Macrophage migration inhibitory factor promotes cell ... - SAGE Journals

Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106 & 2011 ISCBFM All rights reserved 0271-678X/11 $32.00 www.jcbfm.com

Macrophage migration inhibitory factor promotes cell death and aggravates neurologic deficits after experimental stroke Ana R Inaćio1, Karsten Ruscher1, Lin Leng2, Richard Bucala2 and Tomas Deierborg1 1

Department of Clinical Sciences, Laboratory for Experimental Brain Research, Lund University, Lund, Sweden; 2Department of Medicine and Pathology, Yale University, New Haven, Connecticut, USA

Multiple mechanisms contribute to tissue demise and functional recovery after stroke. We studied the involvement of macrophage migration inhibitory factor (MIF) in cell death and development of neurologic deficits after experimental stroke. Macrophage migration inhibitory factor is upregulated in the brain after cerebral ischemia, and disruption of the Mif gene in mice leads to a smaller infarct volume and better sensory-motor function after transient middle cerebral artery occlusion (tMCAo). In mice subjected to tMCAo, we found that MIF accumulates in neurons of the peri-infarct region, particularly in cortical parvalbumin-positive interneurons. Likewise, in cultured cortical neurons exposed to oxygen and glucose deprivation, MIF levels increase, and inhibition of MIF by (S,R)-3(4-hydroxyphenyl)-4,5-dihydro-5-isoxazole acetic acid methyl ester (ISO-1) protects against cell death. Deletion of MIF in Mif / mice does not affect interleukin-1b protein levels in the brain and serum after tMCAo. Furthermore, disruption of the Mif gene in mice does not affect CD68, but it is associated with higher galectin-3 immunoreactivity in the brain after tMCAo, suggesting that MIF affects the molecular/cellular composition of the macrophages/microglia response after experimental stroke. We conclude that MIF promotes neuronal death and aggravates neurologic deficits after experimental stroke, which implicates MIF in the pathogenesis of neuronal injury after stroke. Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106; doi:10.1038/jcbfm.2010.194; published online 10 November 2010 Keywords: experimental stroke; functional recovery; galectin-3; macrophage migration inhibitory factor;

neuroprotection; parvalbumin

Introduction Multiple and interrelated processes, including modulation of intrinsic neuronal properties and communication, and inflammatory/immune cascades contribute to stroke outcome (Wieloch and Nikolich, 2006). Inflammatory mediators upregulated in the brain and systemically after stroke are believed to affect damage progression, as well as the creation of a balanced milieu for neuronal plasticity/recovery, Correspondence: AR Inaćio, Department of Clinical Sciences, Laboratory for Experimental Brain Research, Lund University, BMC A13, Lund 22184, Sweden. E-mail: [email protected] This study was supported by Fundac ¸aõ para a Cieˆncia e a Tecnologia, Portugal; the EU 7th workprogram through the European Stroke Network (grant no. 201024); the Bergvall, Crafoord, G&J Kock, Gyllenstiernska Krapperup, Lars Hierta, Pia Sta˚hls, Swedish Stroke, Tore Nilsson, and Wiberg foundations, Swedish Brain Fund, Swedish Research Council (grant No 8466), The Royal Physiographic Society, Sweden; NIH AI042310, and the Brookdale Foundation, USA. Received 6 June 2010; revised 27 August 2010; accepted 7 October 2010; published online 10 November 2010

depending on the temporal and spatial expression profile (Endres et al, 2008). Conversely, inhibition of pro-inflammatory molecules efficiently diminishes cell death (Mitsios et al, 2006). Also, we have previously shown that interleukin-1b (IL-1b) is downregulated in the peri-infarct region of animals housed in an enriched environment after stroke, supporting the notion that experience-enhanced recovery of function involves the modulation of pro-inflammatory mediators (Ruscher et al, 2009). Macrophage migration inhibitory factor (MIF) is a pleiotropic protein, expressed in a diversity of cell types (Bernhagen et al, 1993; Calandra et al, 1994). Macrophage migration inhibitory factor is a mediator of immunity (Calandra and Roger, 2003), hypoxic adaptation (Maity and Koumenis, 2006), and has been implicated in inflammatory conditions (Morand et al, 2006) and heart ischemia (Miller et al, 2008). It is constitutively expressed in the brain (Bernhagen et al, 1993; Nishibori et al, 1996), particularly in neurons (Bacher et al, 1998), where it can act as a neuromodulator (Sun et al, 2004). Macrophage migration inhibitory factor is upregulated in glial tumor cells upon hypoxic and hypo-

MIF and stroke AR Inaćio et al 1094

glycemic stress (Bacher et al, 2003), and is induced in end-feet astrocytes at the blood–brain barrier after Borna disease virus infection (Bacher et al, 2002). Moreover, MIF deficiency decreases inflammation in the brain after West Nile virus infection (Arjona et al, 2007). Furthermore, MIF exhibits tautomerase activity toward oxidized cathecholamines, and has thereby been implicated in neurodegenerative disorders (Matsunaga et al, 1999). Cultured cerebellar granular neurons from Mif / mice are resistant to excitotoxicity (Nishio et al, 2009), suggesting that MIF can affect neuronal susceptibility to ischemia. Although MIF has been shown upregulated after global (Yoshimoto et al, 1997) and focal ischemia (Wang et al, 2009), the functional relevance of MIF in stroke has not been assessed. The aim of the present investigation was to study the contribution of MIF to neuronal cell death and to the development of neurologic deficits during the first week after transient middle cerebral artery occlusion (tMCAo). Interestingly, we found that MIF is highly expressed in cortical parvalbuminpositive neurons, and that MIF is strongly regulated in peri-infarct neurons after stroke. Importantly, we found that the absence of neuronal MIF prevents ischemic cell death and improves sensory-motor function. The MIF-mediated effects in experimental stroke may also depend on the modulation of the macrophages/microglia response in the injured brain hemisphere, as indicated by a higher galectin-3 (Gal-3) immunoreactivity in Mif / mice.

Materials and methods Animals Animals were housed under diurnal conditions, with free access to water and food. All experiments were performed in accordance with protocols approved by the Malmo¨/ Lund Ethical Committee for Animal Research. Surgeries and animal behavior were conducted in a blinded and randomized manner.

Mouse Strains and Genotyping To determine whether MIF affects the functional outcome after experimental stroke, we used Mif / (MIF-KO) male mice on a pure C57BL/6 background (generation N10; Bozza et al, 1999) and respective Mif + / + (wild type (WT)) littermates, 8 to 36 weeks old, weighing 19.5 to 39.0 g. Animals were bred at Lund University. The genotype was determined by polymerase chain reaction. We used the following primer sequences, specific for WT and mutant alleles, respectively: ACGACATGAACGCTGCCAAC (forward) and ACCGTGGTCTCTTATAAACC (reverse); GAAT GAACTGCAGGACGAGG (forward) and GCTCTTCGTCCA GATCATCC (reverse). We used 9- to 14-week-old C57BL/6 male mice (Taconic, Ry, Denmark), weighing 20.5 to 30.0 g, for the spatialtemporal characterization of MIF expression after tMCAo. Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106

Transient Middle Cerebral Artery Occlusion Focal cerebral ischemia was induced in mice by transient occlusion of the right MCA, using the intraluminal filament placement technique as described previously (Nygren and Wieloch, 2005). Briefly, mice were anesthetized by inhalation of 2.5% isoflurane (IsobaVet, ScheringPlough Animal Health, Milton Keynes, UK) in O2:N2O (30:70), through an open mask. Anesthesia was subsequently reduced to 1.5% to 1.8% isoflurane and sustained throughout the occlusion period. Body temperature was kept at B371C. To monitor regional cerebral blood flow (rCBF), an optical fiber probe (Probe 318-I, Perimed, Ja¨rfa¨lla, Sweden) was fixed to the skull at 2 mm posterior and 4 mm lateral to bregma and connected to a laser Doppler flow meter (Periflux System 5,000, Perimed). A filament composed of a 6-0 polydioxanone suture (PSD II, Ethicon, Norderstedt, Germany) and a silicone tip with a diameter of 225 to 275 mm (adapted to the age of the animal) was inserted into the external carotid artery and advanced into the common carotid artery. The filament was retracted, moved into the internal carotid artery, and advanced until the origin of the MCA, given by the sudden drop in rCBF. After 45 minutes, the filament was withdrawn and reperfusion was observed. The animals were placed in a heating box at 371C for the first 2 hours after surgery and thereafter transferred into a heating box at 351C (overnight), to avoid postsurgical hypothermia. Thirty minutes after the onset of reperfusion, 0.5 mL of 5% glucose were administered subcutaneously. Temperature and gross sensory-motor deficits were evaluated at 1, 2, and 24 hours after surgery (when applicable). Body weight was controlled daily up to the experimental end point (when applicable). In shamoperated animals, the filament was withdrawn from the internal carotid artery before reaching the MCA. Specific experimental outlines are indicated in the Results section.

