OUP user menu

Interferon-α induces nitric oxide synthase expression and haem oxygenase-1 down-regulation in microglia: implications of cellular mechanism of IFN-α-induced depression

Dah-Yuu Lu , Yuk-Man Leung , Kuan-Pin Su
DOI: http://dx.doi.org/10.1017/S1461145712000338 433-444 First published online: 1 March 2013

Abstract

Substantiating evidence for the inflammation theory of depression is that interferon-alpha (IFN-α) induces clinical depression. Despite numerous researches on neurochemical and neuroendocrinological mechanisms from human and animal studies, the direct mechanisms of IFN-α at cellular levels are still lacking. In this study, we aimed to identify the cellular mechanisms for IFN-α-induced neuroinflammatory response with the murine BV-2 microglia cell line. IFN-α potently induced nitric oxide synthase (iNOS) and nitric oxide (NO) release and down-regulated haem oxygenase-1 (HO-1) expression, which could be dampened by Janus kinase 1 (JAK1) and c-Jun NH2-terminal kinase (JNK) inhibition, respectively. IFN-α activated JAK1, JNK, signal transducers and activators of transcription (STAT)1 and STAT3, but not extracellular signal-regulated kinases (ERK) and phosphoinositide 3 (PI3) kinase, signal pathways. The transfection with STAT1 and STAT3 siRNA also inhibited IFN-α-induced iNOS/NO expression and HO-1 down-regulation. The HO-1 activator, CoppIX, reversed iNOS/NO up-regulation and HO-1 down-regulation induced by IFN-α. On the other hand, a knockdown of HO-1 expression enhanced IFN-α-induced iNOS/NO expression. The effects of IFN-α-induced iNOS/NO up-regulation and HO-1 down-regulation in microglia are associated with JAK1/JNK/STAT1 and STAT3 signalling pathways. The different effects between IFN-α and IFN-γ on HO-1 regulation and ERK phosphorylation might provide a possible explanation of different risk in their induction of neuropsychiatric adverse effects in clinical and animal studies. The results from this study add the missing part of direct cellular mechanisms for IFN-α-induced depression.

Key words
  • Haem oxygenase-1 (HO-1)
  • interferon-alpha (IFN-α)
  • microglia
  • nitric oxide (NO)
  • nitric oxide synthase (iNOS)

Introduction

Accumulating evidence suggests that pathophysiology of depression might be associated with activated inflammatory processes (Dantzer et al. 2008; Miller et al. 2009). Clinically depressed patients have been found to have higher levels of pro-inflammatory cytokines and inflammatory markers (Dowlati et al. 2010; Maes, 1999; Schiepers et al. 2005), but the most supportive evidence for inflammation theory of depression is probably the behavioural changes induced by pro-inflammatory cytokines in clinical and animal studies (Felger et al. 2007; Simmons & Broderick, 2005; Song & Wang, 2011). For example, the standard treatment with interferon-alpha (IFN-α) for chronic hepatitis C virus (HCV) infection is associated with the development of major depressive episode (MDE) in up to 45% of HCV patients (Musselman et al. 2001; Raison et al. 2005; Su et al. 2010). In fact, IFN-α-induced depression is one of the excellent models to explore the action of peripheral cytokine administration on human brains and to study the development of depression in a prospective way.

IFN-α is an innate immune cytokine that has both anti-viral and anti-proliferative activities and is a potent stimulator of pro-inflammatory cytokines not only at the periphery but also within the central nervous system (CNS) (Borden et al. 2007). IFN-α has demonstrated CNS regulatory effects, ranging from cerebral genetic expression (Wang et al. 2008) and neuronal excitabilities (Calvet & Gresser, 1979) to circadian rhythm (Ohdo et al. 2001) and animal behaviours (Schrott & Crnic, 1996). In rodents or rhesus monkeys, acute or chronic exposures of IFN-α induce depression-like behaviours (Felger et al. 2007; Makino et al. 2000; Sammut et al. 2001) and influence the function of neurotransmitter metabolism and neuroendocrine linking to depression (Felger et al. 2007; Kitagami et al. 2003; Kumai et al. 2000; Menzies et al. 1996), although not all the findings are consistent (De La et al. 2005). In humans, substantiating the fact that IFN-α indeed induces depression is that IFN-α-induced depression (Raison et al. 2006): (i) is responsive to standard antidepressant therapy (Capuron et al. 2002a; Farah, 2002; Musselman et al. 2001); (ii) is associated with alterations in monoamine metabolism and its metabolic enzyme indoleamine 2,3-dioxygenase (IDO) (Capuron et al. 2002b; Maes et al. 2001; Raison et al. 2009, 2010b; Wichers et al. 2005); and (iii) is associated with alterations in hypothalamic-pituitary-adrenal (HPA) axis function (Capuron et al. 2003; Raison et al. 2010a, c).

