Tumor necrosis factor receptor-associated factor 5 is an essential mediator of ischemic brain infarction

Authors

  • Lang Wang,

    1. Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, PR China
    2. Cardiovascular Research Institute of Wuhan University, Wuhan, PR China
    Search for more papers by this author
    • Lang Wang and Yanyun Lu contributed equally to this work.
  • Yanyun Lu,

    1. Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, PR China
    2. Cardiovascular Research Institute of Wuhan University, Wuhan, PR China
    Search for more papers by this author
    • Lang Wang and Yanyun Lu contributed equally to this work.
  • Hongjing Guan,

    1. Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, PR China
    2. Cardiovascular Research Institute of Wuhan University, Wuhan, PR China
    Search for more papers by this author
  • Dingsheng Jiang,

    1. Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, PR China
    2. Cardiovascular Research Institute of Wuhan University, Wuhan, PR China
    Search for more papers by this author
  • Yu Guan,

    1. Department of Thoracic and Cardiovascular Surgery, Tongji Medical College, Tongji Hospital, Huazhong University of Science and Technology, Wuhan, PR China
    Search for more papers by this author
  • Xin Zhang,

    1. Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, PR China
    2. Cardiovascular Research Institute of Wuhan University, Wuhan, PR China
    Search for more papers by this author
  • Hiroyasu Nakano,

    1. Department of Immunology, Juntendo University Graduate School of Medicine, Tokyo, Japan
    Search for more papers by this author
  • Yan Zhou,

    1. College of Life Sciences, Wuhan University, Wuhan, PR China
    Search for more papers by this author
  • Yan Zhang,

    1. Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, PR China
    2. Cardiovascular Research Institute of Wuhan University, Wuhan, PR China
    Search for more papers by this author
  • Li Yang,

    1. Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, PR China
    2. Cardiovascular Research Institute of Wuhan University, Wuhan, PR China
    Search for more papers by this author
  • Hongliang Li

    Corresponding author
    1. Cardiovascular Research Institute of Wuhan University, Wuhan, PR China
    • Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, PR China
    Search for more papers by this author

Address correspondence and reprint requests to Hongliang Li, MD, PhD, Professor and Director, Renmin Hospital of Wuhan University, Jiefang Road 238, Wuhan 430060, PR China. E-mail: lihl@whu.edu.cn

Abstract

Tumor necrosis factor receptor-associated factor 5 (TRAF5) is an adaptor protein of the tumor necrosis factor (TNF) receptor superfamily and the interleukin-1 receptor/Toll-like receptor superfamily and plays important roles in regulating multiple signaling pathways. This study was conducted to investigate the role of TRAF5 in the context of brain ischemia/reperfusion (I/R) injury. Transient occlusion of the middle cerebral artery was performed on TRAF5 knockout mice (KO), neuron-specific TRAF5 transgene (TG), and the appropriate controls. Compared with the WT mice, the TRAF5 KO mice showed lower infarct volumes and better outcomes in the neurological tests. A low neuronal apoptosis level, an attenuated blood-brain barrier (BBB) disruption and an inhibited inflammatory response were exhibited in TRAF5 KO mice. TRAF5 TG mice exhibited an opposite phenotype. Moreover, the Akt/FoxO1 signaling pathway was enhanced in the ischemic brains of the TRAF5 KO mice. These results provide the first demonstration that TRAF5 is a critical mediator of I/R injury in an experimental stroke model. The Akt /FoxO1 signaling pathway probably plays an important role in the biological function of TRAF5 in this model.

Abbreviations used
BBB

blood–brain barrier

CBF

cerebral blood flow

COX2

cytochrome c oxidase subunit II

KO

knockout

MCA

middle cerebral artery

MCAO

middle cerebral artery occlusion

MCP-1

monocyte chemoattractant protein-1

MMP

matrix metalloproteinase

NTG

non-transgenic

OCT

ornithine carbamyl transferase

PBS

phosphate buffer solution

TG

transgene

TLR

toll-like receptor

tMCAO

transient middle cerebral artery occlusion

TNF

tumor necrosis factor

tPA

tissue plasminogen activator

TRAF5

Tumor necrosis factor receptor-associated factor 5

TTC

2,3,5-triphenyl-2H-tetrazolium chloride

WT

wild type

Stroke is the second leading cause of death and the most frequent cause of adult disability worldwide (Bronner et al. 1995; Tu 2010). Approximately, 80–85% of strokes are ischemic, resulting from a thrombus or embolus that leads to significantly decreased blood flow in a major cerebral artery, commonly the middle cerebral artery (MCA). More importantly, in addition to ischemia, reperfusion injury also induces stroke-related brain damage. Reperfusion strategies with recombinant human tissue plasminogen activator (tPA) have proven to be the most effective therapies for stroke treatment (Baldwin et al. 2010). However, the reperfusion of ischemic brain tissue can have harmful consequences, including an augmented inflammatory response and blood–brain barrier degradation, resulting in cerebral edema and/or brain hemorrhage, neurovascular injury, and neuronal death (Jung et al. 2010; Murray et al. 2010). Currently, only 2–4% of patients receive tPA therapy for acute ischemic stroke worldwide. This fact may be related to the narrowed therapeutic time window (3–4.5 h) needed to lessen the reperfusion injury (Missiou et al. 2010). Thus, identifying new molecular targets for cerebral I/R injury could increase the time window for tPA, decrease the risk of cerebral hemorrhage, and ultimately increase the overall efficacy of reperfusion therapy.

The tumor necrosis factor receptor-associated factor (TRAF) family consists of seven members (TRAF1-7) (Zotti et al. 2012). TRAFs are composed of an N-terminal cysteine/histidine-rich region containing a zinc RING and/or zinc finger motifs, a coiled-coil (leucine zipper) motif, and a C-terminal homology region that defines the TRAF family (Bradley and Pober 2001). TRAFs serve as adapter proteins for the TNF receptor superfamily and the IL-1R/TLR superfamily and participate in multiple biological functions, such as adaptive and innate immunity, tumorigenesis, embryonic development, and the stress response (Au and Yeh 2007; Zhou and Geahlen 2009; Clark et al. 2011; Xie et al. 2011). Recent studies have shown that TRAF proteins are markedly up-regulated in atherosclerotic plaques of human carotid arteries. Several TRAF members participate in atherosclerosis and neointima formation (Zirlik et al. 2007; Donners et al. 2008; Engel et al. 2009; Lutgens et al. 2010; Missiou et al. 2010).

