphorylated IκB-α, IκB-α, phosphorylated p65 (S276, S536), p65, and p50 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), TATA-binding protein (Abcam, Cambridge, MA, USA), phosphorylated cAMP response element-binding protein (CREB S133), and (Epitomics, Burlingame, CA, USA), β-tubulin (BD, San Diego, CA, USA), and JEV NS3. Finally, the blots were developed using enhanced chemiluminescence western blotting reagents. The intensity of each signal was determined using a computer image analysis system (IS1000; Alpha Innotech Corporation, Randburg, SA, USA).

Leukocyte adhesion and migration measurement

Blood was collected from the abdominal aorta of pentobarbital-anesthetized adult male Sprague–Dawley rats. Neutrophils and peripheral blood mononuclear cells (PBMC) were purified (more than 95% purity) by dextran sedimentation, centrifugation through Ficoll–Hypaque, and hypotonic lysis of erythrocytes (Pan et al. 2010). For cell adhesive assay, calcein-AM-labeled leukocytes (1 × 106) were loaded onto tested brain microvascular endothelial cell monolayers for 30 min. The unbound labeled cells were removed by gentle washing with medium and the retained fluorescent signals were measured using a fluorometer (Ex495 nm and Em515 nm). The number of signals was subtracted from the number of signals measured without cells. The chemotactic response was evaluated using a modified 24-well Transwell (Pan et al. 2010). One hundred microliters of calcein-AM-labeled leukocyte (1 × 106) suspension were added to the upper well of the chamber. The lower well received 250 μL of fresh medium and 250 μL of tested media. The upper and lower wells were separated by 3-μm pore polycarbonate filters, and the chamber was incubated for 1 h at 37°C. After incubation, non-migrating cells were scraped off from the upper surfaces of the filters. The migrating calcein-AM-labeled leukocytes, attached to the lower surfaces, were detected by a fluorometer (Ex495 nm and Em515 nm). The number of signals was subtracted from the number of signals measured with fresh medium.

Transendothelial electrical resistance (TEER)

The culture medium was aspirated, then washed three times with medium. After the insert was dropped into medium, the barrier function of the endothelial monolayer was estimated by measuring the transendothelial electrical resistance using a Millicell ERS ohmmeter (Millipore, Billerica, MA, USA), as previously reported Tedelind et al. 2003). The values were corrected for the background resistance measured across the filter without cells.

Transendothelial permeability assay

Transendothelial permeability assay was carried out according to previously reported methods with some modifications (Xu et al. 2012). Brain microvascular endothelial cells were grown on 3-μm pore Transwell filters until confluent. After treatments, fluorescein isothiocyanate (FITC)-dextran was applied apically at 0.1 μg/mL for 30 min. Samples were removed from the lower chamber for fluorescence measurements and compared to control monolayers. Fluorescence was measured using a fluorometer (Ex 492 nm and Em 520 nm).

Enzyme-linked immunosorbent assay (ELISA)

The levels of tumor necrosis factor-α (TNF-α), IL-1β, IL-6, regulated-on-activation normal T-cell expressed and secreted (RANTES), cytokine-induced neutrophil chemoattractant 1 (CINC-1), and intercellular adhesion molecule 1 (ICAM-1) in the supernatants were measured using an ELISA kit according to the manufacturer's instructions (R&D Systems, Minneapolis, MN, USA).

Preparation of nuclear extracts and electrophoretic mobility shift assay (EMSA)

Nuclear extracts were prepared as described previously (Chen et al. 2012). The oligonucleotides of NF-κB (5′-AGTTGAGGGGACTTTCCCAGGC) and CREB (5′-AGAGATTGCCTGACGTCAGAGAGCTAG) were synthesized and 5′ labeled with biotin according to the recommendations of the manufacturer (Panomics, Fremont, CA, USA). Nuclear extract (5 μg) was used for EMSA. The DNA–protein complex was analyzed on 6% native polyacrylamide gels and electrically transferred to nylon membranes. The labeled oligonucleotides were reacted with horseradish peroxidase-labeled streptavidin and detected using chemiluminescence reagents.

Statistical analysis

The data are expressed as mean values ± standard deviation. Statistical analysis was carried out using one-way analysis of variance (anova), followed by Dunnett's test to assess the statistical significance between treated and untreated groups in all experiments. A level of p < 0.05 was considered statistically significant.