Exclusion Criteria All the animals that did not exhibit an immediate reduction in rCBF upon MCAo and reperfusion were excluded from the study (B8%). Intracerebral hemorrhage was confirmed in all the cases by a posteriori brain analysis. One MIF-KO animal was killed at day 3 of recovery due to signs of pain and general weakness, and included in the mortality rate at 7 days.

Behavioral Tests in Mice At 2 and 7 days of recovery, we used the rotating pole test as described before (Nygren and Wieloch, 2005), with a few modifications: the animals were trained the day before surgery (day 1) at 0, 3, and 10 rotations per minute (r.p.m.), to the left and right, and scored in a final run for each of these pole rotations. The tests were video recorded for a final analysis. Additionally, we used the grip strength test (Bioseb, Vitrolles, France), in which a mouse held by the tail is allowed to grab and pull a horizontal bar with the front paws. We measured the maximum value in five consecutive trials. We analyzed the average out of the five trials. Baseline values were collected at day 1 before


surgery (day 1). The postsurgical decrease in muscular strength reflects the deficit in the left paw.

Carbon-Black Perfusion Naive and sham-operated (day 7 after surgery) WT and MIFKO mice were perfused, essentially as reported before (Berry et al, 1975). In summary, animals were anesthetized with isoflurane and perfused with NaCl, followed by 4% formaldehyde and finally carbon-black solution. Brains were dissected immediately and kept at 41C, overnight, in fixative. Vessel conformation was analyzed, and pictures were acquired with a MicroPublisher 3.3 RTV CCD (QImaging, Surrey, BC, Canada) camera under standard conditions.

Infarct Volume We perfused WT and MIF-KO mice with 4% formaldehyde at day 7 after tMCAo. Brains were cut in 30 mm thick coronal slices, in eight subsequent series (240 mm between consecutive sections within each series). We used the first series for NeuN immunohistochemistry. Primary antibody was detected by the biotin-avidin-horseradish peroxidase system (Vector Laboratories, Peterborough, UK). The entire area of the contralateral (Contra) and ipsilateral (Ipsi) hemispheres, and infarct area (IA) of each slice were encircled using ImageJ software (National Institutes of Health, Bethesda, MD, USA). Infarct area was corrected for shrinkage: IAcorrected = Contra/Ipsi IA. Infarct volume was obtained by volumetric integration.

Interleukin-1b Immunosorbent Assay At 48 hours after surgery (tMCAo and sham), WT and MIFKO mice were anesthetized with isoflurane and blood was collected by heart puncture; isolated serum was stored at 801C. Thereafter and following a 2-minute transcardial perfusion with NaCl, brains were obtained and immediately frozen by immersion in isopentane at 401C, further cooled down to 70 to 801C, and stored at 801C. We dissected infarct core (IC) and peri-infarct region (PI) from consecutive 2, 1, and 2 mm thick brain sections (starting at 2 mm and ending at 7 mm from the olfactory bulb), at 151C. At this temperature, the IC can easily be detached from the remaining ipsilateral tissue, considered as PI. In shamoperated animals, equivalent brain regions were dissected, also designated as IC and PI (a schematic representation of IC and PI is included in the Results section). Whole cellular protein extractions from mice brains were performed as in Rickhag et al (2008). Interleukin-1b levels were determined by an electrochemiluminescence-coupled immunosorbent assay according to the manufacturer’s instructions and using a SECTOR Imager 6,000 (Meso Scale Discovery, Gaithersburg, MD, USA). Serum was assayed undiluted; for each brain sample, the results are presented as protein concentration (pg/mL) when using 150 mg of protein.

brain slices were subsequently generated. CD68 and Gal-3 were detected by immunohistochemistry. We used three brain sections per animal, and per antigen. With respect to CD68, we used sections corresponding to the following levels: B0.86, 0.22, and 1.06 mm, in relation to bregma. Galectin3 immunohistochemistry was performed using three additional brain sections: B1.18, 0.62, and 1.22 mm, in relation to bregma. Anti-CD68 and anti-Gal-3 primary antibodies were detected by a donkey anti-rat Cy3 conjugate (Jackson ImmunoResearch, Suffolk, UK) and streptavidin-Alexa555 (Molecular Probes, Invitrogen, Paisley, UK), respectively. We acquired monochromatic 10-fold magnification micrographs covering the entire right brain hemisphere within each section using a Nikon Eclipse 80i microscope and the NIS-Elements software (Nikon, Solna, Sweden). Micrographs were acquired under standardized conditions; exposure level was set to minimize loss of information (under-exposure or saturation). A final tile figure for each section was subsequently generated (NIS-Elements software, Nikon). The visually detectable fluorescence within each generated tile figure was encircled, and the average intensity level (0 to 255) per pixel was determined using the ImageJ software (National Institutes of Health), and designated as average immunoreactivity.

Primary Neuronal Cell Cultures We dissected cerebral cortices from Wistar rats (Harlan, Horst, the Netherlands) at the embryonic day 17. We incubated the intact cortices with 0.05% (w/v) trypsin and 0.02% (w/v) ethylene diamine tetraacetic acid (EDTA) in phosphate-buffered saline (PBS) (without calcium and magnesium), for 15 minutes at 371C. After two rinses with a serum-containing medium (Dulbecco’s Modified Eagle Medium (DMEM), 10% fetal bovine serum, 100 U/mL penicillin/ streptomycin, and 0.5 mmol/L glutamine), tissue was gently dissociated in culture medium (neurobasal medium, 2% B27, 100 U/mL penicillin/streptomycin, 0.5 mmol/L glutamine, and 2.5 mmol/L glutamate). After centrifugation (210 g, 2 minutes, 41C), cells were plated at a density of 225,000 cells/cm2 on 0.01% poly-L-lysine-coated coverslips in 24-well plates, six-well plates, and live-cell imaging chambers. Cells were maintained at 371C in a humidified incubator, under normoxia (5% CO2). We replaced half of the medium by culture medium without glutamate every fourth day. Media and supplements were purchased from Invitrogen.

Cell Transfection and Live-Cell Imaging At day in vitro four, cells were transfected to express eGFPMIF fusion protein, using lipofectamine 2,000 (Invitrogen), following the manufacturer’s instructions. We used two different constructs: pEGFP-C2-Mif and pEGFP-N2-Mif (0.5 to 1 mg per culture chamber). Transfected cells were imaged at day in vitro 10 to 15 using an inverted laser-scanning microscope (LSM 510, Zeiss, Jena, Germany), under controlled conditions (371C, humidified air with 5% CO2).