IFN-α is relatively a large molecule (15–25 kDa) that does not pass freely through the blood–brain barrier (BBB). In addition, no transport system across BBB for IFN-α has been described to date (Raison et al. 2009). To administrate IFN-α peripherally, however, does increase IFN-α levels in the CNS in humans (Raison et al. 2009), rhesus monkeys (Felger et al. 2007), and rodents (Wang et al. 2008), suggesting that IFN-α either enters the brain via passage through leaky regions in the BBB (Vitkovic et al. 2000) or alternatively activates cells at the BBB to induce local IFN-α production (Indraccolo et al. 2007; Shehata et al. 2000).

Excessive secretion of macrophage pro-inflammatory cytokines is proposed as the cause of depression (Smith, 1991). Microglia are the resident macrophages of the brain and spinal cord, and thus act as the main form of active immune defence in the CNS (Hanisch & Kettenmann, 2007). It has been proposed that neuroinflammatory processes may cause neuropsychiatric disorders (e.g. Alzheimer's disease, Parkinson's disease, and depression) through microglial activation (Hanisch & Kettenmann, 2007; Maes et al. 2011). Engagement of the immune-to-brain communication pathways by IFN-α or other pro-inflammatory cytokines [e.g. IFN-γ, interleukin (IL)-1] ultimately lead to microglial activations (Dantzer et al. 2008; Hanisch & Kettenmann, 2007) and trigger several inflammatory signalling pathways, including signal transducers and activators of transcription (STAT)1α, NF-κB and mitogen-activated protein kinases (MAPKs) (Ferreira et al. 2010; Fujigaki et al. 2006; Li et al. 2003; Lu et al. 2010b; Shen et al. 2005). Upon activation, microglia up-regulate the expression of detrimental factors of reactive oxygen species [e.g. nitric oxide (NO) via inducible nitric oxide synthase, iNOS], which induce oxidative stress (Lu et al. 2010b; Shen et al. 2005) and contribute to pathogenesis of neuropsychiatric disorders (Hanisch & Kettenmann, 2007; Maes et al. 2011). On the other hand, the expression of anti-oxidative enzymes, e.g. haem oxygenase-1 (HO-1), are neuroprotective in response to oxidative stress (Choi & Kim, 2008; Gozzelino et al. 2010; Jazwa & Cuadrado, 2010; Jeong et al. 2011; Le et al. 1999) and might characterize the antidepressant mechanisms (Chen et al. 2005; Krishnan & Nestler, 2008; Lu et al. 2010b; Oh et al. 2004). In addition, neuroinflammation also reduces the survival of serotonergic neurons (Hochstrasser et al. 2011) and decreases neurogenesis (Song & Wang, 2011), while antidepressants exert neuroprotection against microglia-mediated neurotoxicity (Zhang et al. 2012).

Despite extensive studies of IFN-γ, IL-1 and tumour necrosis factor (TNF)-α on microglial cells, the study of IFN-α is still lacking. The goal of this study is to identify the mechanism of IFN-α-induced neuroinflammatory response in microglia and to provide mechanistic implications of IFN-α-induced depression.

Materials and methods

Materials

Recombinant human and mouse IFN-α were purchased from PBL Biomedical Labs (USA). Fetal bovine serum (FBS), Dulbecco's modified Eagle's medium (DMEM), and OPTI-MEM were purchased from Gibco-Invitrogen (USA). Goat anti-mouse and anti-rabbit horseradish peroxidase-conjugated IgG, primary antibodies against PCNA, β-actin and c-Jun NH2-terminal kinase (JNK) were purchased from Santa Cruz Biotechnology (USA). iNOS antibody was purchased from BD Pharmingen Transduction Laboratories (USA). HO-1 antibody was purchased from StressGen Biotechnologies (USA). Rabbit polyclonal antibodies against JNK phosphorylated at Thr183/Tyr185, STAT1 phosphorylated at Tyr701, STAT3 phosphorylated at Ser727, phospho-Jak1, phospho-Jak2 and phospho-TYK2 were purchased from Cell Signaling and Neuroscience (USA). Jak2 inhibitor AG490 and Jak1 inhibitor Pyridone 6 were purchased from Calbiochem (USA). The dominant-negative mutant of JNK (DN-JNK) was provided by Dr C.-H. Tang (Department of Pharmacology, China Medical University). All other chemicals were obtained from Sigma-Aldrich (USA).