In 1996, Ishida TK et al. and Nakano H et al. (Ishida et al. 1996; Nakano et al. 1996) identified TRAF5 as a putative signal transducer for CD40 and the lymphotoxin-β receptor to mediate nuclear factor-κB (NF-κB) activation. Subsequent studies have demonstrated that TRAF5 participates in T-cell expansion and innate immunity signaling following viral infection and regulates c-Jun N-terminal kinase (JNK) activation during atherogenesis (Au and Yeh 2007; Kraus et al. 2008; Missiou et al. 2010; Tang and Wang 2010). Recent evidence has shown that TRAF5 expression in glial cells and neurons is induced after spinal cord injury (Huan et al. 2012). In vitro study has revealed TRAF5 is involved in the regulation of glioma cell migration via NF-κB signaling (Tao et al. 2012). These results indicate that TRAF5 may play an important role in neurobiological diseases. In this study, we demonstrated that TRAF5 deficiency protects against brain I/R injury by attenuating blood-brain barrier (BBB) degradation and inflammation, whereas constitutive expression of TRAF5 in the brain exhibited the opposite phenotype. The results of our study suggest that TRAF5 is a crucial mediator of brain I/R injury.

Materials and methods

Animals

Experiments were performed on 10- to 12-week-old, male, wild-type (WT), TRAF5 knockout (KO), neuron-specific TRAF5 transgenic mice (TG) and its WT littermates (NTG), all the mice were on C57BL/6 background. The animals were fed on a normal chow diet and had ad libitum access to food and water and were maintained at 24°C under a 12 h light/dark cycle. The Animal Care and Use Committee of Renmin Hospital of Wuhan University approved the animal procedures used in this study. The animal experiments were performed in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80–23), revised in 1996. The TRAF5 KO mice have been previously described (Oh et al. 1999).

To generate TRAF5 transgenic mice, full-length mouse TRAF5 cDNA was cloned downstream of the neuron-specific promoter platelet-derived growth factor (PDGF). The resulting construct drives the preferential expression of TRAF5 within neuronal cell bodies in the cortex, hippocampus, and cerebellum. TG mice were produced via microinjection of the PDGF-TRAF5 construct into fertilized mouse embryos (C57BL/6 background). Four independent transgenic lines were established and examined. The TG mice were identified using PCR analysis of tail genomic DNA (forward primer: 5′-AAGGGTGGCAACTTCTCCTC-3′; reverse primer: 5′-ATAAGGAATGGACAGCAGGG-3′). The functional data and gene expression levels were analyzed in pairs of PDGF-TRAF5 (TG) and non-transgenic (NTG) male littermates that ranged in age from 7 to 8 weeks.

There was no macroscopic difference in the vessels of the circle of Willis and their branches among TRAF5 KO mice, WT mice, TRAF5 TG mice, and NTG mice as revealed by Indian ink staining (data not show).

Mouse cerebral ischemia/reperfusion model

A mouse cerebral I/R model was established using transient middle cerebral artery occlusion (tMCAO) procedures as previously described (Wang et al. 2012a, b). Briefly, the animals were anesthetized with 2.5–3% isoflurane in O2. The rectal temperature was maintained at 37 ± 0.5°C using a heating pad. A probe was affixed to the skull (2 mm posterior and 5 mm lateral to the bregma) and connected to a laser Doppler flowmetry instrument (Periflux System 5010; Perimed, Stockholm, Sweden) for continuous monitoring of the CBF. To achieve tMCAO, a 6-0 Silicon-coated monofilament surgical suture (Doccol, Redland, CA, USA) was inserted into the left external carotid artery, advanced into the internal carotid artery, and wedged into the cerebral arterial circle to obstruct the origin of the MCA. A > 80% decline in relative cerebral blood flow (CBF) confirmed an interruption of the CBF in the MCA. The filament was left in place for 45 or 60 min and withdrawn. A return to > 70% of basal CBF within 10 min of suture withdrawal confirmed a reperfusion of the MCA territory. Middle cerebral artery occlusion (MCAO) experiments were performed blindly.

A total of 150 mice were included in this study: WT mice (n = 73), KO mice (n = 45), NTG mice (n = 18), TG mice (n = 14) (see Table 1 for further details). The WT and KO mice underwent a 60-min tMCAO operation, whereas the NTG and TG mice underwent a 45-min tMCAO operation. The mortalities of WT mice, TRAF5 KO mice, NTG mice and TRAF5 TG mice are 27%, 16.67%, 5.26%, and 6.67%, respectively. In addition, sham operations were performed with WT mice (n = 12) and TRAF5 KO mice (n = 12). For the sham operation, mice were operated in the same way, but without occlusion of MCA. There was no death occurred in sham group.

Table 1. The number of animals for different experimental groups and various parameters
 WTTRAF5 KONTGTRAF5 TG
  1. a

    TRAF5 expression was only detected in WT mice. The same animals were used for detection of CBF, TTC staining and neurological score. The same animals were used for detection of TUNEL and Immunofluorescence staining of bax, bcl2, collagen IV, laminin, MMP2, MMP9, p65, p-65, TRAF5 at 24 h after tMCAO. The same animals were used for immunofluorescence staining of F4-80, 7/4, GFAP, TRAF5 at 72 h after tMCAO. Another 12 WT mice and 12 TRAF5 mice were used to perform sham operation and served as sham control for quantitative real-time PCR and western blot analysis.