CD68 and Galectin-3 Immunoassays

Combined Oxygen–Glucose Deprivation

At day 7 postocclusion, WT and MIF-KO mice were perfused transcardially with 4% formaldehyde; 30 mm thick coronal

We exposed neurons to combined oxygen–glucose deprivation (OGD) at day in vitro 10 to 15. Briefly, culture Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106


medium was removed and kept (conditioned medium). After gently rinsing with PBS, medium was replaced by a deoxygenated basic salt solution without glucose, in a hypoxic chamber (oxygen below 1 mm Hg; Electrotek, Shipley, UK) at 371C. After 5 minutes (sublethal OGD) or 10 minutes (lethal OGD), we replaced the deoxygenated aglycemic solution by conditioned medium and allowed cells to recover under normoxia (5% CO2) at 371C. In control conditions (Ctrl), neurons were rinsed with PBS and incubated in a normoxic basic salt solution containing 4.5 g/L glucose, in a humidified cell culture incubator for 5 or 10 minutes, respectively. Composition of basic salt solution: 143.8 mmol/L Na + , 5.5 mmol/L K + , 1.8 mmol/L Ca2 + , 0.8 mmol/L Mg2 + , 125.3 Cl mmol/L, 26.2 mmol/L HCO3 , 1.0 mmol/L PO34 , 0.8 mmol/L SO24 (pH 7.4).

Pharmacology and Lactate Dehydrogenase Assay We treated neurons with 1 mmol/L MG-132 (Tocris, Bristol, UK), final concentration in the culture medium, immediately after sublethal OGD/Ctrl. Vehicle-treated cells (dimethyl sulfoxide (DMSO), 1% of the cell culture medium) served as control (n = 3 to 6, out of four independent neuronal preparations). We applied the MIF-inhibitor (S,R)-3-(4-hydroxyphenyl)4,5-dihydro-5-isoxazole acetic acid methyl ester (ISO-1), at the final concentrations of 1, 10, and 100 mmol/L, or vehicle (DMSO:H2O, 1:3, 0.4% of the cell culture medium) to cultures immediately after lethal OGD/Ctrl. Cell death was assessed by measurement of lactate dehydrogenase (LDH) activity in the cell culture medium using a colorimetric assay (Biovison, Mountain View, CA, USA). Lactate dehydrogenase activity (U/mL) was calculated using a standard curve. Specific experimental outlines are indicated in the Results section.

Protein Extraction and Western Blotting Cytoplasmic and nuclear protein fractionations were done according to Dignam et al (1983). Proteins were separated in 10% to 20% gels (Novex Tricine system, Invitrogen). Transfer was performed onto polyvinyldifluoride membranes. We performed immunoblotting using the Snap i. d. system, according to the manufacturer’s instructions (Millipore, Billerica, MA, USA). We used a polyclonal anti-MIF antibody (1:1,000, Torrey Pines Biolabs, East Orange, NJ, USA) and a secondary antibody horseradish peroxidase conjugate (1:5,000, Sigma-Aldrich, Deisenhofen, Germany). The signal was visualized using a chemiluminescence kit (Millipore) and a CCD camera (Fujifilm LAS 1,000). Membranes were stripped and reprobed for b-actin using a monoclonal antibody directly conjugated with horseradish peroxidase (1:75,000, Sigma-Aldrich). After densitometric analysis, the expression of MIF was normalized according to the expression of b-actin.

Immunohistochemistry/Immunocytochemistry Formaldehyde-fixed free-floating brain slices or cells were rinsed three times with PBS and incubated with 5% blocking solution (normal serum, Jackson ImmunoResearch, Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106

and 0.25% Triton X-100, in PBS), for 1 hour at room temperature. Next, we incubated with primary antibody diluted in 2% blocking solution, overnight at 41C. After rinsing, secondary antibody was applied (except for the GFAP-Cy3 direct conjugate). We used the avidin-biotinhorseradish peroxidase system or secondary antibodies conjugated with fluorescent dyes. We used Cy-conjugated and DyLight488-conjugated secondary antibodies (Jackson ImmunoResearch) at a dilution of 1:400 in 2% blocking solution, for 2 hours at room temperature. Alexa488-conjugated secondary antibody and streptavidin-Alexa555 (Molecular Probes) were also used at a dilution of 1:400. Bright-field pictures were acquired using an Olympus BX60 microscope (Solna, Sweden), under standard conditions. Fluorescent dyes were imaged using a laser-scanning microscope (Zeiss, LSM 510), unless otherwise specified. The following primary antibodies and respective dilutions were used in this study: rabbit anti-MIF (Torrey Pines Biolabs, 1:300 to 1:1,500), rabbit anti-MIF (clone R102, 1:300 to 1:1,500), mouse anti-NeuN (Sigma Aldrich, 1:800), mouse anti-Tuj-1 (Sigma-Aldrich, 1:500 to 1:800), mouse antiparvalbumin (Sigma-Aldrich, 1:4,000 to 1:6,000), mouse anti-GST-p (Sigma-Aldrich, 1:500); rat anti-Gal-3 biotinylated (Acris Antibodies, Herford, Germany, 1:500), mouse antiGFAP (Sigma-Aldrich, 1:1,000 to 1:2,000), mouse anti-GFAP Cy3 conjugate (Sigma-Aldrich, 1:5,000), rat anti-CD74 (BD Pharmingen, Stockholm, Sweden, 1:200), and rat anti-CD68 (Serotec, Du¨sseldorf, Germany, 1:200).

Statistics Data are summarized as median and percentiles or as mean±s.d. (as indicated in the legends to figures). For all the tests, P < 0.05 was considered significant. To compare the infarct volume, we used the Student’s t-test. To compare the performance in several behavioral tests between WT and MIF-KO littermates, we used the Mann–Whitney U-test and the Fisher’s exact test (for frequencies). Differences in physiological parameters between WT and MIF-KO mice were assessed by the Student’s t-test (percent drop in rCBF and reperfusion), and one-way analysis of variance for repeated measurements (temperature and body weight). To compare the mortality at 48 hours and 7 days between the mice groups, we used Fisher’s exact test. To study the effect of ischemia/MG-132 on the expression of MIF, we used analysis of variance. The effect of ISO-1 on neuronal cell death after OGD was analyzed by one-way analysis of variance. Differences in IL-1b levels were assessed by analysis of variance. Differences in CD68 and Gal-3 immunoreactivities were determined by the Student’s t-test. Bonferroni correction was used for multiple comparisons.

Results Macrophage Migration Inhibitory Factor-KO Mice Exhibit a Smaller Infarct Volume and Better Functional Outcome After Transient Middle Cerebral Artery Occlusion than Wild Type Littermates

To determine whether MIF participates in cell death or functional recovery after stroke, we subjected WT


Figure 1 Genetic deletion of macrophage migration inhibitory factor (MIF) decreases infarct volume and promotes recovery of neurologic function after transient middle cerebral artery occlusion (tMCAo) in mice. (A) Experimental outline. (B) Representative photographs of brain vasculature in wild type (WT) and MIF-KO mice, after perfusion with a carbon-black solution. (C) Infarct volume of WT and MIF-KO mice at 7 days after 45 minutes tMCAo (in mm3; mean±s.d.). (D) Rotating pole test. Here, we present the functional score of WT and MIF-KO mice when traversing a horizontal pole at 10 r.p.m. to the left, at days 2 and 7 after tMCAo (individual data points are included). (E) Grip strength test. The box plots illustrate the average of the maximum force achieved in five consecutive trials by WT and MIF-KO mice, when grabbing a T-shaped horizontal bar (in g). *P < 0.05; **P < 0.01 (using the Fisher’s exact test).