Cell culture

The murine microglia cell line BV-2 is a useful model to investigate mechanisms of microglial activation and neuroinflammation (Henn et al. 2009). The cell line has been generated by infecting primary microglial cell cultures with a v-raf/v-myc oncogene carrying retrovirus (J2). Phenotypically, BV-2 cells resulted in positive for macrophage antigen complex (MAC)1 and MAC2 antigens. Since BV-2 cells retain most of the morphological, phenotypical and functional properties described for freshly isolated microglial cells they can be considered as immortalized active microglial cells (Blasi et al. 1990). Cells were cultured in DMEM (Gibco-Invitrogen) with 10% FBS at 37 °C in a humidified incubator in an atmosphere of 5% CO2 and 95% air. Confluent cultures were passaged by trypsinization.

Western blot analysis

BV-2 microglia cells were treated with IFN-α and signal transduction inhibitors for various time periods and then washed with cold PBS that had been lysed for 30 min on ice with radioimmunoprecipitation assay buffer. The nuclear extracts were prepared as described previously (Lin et al. 2011). Cells were rinsed with PBS and suspended in hypotonic buffer A [10 mm Hepes (pH 7.6), 10 mm KCl, 1 mm DTT, 0.1 mm EDTA, and 0.5 mm phenylmethylsulfonyl fluoride (PMSF)] for 10 min on ice and vortexed for 10 s. The lysates were separated into cytosolic and nuclear fractions by centrifugation at 12 000 g for 10 min. The supernatants containing cytosolic proteins were collected. A pellet containing nuclear fraction was resuspended in buffer C [20 mm Hepes (pH 7.6), 1 mm EDTA, 1 mm DTT, 0.5 mm PMSF, 25% glycerol, and 0.4 m NaCl] for 30 min on ice. The supernatants containing nuclear proteins were collected by centrifugation at 13 000 g for 20 min and stored at −80 °C.

Protein samples were separated by sodium dodecyl sulphate (SDS)–polyacrylamide gels and transferred to polyvinylidene fluoride membranes. The membranes were blocked with 5% non-fat milk in PBS for 1 h at room temperature and then probed with primary antibodies. After undergoing three PBS washes, the membranes were incubated with secondary antibodies. The blots were visualized by enhanced chemiluminescence using Kodak X-OMAT LS film (Eastman Kodak, USA). The blots were subsequently stripped through incubation in stripping buffer [62.5 mm Tris (pH 6.8), 2% SDS, and 0.1 m β-mercaptoethanol] and reprobed for β-actin (cytoplasmic) or PCNA (nuclear) as a loading control. Quantitative data were obtained using a computing densitometer and ImageQuant software (Molecular Dynamics, USA).

Transfection

The protocol of transient transfection was applied according to our previously report (Chen et al. 2011). The siRNAs against JAK1, JAK2, STAT1, STAT3 and HO-1 and the control siRNA (at a final concentration of 100 nm) were purchased commercially from Santa Cruz Biotechnology (USA). Plasmid DNAs or siRNAs were premixed with LF2000 in OPTI medium (Invitrogen Life Technologies, USA) for 20 min and applied to the cells. Medium containing 20% FBS was added 4–6 h later. After transfection for 24 h, LF2000-containing medium was replaced with fresh serum-free medium.

Assay of nitric oxide

Synthesis of NO was determined by assay of culture supernatants for the nitrite levels of the stable nitric oxide metabolite, which reached maximum production after microglia stimulated for 24 h (Pahan et al. 2001). Briefly, accumulation of nitrite in the medium was determined by colorimetric assay with Griess reagent. Cells (2 × 105 cells per well) in 24-well plates in 500 µl culture medium were stimulated with IFN-α for 24 h. One hundred µl of culture supernatant reacted with an equal volume of Griess reagent (one part 0.1% naphthylethylenediamine and one part 1% sulfanilamide in 5% H3PO4) in 96-well culture plates for 10 min at room temperature in the dark. The absorbance at 550 nm was determined using a microplate reader (Bio-Tek, USA). Each experiment was performed in duplicate.