Results

JEV infected brain microvascular endothelial cells

More than 95% of cultured brain microvascular endothelial cells showed positivity of CD31 immunoreactivity (data not shown). Although JEV (multiplicity of infection, MOI 20) replicated in cultured brain microvascular endothelial cells, the production of infectious virus particles (Fig. 1a) and the expression of viral non-structural protein NS3 (Fig. 1b) were not as elevated as those in BHK21 cells (MOI 5). Unlike the result with BHK21 cells (Fig. 1c), JEV infection had a negligible effect on the viability of brain microvascular endothelial cells (Fig. 1d).

Figure 1.

Characterization of Japanese encephalitis virus (JEV) replication and toxicity. Brain microvascular endothelial cells (20 MOI) and BHK21 cells (5 MOI) were infected with mock or JEV over time. The supernatants were collected and subjected to plaque assay for the determination of viral titers (a). Cellular proteins were isolated and subjected to western blot with antibodies against JEV, NS3, and β-tubulin (b). Cell viability of BHK21 cells (c) and brain microvascular endothelial cells (d) was measured by MTS reduction. **p < 0.01 versus each mock control, n = 4.

JEV infection activated brain microvascular endothelial cells

TEER and permeability to FITC-dextran were utilized as evaluation indices to determine the effects of JEV infection on endothelial function (Tedelind et al. 2003; Xu et al. 2012). Phenylarsenoxid, a tight junction-destroying agent (Lohmann et al. 2004), disrupted established electrical resistance (Fig. 2a) and impermeability (Fig. 2b), indicating the functional execution of cultured brain microvascular endothelial cells in barrier integrity. Up to 72 h after infection, the TEER (Fig. 2a) and transendothelial permeability (Fig. 2b) were not significantly changed in JEV-infected cells when compared with mock-infected cells. Without the presence of brain microvascular endothelial cells, large amounts of virus particles could be collected in the lower chambers (~106 pfu/mL) 24 h after the loading of virus (data not shown). Although the levels were low, there was a progressive production of infectious JEV particles detected in the lower chambers when the filters were covered with confluent brain microvascular endothelial cells (Fig. 2c). Next, the adhesion and chemotactic migration of leukocytes were observed to evaluate the effects of JEV infection on the activation of endothelial cells. The adhesion of naïve neutrophils (Fig. 3a) and PBMCs (Fig. 3b) onto endothelial monolayers was promoted by JEV infection and this effect was abrogated after virus inactivation. Supernatants from JEV-infected brain microvascular endothelial cells exhibited a significantly enhanced capacity to stimulate neutrophil (Fig. 3a) and PBMC (Fig. 3b) chemotaxis compared with the mock-infected control. The chemotactic activity of supernatants from JEV-infected cells was almost completely lost when the virus was heat- or UV-inactivated (Fig. 3a and b).

Figure 2.

Effects of Japanese encephalitis virus (JEV) infection on TEER and transendothelial permeability. Confluent monolayers of brain microvascular endothelial cells were infected with mock or JEV (20 MOI) over time or treated with phenylarsenoxid (10 μM). The TEER (a), transendothelial permeability to FITC-dextran (b), and the virus titers in the supernatants collected from the lower chamber (24-well, 0.5 mL medium) (c) were measured at the indicated times. **p < 0.01 versus each control, n = 4.

Figure 3.

Effects of Japanese encephalitis virus (JEV) infection on leukocyte adhesion and migration. Confluent monolayers of brain microvascular endothelial cells were infected with mock, wild-type JEV (20 MOI), heat-inactivated JEV, or UV-inactivated JEV for 24 h. The resultant cells were pre-treated with medium, control IgG, and intercellular adhesion molecule 1 (ICAM-1), cytokine-induced neutrophil chemoattractant 1 (CINC-1), and regulated-on-activation normal T-cell expressed and secreted (RANTES)-neutralizing antibodies (10 μg/mL) for 30 min before being subjected to leukocyte adhesion assay with calcein-AM-labeled naïve neutrophils (a) and peripheral blood mononuclear cells (PBMCs) (b). Supernatants isolated from infected cells were modified by neutralization with medium, control IgG, and ICAM-1, CINC-1, and RANTES-neutralizing antibodies (10 μg/mL) for 30 min before being subjected to chemotactic assay with calcein-AM-labeled naïve neutrophils (a) and PBMCs (b). The fluorescent signals were measured by a fluorometer. **p < 0.01, n = 4.