Cerebral blood flow, TTC staining and neurological score8988

Evans blue

Extravasation

109
TUNEL staining and immunofluorescence staining (for bax, bcl2, collagen IV, laminin, MMP2, MMP9, p65, p-65, TRAF5)6764
Immunofluorescence staining (for F4-80, 7/4, GFAP, aTRAF5)5444
Quantitative real-time PCR66
Western blot analysis66
Western blot analysis for aTRAF5 expression after tMCAO18 
In situ zymography44
Total73451814

Neurological deficit scores

At 24 h or three days after the tMCAO, neurological deficits were assessed using a 9-point scale (Won et al. 2006). No neurological deficit was scored as 0; left forelimb flexion upon suspension by the tail or failure to fully extend the right forepaw was scored as 1; left shoulder adduction upon suspension by the tail was scored as 2; reduced resistance to a lateral push toward the left was scored as 3; spontaneous movement in all directions, with circling to the left exhibited only if pulled by the tail, was scored as 4; circling or walking spontaneously only to the left was scored as 5; walking only when stimulated was scored as 6; no response to stimulation was scored as 7; and stroke-related death was scored as 8. Neurological deficit measurements were performed in a blinded way.

Measurement of the infarct volume

The infarct volume and swelling were measured at 24 h following tMCAO using 2,3,5-triphenyl-2H-tetrazolium chloride (TTC) staining. The brains were dissected into 1 mm thick coronal sections using a mouse brain matrix. The sections were stained with 2% TTC in phosphate buffer (pH 7.4) for 15 min at 37°C. The sections were subsequently transferred to a 10% formalin solution and fixed overnight. The fixed sections were photographed, and the volume of the infarct area was quantified using Image-Pro Plus 6.0 (Media Cybernatics, Bethesda, MD). To correct for the effects of edema, the area of the infarction was measured by subtracting the area of the non-lesioned ipsilateral hemisphere from that of the contralateral hemisphere. The volume of the infarction was calculated by integrating the lesion areas at the seven measured levels of the brain. The measurements of infarct volume and edema volume were performed blindly.

Quantitative evaluation of Evans blue extravasation

The vascular permeability was quantitatively evaluated using fluorescent detection of extravasated Evans blue dye (Uyama et al. 1988). Briefly, 2% Evans blue in phosphate buffer solution (PBS) was infused (4 mL/kg, i.v.) as a BBB permeability tracer at the onset of reperfusion. At 23 h after reperfusion, the mice were deeply anesthetized with sodium pentobarbital and transcardially perfused with ice-cold PBS (100 mm Hg, 5 min) to remove the intravascular dye. The brains were removed and divided into ipsilateral ischemic and contralateral non-ischemic hemispheres. The ipsilateral ischemic hemispheres were immediately frozen in liquid nitrogen and stored at −80°C until further analysis. The brain samples were homogenized in 1 mL of 50% trichloroacetic acid and centrifuged (7 378 g, 20 min). The supernatant was diluted fourfold with ethanol. A fluorescent plate reader (620-nm excitation, 680-nm emission) was used to quantify the dye concentrations. The calculations were based on external standards (50–1000 ng/mL) dissolved in the same solvent (1 : 3; 50% trichloroacetic acid : ethanol). The amount of extravasated Evans blue was quantified as milligrams of Evans blue per gram of ischemic hemisphere tissue. This experiment was performed in a blind manner.

Immunofluorescence staining

The mice were anesthetized at 24 or 72 h after tMCAO using sodium pentobarbital. The mice were subsequently perfused via the left ventricle with 0.1 mol/L sodium phosphate buffer under 100 mm Hg of pressure for 5 min, followed by perfusion with a fixative solution containing 4% paraformaldehyde in a 0.1 mol/L phosphate buffer (pH 7.4) for 15 min. The brains were carefully removed and post-fixed for 6–8 h in the same fixative solution at 25°C. Subsequently, the brains were immersed overnight at 4°C in a 0.1 mol/L phosphate buffer containing 30% sucrose. The brains were embedded in ornithine carbamyl transferase (OCT), and serial frontal sections were cut with a cryostat microtome. For immunofluorescence staining, the sections were washed in PBS containing 10% goat serum. The sections were incubated overnight with primary antibody solutions at 4°C. The following primary antibodies were used: anti-NeuN (1 : 200 dilution, Millipore, Bedford, MA, USA), anti-matrix metalloproteinase (MMP) 2 (1 : 50 dilution), anti-MMP9 (1 : 50 dilution), anti-phosphorylated p65 (1 : 50 dilution), anti-p65 (1 : 50 dilution), anti-bax (1 : 50 dilution) (all from Cell Signaling Technology, Danvers, MA, USA), anti-collagen IV (1 : 50 dilution), anti-bcl2 (1 : 50 dilution), anti-7/4 (1 : 50 dilution) (all from Abcam, Cambridge, MA, USA), anti-TRAF5 (1 : 50 dilution, Santa Cruz, Santa Cruz, CA, USA), anti-glial fibrillary acidic protein (GFAP) (1 : 100 dilution, Epitomics, Burlingame, CA, USA), anti-laminin (1 : 50 dilution, Sigma, St. Louis, MO, USA), and anti-F4/80 (1 : 50 dilution, Serotec, Raleigh, NC, USA). After the sections were washed in PBS, they were incubated in the appropriate secondary antibody for 1 h. The secondary antibodies included (i) anti-rat IgG Alexa Fluor 555 Conjugate (Cell Signaling Technology), (ii) anti-mouse IgG Alexa Flour 568 Conjugate (Invitrogen, Carlsbad, CA, USA), and (iii) anti-rabbit IgG Alexa Flour 568 Conjugate (Invitrogen). Lastly, the nuclei were labeled with 4′,6-diamidino-2-phenylindole (DAPI). The visualization was performed using fluorescence microscopy (OLYMPUS DX51, Olympus, Tokyo, Japan) with the DP2-BSW software (version 2.2). The image analysis was performed using Image-Pro Plus 6.0. The image analysis was performed blindly.