Reverse transcriptase (RT)–PCR and quantitative real-time PCR (qPCR)

Total RNA was extracted from cells using a TRIzol kit (MDBio Inc., Taiwan). The reverse transcription reaction was performed using total RNA (2 µg) that was reverse-transcribed into cDNA and then amplified using oligonucleotide primers as follows:

  • IFNAR1: 5′-AAGACGATGCTCGCTGTCGTG-3′ and 5′-ACCTGTCAGCAGAGAAGAGCTG-3′;

  • GAPDH: 5′-ACCACAGTCCATGCCATCAC-3′ and 5′-TCCACCACCCTGTTGCTGTA-3′.

PCR reaction procedures were performed following our previous reports and products were then separated electrophoretically in an agarose gel and stained with ethidium bromide (Lu et al. 2010a). Quantitative real-time PCR was performed using the Step One Real Time System (Applied Biosystems, USA) according to our previous report (Lu et al. 2010b). Sequence-specific oligonucleotide primers were:

  • iNOS: 5′-CCCAGAGTTCCAGCTTCTGG-3′ and 5′-CCAAGCCCCTCACCATTATCT-3′;

  • HO-1: 5′-GGGTGACAGAAGAGGCTAAGACC-3′ and 5′-AGATTCTCCCCTGCAGAGAGAAG-3′;

  • GAPDH: 5′-CTCAACTACATGGTCTACATGTTCCA -3′ and 5′-CTTCCCATTCTCAGCCTTGACT-3′.

Statistics

Statistical analysis was performed using Graphpad Prism 4.01 software (Graph Pad Software Inc., USA). The values given are means±s.e.m. Statistical analysis between two samples was performed using Student's t test. Statistical comparisons of more than two groups were performed using one-way ANOVA with Bonferroni's post-hoc test. In all cases, a value of p < 0.05 was considered as significant.

Results

IFN-α increases iNOS/NO expression and HO-1 down-regulation in microglia

We first identified the effects of IFN-γ and IFN-α on the regulation of iNOS, NO and HO-1 in BV-2 microglia. Stimulation of IFN-γ for various time periods (0, 2, 4, 6, 16 or 24 h) and at various concentrations (0.3, 1, 3, 10 or 30 ng/ml for 24 h) increased iNOS protein expression, while HO-1 protein levels remained unchanged (Fig. 1 a, b). On the other hand, stimulation of IFN-α for various time periods at various concentrations increased iNOS protein expression (see Fig. 1 c, d). Moreover, IFN-α down-regulated HO-1 expression in both a time- and concentration-dependent manner (Fig. 1 c, d). In addition, recombinant mouse IFN-α also increased iNOS expression and down-regulated HO-1 expression (Supplementary Fig. S1A, B). The dose–response curves of iNOS expression between mouse and human IFN-α increases are very similar (Supplementary Fig. S1C), indicating that the effects did not have species specificity in our in vitro model. IFN-α increased NO production initiated at a concentration of 3000 U/ml (Supplementary Fig. S2A); therefore, 3000 U/ml IFN-α was used in the following experiments. It has been reported that lipopolysaccharides (LPS) increase iNOS and NO production in BV-2 microglia (Lu et al. 2007, 2010b). Here, we also showed that LPS increased iNOS expression and NO production (up to 5-fold) but did not affect HO-1 expression (Fig. 1 d). To further determine whether IFN-α-induced iNOS/NO expression involved in transcription and translation, cells were stimulated with IFN-α in the presence of inhibitors at transcription (actinomycin D) and translation (cycloheximide) levels. IFN-α-mediated induction of NO production was abolished by actinomycin D and cycloheximide in a concentration-dependent manner (Supplementary Fig. S2B). BV-2 microglia treated with IFN-α (3000 U/ml) for various time periods (0, 2, 4, 8 or 24 h) increased mRNA expression in iNOS levels (Supplementary Fig. S2C), but did not significantly change HO-1 levels (Supplementary Fig. S2D). Taken together, IFN-α down-regulates HO-1 protein expression and induces iNOS/NO via de novo protein synthesis.

