Suppression of stroke-induced progenitorproliferation in adult subventricular zone by tumornecrosis factor receptor 1

Robert E Iosif1,4,5, Henrik Ahlenius2,4,5, Christine T Ekdahl1,4, Vladimer Darsalia2,4,Par Thored1,4, Stefan Jovinge3,4, Zaal Kokaia2,4 and Olle Lindvall1,4

1Laboratory of Neurogenesis and Cell Therapy, Section of Restorative Neurology, Wallenberg NeuroscienceCenter, University Hospital, Lund, Sweden; 2Laboratory of Neural Stem Cell Biology, Section of RestorativeNeurology, Stem Cell Institute, University Hospital, Lund, Sweden; 3Hematopoietic Stem Cell Laboratory,Stem Cell Institute, University Hospital, Lund, Sweden; 4Lund Strategic Research Center for Stem CellBiology and Cell Therapy, Lund, Sweden

Stroke induced by middle cerebral artery occlusion leads to transiently increased progenitorproliferation in the subventricular zone (SVZ) and long-lasting striatal neurogenesis in adult rodents.Tumor necrosis factor-a (TNF-a) is upregulated in stroke-damaged brain. Whether TNF-a and itsreceptors influence SVZ progenitor proliferation after stroke is unclear. Here we show that theincreased proliferation 1 week after stroke occurred concomitantly with elevated microglia numbersand TNF-a and TNF receptor-1 (TNF-R1) gene expression in the SVZ of wild-type mice. TNF receptor-1 was expressed on sorted SVZ progenitor cells from nestin-green fluorescent protein reporter mice.In animals lacking TNF-R1, stroke-induced SVZ cell proliferation and neuroblast formation wereenhanced. In contrast, deletion of TNF-R1 did not alter basal or status epilepticus-stimulated cellproliferation in SVZ. Addition of TNF-a reduced the size and numbers of SVZ neurospheres througha TNF-R1-dependent mechanism without affecting cell survival. Our results provide the firstevidence that TNF-R1 is a negative regulator of stroke-induced SVZ progenitor proliferation.Blockade of TNF-R1 signaling might be a novel strategy to promote the proliferative response in SVZafter stroke.Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587; doi:10.1038/jcbfm.2008.47; published online21 May 2008

Keywords: neural progenitor; proliferation; stroke; subventricular zone; TNF-a; TNF-R1

Introduction

Ischemic stroke, induced by middle cerebral arteryocclusion (MCAO) in adult rodents, leads toincreased proliferation of progenitor cells in thesubventricular zone (SVZ) and migration of newlyformed neuroblasts into the damaged striatum (forreferences, see Lindvall and Kokaia, 2008). Many ofthe stroke-generated neuroblasts differentiate intomature neurons with the phenotype of striatalprojection neurons (Arvidsson et al, 2002; Parentet al, 2002). Striatal neurogenesis after stroke is long-

lasting and continues during several months afterthe insult (Thored et al, 2006). Recent findingssuggest increased progenitor proliferation and for-mation of immature neurons also in the humanbrain (Jin et al, 2006). From a clinical-therapeuticperspective, the neurogenic response from SVZprogenitor cells after stroke is probably insufficientand only a fraction of the dead neurons are replaced(Arvidsson et al, 2002). Optimization of neurogen-esis after stroke will require identification of themolecular mechanisms regulating its various steps,that is, proliferation, survival, migration, differentia-tion, and functional integration. Overall cell pro-liferation in the SVZ is transiently increased 1 to 2weeks after MCAO (Arvidsson et al, 2002; Jin et al,2001). When supplied to the SVZ, many factorsincrease cell proliferation after stroke (Lindvall andKokaia, 2008). However, the intrinsic regulation ofstroke-induced proliferation is poorly understood,and no endogenous mechanism that suppresses the

Received 1 February 2008; revised and accepted 21 April 2008;published online 21 May 2008

Correspondence: Dr O Lindvall, Laboratory of Neurogenesis andCell Therapy, Section of Restorative Neurology, WallenbergNeuroscience Center, University Hospital, SE 221 84 Lund,Sweden.E-mail: Olle.Lindvall@med.lu.se5These authors contributed equally to this work.

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587& 2008 ISCBFM All rights reserved 0271-678X/08 $30.00

www.jcbfm.com

proliferative response in the SVZ has yet beenidentified.

Tumor necrosis factor-a (TNF-a) is a major playerin many neurodegenerative diseases includingstroke (Hallenbeck, 2002), its main source beingactivated microglia (Gebicke-Haerter, 2001). Tumornecrosis factor-a influences cell survival through theaction on two different receptor subtypes, TNF-R1and TNF-R2 (Wajant et al, 2003): TNF-R1 contri-butes to neuronal death, whereas TNF-R2 can beneuroprotective (Marchetti et al, 2004). We recentlyidentified TNF-R1 signaling as a negative regulatorof progenitor proliferation in hippocampal neuro-genesis (Iosif et al, 2006). Cell proliferation in thedentate subgranular zone (SGZ) was enhanced inTNF-R1�/� mice both under basal conditions andafter status epilepticus (SE).