TUNEL staining

Twenty-four hour following the onset of ischemia, the brains were collected and sectioned as described above. For NeuN immunofluorescence staining, the sections were washed in a PBS containing 10% goat serum and 0.1% Triton X-100. The sections were subsequently incubated with the anti-NeuN antibody for 2 h, followed by an additional incubation with secondary antibody for 1 h at 37°C. Following the completion of NeuN immunofluorescence staining, Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) was performed using the In Situ Cell Death Detection Kit (Roche Diagnostics, Indianapolis, IN, USA), following the manufacturer's protocol. The nuclei were labeled with DAPI, and DNA fragmentation was quantified under high-power magnification (200x). The percentages of TUNEL-positive cells relative to the DAPI-positive cells were counted by an investigator blinded to the experimental groups. The image analysis was performed blindly.

Tissue preparation

For quantitative real-time PCR and western blot analysis, the mice were anesthetized with sodium pentobarbital and perfused via the left ventricle with cold sodium phosphate; the brains were then immediately removed. To collect the tissue in an unbiased manner that reflected the global extent of the infarcts, the olfactory bulbs and 1-mm sections of anterior and posterior brain tissue were excised from the mice, then the left hemispheres of the remaining brains were collected. The brain tissue was immediately frozen in liquid nitrogen and subsequently transferred to −80°C for storage.

Quantitative real-time PCR

Total RNA was prepared from the frozen tissue using TRIZOL reagent (Invitrogen), and the cDNA was reverse transcribed from 2 μg of RNA from each sample using the Transcriptor First Strand cDNA Synthesis Kit (Roche). The specific mRNA expression levels of vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), monocyte chemoattractant protein-1 (MCP-1), interleukin (IL) -1β, TNF-α, inducible nitric oxide synthases (iNOS), cytochrome c oxidase subunit II (COX2), MMP2, and MMP9 were normalized to the levels of GAPDH mRNA to calculate the relative expression levels. Quantitative RT-PCR analysis was performed using the LightCycler 480 SYBR Green 1 Master Mix (Roche) and the LightCycler 480 QPCR System (Roche). The sequence-specific primers for TNF-α, MCP-1, IL-1β, ICAM, and GAPDH have been described in our previous study (Wang et al. 2012b). The following primer sequences were used for to analyze the MMP2, MMP9, iNOS, and COX2 mRNA levels:

MMP2 forward: 5′-TTTGCTCGGGCCTTAAAAGTAT-3′;

MMP2 reverse: 5′-CCATCAAACGGGTATCCATCTC-3′;

MMP9 forward: 5′-CGGACCCGAAGCGGACAT-3′;

MMP9 reverse: 5′-GGGGCACCATTTGAGTTT-3′;

iNOS forward: 5′-TGCGCCTTTGCTCATGACATCGA-3′;

iNOS reverse: 5′-ATGGATGCTGCTGAGGGCTCTGTT-3′;

COX2 forward: 5′-GGTCTACAAGACGCCACATCCCCT-3′;

COX2 reverse: 5′-TCGGCCTGGGATGGCATCAGTT-3′.

For statistic analysis, the mRNA levels of WT mice with sham operation were defined as 1, the quantitative values of other groups were expressed by there fold changes compared with WT-sham group.

Western blot analysis

Western blotting was conducted to determine the protein levels of phosphorylated IKK α/β (ser176/180, 1 : 1000 dilution), IKK β (1 : 1000 dilution), NF-κB p65 (1 : 1000 dilution), cleaved caspase3 (Asp175, 1 : 500 dilution), Akt (1 : 1000 dilution), phosphorylated Akt (Ser473, 1 : 1000 dilution), FoxO1 (1 : 1000 dilution), and phosphorylated FoxO1 (Ser256, 1 : 1000 dilution) (all from Cell Signaling Technology), phosphorylated IκBα (Ser32/36, 1 : 1000 dilution), phosphorylated NF-κB p65 (ser536, 1 : 1000 dilution) (both from Bioworld Technology, Minneapolis, MN, USA). The western blot analyses were performed using 50 μg of extracted protein that was separated on 8–12% SDS-PAGE gels. The proteins were subsequently transferred to polyvinylidene difluoride (PVDF) membranes (Millipore). The blocking of the membranes (with 5% skim milk powder), the PBS washes and the secondary antibody (Goat anti-Rabbit IRDye 800CW or Goat anti-Mouse IRDye 800CW; Licor Biosciences, Lincoln, NE, USA) incubations were all performed at 25°C for 1 h. The membranes were incubated with the primary antibodies overnight at 4°C. The signals were detected using the Odyssey infrared imaging system (Licor Biosciences). The specific protein expression levels were normalized to the GAPDH protein levels (Cell Signaling Technology).

In situ zymography

To localize the net gelatinolytic activity of MMPs by in situ zymography, FITC-labeled intramolecularly quenched intramolecular DQ gelatin [available in a gelatinase/collagenase assay kit (EnzChek; Molecular Probes, Eugene, OR, USA)] was used as a substrate for gelatinase degradation (Nakano et al. 1999). Gelatinase-induced proteolysis yields cleaved gelatin-FITC fluorescent peptides. The localization of the fluorescence indicates the sites of net gelatinolytic activity. Twenty-four hours following tMCAO, the brains were dissected and rinsed with cold PBS. The brains were subsequently immersed in an OCT compound (Tissue-Tek, Torrance, CA, USA) and snap frozen into a block on dry ice. Once embedded in the OCT block, the optic nerves were cut into 20-μm sections using a cryostat (Leica, Wetzler, Germany) and sequentially collected. The brain sections were stored at −80°C until further use for in situ zymography. For in situ zymography, the sections were thawed and incubated overnight in a reaction buffer (0.05 M Tris-HCl, 0.15 M NaCl, 5 mM CaCl2, and 0.2 mM NaN3, pH 7.6) containing 40 mg/mL DQ gelatin. At the end of the incubation period and without fixation or washes, the gelatinolytic activity of MMPs was localized and photographed using fluorescence microscopy. The images were acquired using a Spot digital camera with a computer imaging program (Image-Pro Plus).