and MIF-KO mice to 45 minutes tMCAo and evaluated the outcome in terms of infarct volume and sensory-motor function (Figure 1A). Putative differences in vessel conformation and physiological parameters caused by the deletion of Mif were first analyzed. By transcardial perfusion of a gelatinous carbon-black solution, we observed a similar MCA configuration in WT (n = 5) and MIF-KO mice (n = 3) (Figure 1B). We did not find a difference in rCBF between WT and MIF-KO mice at occlusion of the MCA (29%±13% and 24%±13% of baseline, respectively) and at reperfusion (85%± 15% and 79%±24% of baseline, respectively). Additionally, no differences were found in body temperature and body weight between the two experimental groups, WT (n = 13) and MIF-KO (n = 9) (Table 1). No significant difference in mortality was observed at the 7 days end point (6 out of 19 WT and 2 out of 11 MIF-KO mice). We assessed infarct size at 7 days postocclusion, which is when infarct growth has subsided. The MIF-KO animals exhibited smaller infarct volume (16.6±3.6 mm3, n = 9) than the respective WT littermates (21.6±6.3 mm3, n = 13), measured as the loss of NeuN immunoreactivity, P < 0.05 (Figure 1C). The smaller infarct of the MIF-KO animals was associated with improved neurologic function. Two days after occlusion, we observed a marked

Table 1 Physiological parameters in WT and MIF-KO mice Parameter Temperature (1C)a Body weight

Time point

WT

MIF-KO

1 hour 2 hours 24 hours Beforeb Day 7a

36.1±0.4 36.4±0.5 36.0±0.3 25.3±2.6 22.9±3.2

35.9±0.5 36.2±0.4 36.0±0.4 24.8±6.8 23.2±6.0

MIF, macrophage migration inhibitory factor; WT, wild type. a After surgery. b Before surgery.

difference in performance when traversing the rotating pole at 10 r.p.m. to the left: MIF-KO animals scored relatively high (with a median value of 3), whereas WT animals revealed severe difficulties in traversing the pole (median value of 2), P < 0.05. At 7 days after occlusion, although the majority of WT mice scored lower than 4, all the MIF-KO animals scored 4 or 5, denoting a critical difference between the groups in terms of lower scores (0 to 3) and higher scores (4 to 6); P < 0.01 (Figure 1D). We also found a significantly higher grip strength value in MIF-KO animals at 2 days after occlusion (Figure 1E), with median values of 127.5 g (MIF-KO mice) and 87.9 g (WT mice), P < 0.05. At 7 days, we Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106


observed a trend toward a higher average grip strength value in MIF-KO mice (P = 0.14). We allowed the mice to grab the horizontal T-shaped bar with both front paws to avoid bias of paw immobilization and lack of motivation to perform the test, on the cost of loss of sensitivity at later times of recovery. Baseline values (collected at day 1) were not significantly different between WT and MIF-KO animals (median values of 135.1 and 145.2 g, respectively). Macrophage Migration Inhibitory Factor Accumulates in Neurons of the Peri-Infarct Cortex Early After Transient Middle Cerebral Artery Occlusion in Mice

After experimental focal ischemia, MIF protein expression is upregulated in the ipsilateral hemisphere (Wang et al, 2009), but the underlying cellular distribution is unknown. We subjected C57BL/6 mice to 45 minutes tMCAo and collected perfused brains at 0, 3, 6, 12, 18, 24, 48, and 72 hours after occlusion. The spatial-temporal expression of MIF protein was determined by immunohistochemistry (n = 3 to 6 per time point), including naive (n = 3) and sham-operated (n = 3) mice. The two different polyclonal anti-MIF antibodies used yielded similar results (data not shown). Figures 2A and 2B show the expression of MIF within different cell layers of the motor and somatosensory cortex. In brains of control and sham-operated mice, MIF immunoreactivity (MIF-ir) is evident in neurons (Figures 2A1 and 2B1); this was confirmed by the colocalization with the neuronal markers Tuj-1 and NeuN (Figure 2C) but MIF-ir was not homogenously distributed among neurons. Apart from regional differences, some neurons are particularly rich in MIF; for instance parvalbumin-positive (PV + ) cortical interneurons exhibit a relatively high-MIF signal (Figure 2C). Within neurons, MIF is found predominantly in the soma as well as processes (Figure 2C). At 3 hours after tMCAo, we observed a dramatic loss of MIF in the expanding infarct core (Figures 2A2 and 2B2). This early loss was complete in the affected striatum and partial in the affected cortex. We found scattered cortical cells positive for MIF at 3 hours after occlusion (Figure 2B2) but not after 18 to 24 hours (Figure 2B3). We identified these cells as PV-containing interneurons (Figure 2D), exhibiting a delayed cell death in the infarct core. Concomitantly with the loss in the infarct core, MIF-ir increases in the peri-infarct region, particularly in the peri-infarct cortex. The cellular accumulation of MIF is evident as early as 3 hours after tMCAo (Figure 2A2) and progressively increases in subsequent time points (Figure 2A3). Surprisingly, we found that the upregulation of MIF in proximal peri-infarct cortex of mice occurs in neurons, particularly in PV-expressing interneurons (Figure 2E). The regional and cellular distribution of MIF is identical up to 72 hours after occlusion (Supplementary Figure 1). We did not detect MIF in cortical peri-infarct GFAP + astrocytes, GST-p + oligodendrocytes, or Gal-3 + microglia (data Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106

not shown). MIF is a signaling ligand for cell surface receptor, CD74 (Leng et al, 2003). We did not detect CD74 expression in brain cells at any of the time points analyzed (data not shown). Antibody specificity was confirmed by lack of MIF-ir in brains of MIFKO mice (Supplementary Figure 2). Isolated Cortical Neurons Regulate Cytoplasmic Macrophage Migration Inhibitory Factor Levels After Sublethal Oxygen–Glucose Deprivation

We next sought to determine whether neurons regulate the expression of MIF after ischemia and if the inhibition of neuronal MIF expression is protective upon ischemia. Isolated cortical neurons were cultured, exposed to sublethal OGD, and processed at several time points of postischemic recovery, as indicated in Figure 3A. In control cultures, MIF-ir is mainly located to the cytoplasm of neurons (Tuj-1 + cells), with low fluorescence intensity in the cell nucleus (Figure 3B). We did not detect major changes in MIF compartmentalization after OGD (data not shown). Live-cell imaging of neurons transfected with a construct encoding for an eGFP-MIF fusion protein also revealed the accumulation of the fluorescent signal in the cytoplasm (Supplementary Figure 3). We did not detect changes in eGFP-MIF compartmentalization following exposure to OGD (data not shown). Endogenous MIF (detected by immunocytochemistry) colocalized with eGFP-MIF, both under control conditions and after OGD. We used two different constructs encoding for eGFP-MIF and obtained equivalent results (data not shown). By Western blotting of cytoplasmic protein fractions, we found a significant upregulation of MIF at 3 and 24 hours after OGD (by B70% and 43%, respectively) when compared with control conditions, P < 0.05 (Figure 3C). In parallel, we selectively inhibited the proteasome system by application of 1 mmol/L MG-132 to cultured neurons immediately after OGD. MG-132 promotes a nonsignificant increase in cytoplasmic MIF levels under control conditions, but there was no difference between MG132-treated control cells and OGD cells (vehicle or MG treated) at 3 hours of recovery. MG-132 did not affect MIF protein levels in neurons at 1 and 3 hours after OGD, but it led to a significant decrease in MIF levels at 24 hours after OGD, compared with vehicletreated cells, P < 0.05 (Figure 3C). In addition, we did not detect MIF in nuclear fractions by Western blotting (Supplementary Figure 3). Inhibition of Neuronal Macrophage Migration Inhibitory Factor Is Protective After Oxygen–Glucose Deprivation

We next applied the small molecule MIF-inhibitor, ISO-1 (1, 10, or 100 mmol/L) (Lubetsky et al, 2002), or vehicle (DMSO:water, 1:3) to neurons immediately after lethal OGD or control condition. ISO-1 is a