Fig. 1

Both IFN-α and IFN-γ increase iNOS expression but only IFN-α down-regulates HO-1 in BV-2 microglia. (a) Cells were incubated with IFN-γ (10 ng/ml) for various time periods (2, 4, 6, 16 or 24 h) or (b) at various IFN-γ concentrations (0.3, 1, 3,10 or 30 ng/ml) for 24 h. (c) Cells were incubated with IFN-α (3000 U/ml) for various time periods (2, 4, 6, 16 or 24 h) or (d) at various IFN-α concentrations (0, 1000, 3000, 10 000 or 30 000 U/ml) for 24 h. Treatment of LPS (100 ng/ml) for 24 h was as a positive control. The protein levels of iNOS and HO-1 were analysed by Western blot, and similar results were obtained from four independent experiments.

IFN-α-stimulated iNOS/NO expression and HO-1 down-regulation are mediated through the IFN-α receptor (IFNAR)/JAK1 signalling pathway

The binding of type I IFN (e.g. IFN-α) to its subunits of IFNAR1 and IFNAR2 induces subsequent phosphorylation of TYK2 and JAK1, respectively; while the binding of type II IFN (e.g. IFN-γ) receptors induces phosphorylation of JAK1and JAK2 (Borden et al. 2007). It has been reported that TYK2 deficiency reduces LPS-induced NO production in murine macrophages (Karaghiosoff et al. 2000). We therefore examined whether specific IFNARs and their associated proteins, Janus kinases (JAK), are present in BV-2 microglia and also involved in the signal transduction pathways leading to IFN-α-induced changes in iNOS, NO and HO-1. IFNAR1 mRNA was expressed in BV-2 microglia (Fig. 2 a). After IFN-α treatment, both JAK1 and TYK2 phosphorylation were expressed in a time-dependent manner (Fig. 2 b). Treatment with the JAK1 inhibitor (Pyridone 6) attenuated IFN-α-induced iNOS expression and reversed the HO-1 down-regulation induced by IFN-α (Fig. 2 c). Treatment with the JAK1 inhibitor could also attenuate NO production (Supplementary Fig. S3A). Correspondingly, transfection with siRNAs of JAK1 also attenuated IFN-α-induced iNOS (Fig. 2 d) and NO expression (Supplementary Fig. S3B). Taken together, these findings demonstrate that the IFNAR/JAK1 signalling pathway is specifically involved in IFN-α-induced changes of expression in iNOS, NO and HO-1.

Fig. 2

Involvement of IFN receptor signal in IFN-α-mediated iNOS expression and HO-1 down-regulation in BV-2 microglia. (a) Cells were incubated with IFN-α (3000 U/ml) for various time periods (4 or 8 h). IFNR1 expression was determined by RT–PCR. (b) Cells were incubated with IFN-α for various time periods (5, 10, 30, 60 or 120 min). The phosphorylation levels of JAK1 and TYK2 were analysed by Western blot. (c) Cells were pre-incubated with JAK1 inhibitor for 30 min followed by stimulation with IFN-α (3000 U/ml) for 24 h. The protein levels of iNOS and HO-1 were analysed by Western blot. Similar results were obtained from four independent experiments. (d) Cells were pre-transfected with JAK1 or controlled siRNA (100 nm) for 24 h followed by stimulation with IFN-α (3000 U/ml) for another 24 h, the protein levels of iNOS were analysed by Western blot, and similar results were obtained from three independent experiments.

IFN-α-stimulated iNOS/NO expression and HO-1 down-regulation mediated through the JNK signalling pathway

IFN-α-increased iNOS expression (Fig. 3 a) and NO production (Supplementary Fig. S4A) could be inhibited by JNK inhibitor (SP6100125), but not inhibitors of extracellular signal-regulated kinases (ERK) (PD98059), p38 (SB203580), PI3 kinase (LY294002), or AKT (AKTI). In Fig. 3 c, IFN-α resulted in JNK phosphorylation. Furthermore, JNK inhibitor SP6100125 attenuated IFN-α-induced iNOS expression (Fig. 3 c) and NO production (Supplementary Fig. S4B). The pre-transfection with JNK-DN also attenuated IFN-α-induced iNOS expression (Fig. 3 d) and NO production (Supplementary Fig. S4C) in a dose-dependent manner. Interestingly, treatment with JNK inhibitor (Fig. 3 c) or pre-transfection with JNK-DN (Fig. 3 d) also reversed IFN-α-induced HO-1 down-regulation. Taken together, these findings demonstrate that IFN-α-induced NOS/NO expression and HO-1 down-regulation are caused by JNK activation in microglia.