Statistical analysis

The data were expressed as the means ± SD except neurological deficit scores. Differences among the groups were determined using anova, followed by a post hoc Tukey test. Comparisons between the two groups were performed using the unpaired Student's t-test. Neurological deficit scores were presented as a scatterplot with a horizontal line to indicate the median. A rank-sum test was used to compare the data of neurological deficit scores. A p-value of < 0.05 was accepted as statistically significant.

Results

Endogenous TRAF5 expression is up-regulated in the ischemic brain

Immunofluorescence staining showed that TRAF5 was minimally expressed in the contralateral hemisphere of WT mice. The cerebral I/R injury resulted in a significant increase in TRAF5 protein expression in the neurons (Fig. 1a). The TRAF5 expression increased by nearly twofold after 24 h and further increased by threefold after 72 h (Fig. 1b).Western blot analysis showed a consistent change (Fig. 1c). We further examine the TRAF5 expression in TRAF5 KO mice and TG mice. As showed in (Fig. 1d), the TRAF5 expression was absent in KO mice whether in ischemic hemisphere or in contralateral hemisphere. The TG mice exerted an up-regulated expression of TRAF5 in both ischemic hemisphere and contralateral hemisphere compared with NTG mice.

Figure 1.

Tumor necrosis factor receptor-associated factor 5 (TRAF5) expression is up-regulated in the ischemic brain. Immunofluorescence staining showed that TRAF5 is minimally expressed in the contralateral hemisphere and significantly induced in the neurons of peri-infarct area after I/R injury (a). TRAF5 expression was doubled at 24 h following the ischemic onset and tripled at 72 h (b) in the ischemic brain of wild-type (WT) mice (n = 4–6, *< 0.05 vs. Ctrl.). Western blot analysis also showed that TRAF5 expression doubled at 24 h and tripled at 72 h after ischemic onset in WT mice (c) (n = 6, *< 0.05 vs. Sham). Immunofluorescence staining demonstrated that TRAF5 expression was absent in knockout (KO) mice whether in ischemic hemisphere or in contralateral hemisphere. The transgene (TG) mice exerted an up-regulated expression of TRAF5 in both ischemic hemisphere and contralateral hemisphere compared with NTG mice (d).

TRAF5 deficiency improves stroke outcome following cerebral ischemia/reperfusion

A tMCAO protocol was performed on both the TRAF5 KO and WT mice. Laser Doppler measurements revealed no difference in regional CBF before MCAO between TRAF5 KO and WT mice and showed a similar alteration upon MCAO and reperfusion (Fig. 2a). The TRAF5 KO mice exhibited much smaller infarct and edema volumes (Fig. 2b) at 24 h following the ischemia/reperfusion injury compared with the WT mice (26% and 47% reductions, p = 0.02 and p = 0.039, respectively). In addition, the TRAF5 KO mice exhibited better scores in the neurological assessment scale following the tMCAO (p < 0.05) (Fig. 2b).

Figure 2.

The effects of tumor necrosis factor receptor-associated factor 5 (TRAF5) on brain injury following I/R injury. Laser Doppler measurements revealed no difference in regional cerebral blood flow (CBF) before middle cerebral artery occlusion (MCAO) between wild-type (WT) and TRAF5 knockout (KO) mice and showed a similar alteration upon MCAO and reperfusion (a). The infarct and edema volumes at 24 h following the tMCAO were determined using 2,3,5-triphenyl-2H-tetrazolium chloride (TTC) staining (b). The TRAF5 KO mice exhibited lower infarct and edema volumes and better outcomes based on neurological assessments (c). (n = 8–9, *< 0.05 vs. WT).

TRAF5 deficiency suppresses neuronal apoptosis

Neuronal apoptosis was evaluated using TUNEL and NeuN double staining at 24 h following cerebral I/R. Compared with the WT group, the ratio of TUNEL-positive cells to the total number of cells in the TRAF5 KO mice was considerably lower (24.99 ± 2.28% vs. 39.93 ± 5.35%, p < 0.001) following tMCAO (Fig. 3a). The pro-apoptotic gene Bax and anti-apoptotic gene Bcl-2, as well as other apoptotic markers such as Fas, FasL, cleaved caspase3 (c-caspase3) were estimated by immunofluorescence staining or quantitative real-time PCR or western blot analysis in the sham/control brain and the ischemic brain. There was no baseline difference between WT and KO mice in the sham/contralateral brain. The mRNA and protein levels of Bax were decreased in the ischemic brain of the KO mice, whereas the expression of the B-cell lymphoma-2 (Bcl-2) was increased (Fig. 3b–d) compared with WT mice. The mRNA levels of Fas and FasL were also down-regulated in the TRAF5 KO mice. Correspondingly, the protein level of cleaved caspase3 (c-caspase3) was decreased in the ischemic brain of TRAF5 KO mice (Fig. 3d).

Figure 3.

The effect of tumor necrosis factor receptor-associated factor 5 (TRAF5) on neuronal apoptosis in the ischemic brain. At 24 h following the ischemic onset, most of the observed apoptosis (TUNEL-positive signal) was observed in neurons (NeuN-positive) in the peri-infarct area. The percentage of TUNEL-positive cells was lower in the TRAF5 KO mice (a) (n = 5–6, *< 0.05 vs. WT). The TRAF5 KO mice exhibited decreased protein level of bax and cleaved-caspase 3 and increased Bcl-2 expression (b and d) in ischemic brain (n = 5–6, *< 0.05 vs. WT), but no difference can be detected between WT-sham group and KO-sham group. The mRNA levels of Fas, FasL, and bax were elevated after I/R injury, TRAF5 deficiency significantly inhibited the elevations. The mRNA level Bcl-2 was increased in the ischemic brains of TRAF5 KO mice (c) (n = 6, *< 0.05 vs. WT, #< 0.05 vs. Sham).