JEV infection induced expression of adhesion molecules and chemokines

There is evidence showing that activation of endothelial cells is a key event that stimulates extensive leukocyte adhesion and chemotaxis by up-regulating expressions of cytokines, chemokines, and adhesion molecules (Dietrich 2002; Ubogu et al. 2006; Connolly-Andersen et al. 2011). We assessed whether JEV infection induces elevated expressions of cytokines, chemokines, and adhesion molecules which participate in up-regulating leukocyte adhesion and chemotaxis. Our results showed that wild-type but not heat- or UV-inactivated JEV infection caused robust ICAM-1, CINC-1, and RANTES release from brain microvascular endothelial cells (Fig. 4). In contrast, there was no detectable production of TNF-α, IL-1β, and IL-6 during the course of JEV infection (data not shown). Pre-treatment of brain microvascular endothelial cells with ICAM-1-neutralizing antibody had an inhibitory effect on neutrophil (Fig. 3a) and PBMC (Fig. 3b) adhesion. This adhesive inhibition was not observed after pre-treatment with IgG, CINC-1-, or RANTES-neutralizing antibody (Fig. 3a and b). The chemotactic activity of supernatants from JEV-infected brain microvascular endothelial cells was partially reduced when CINC-1- or RANTES-neutralizing antibodies was added to the supernatants before the chemotaxis assay. There was a negligible effect on chemotactic activity when control IgG or ICAM-1-neutralizing antibody was added (Fig. 3a and b).

Figure 4.

Effects of Japanese encephalitis virus (JEV) infection on inflammatory gene production. Confluent monolayers of brain microvascular endothelial cells were infected with mock, wild-type JEV (20 MOI), heat-inactivated JEV, or UV-inactivated JEV over time. Supernatants isolated from infected cells were subjected to ELISA for the measurement of intercellular adhesion molecule 1 (ICAM-1), cytokine-induced neutrophil chemoattractant 1 (CINC-1), and regulated-on-activation normal T-cell expressed and secreted (RANTES). **p < 0.01 versus each mock control, n = 4.

Transcription factors NF-κB and CREB were involved in JEV-induced gene expression

The activation and induction of inflammatory gene expression in endothelial cells have been previously shown to be dependent on NF-κB and CREB activation (Hadad et al. 2011). Intensive bands representing a DNA–protein complex of NF-κB and CREB (Fig. 5a) were observed in JEV-infected cells. The involvement of NF-κB in regulating endothelial ICAM-1, CINC-1, and RANTES expression after JEV infection was demonstrated by the inhibitory effects of an NF-κB inhibitor, [3-chloro-4-nitro-N-(5-nitro-2-thiazolyl)-benzamide; SM-7368], on NF-κB DNA-binding activity (Fig. 5a), protein release (Fig. 5b), and mRNA expression (Fig. 5c). The results of western blotting further revealed the characteristics of JEV-stimulated NF-κB involvement in IKK-α/β phosphorylation, IκB-α phosphorylation, IκB-α degradation, p65 phosphorylation (S536 and S276) (Fig. 6a), and p65 and p50 nuclear translocation (Fig. 6b).

Figure 5.

Effects of agents on Japanese encephalitis virus (JEV)-induced transcription factors and cytokine expression. Confluent monolayers of brain microvascular endothelial cells were infected with mock or JEV (20 MOI) in the presence of U0126, PD98059, methyl arachidonyl fluorophonate (MAFP), AACOCF3, or NF-κB inhibitor. Nuclear extracts (3 h) were isolated and subjected to EMSA for the measurement of NF-κB and cAMP response element-binding protein (CREB). One of three independent experiments is shown (a). Supernatants (24 h) isolated from infected cells were subjected to ELISA for the measurement of intercellular adhesion molecule 1 (ICAM-1), cytokine-induced neutrophil chemoattractant 1 (CINC-1), and regulated-on-activation normal T-cell expressed and secreted (RANTES). **p < 0.01 versus each mock control and ##p < 0.01 versus JEV control, n = 4 (b). Total cellular RNAs (12 h) were isolated and subjected to RT-PCR for the measurement of ICAM-1, CINC-1, RANTES, and β-actin (c). One of three independent experiments is shown. Supernatants (24 h) isolated from infected cells were subjected to plaque assay for the measurement of virus titer. n = 4 (d). Total cellular RNAs (24 h) were isolated from infected cells and subjected to quantitative real-time RT-PCR for the measurement of JEV RNA. n = 4 (e).