Fig. 3

JNK signalling pathway is involved in IFN-α-induced iNOS expression and HO-1 down-regulation in BV-2 microglia. (a) Cells were pre-treated with PD98059 (PD, 30 µm), SB203580 (SB, 10 µm), SP600125 (SP, 10 µm), or LY294002 (LY, 10 µm) for 30 min before application of IFN-α (3000 U/ml) for 24 h. The protein levels of iNOS were analysed by Western blot. (b) Cells were incubated with IFN-α (3000 U/ml) for various time periods (5, 10, 30, 60 or 120 min), and cell lysates were separated by SDS–PAGE and immunoblotted with anti-phospho(p)-JNK. (c) Cells were pre-treated with various concentrations (2, 5 or 10 µm) of SP600125 for 30 min before application of IFN-α (3000 U/ml) for 24 h. The protein levels of iNOS and HO-1 were analysed by Western blot. (d) Both iNOS and HO-1 expression are significantly different between IFN-α/- and IFN-α/SP600125 groups (one-way ANOVA with Bonferroni's correction). Cells were pre-transfected with various concentrations of DN-JNK for 24 h before application of IFN-α (3000 U/ml) for another 24 h. The protein levels of iNOS and HO-1 were analysed by Western blot. The levels of iNOS expression and HO-1 degradation are significantly different between IFN-α/- and IFN-α/DN-JNK groups (one-way ANOVA with Bonferroni's correction). Similar results were obtained from three independent experiments.

STAT1 and STAT3 signalling pathways are involved in IFN-α stimulation in BV-2 microglia

IFN binds IFNR to initiate signal transduction through phosphorylation of a family of signal transducers and activation of transcription (STAT) proteins (Leaman et al. 1996). We then investigated the involvement of STAT proteins in the effects of IFN-α on BV-2 microglia. IFN-α treatment resulted in both STAT1 (Tyr701) and STAT3 (Ser727) phosphorylation in a time-dependent manner (Fig. 4 a). Pre-transfection with STAT1 and STAT3 siRNA attenuated IFN-α-induced iNOS expression (Fig. 4 b) and reversed IFN-α-induced HO-1 down-regulation (Fig. 4 c). Taken together, IFN-α-induced iNOS expression and HO-1 down-regulation were via activation of STAT1 and STAT3 signalling pathways.

Fig. 4

Involvement of STAT1 and STAT3 in IFN-α-induced iNOS expression and HO-1 down-regulation in BV-2 microglia. Cells were incubated with IFN-α (3000 U/ml) for various time periods and cell lysates were separated by SDS–PAGE and determined with phospho(p)-STAT1 and p-STAT3. Cells were pre-transfected with STAT1, STAT3, or controlled siRNA for 24 h followed by stimulation with IFN-α (3000 U/ml) for another 24 h. The protein levels of (b) iNOS and (c) HO-1 were analysed by Western blot. Similar results were obtained from three independent experiments.

HO-1 regulates IFN-α-induced iNOS/NO expression in BV-2 microglia

Previous reports have shown that HO-1 is a critical regulator of NO (Tsoyi et al. 2008; Vareille et al. 2008) and plays a therapeutic role in neuropsychiatric diseases (Gozzelino et al. 2010; Jazwa & Cuadrado, 2010; Jeong et al. 2011; Lu et al. 2010b). Here, we have shown that treatment with the HO-1 activator CoppIX reversed IFN-α-induced HO-1 down-regulation (Fig. 5 a). The HO-1 activator markedly reversed IFN-α-induced iNOS (Fig. 5 b) and NO (Supplementary Fig. S5A) over-expression while pre-transfection with HO-1 siRNA aggravated IFN-α-induced iNOS (Fig. 5 b) and NO (Supplementary Fig. S5B) over-expression. In addition, treatment with the HO-1 inhibitor ZnppIX also increased iNOS (Fig. 5 c) and NO (Supplementary Fig. S5C) expression. Furthermore, our results also showed that the antidepressants desipramine and fluoxetine reduced IFN-α-induced iNOS expression (Fig. 5 d) and increased HO-1 expression (Fig. 5 e).