Figure 2 Early and gradual accumulation of macrophage migration inhibitory factor (MIF) in neurons of the peri-infarct cortex in mice. (A) Expression of MIF in the intact somatosensory-motor cortex (SMC) in sham-operated animals (A1), and in the peri-infarct/ infarct SMC at 3 and 24 hours after transient middle cerebral artery occlusion (tMCAo) (A2 and A3, respectively). (B) Expression of MIF in the intact somatosensory cortex in sham-operated animals (B1), and in the evolving infarct core at 3 and 24 hours after occlusion (B2 and B3, respectively). (C) Colocalization of MIF (Cy3, red) with the neuronal markers Tuj-1 and NeuN (both Cy5, in green), and with the calcium-binding protein parvalbumin (PV) (Cy5, in green), in sham-operated mice. (D) Resistant PV + cells (Cy5, green) in the evolving cortical core at 6 hours after tMCAo colocalize with MIF (Cy3, red), but exhibit an irregular shape. (E) Colocalization of MIF (Cy3, red) with Tuj-1 (Cy2, green) and PV (Cy5, green) at 24 hours of recovery. (A, B) Bright-field micrographs. Scale bars: 100 mm (lower magnifications) and 50 mm (for higher magnifications in A). (C-E) Confocal images. Scale bars: 50 mm. IC, infarct core.

cell-permeable isoxazoline. Twenty-four hours after treatment, neuronal cell death was assessed by LDH activity in the cell culture medium (n = 8 to 32 per condition). As shown in Figure 4, there was a significant decrease in OGD-induced LDH release, at 100 mmol/L of ISO-1, P < 0.01. At this concentration of ISO-1, the levels of LDH activity were similar after Ctrl and OGD. These results indicate that the inhibition of MIF after OGD is protective against cell death, and taken together with the in vivo data, suggest that neuronal MIF is detrimental for brain integrity after experimental stroke.

Interleukin-1b Protein Levels Are Similar in Wild Type and Macrophage Migration Inhibitory Factor-KO Mice After Transient Middle Cerebral Artery Occlusion

We then addressed the putative involvement of MIF in mechanisms of poststroke inflammation, at a time point postocclusion when we first observed that sensory-motor performance was markedly better in mice lacking MIF. Here, we quantified IL-1b protein levels in the serum and brains of WT mice (n = 8 to 11) and MIF-KO (n = 8 to 12) littermates at 48 hours after 45 minutes tMCAo. Sham-operated Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106


Figure 3 Macrophage migration inhibitory factor (MIF) is upregulated in the cytoplasm of cultured cortical neurons after sublethal oxygen–glucose deprivation (OGD). (A) Experimental outline. (B) Intracellular localization of MIF in cultured cortical neurons. Confocal micrographs. Colocalization of MIF (Cy5, represented in green) with the neuronal marker Tuj-1 (Cy3, red), at 24 hours after control conditions (Ctrl) (upper panel). Colocalization of MIF (Cy3, red) with NeuN (DyLight488, green), at 24 hours after Ctrl (right). Scale bars: 20 mm. (C) Levels of MIF in cytoplasm/membrane fractions of neurons at 1, 3, and 24 hours after OGD; representative Western blots (left), where the first lane corresponds to the protein marker (9 kDa); quantification (right; mean±s.d.). Effect of 1 mmol/L MG-132 (abbreviated as MG) immediately after Ctrl/OGD; vehicle-Veh. aP < 0.05 versus Ctrl. bP < 0.05 versus Ctrl-Veh and Ctrl-MG. cP < 0.05 versus OGD-Veh.

Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106


In parallel, we found similar serum IL-1b protein levels in WT and MIF-KO sham-operated mice: 2.5±0.6 pg/mL (n = 4) and 1.3±0.4 pg/mL (n = 4), respectively (Figure 5B). Additionally, we did not detect a significant difference in serum IL-1b levels between WT and MIF-KO mice after tMCAo: 10.0±21.0 pg/mL (n = 8) and 1.5±0.2 pg/mL (n = 8). Importantly, IL-1b is not upregulated after tMCAo versus baseline, for both genotypes. In summary, we observed an increase in IL-1b levels in the infarct core of mice subjected to experimental stroke in comparison with an anatomically equivalent region collected from sham-operated animals. However, the inactivation of the Mif gene has no effect on the accumulation of IL-1b, a major inflammatory mediator, after experimental stroke. Figure 4 Pharmacological inhibition of macrophage migration inhibitory factor (MIF) is neuroprotective after lethal oxygen– glucose deprivation (OGD) in cultured cortical neurons. (A) Experimental outline. (B) Lactate dehydrogenase (LDH) activity in the culture medium after Ctrl/OGD and after application o f 0 mmol/L (vehicle), 1, 10, and 100 mmol/L of the MIF-inhibitor (S,R)-3-(4-hydroxyphenyl)-4,5-dihydro-5-isoxazole acetic acid methyl ester (ISO-1) (units/mL; mean±s.d.). aP < 0.01, when comparing Ctrl and OGD. bP < 0.01 when comparing OGD 100 mmol/L with OGD 0, 1, and 10 mmol/L. cP < 0.01 in Ctrl 100 mmol/L versus Ctrl 0 and 1 mmol/L.

Galectin-3 Immunoreactivity in the Brain After Transient Middle Cerebral Artery Occlusion Is Higher in Macrophage Migration Inhibitory Factor-KO Mice than in Wild Type Littermates

We finally evaluated the effect of MIF on the macrophages/microglia response after stroke. Accordingly, we quantified CD68 immunoreactivity (CD68 IR) and Gal-3 immunoreactivity (Gal-3 IR) in the brain of WT (n = 6) and MIF-KO (n = 6) mice subjected to tMCAo (by immunohistochemistry). Negative controls omitting the primary antibody WT (n = 4 to 6) and MIF-KO (n = 4 to 6) mice, killed at were included. In all the brains analyzed, we 48 hours after surgery, were included (baseline). observed a clear induction of CD68 and Gal-3 within In accordance with the previous results, we did the injured hemisphere. Moreover, we found a not detect differences in rCBF at MCA occlusion and similar average CD68 IR, given by the average reperfusion between WT and MIF-KO mice. Body fluorescence intensity within the entire CD68-positemperature and body weight were similar between tive area, in the injured brain hemisphere of WT and genotypes. Consistently, we did observe a difference MIF-KO mice (Figure 6A). At 7 days after tMCAo, in mortality between genotypes after tMCAo (4 out of CD68 IR is heterogeneously distributed within the 15 WT and 3 out of 15 MIF-KO mice) and after sham injured brain hemisphere, and appears denser within treatment (0% for both groups). its border zone and sparser within more centrally We found similar IL-1b levels in the infarct core affected areas (Figure 6A). However, no difference in of WT and MIF-KO sham-operated animals: local CD68 IR was detected between experimental 0.17±0.19 pg/mL (n = 6) and 0.26±0.30 pg/mL groups, as illustrated by higher magnification micro(n = 6), respectively (Figure 5A). Similarly, we did graphs of two different cortical areas in Figure 6A. not detect a significant difference in the peri-infarct Furthermore, CD68 + macrophages/microglia exhibit area: 0.40±0.20 pg/mL in WT (n = 6) and 0.17± a diverse morphology, both ramified and amoeboid, 0.15 pg/mL in KO (n = 6) mice. We found that IL-1b and no clear difference was found between the is robustly upregulated in the infarct core of WT groups. By contrast, we found a significantly higher mice subjected to tMCAo (6.11±2.99 pg/mL, n = 11) average Gal-3 IR in the brains of MIF-KO mice, when versus WT sham-operated mice; P < 0.05. In the same compared with those of WT mice (Figure 6B), manner, IL-1b is upregulated in the infarct core of P < 0.05. Higher magnification micrographs show a MIF-KO animals after occlusion (7.44±5.50 pg/mL) higher density of Gal-3 + cells consistently found in in comparison to baseline; P < 0.05. No difference the brains of MIF-KO versus WT mice. In terms of was detected between genotypes. The average IL-1b respective morphology, no critical difference was level is elevated in the peri-infarct region of WT observed between WT and MIF-KO mice (Figure 6B). mice subjected to tMCAo (0.80±0.46 pg/mL) versus WT sham-operated mice, but not significantly. Similarly, IL-1b is not significantly upregulated Discussion in the peri-infarct region of MIF-KO animals at 48 hours postocclusion (0.86±5.50 pg/mL) versus We will discuss the following three major findings of baseline. Importantly, there is no difference between this study: (1) MIF promotes cell death and aggravates genotypes. neurologic symptoms after experimental stroke; (2) Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106