Figure 6.

Effects of agents on Japanese encephalitis virus (JEV)-induced signaling. Confluent monolayers of brain microvascular endothelial cells were infected with mock or JEV (20 MOI) in the presence of U0126 (10 μM), PD98059 (10 μM), methyl arachidonyl fluorophonate (MAFP) (20 μM), or AACOCF3 (10 μM) for 3 h. Total cellular proteins were isolated and subjected to western blot with antibodies against P-IKK-α/β, IKK-α/β, P-IκB-α, IκB-α, P-p65 (S276), P-p65 (S536), p65, P-CREB, cAMP response element-binding protein (CREB), P-ERK, ERK, P-cPLA2, cPLA2, and β-tubulin (a). The nuclear extracts and cytosolic extracts were isolated and subjected to western blot with antibodies against p50, p65, TATA-binding protein, and β-tubulin (b). One of three independent experiments is shown.

ERK was involved in JEV-induced gene expression

The increased DNA-binding activity of CREB and the activation of NF-κB can be promoted by phosphorylation and ERK is one of the upstream regulators activated through phosphorylation. Besides, the NF-κB transcriptional activity is alternatively regulated by its phosphorylation of the p65 subunit of NF-κB (Hu et al. 2004; Lee et al. 2011; Lin et al. 2011). Elevated phosphorylation of ERK, CREB, IKK-α/β, IκB-α, and p65 (S536 and S276) was observed in JEV-infected brain microvascular endothelial cells and this increased phosphorylation was attenuated by ERK inhibitors U0126 and PD98059 (Fig. 6a). In parallel with the attenuation of phosphorylation, ERK inhibition by U0126 and PD98059 caused significant reductions in JEV-elevated p65/p50 nuclear translocation (Fig. 6b) and NF-κB/CREB DNA-binding activities (Fig. 5a), ICAM-1, CINC-1, and RANTES mRNA expression (Fig. 5c) and protein release (Fig. 5b). These findings indicate that a close relationship existed between ERK, NF-κB, CREB, and ICAM-1/CINC-1/RANTES gene expression after JEV infection in brain microvascular endothelial cells.

cPLA2 was involved in JEV-induced gene expression

Evidence suggests that the activation of cPLA2 also play an important role in the signal transduction cascade event during induction of endothelial gene transcription (Hadad et al. 2011). To examine the role of cPLA2 signaling in JEV infection-induced endothelial gene expression, the cPLA2 activity was assessed by measuring its phosphorylation (Liao et al. 2011). As shown in Fig. 6a, JEV infection caused an increased phosphorylation of cPLA2 in brain microvascular endothelial cells. In the presence of cPLA2 inhibitors, such as methyl arachidonyl fluorophonate (MAFP) and AACOCF3, the levels of JEV-induced ICAM-1, CINC-1, and RANTES mRNA expression (Fig. 5c) and protein release (Fig. 5b) were reduced. To get a clear insight into the association between cPLA2 and endothelial gene expression, the effects of cPLA2 inhibitors on regulatory signaling molecules and transcription factors were examined. Addition of MAFP and AACOCF3 blocked JEV-elevated phosphorylations of IKK-α/β, IκB-α, p65 (S536 and S276), and CREB (Fig. 6a), p50 and p65 nuclear translocation (Fig. 6b), and DNA-binding activities of NF-κB and CREB (Fig. 5a). ERK inactivation blocked JEV-induced cPLA2 phosphorylation, whereas MAFP and AACOCF3 treatments left the ERK phosphorylation unchanged (Fig. 6a). None of those pharmacological inhibitors showed significant inhibitory effect on infectious viral particle production (Fig. 5d) and JEV replication (Fig. 5e).