Fig. 5

Regulation effect of HO-1 expression in IFN-α-induced iNOS expression in BV-2 microglia. (a) Cells were pre-incubated with various concentrations of CoppIX for 30 min followed by stimulation with IFN-α for 24 h. The protein levels of iNOS and HO-1 were analysed by Western blot. (b) The levels of iNOS expression and HO-1 degradation are significantly different between IFN-α/- and IFN-α/CoppIX groups (one-way ANOVA with Bonferroni's correction). Cells were (b) pre-transfected with HO-1 or controlled siRNA for 24 h, or (c) pre-treated with ZnppIX (1 µm) for 30 min followed by stimulation with IFN-α for another 24 h. The protein levels of iNOS were analysed by Western blot. Similar results were obtained from three independent experiments. (d) Cells were pre-incubated with various concentrations of desipramine (Des) or fluoxetine (Flx) for 30 min followed by stimulation with IFN-α for 24 h, the protein levels of iNOS were analysed by Western blot. (e) Cells were pre-incubated with various concentrations of desipramine (Des) or fluoxetine (Flx) for 24 h, the protein levels of HO-1 was analysed by Western blot.

Discussion

Substantiating evidence reveals that IFN-α induces clinical depression in patients receiving cytokine therapy for cancers and viral infections. Despite much research on neurochemical and neuroendocrinological mechanisms from human and animal studies, direct molecular mechanisms of IFN-α at the cellular level are still lacking. The main finding of this study is that IFN-α potently induces iNOS and NO releases and down-regulates HO-1 expression via JAK1/JNK/STAT1 and STAT3 signalling pathways. To our knowledge, this is the first report to demonstrate the neuroinflammatory mechanisms of IFN-α in microglial cells.

First, the microglial activation induced by IFN-α contributes to the over-production of iNOS and NO, which is consistent with neuropathological findings characterizing depression (Maes et al. 2011). In addition, it has been reported that the over-expression of NO might be involved in chronic stress-induced depression-like behaviour in rats (Sevgi et al. 2006) and in IFN-α-induced depression in patients with chronic HCV infection (Suzuki et al. 2003). Consistently, NOS inhibitors demonstrated antidepressant effects in animal models of depression (da Silva et al. 2000; Ergun et al. 2006; Harkin et al. 1999). In addition, antidepressant agents, including specific serotonin reuptake inhibitors (SSRIs) (Finkel et al. 1996; Harkin et al. 2004), tricyclic antidepressants (TCAs) (Tai et al. 2006), a norepinephrine-dopamine reuptake inhibitor (NDRI, bupropion) (Dhir & Kulkarni, 2007), and omega-3 fatty acids (Lu et al. 2010b) have been found to attenuate the oxidative stress and inflammatory reaction by suppressing the expression of iNOS, NO and several pro-inflammatory cytokines, which are all consistent with the findings of our current study, i.e. that the antidepressants desipramine or fluoxetine reduce IFN-α-induced iNOS expression and increase HO-1 expression. Taken together, the findings from previous studies and our current study suggest that iNOS/NO pathway might be involved in IFN-α-induced depression.

Second, treatment with IFN-α, but not IFN-γ, down-regulated HO-1 expression in BV-2 microglia. The induction of HO-1 in brain neuronal and non-neuronal cells is important in neuroprotection and neuroplasticity (Choi & Kim, 2008; Gozzelino et al. 2010; Jazwa & Cuadrado, 2010; Jeong et al. 2011; Le et al. 1999), which plays a potential role in antidepressant mechanisms (Chen et al. 2005; Kim et al. 2008; Lu et al. 2010b; Oh et al. 2004; Shin et al. 2009; Tai et al. 2009). The HO-1 down-regulation induced by IFN-α might be further detrimental to the pathogenesis of depression. Therefore, the different effects on HO-1 with IFN-α and IFN-γ might provide a possible explanation that occurrence of neuropsychiatric adverse effects from IFN-α treatment is higher than those from IFN-γ treatment in clinical observations (Bocci, 1988, 1994; Meyers et al. 1991) and in animal models of depression (Makino et al. 1998, 2000).