of MIF in mice results both in a smaller infarct size and in a better functional outcome after tMCAo, strongly implicating MIF in mechanisms of acute brain injury. Importantly, we observed an early and sustained upregulation of MIF in neurons of the evolving core/peri-infarct border zone, particularly in cortical areas (motor and somatosensory cortex). This is in line with our finding that MIF is upregulated in cultured cortical neurons exposed to sublethal OGD. Collectively, these results show that neurons actively regulate MIF expression upon ischemic conditions and further suggest an important role of neuronal MIF in MIF-mediated effects after stroke. Indeed, we found that the MIF-inhibitor ISO-1 is neuroprotective against OGD-induced damage, demonstrating a detrimental action of neuronal MIF under ischemic conditions over the sublethal threshold. Similarly, in a prior study of spinal cord injury, the deletion of MIF resulted in neuroprotection (Nishio et al, 2009). In addition, Nishio et al (2009) found that recombinant human MIF exacerbated glutamate-induced cell death of cerebellar granular neurons from MIF-KO mice, and proposed an involvement of MIF in p53-dependent pathways upon excitotoxic stimuli. Macrophage Migration Inhibitory Factor Signal Initiation in Neurons

Figure 5 Unaltered levels of interleukin-1b (IL-1b) protein in the brain and serum of macrophage migration inhibitory factor (MIF)-KO versus wild type (WT) mice. (A) Schematic representation of dissected brain areas at 48 hours after transient middle cerebral artery occlusion (tMCAo) or sham operation: IC, infarct core (white); PI, peri-infarct region (delineated). (B) IL-1b protein levels in the IC and PI of WT and MIF-KO mice (pg/ mL; mean±s.d.); white bars, WT mice; gray bars, MIF-KO mice. (C) Serum IL-1b (pg/mL; mean±s.d.). *P < 0.05 tMCAo_IC versus Sham_IC.

the specific regulation of MIF in neurons postischemia and the contribution of neuronal MIF to ischemic cell death, including the particular expression of MIF in PV + interneurons; and (3) the absence of MIF is associated with an altered Gal-3 immunoreactivity in the injured brain hemisphere after experimental stroke.

Relevance of Neuronal Macrophage Migration Inhibitory Factor in Cell Death and Functional Outcome After Experimental Stroke

Previous reports show that brain ischemia induces MIF mRNA and protein expression (Yoshimoto et al, 1997; Wang et al, 2009), suggesting a functional role in stroke. We show for the first time that abrogation Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106

We did not detect CD74 in brain cells, excluding the possibility of MIF-CD74 initiated cascades among neuronal populations and putatively among neurons and other brain cells. Release of MIF and subsequent receptor binding or internalization (Kleemann et al, 2002) may not be necessary for MIF activity in neurons. For instance, in a study by Sun et al (2004), stimulation of neurons resulted in increased MIF mRNA and protein levels, but did not promote the release of MIF. Furthermore, the authors show that MIF modulates neuronal firing rates when applied intracellularly, but not when applied extracellularly. Therefore, it seems likely that the observed increase in MIF protein levels in neurons in vitro after OGD and in vivo after tMCAo result from transcriptional activation, and that overexpression of intracellular MIF may promote neuronal cell death or renders the neuronal networking more susceptible to ischemic cell death. Regulation of Macrophage Migration Inhibitory Factor in Neurons

The mechanisms of MIF induction after ischemia are not fully understood, but MIF may be regulated at the transcriptional level or by degradation. We did not see any appreciable effect of proteasome inhibition on MIF levels after OGD (1, 3, and 24 hours). On the other hand, MIF is induced by the transcription factor HIF-1, activated upon hypoxic conditions in vitro (Baugh et al, 2006). Indeed, the observed


Figure 6 CD68 and galectin-3 (Gal-3) immunoreactivities in the brains of wild type (WT) and macrophage migration inhibitory factor (MIF)-KO mice after transient middle cerebral artery occlusion (tMCAo). (A) Average CD68 immunoreactivity (CD68 IR) in the brain of WT and MIF-KO mice at 7 days after tMCAo, given by the average Cy3 fluorescence intensity per pixel (0 to 255) within the area considered as CD68 + (I/pixel; mean±s.d.). (A.a) Representative tile image of CD68 IR in the injured brain hemisphere of WT mice; I and II exemplify relatively high and low CD68 IR regions, respectively. (A.b) CD68 (Cy3, red) in WT and MIF-KO mice, and in cortex I and II, as indicated. (B) Average Gal-3 immunoreactivity (Gal-3 IR) in WT and MIF-KO mice at 7 days after tMCAo (I/pixel; mean±s.d.). (B.a) General view of Gal-3 IR in the right brain hemisphere of WT mice. (B.b) Gal-3 (Alexa555, red) in WT and MIFKO mice, in cortex II (upper panel); negative control (bottom panel). (A.a, B.a) Epifluorescence micrographs. (A.b, B.b) Confocal images. Scale bars: 50 mm. The color reproduction of this figure is available on the html full text version of the manuscript.



sustained upregulation of MIF during 3 days of recovery after tMCAo correlates with an earlier reported similar expression pattern of HIF-1 (Chen et al, 2007), suggesting that the upregulation is due mainly to transcriptional activation of the gene. Cell Death Promoting Action of Macrophage Migration Inhibitory Factor

The significance of MIF expression in the brain under physiological and pathological conditions, particularly in neurons, is largely unexplored. Macrophage migration inhibitory factor may be involved in neuronal intracellular signaling, as it reduces the basal firing rate in cultured rat hypothalamic neurons (Sun et al, 2004). However, the precise mechanism whereby MIF promotes neuronal cell death can only be speculated upon at present. Outside the central nervous system, MIF has been proposed as an antagonist of the p53 pathway (Hudson et al, 1999), preventing apoptosis, and p53 may be an intracellular MIF-binding partner (Jung et al, 2008). The role of p53 in neuronal apoptosis is controversial, and p53 may have a regulatory role in the postischemic brain (Tomasevic et al, 1999). Moreover, it was recently shown that in cultured dermal fibroblasts, the lack of MIF is associated with a reduced nuclear localization of p21 (Taranto et al, 2009), a cyclin-dependent kinase inhibitor (1A) that participates in cell-cycle arrest, suggesting the involvement of MIF in p53-independent mechanisms. Apart from its function in the cell cycle, there is evidence that p21 is a direct inhibitor of apoptosis (Suzuki et al, 1998). Moreover, Harms et al (2007) show that activation of p21 via phosphorylation and translocation into the cytoplasm in neurons prevents apoptosis. Hence, an upregulation of nuclear p21 by MIF can be envisaged to increase neuronal cell death upon ischemia. Macrophage migration inhibitory factor was shown to be protective after heart ischemia, via a mechanism that includes binding to the receptor CD74, subsequent activation of AMP-activated protein kinase, and recruitment of glucose transporters to the cardiomyocytes surface (Miller et al, 2008). This mechanism may not apply to neurons after ischemia. First, neurons do not express CD74. Second, activation of AMP-activated protein kinase was shown to have deleterious effects in ischemic stroke (Li et al, 2007). Third, GLUT4 is an insulin-sensitive glucose transporter mostly expressed in peripheral tissues, whereas its expression in the brain is limited to the olfactory bulb, dentate gyrus of the hippocampus and cerebellum (Vannucci et al, 1998). Expression in Parvalbumin-Positive Interneurons