BBB disruption is attenuated in TRAF5 KO mice after cerebral ischemia/reperfusion

Fluorescence spectrophotometric analysis of Evans blue standards showed a linear correlation between fluorescence intensity and dye concentration within the 50–1000 ng/mL range (data not shown), thus allowing for a quantitative measurement of dye concentrations in the ischemic mouse brains. In all mice reperfused with Evans blue, leakage of the dye into brain parenchyma was observed at 24 h after ischemia. The severity of BBB disruption in the ischemic brains, expressed as Evans blue extravasation per gram tissue in the ischemic hemisphere, was significantly reduced in the KO mice compared with the WT mice (μg/g tissue in hemisphere) (Fig. 4a).

Figure 4.

The effect of tumor necrosis factor receptor-associated factor 5 (TRAF5) on the blood-brain barrier (BBB) disruption after I/R injury. Evans blue extravasation assessment showed that the BBB disruption was significantly reduced in the TRAF5 KO mice compared with wild-type mice (a) (n = 9–10, *< 0.05 vs. WT). The ischemic degradation of collagen IV and laminin was significantly reduced in the TRAF5 KO mice compared with WT mice (b) (n = 4–5, *< 0.05 vs. WT). The protein level of occludin was higher in the ischemic brain of the TRAF5 KO mice compared with WT mice (c). (n = 6, *< 0.05 vs. WT).

The basal lamina components collagen IV and laminin were estimated using immunofluorescence staining. I/R injury resulted in a degradation of collagen IV and laminin at 24 h after ischemia (compared with the contralateral hemisphere). The ischemic degradation of collagen IV and laminin was significantly reduced in the KO mice compared with the WT mice (Fig. 4b), and the continuity of collagen IV and laminin was better in the KO mice. The tight junction protein occludin was detected using western blot analysis. The protein level of occludin was higher in the ischemic brain of KO mice compared with WT mice (Fig. 4c). There was no difference between KO and WT in the sham brain.

Expression and activity of MMPs are decreased in TRAF5 KO mice

MMP-2 and MMP-9 facilitate inflammatory recruitment and mediate brain injury during cerebral ischemia. The mRNA and protein levels of MMP2 and MMP9 were detected using real-time quantitative PCR and western blot analysis, respectively. The TRAF5 deficiency was associated with a decrease in MMP2 and MMP9 mRNA and protein levels in the infarct brain (Fig. 5a–c). However, there was no difference between WT and KO mice with sham operation. In situ zymography indicate that the gelatinolytic activities of the MMPs were inhibited in the KO mice following the brain I/R injury (Fig. 5d).

Figure 5.

The effect of tumor necrosis factor receptor-associated factor 5 (TRAF5) on matrix metalloproteinase (MMP) expression and activity. Immunofluorescence staining indicated that MMP2 and MMP9 expression in the peri-infarct area was reduced in TRAF5 KO mice following the ischemic insult (a) (n = 3–4, *< 0.05 vs. WT). The mRNA (b) levels of MMP2 and MMP9 were elevated after I/R injury, TRAF5 deficiency markedly mitigated the elevation (n = 6, *< 0.05 vs. WT, #< 0.05 vs. Sham). Protein levels (c) of MMP2 and MMP9 were also down-regulated in the ischemic brains of TRAF5 KO mice (n = 6, *< 0.05 vs. WT). The gelatinolytic activity of MMPs, as shown using in situ zymography, was inhibited in the TRAF5 KO mice (d) (n = 3, *p < 0.05 vs. WT).

The deletion of TRAF5 inhibits ischemia-induced inflammation and NF-κB activity

We analyzed the inflammatory cell recruitment and expression of pro-inflammatory markers in brain homogenates isolated from the ischemic hemisphere of I/R-injured mice. Inflammatory cell recruitment was detected using immunofluorescence staining. An anti-7/4 antibody was used to detect neutrophils, an anti-F4/80 antibody was used to detect macrophages/microglia, and an anti-GFAP antibody was used to detect astrocytes. Inflammatory cell recruitment was dramatically reduced in the ischemic brain of KO mice relative to the WT controls (Fig. 6a). Acute expression of inflammatory mediators, such as iNOS, COX-2, and pro-inflammatory cytokines, including TNF-α, IL-1β, MCP-1, and ICAM-1, participate in the mediation of brain damage following stroke. The TRAF5 deficiency resulted in decreased mRNA levels of iNOS, COX-2, TNF-α, IL-1β, MCP-1, and ICAM-1 at 24 h following the ischemic onset (Fig. 6b). There was rare inflammatory cell recruitment in the contralateral hemisphere, and minimal inflammatory gene expression in the sham brains in both WT and KO mice.

Figure 6.

The effect of tumor necrosis factor receptor-associated factor 5 (TRAF5) on the ischemia-induced inflammatory response. The recruitment of neutrophils, microglia, and astrocytes was inhibited in the peri-infarct area of the TRAF5 KO mice (a) (n = 4–5, *< 0.05 vs. WT). The mRNA levels of tumor necrosis factor (TNF)α, monocyte chemoattractant protein-1 (MCP-1), IL-1β, ICAM-1, cytochrome c oxidase subunit II (COX-2), and iNOS were elevated after cerebral I/R injury. TRAF5 KO mice exhibited a suppressed elevation of mRNA levels of inflammatory genes compared with WT mice (b) (n = 6, *< 0.05 vs. WT, #< 0.05 vs. Sham). The expression levels of p-p65 and p65 in the peri-infarct area nuclei were also lower in the TRAF5 KO mice (c) (n = 4–5, *< 0.05 vs. WT). The phosphorylation of IκBα, IKKβ, and p65 was inhibited in the ischemic brains of TRAF5 KO mice (d) (n = 6, *< 0.05 vs. WT).