The biological activities of IFNs are mediated by the interaction with specific receptors to initiate signalling cascades via JAK and the signal transducers and activators of transcription (STAT), known as JAK/STAT pathways (Borden et al. 2007). The binding of IFN-α to its subunits IFNAR1 and IFNAR2 induces subsequent phosphorylation of TYK2 and JAK1, respectively; while the binding of IFN-γ receptors induces phosphorylation of JAK1and JAK2 (Borden et al. 2007). Previous reports have shown that IFN-α activates ERK/MAPKs and PI3-kinase/AKT pathways in non-neural cells (Kalvakolanu, 2003; Platanias, 2003; Thyrell et al. 2004; Zhao et al. 2011). Here, we demonstrate for the first time the involvement of JAK1/JNK/STAT1 and STAT3, but not ERK and PI3-kinase/AKT, signal pathways in the effect of IFN-α on NO expression in murine BV-2 microglial cells. In our current study, both human and mouse IFN-α increased iNOS expression and down-regulated HO-1 expression and the dose–response curves of iNOS expression between mouse and human IFN-α increases are very similar, indicating that the effects did not have species specificity in our in vitro model. An important research question for future studies is whether there are different molecular responses in ex vivo or in vivo regulation of NO/iNOS/OH-1 by human and mouse IFN-α in freshly isolated (or primary cultured) microglia or macrophages.

It has been reported that IFN-γ stimulates iNOS expression via modulation of JAK1/JNK/STAT1 and STAT3, as well as the ERKs (signalling pathways in microglia) (Jung et al. 2010; Shen et al. 2005). Our results, however, did not support the involvement of an ERK pathway for IFN-α action. Interestingly, the activation of ERK phosphorylation seems to be associated with mechanisms of antidepressant effects of SSRIs (Li et al. 2008a, b; Mercier et al. 2004) and off-label agents used for treatment of depression (e.g. lithium, valproate, omega-3 PUFAs, and atypical antipsychotics) (Chen et al. 1999; Lu et al., 2004, 2010b; Lu & Dwyer, 2005). The different effects of IFN-α and IFN-γ on ERK phosphorylation also provide a possible explanation of different degree in their induction of neuropsychiatric adverse effects in clinical (Bocci, 1988, 1994; Meyers et al. 1991) and animal (Makino et al. 1998, 2000) studies.

In conclusion, this study demonstrates the regulatory mechanisms of IFN-α-induced iNOS/NO up-regulation and HO-1 down-regulation via JAK1/JNK/STAT1 and STAT3 signalling pathways in microglia. The different effects on HO-1 regulation and ERK phosphorylation might provide a possible explanation of different degree between IFN-α and IFN-γ in their induction of neuropsychiatric adverse effects in clinical and animal studies. The different molecular response to IFN-α in different cell types is an important research question for the future. The results from this study provide cellular mechanisms for inflammation theory in IFN-α-induced depression. The understanding of these regulatory signalling pathways might be helpful for the development of novel therapeutic strategies for clinical depression.

Supplementary material

For supplementary material accompanying this paper, visit http://dx.doi.org/10.1017/S1461145712000338.

Supplementary information supplied by authors.

Supplementary information supplied by authors.

Supplementary information supplied by authors.

Supplementary information supplied by authors.

Supplementary information supplied by authors.

Supplementary information supplied by authors.

Statement of Interest

None.

Acknowledgements

We thank Dr J.-S. Hong at the National Institute of Environmental Health Sciences, USA, to provide the BV-2 cell line and Dr M.-C. Tseng for technical support. The work was supported by the following grants: NSC 100-2627-B-039-005, NSC 100-2320-B-039-010, NSC 98-2320-B-039-009-MY2, NSC 98-2627-B-039-005, NSC-99-2911-I-039-002, NSC-98(99&100)-2627-B-039-003 and NSC 98-2628-B-039-020-MY3 from the National Science Council in Taiwan; NHRI-EX101-10144NI from the National Health Research Institute in Taiwan; CMU99-ASIA-20, CMU99-N1-18, CMU99-S-12, CMR99-114 and CMU97-336 from the China Medical University in Taiwan; and DOH100-TD-B-111-004 from the Department of Health Clinical Trial and Research Center of Excellence in Taiwan.

References

View Abstract