Macrophage migration inhibitory factor is generally expressed in neurons throughout the cortex, particularly in PV + interneurons. Parvalbumin is a small Journal of Cerebral Blood Flow & Metabolism (2011) 31, 1093–1106

calcium-binding protein expressed in subtypes of cortical interneurons, including basket and chandelier cells, which typically form inhibitory synapses with the perisomatic region and with axons of pyramidal cells, respectively (Markram et al, 2004). Although the exact mode of action of PV in the central nervous system is not completely understood, PV + cortical interneurons are often inhibitory and fast spiking, and are critical regulators of cortical output and information processing (Freund and Katona, 2007). The PV + interneurons are also known to be relatively resistant to ischemia (Medina et al, 1996), and may consequently appear to retain MIF-ir up to later time points than other subtypes of neurons in the expanding infarct core after tMCAo. Nevertheless, the strong coexpression of MIF and PV may be of significance for the detrimental effect of MIF on both cell death and also for mechanisms of functional recovery after experimental stroke. This may be accomplished by altered spiking frequency by MIF in PV-containing cells via calcium-mediated mechanisms. In particular, the early accumulation in the dynamic infarct core/peri-infarct region border zone may indicate a detrimental intraneuronal action or detrimental action on the neuronal networking, rendering infarct expansion to a larger extent than in MIF-KO mice. Additionally, the same mechanisms may be generally relevant in other neuronal subtypes after ischemia. Importantly, PV + cells are also found in areas of tissue reorganization important for recovery of lost function after stroke (Wieloch and Nikolich, 2006; Murphy and Corbett, 2009). Macrophage Migration Inhibitory Factor and the Inflammatory Response After Transient Middle Cerebral Artery Occlusion

Inflammatory cascades are activated in the brain and periphery after stroke, and are believed to influence damage expansion and recovery of lost neurologic function (Endres et al, 2008). Importantly, neurons may regulate the inflammatory response after stroke (Qiu et al, 2008). Macrophage migration inhibitory factor is a major mediator of immunity (Calandra and Roger, 2003), and has been extensively implicated in inflammatory conditions, mostly outside the central nervous system (Morand et al, 2006), where it has essentially a deleterious action. In the brain, mRNA levels of major pro-inflammatory cytokines are decreased in MIF-KO mice upon West Nile Virus infection (Arjona et al, 2007). We therefore addressed a putative change in the inflammatory milieu in brains and serum of WT and MIF-KO mice, by measuring IL-1b protein levels. We found a clear upregulation of IL-1b in the ipsilateral hemisphere of mice subjected to tMCAo, in accordance with previous studies (Ruscher et al, 2009). The lack of MIF, however, did not affect the levels of IL-1b in the infarct core and peri-infarct region in mice after tMCAo, strongly suggesting that MIF and IL-1b are independently upregulated in the brain after experi-

MIF and stroke AR Inaćio et al

mental stroke. Interestingly, we found a higher average Gal-3 immunoreactivity in the brains of MIF-KO mice subjected to tMCAo (versus WT littermates), related to a higher density of Gal-3 + cells. These results indicate that MIF affects the nature of the macrophages/microglia response after experimental stroke. This may occur via direct or indirect alterations in the expression of Gal-3 or proliferation of Gal-3 + cells, or eventually through the recruitment of monocytes into the brain parenchyma (Arjona et al, 2007; Bacher et al, 2002). Moreover, a higher density of Gal-3 + cells in the injured brain hemisphere is associated with a better functional outcome after tMCAo. These results are in convergence with the concept that subpopulations of macrophages/microglia, integrated in space and time, may play a beneficial role at the locus of injury, namely in wound healing and repair (Shechter et al, 2009). In particular, Gal-3 + microglial cells have been shown to express growth factors (Lalancette-He´bert et al, 2007; Venkatesan et al, 2010), and may contribute to postischemic brain remodeling (Yan et al, 2009). Conclusion

In conclusion, our results show that MIF promotes neuronal death, worsens the neurologic deficits, and may depress the recovery of sensory-motor function after stroke, by intra or interneuronal mechanisms, and likely via the modulation of the macrophages/ microglia response. These results imply that MIF can be targeted in the development of future stroke therapies.

Acknowledgements The authors thank Professor Tadeusz Wieloch for valuable discussions and suggestions. The authors also thank Carin Sjo¨lund, Kerstin Beirup, and Sol da Rocha for technical assistance.

Disclosure/conflict of interest The authors declare no conflict of interest.

References Arjona A, Foellmer HG, Town T, Leng L, McDonald C, Wang T, Wong SJ, Montgomery RR, Fikrig E, Bucala R (2007) Abrogation of macrophage migration inhibitory factor decreases West Nile virus lethality by limiting viral neuroinvasion. J Clin Invest 117:3059–66 Bacher M, Meinhardt A, Lan HY, Dhabhar FS, Mu W, Metz CN, Chesney JA, Gemsa D, Donnelly T, Atkins RC, Bucala R (1998) MIF expression in the rat brain: implications for neuronal function. Mol Med 4:217–30 Bacher M, Schrader J, Thompson N, Kuschela K, Gemsa D, Waeber G, Schlegel J (2003) Up-regulation of macrophage migration inhibitory factor gene and protein

expression in glial tumor cells during hypoxic and hypoglycemic stress indicates a critical role for angiogenesis in glioblastoma multiforme. Am J Pathol 162:11–7 Bacher M, Weihe E, Dietzschold B, Meinhardt A, Vedder H, Gemsa D, Bette M (2002) Borna disease virus-induced accumulation of macrophage migration inhibitory factor in rat brain astrocytes is associated with inhibition of macrophage infiltration. Glia 37:291–306 Baugh JA, Gantier M, Li L, Byrne A, Buckley A, Donnelly SC (2006) Dual regulation of macrophage migration inhibitory factor (MIF) expression in hypoxia by CREB and HIF-1. Biochem Biophys Res Commun 347:895–903 Bernhagen J, Calandra T, Mitchell RA, Martin SB, Tracey KJ, Voelter W, Manogue KR, Cerami A, Bucala R (1993) MIF is a pituitary-derived cytokine that potentiates lethal. Nature 365:756–9 Berry K, Wisniewski HM, Svarzbein L, Baez S (1975) On the relationship of brain vasculature to production of neurological deficit and morphological changes following acute unilateral common carotid artery ligation in gerbils. J Neurol Sci 25:75–92 Bozza M, Satoskar AR, Lin G, Lu B, Humbles AA, Gerard C, David JR (1999) Targeted disruption of migration inhibitory factor gene reveals its critical role in sepsis. J Exp Med 189:341–6 Calandra T, Bernhagen J, Mitchell RA, Bucala R (1994) The macrophage is an important and previously unrecognized source of macrophage migration inhibitory factor. J Exp Med 179:1895–902 Calandra T, Roger T (2003) Macrophage migration inhibitory factor: a regulator of innate immunity. Nat Rev Immunol 3:791–800 Chen C, Hu Q, Yan J, Lei J, Qin L, Shi X, Luan L, Yang L, Wang K, Han J, Nanda A, Zhou C (2007) Multiple effects of 2ME2 and D609 on the cortical expression of HIF1alpha and apoptotic genes in a middle cerebral artery occlusion-induced focal ischemia rat model. J Neurochem 102:1831–41 Dignam JD, Lebovitz RM, Roeder RG (1983) Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res 11:1475–89 Endres M, Engelhardt B, Koistinaho J, Lindvall O, Meairs S, Mohr JP, Planas A, Rothwell N, Schwaninger M, Schwab ME, Vivien D, Wieloch T, Dirnagl U (2008) Improving outcome after stroke: overcoming the translational roadblock. Cerebrovasc Dis 25:268–78 Freund TF, Katona I (2007) Perisomatic inhibition. Neuron 56:33–42 Harms C, Albrecht K, Harms U, Seidel K, Hauck L, Baldinger T, Hubner D, Kronenberg G, An J, Ruscher K, Meisel A, Dirnagl U, von HR, Endres M, Hortnagl H (2007) Phosphatidylinositol 3-Akt-kinase-dependent phosphorylation of p21(Waf1/Cip1) as a novel mechanism of neuroprotection by glucocorticoids. J Neurosci 27:4562–71 Hudson JD, Shoaibi MA, Maestro R, Carnero A, Hannon GJ, Beach DH (1999) A proinflammatory cytokine inhibits p53 tumor suppressor activity. J Exp Med 190:1375–82 Jung H, Seong HA, Ha H (2008) Critical role of cysteine residue 81 of macrophage migration inhibitory factor (MIF) in MIF-induced inhibition of p53 activity. J Biol Chem 283:20383–96 Kleemann R, Grell M, Mischke R, Zimmermann G, Bernhagen J (2002) Receptor binding and cellular uptake studies of macrophage migration inhibitory