NF-κB is a transcription factor that plays a key role in mediating the expression of a variety of genes involved in inflammatory responses. As a sign of NF-κB activation, the protein levels of p65 and phosphorylated p65 were determined at 6 h following ischemic onset. The TRAF5 deficiency inhibited NF-κB activity, as the nuclear translocation and phosphorylation of p65 were attenuated (Fig. 6c and d), and the cytoplasmic levels of phosphorylated IκBα and IKKβ, which also indicate NF-κB activation, were decreased (Fig. 6d). There was no difference between WT mice and KO mice in the baseline activity of NF-κB signaling in sham groups.

The deletion of TRAF5 enhances the Akt/FoxO1 pathway in the ischemic brain

The brains of KO and WT mice were isolated for western blot analysis to explore the potential molecular mechanisms underlying TRAF5 deficiency-mediated neuronal protection in the cerebral I/R model. We observed that the phosphorylation of Akt and its downstream transcription factor Foxo1 was up-regulated in the TRAF5 KO mice compared with the WT mice (Fig. 7). Akt is activated by phosphorylation at Ser473, a residue that lies within its carboxy terminus. FoxO1 can be inactivated through the phosphorylation by Akt at Thr24, Ser256, and Ser319, which results in nuclear export and inhibition of transcription factor activity.

Figure 7.

The effect of tumor necrosis factor receptor-associated factor 5 (TRAF5) on Akt/FoxO1 signaling activation in the ischemic brain. The protein levels of phosphorylated Akt and FoxO1 were elevated in the ischemic brain of TRAF5 KO mice. The total protein levels of Akt and FoxO1 did not change. There was no difference in the Akt/FoxO1 signaling activation between sham groups (n = 6, *< 0.05 vs. WT).

Neuron-specific TRAF5 transgene expression potentiates brain injury in mice following focal cerebral ischemia/reperfusion

The above findings suggested that TRAF5 deletion mitigates brain injury following cerebral ischemia. To confirm this hypothesis, we generated TG mice that constitutively express full-length mouse TRAF5 cDNA under the control of the PDGF promoter (Fig. 8a). Four lines of TG mice were confirmed by PCR analysis (Fig. 8b). All of the experiments that are reported in this article were performed using male mice that were 78 weeks old. We analyzed TRAF5 protein levels in various tissues by western blot analysis using an anti-TRAF5 antibody. We observed a robust expression of TRAF5 protein in the brain, but the expression in other organs was very low (Fig. 8c).

Figure 8.

Ischemic outcomes in neuron-specific tumor necrosis factor receptor-associated factor 5 (TRAF5) transgenic mice. Full-length mouse TRAF5 cDNA was produced under the control of the PDGF promoter (a). Four lines of TG mice were confirmed by PCR analysis (b). Representative blots for TRAF5 expression in various tissues of TG mice indicated robust TRAF5 protein expression in the brain, whereas low TRAF5 expression was observed in other organs (c). Representative blots for TRAF5 protein levels in heart tissue from four lines of both TG and NTG mice (d). The TRAF5 TG mice exhibited larger infarct and edema volumes. The TRAF5 TG mice also exhibited poorer outcomes based on neurological assessments (e) (n = 8, *< 0.05 vs. NTG). Neuronal apoptosis (f and g), blood-brain barrier (BBB) disruption (h) and inflammatory cell recruitment (i) were increased in the TRAF5 TG mice (n = 4–6, *< 0.05 vs. NTG).

Among the four established lines of TG mice, the TG1 and TG4 mouse line were used to perform a 45-min tMCAO model (Fig. 8d). There were no differences in the infarct or edema volumes or the neurological assessment score between the two TG lines (data was no showed). We used the TG1 line in the following investigations. The TRAF5 TG mice exhibited much larger infarct and edema volumes at 24 h following ischemic injury compared with the NTG mice. Furthermore, the TRAF5 TG mice exhibited poorer neurological assessment scores following tMCAO (Fig. 8e).

TUNEL and NeuN double staining revealed that the percentage of TUNEL-positive cells in the peri-infarct area of the TRAF5 TG mice was twice that of the WT mice (Fig. 8f). The expression of Bcl-2 was down-regulated, whereas Bax expression was up-regulated (Fig. 8g). The immunofluorescence staining revealed that the degradation of collagen IV and laminin was enhanced in the ischemic brain of the TRAF5 TG mice, suggesting that the BBB disruption was aggravated (Fig. 8h). The recruitment of neutrophils, macrophages/microglia and astrocytes in the peri-infarct area was increased in the TRAF5 TG mice compared with the WT mice (Fig. 8i). These results indicate that neuronal apoptosis, BBB disruption and post-ischemic inflammation are potentiated by TRAF5.

Discussion

In this study, we identify TRAF5 as a critical mediator of ischemic brain injury in an experimental model of cerebral I/R injury. The following major novel findings were revealed in this study: (i) the expression of TRAF5 is markedly induced in the ischemic brain; (ii) the deletion of TRAF5 improves stroke outcome following transient cerebral ischemia; (iii) BBB degradation and post-ischemic inflammation are inhibited in TRAF5 KO mice; (iv) TRAF5 deficiency leads to a significant activation of the Akt/FoxO1 pathway in the ischemic brain; and (v) a neuron-specific TRAF5 transgene markedly aggravates neuronal apoptosis, BBB degradation and post-ischemic inflammation, leading to exacerbated injury following cerebral I/R injury.

It has previously been demonstrated that TRAF5 expression is increased after spinal cord injury (Akiba et al. 1998; Brown et al. 2011; Huan et al. 2012). Here, we observed that TRAF5 expression is obviously induced in neurons at 24 and 72 h in the peri-infarct area following brain ischemic injury. We examined TRAF5 KO and neuron-specific TRAF5 TG mice to investigate the function of TRAF5 in cerebral ischemia. The deletion of TRAF5 revealed a protective effect after cerebral I/R injury by impairing neuronal apoptosis, attenuating BBB disruption, and inhibiting the inflammatory response, whereas TRAF5 over-expression resulted in aggravated brain injury. These results strongly suggest that TRAF5 plays an important role in the pathological processes of cerebral I/R injury.