Discussion

The presence of viral particles in BBB endothelial cells and signs of endothelial damage and disruption of the BBB have been observed in a mouse model of Japanese encephalitis (Mathur et al. 1992; Liou and Hsu 1998; German et al. 2006; Mishra et al. 2009). Cultured brain microvascular endothelial cells were susceptible to JEV infection, but with a limited efficacy. Despite the low efficacy, JEV was able to cross the brain microvascular endothelium monolayers. Unfortunately, the significance of endothelial cell damage and barrier disruption was not observed in this in vitro monoculture model. The results of current monoculture model suggest that direct infection of endothelial cells with JEV is not the determining event in regulating endothelial viability and barrier activity. Recently, we found that lower titers of JEV infection (5 and 10 MOI) were able to cause disruption of barrier integrity on cocultures of brain microvascular endothelial cells with astrocytes and pericytes (ongoing study). These preliminary observations indicated that coculture model might be better than monoculture model for the study of JEV's effect on BBB. Currently, this model system is under investigation in our laboratory.

Adhesion molecules and chemokines exert their effects via regulatory signals for the cerebral recruitment and extravasation of leukocytes. Expression of adhesion molecules has been found to attract circulating leukocytes to endothelial cells and increase infiltration of immune cells to the brain parenchyma (Dietrich 2002). Chemokines are able to promote leukocyte recruitment and migration across the BBB (Ubogu et al. 2006). Without compromising the barrier integrity, JEV infection caused endothelial release of chemotactic molecules and expression of adhesion molecules. Current findings show the potential involvement of JEV-infected endothelial cells in some steps of leukocyte trafficking into the CNS via release of ICAM-1, CINC-1, and RANTES. Similar characteristics were found in West Nile virus, another neurotropic flavivirus (Verma et al. 2009).

Our data clearly showed an association between transcription factors NF-κB and CREB and JEV-induced ICAM-1, CINC-1, and RANTES expression. ERK is one potential upstream regulator in transduction of signals to NF-κB and CREB through phosphorylation (Hu et al. 2004; Lee et al. 2011; Lin et al. 2011). In the current study, the formation of NF-κB and CREB DNA–protein complexes and phosphorylation of ERK were activated in JEV-infected cells. The best-characterized NF-κB regulation pathway applies to the p65-p50 heterodimer, which is held captive in the cytoplasm by specific inhibitory IκB proteins. The activated IKK complex phosphorylates NF-κB-bound IκB allowing the liberated NF-κB dimer to translocate to the nucleus. Alternatively, the nuclear translocation, DNA binding, and transcriptional activity of NF-κB are promoted by phosphorylation of p65 at S536 and S276. In addition to classical IKK, the phosphorylation of p65 at S536 and S276 is also mediated by other signal-induced kinases (Sakurai et al. 1999; Hu et al. 2004; Hayden and Ghosh 2008; Lin et al. 2011). Similar activation of phosphorylation also occurs in CREB by phosphorylation at S133 (Lee et al. 2011). It has been shown that the phosphorylation of IKK, p65 S536, p65 S276, and CREB S133 can be mediated by ERK (Hu et al. 2004; Lee et al. 2011; Lin et al. 2011). The phosphorylation of IKK-α/β, IκB-α, p65 S536, p65 S276, and CREB S133 in JEV-infected cells was accompanied by ERK activation and decreased by ERK inhibitors. Multiple distinct signaling pathways are activated by virus infection, leading to the modulation of transcription factor activation and gene expression. Previously, we found that the activation of ERK, NF-κB, and CREB can trigger expression of TNF-α, IL-1β, IL-6, and RANTES in JEV-infected microglia. Lipid rafts, Src, protein kinase c, and oxidative stress serve as upstream regulators to transduce signals to ERK (Chen et al. 2004, 2010, 2012). Our current findings suggest that the canonical activation of NF-κB through IKK signaling, alternative activation of NF-κB through p65 phosphorylation, and CREB activation through phosphorylation were mediated partly by ERK and these signaling cascades were active participants which contributed to endothelial ICAM-1, CINC-1, and RANTES expression following JEV infection. However, the detailed characterization of the relationship between kinases and substrates in these phosphorylation events was not addressed.