1105



factor (MIF): use of biologically active labeled MIF derivatives. J Interferon Cytokine Res 22:351–63 Lalancette-He´bert M, Gowing G, Simard A, Weng YC, Kriz J (2007) Selective ablation of proliferating microglial cells exacerbates ischemic injury in the brain. J Neurosci 27:2596–605 Leng L, Metz CN, Fang Y, Xu J, Donnelly S, Baugh J, Delohery T, Chen Y, Mitchell RA, Bucala R (2003) MIF signal transduction initiated by binding to CD74. J Exp Med 197:1467–76 Li J, Zeng Z, Viollet B, Ronnett GV, McCullough LD (2007) Neuroprotective effects of adenosine monophosphateactivated protein kinase inhibition and gene deletion in stroke. Stroke 38:2992–9 Lubetsky JB, Dios A, Han J, Aljabari B, Ruzsicska B, Mitchell R, Lolis E, Al-Abed Y (2002) The tautomerase active site of macrophage migration inhibitory factor is a potential target for discovery of novel anti-inflammatory agents. J Biol Chem 277:24976–82 Maity A, Koumenis C (2006) HIF and MIF—a nifty way to delay senescence? Genes Dev 20:3337–41 Markram H, Toledo-Rodriguez M, Wang Y, Gupta A, Silberberg G, Wu C (2004) Interneurons of the neocortical inhibitory system. Nat Rev Neurosci 5:793–807 Matsunaga J, Sinha D, Pannell L, Santis C, Solano F, Wistow GJ, Hearing VJ (1999) Enzyme activity of macrophage migration inhibitory factor toward oxidized catecholamines. J Biol Chem 274:3268–71 Medina L, Figueredo-Cardenas G, Reiner A (1996) Differential abundance of superoxide dismutase in interneurons versus projection neurons and in matrix versus striosome neurons in monkey striatum. Brain Res 708:59–70 Miller EJ, Li J, Leng L, McDonald C, Atsumi T, Bucala R, Young LH (2008) Macrophage migration inhibitory factor stimulates AMP-activated protein kinase in the ischaemic heart. Nature 451:578–82 Mitsios N, Gaffney J, Kumar P, Krupinski J, Kumar S, Slevin M (2006) Pathophysiology of acute ischaemic stroke: an analysis of common signalling mechanisms and identification of new molecular targets. Pathobiology 73:159–75 Morand EF, Leech M, Bernhagen J (2006) MIF: a new cytokine link between rheumatoid arthritis and atherosclerosis. Nat Rev Drug Discov 5:399–410 Murphy TH, Corbett D (2009) Plasticity during stroke recovery: from synapse to behaviour. Nat Rev Neurosci 10:861–72 Nishio Y, Koda M, Hashimoto M, Kamada T, Koshizuka S, Yoshinaga K, Onodera S, Nishihira J, Okawa A, Yamazaki M (2009) Deletion of macrophage migration inhibitory factor attenuates neuronal death and promotes functional recovery after compression-induced spinal cord injury in mice. Acta Neuropathol 117:321–8 Nishibori M, Nakaya N, Tahara A, Kawabata M, Mori S, Saeki K (1996) Presence of macrophage migration inhibitory factor (MIF) in ependyma, astrocytes and neurons in the bovine brain. Neurosci Lett 213:193–6 Nygren J, Wieloch T (2005) Enriched environment enhances recovery of motor function after focal ischemia in mice, and downregulates the transcription factor NGFI-A. J Cereb Blood Flow Metab 25:1625–33

Qiu J, Nishimura M, Wang Y, Sims JR, Qiu S, Savitz SI, Salomone S, Moskowitz MA (2008) Early release of HMGB-1 from neurons after the onset of brain ischemia. J Cereb Blood Flow Metab 28:927–38 Rickhag M, Deierborg T, Patel S, Ruscher K, Wieloch T (2008) Apolipoprotein D is elevated in oligodendrocytes in the peri-infarct region after experimental stroke: influence of enriched environment. J Cereb Blood Flow Metab 28:551–62 Ruscher K, Johannesson E, Brugiere E, Erickson A, Rickhag M, Wieloch T (2009) Enriched environment reduces apolipoprotein E (ApoE) in reactive astrocytes and attenuates inflammation of the peri-infarct tissue after experimental stroke. J Cereb Blood Flow Metab 29:1796–805 Shechter R, London A, Varol C, Raposo C, Cusimano M, Yovel G, Rolls A, Mack M, Pluchino S, Martino G, Jung S, Schwartz M (2009) Infiltrating blood-derived macrophages are vital cells playing an anti-inflammatory role in recovery from spinal cord injury in mice. PLoS Med 6:e1000113 Sun C, Li H, Leng L, Raizada MK, Bucala R, Sumners C (2004) Macrophage migration inhibitory factor: an intracellular inhibitor of angiotensin II-induced increases in neuronal activity. J Neurosci 24:9944–52 Suzuki A, Tsutomi Y, Akahane K, Araki T, Miura M (1998) Resistance to Fas-mediated apoptosis: activation of caspase 3 is regulated by cell cycle regulator p21WAF1 and IAP gene family ILP. Oncogene 17:931–9 Taranto E, Xue JR, Morand EF, Leech M (2009) Modulation of expression and cellular distribution of p21 by macrophage migration inhibitory factor. J Inflamm (Lond) 6:24 Tomasevic G, Kamme F, Stubberod P, Wieloch M, Wieloch T (1999) The tumor suppressor p53 and its response gene p21WAF1/Cip1 are not markers of neuronal death following transient global cerebral ischemia. Neuroscience 90:781–92 Vannucci SJ, Koehler-Stec EM, Li K, Reynolds TH, Clark R, Simpson IA (1998) GLUT4 glucose transporter expression in rodent brain: effect of diabetes. Brain Res 797:1–11 Venkatesan C, Chrzaszcz M, Choi N, Wainwright MS (2010) Chronic upregulation of activated microglia immunoreactive for galectin-3/Mac-2 and nerve growth factor following diffuse axonal injury. J Neuroinflammation 7:32 Wang L, Zis O, Ma G, Shan Z, Zhang X, Wang S, Dai C, Zhao J, Lin Q, Lin S, Song W (2009) Upregulation of macrophage migration inhibitory factor gene expression in stroke. Stroke 40:973–6 Wieloch T, Nikolich K (2006) Mechanisms of neural plasticity following brain injury. Curr Opin Neurobiol 16:258–64 Yan YP, Lang BT, Vemuganti R, Dempsey RJ (2009) Galectin-3 mediates post-ischemic tissue remodeling. Brain Res 1288:116–24 Yoshimoto T, Nishihira J, Tada M, Houkin K, Abe H (1997) Induction of macrophage migration inhibitory factor messenger ribonucleic acid in rat forebrain by reperfusion. Neurosurgery 41:648–53

Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website (http:// www.nature.com/jcbfm)