Neuronal apoptosis is an important pathological process of ischemic stroke (Doyle et al. 2008; Brouns and De Deyn 2009). After ischemia onset, neurons in the ischemic core are irreversibly injured within minutes, whereas the tissue in the penumbra is damaged but not dead. The fate of neurons in the peri-infarct area is critical for stroke outcomes. Although reperfusion rescues the tissues in the penumbra by restoring the blood flow, this process can paradoxically result in secondary reperfusion injury. In this study, neuronal apoptosis in the peri-infarct area was decreased in the ischemic brain of TRAF5 KO mice but increased in the neuron-specific TRAF5 TG mice, and a corresponding change in the expression of molecular markers for apoptosis was also observed. These results indicate that TRAF5 is important for apoptotic neuronal death after I/R injury.

BBB degradation is another pathological process of cerebral I/R that can result in cerebral edema, brain hemorrhage, and neuronal death through apoptosis/necrosis (Jin et al. 2010). A number of studies using in vivo and in vitro models have indicated that hypoxia/reoxygenation or I/R leads to increased permeability and/or degradation of BBB tight junctions and basal lamina components (Dimitrijevic et al. 2006; McColl et al. 2008; Sandoval and Witt 2008). In this study, an Evans blue leakage analysis showed that the BBB permeability was reduced in the ischemic brain of TRAF5 KO mice. The degradation of the basal lamina components collagen IV and laminin was inhibited in the TRAF5 KO mice, and the level of the tight junction protein occludin was higher in the TRAF5 KO mice. However, the constitutive expression of TRAF5 in the brain resulted in increased BBB permeability and promoted collagen IV and laminin degradation. These findings strongly suggest that TRAF5 is a critical mediator of BBB degradation after cerebral I/R.

Accumulating evidence suggests that MMP2 and MMP9 are up-regulated after ischemia and play important roles in BBB disruption (Asahi et al. 2001; Rosell et al. 2008). In the ischemic brain, MMP2 and MMP9 are synthesized in both neuronal and non-neuronal cells. Collagen IV acts as a substrate of both MMP2 and MMP9, while laminin is only a substrate of MMP9. It has also been demonstrated that MMP2 and/or MMP9 can degrade tight junction proteins, including ZO-1, occludin and claudin5 (Feng et al. 2011). In this study, the down-regulated expression and activity of MMP2 and MMP9 in TRAF5 KO mice probably account for the attenuated BBB disruption. MMP2 and MMP9 can also directly act on neurons and induce neuronal apoptosis. Thus, the decreased neuronal apoptosis in TRAF5 KO mice may partially attribute to the down-regulation of MMP2 and MMP9.

Post-ischemic inflammation can aggravate both BBB degradation and neuronal apoptosis (Terao et al. 2008; Lakhan et al. 2009). Post-ischemic inflammation is initiated through the synthesis of adhesion molecules and chemokines in injured brain cells and is exacerbated by the recruitment of circulating inflammatory cells (Jin et al. 2010; Supanc et al. 2011). Injured neurons, activated glial cells, and recruited leukocytes secrete cytokine and MMPs that further exacerbate tissue injury (Yang et al. 2010; Feng et al. 2011; del Zoppo et al. 2012). In this study, the TRAF5 deficiency inhibited inflammatory gene expression and inflammatory cell recruitment. In contrast, the over-expression of TRAF5 promoted the inflammatory response following cerebral I/R. The deletion of TRAF5 also inhibits the activity of NF-κB, a crucial transcription factor that mediates the expression of inflammatory genes and MMP2 and MMP9. Previous studies have shown that TRAF5 is involved in NF-κB activation through CD27, CD30, lymphotoxin-β receptor, and the tumor necrosis factor-like weak inducer of apoptosis (TWEAK) signal transduction pathways in multiple cell types (Akiba et al. 1998; Saitoh et al. 2002, 2003). However, this is the first time that TRAF5 has been demonstrated as a regulator for NF-κB activation in a pathological state of I/R.

The finding that TRAF5 deficiency enhances the Akt/FoxO1 pathway was also demonstrated in this study. Activated FoxO1 displayed increased DNA binding activity for the FOXO1-responsive element on the Fas ligand (FasL) promoter and enhanced the expression of the downstream targets FasL and Bcl-2-interacting mediator of cell death (Bim), thereby inducing cellular apoptosis (Lu et al. 2011). Akt tightly regulates FoxO1 activity, which leads to phosphorylation, cytoplasmic retention and the inactivation of FoxO1. The Akt/FoxO1 pathway plays an important role in neuroprotection against global and focal cerebral ischemia (Won et al. 2006; Zhan et al. 2010). In addition to mediating apoptosis, FoxO1 is also involved in inflammation by stimulating the expression of pro-inflammatory cytokines, including IL-1β, CCL20, and L-selectin (Ito et al. 2009). FoxO1 knockdown or deletion attenuates inflammatory cytokine expression and inhibits the toll-like receptor (TLR) -mediated inflammatory response (Fan et al. 2010; Brown et al. 2011). These results suggest that the elevated phosphorylation of Akt/FoxO1 is potentially associated with the overall protective effect observed in the ischemic brain of TRAF5 KO mice in this study.

In conclusion, our present work provides the first evidence that TRAF5 mediates cerebral I/R injury through the regulation of neuronal apoptosis, BBB degradation, and inflammation. Inhibited NF-κB signaling and enhanced Akt/FoxO1 activation likely account for the biological function of TRAF5 in this cerebral I/R model. We propose that targeting of TRAF5 might provide novel and promising strategies for the treatment of ischemic stroke.

Acknowledgements

The TRAF5 knockout mice were provided by the RIKEN BRC through the National Bio-Resource Project of The MEXT Japan. This study was supported by National Natural Science Foundation of China (No. 81100230 and No. 81070089), National Science and Technology Support Project (No. 2011BAI15B02 and No. 2012BAI39B05), National Basic Research Program of China (No. 2011CB503902), Special Funds for Basic Research and Operating Expenses of the Central Universities (No. 4101032). The authors declare that they have no conflicts of interest in the research.

Ancillary