Other interesting findings in this study were that cPLA2 participated in NF-κB and CREB activation and the consequences of ICAM-1, CINC-1, and RANTES expression in JEV-infected brain microvascular endothelial cells. Studies suggest that the involvement of cPLA2 in inflammatory cytokine expression in endothelial cells and immune cells is assumed to trigger activation of transcription factors such as NF-κB and CREB (Szaingurten-Solodkin et al. 2009; Hadad et al. 2011). By extending the scope of these studies, we have demonstrated that the activation of NF-κB and CREB via stimulation of phosphorylations in IKK-α/β, IκB-α, p65 S536, p65 S276, and CREB S133 is an alternative regulatory mechanism of endothelial ICAM-1, CINC-1, and RANTES induction downstream of cPLA2 activation. Although the biochemical data support this conclusion, these findings raise an intriguing question. Unlike ERK, cPLA2 lacks kinase activity. Thus, through what mechanisms does cPLA2 control the phosphorylation of NF-κB subunits and CREB? cPLA2 catalyzes the hydrolysis of the sn-2 substituent from glycerophospholipid substrates to yield arachidonic acid and the consequences of eicosanoid generation. Evidence suggests that PGE2, one of the biologically active eicosanoids, plays an important role in the signal transduction cascade events during induction of gene transcription involving its engagement with receptors and downstream effector protein kinase A (PKA) (Chen and Hughes-Fulford 2000). Additional evidence shows that IKK-α/β, p65 S276, and CREB S133 are potential substrates of PKA (Chen and Hughes-Fulford 2000; Hu et al. 2004; Hadad et al. 2011; Lee et al. 2011; Lin et al. 2011). Current findings show the substantial involvement of cPLA2 in JEV-induced endothelial NF-κB/CREB activation and ICAM-1, CINC-1, and RANTES induction. Although we did not have the precise experimental data to demonstrate its action mechanisms, PKA might serve as a potential switching molecule to link between cPLA2 and NF-κB/CREB. However, the specific mechanisms involved remain to be determined.

Brain homeostasis is maintained by the structure and function of the BBB, which plays a key role in the pathogenesis of neurotropic viruses by regulating the entry of circulating molecules, immune cells, or viruses into the CNS. In Fig. 7, we show a potential mechanism for elucidating the trafficking of leukocytes during the course of JEV infection through the activation of endothelial cells. JEV infection selectively triggers endothelial release of ICAM-1, CINC-1, and RANTES. Under pathophysiological conditions, the consequences of the released ICAM-1, CINC-1, and RANTES are to attract circulating leukocytes and promote leukocyte adhesion onto endothelial surfaces. Transcription factors represent a group of important effectors that could cause the convergence of multiple extrinsic and intrinsic signals resulting in the regulation of gene expression. Our data demonstrate that JEV-stimulated ERK is critical for ICAM-1, CINC-1, and RANTES induction through initiation of ERK/NF-κB, CREB and ERK/cPLA2/NF-κB, CREB in brain microvascular endothelial cells. Our findings show that brain microvascular endothelial cells can be a target for JEV infection and can be one of the mechanisms which facilitate recruitment of peripheral immune cells before transendothelial migration. Collectively, these data suggest that JEV infection could activate brain microvascular endothelial cells and modify their characteristics without compromising the barrier integrity, making them favorable for the recruitment and adhesion of circulating leukocytes, thereby together with other unidentified barrier-disrupting mechanisms contributing to Japanese encephalitis and associated neuroinflammation.

Figure 7.

A possible signaling mechanism of gene induction elicited by Japanese encephalitis virus (JEV) is proposed. This schematic diagram indicates the signaling molecules and cascades used in mediating endothelial activation of NF-κB and cAMP response element-binding protein (CREB) and the consequences of intercellular adhesion molecule 1 (ICAM-1), cytokine-induced neutrophil chemoattractant 1 (CINC-1), and regulated-on-activation normal T-cell expressed and secreted (RANTES) gene induction after JEV infection. Some additional signaling molecules and cascades have been omitted for the sake of clarity. The solid label → indicates a direct stimulating effect and the dotted label → indicates an indirect stimulating effect. IκB, CRE, Ub, and P represent the κB-binding site, cAMP response elements, ubiquitination, and phosphorylation, respectively.

Acknowledgements

This work was supported by a grant from the National Science Council (NSC100-2314-B-075A-004) and a joint grant from Taichung Veterans General Hospital and Central Taiwan University of Sciences and Technology (TCVGH-CTUST1007702), Taiwan. The authors have no conflicts of interest to declare. CYL, YCO, CYC, and HCP contributed to experimental execution, acquisition of data, and analysis and interpretation of data. CJChang contributed to grant application and interpretation of data. SLL contributed to the preparation of endothelial cells. HLS contributed to draft preparation. CJChen contributed to grant application, experimental design, and draft preparation.

Ancillary