Whether TNF-R1 signaling influences progenitorproliferation in SVZ after stroke is unclear. Neuro-sphere cultures from rat SVZ express TNF-R1(Widera et al, 2006) and indirect evidence suggeststhat TNF-a might stimulate SVZ cell proliferation.Intraventricular delivery of TNF-a in rats increasedbromodeoxyuridine (BrdU) incorporation in SVZcells (Wu et al, 2000). Widera et al (2006) reportedthat TNF-a increased volume, BrdU incorporation,and total cell numbers in neurosphere cultures fromintact rat SVZ. Finally, Katakowski et al (2007)showed upregulation of TNF-a-converting enzymein SVZ after stroke and decreased stroke-inducedSVZ proliferation after infusion of an inhibitor ofthis protease.

Here we have used mouse models with loss ofTNF-R1 and TNF-a function, combined with in vitroand in vivo analyses. The main objectives were firstto determine whether signaling through TNF-R1influences SVZ progenitor proliferation after strokeinduced by 40 mins MCAO. This insult, whichcauses damage to striatum and cerebral cortex andmassive microglia activation, is associated withTNF-R1 and TNF-a upregulation in the ischemichemisphere (Lambertsen et al, 2005, 2007). Thesecond objective was to explore if the enhanced SVZprogenitor proliferation caused by SE, that is, apathologic condition that does not lead to extensivestriatal and cortical damage, is affected by TNF-R1signaling. The third objective was to analyzewhether TNF-R1 and TNF-a are expressed in mouseSVZ progenitor cells and if gene levels are alteredafter stroke and, finally, to establish any TNF-R1-mediated effect on SVZ cell proliferation in vitro.

Materials and methods

Animals and Surgery

All experimental procedures followed guidelines setby the Malmo-Lund Ethical Committee. We generatedTNF-R1�/� mice by crossing TNF-R1/R2�/� to wild-typeC57BL/6 mice (B&K Universal, Stockholm, Sweden). TheTNF-R1/R2�/� mice (Peschon et al, 1998), back-crossed for

five generations into wild-type C57BL/6 mice, werepurchased from Jackson Laboratories (Bar Harbor, ME,USA). Interbreeding of heterozygous offspring was fol-lowed by selection of appropriate genotypes using PCR.TNF-a�/� mice (Pasparakis et al, 1996), back-crossed for 10generations into wild-type C57BL/6 mice, were purchasedfrom Jackson Laboratories. Wild-type C57BL/6 mice wereused as controls.

A total of 33 TNF-R1�/�, 12 TNF-a�/�, and 45 wild-typeadult mice, weighing 22 to 24 g, were used for in vivo andin vitro studies. In addition, five adult nestin-greenfluorescent protein (GFP) mice (G Enikolopov, ColdSpring Harbor Laboratory, NY, USA), weighing 28 to30 g, were used for progenitor sorting. Animals werehoused separately under 12 h light/dark conditions withad libitum access to food and water.

Nineteen mice (TNF-R1�/�, n = 10; wild type, n = 9) wereanaesthetized with halothane and implanted with atwisted, insulated stainless-steel stimulating/recordingelectrode (Plastics One, Roanoke, VA, USA) unilaterallyinto the ventral hippocampal CA1–CA3 region (coordi-nates: 2.9 mm caudal and 3.0 mm lateral to bregma,3.0 mm ventral from dura, tooth bar at �3.3 mm).

Induction of Stroke

The intraluminal filament model was performed on 23mice (TNF-R1�/�, n = 8; wild-type, n = 4; TNF-a�/�, n = 5;wild-type, n = 6) (Hara et al, 1996). Briefly, right carotidarteries were exposed, external carotid was ligated, andtemporary sutures were placed around the common andinternal carotid arteries. A small incision was made in theexternal carotid artery and an 8-0 monofilament (AlconSurgical Inc., Stockholm, Sweden) coated with silicone(Xantopren L, Heraeus, Germany) was advanced throughthe internal carotid artery until it blocked the blood flowin the middle cerebral artery. The filament was secured,wounds temporarily closed, and animals allowed to wakeup. For reperfusion, mice were reanesthetized after40 mins of occlusion and the filament was removed. Bodytemperature was maintained between 36.51C and 37.51Cwith a heating lamp during surgery and ischemia, and for2 h thereafter. To assess whether the MCAO had beensuccessful in individual mice, we performed a contral-ateral rotation test (Bederson et al, 1986) and a forelimband hindlimb placing test (De Ryck et al, 1989) 2 h afterthe onset of reperfusion. Animals without motor impair-ment were excluded from the study. At 1 week after stroke,body weight did not differ between wild-type (28±0.4 g),TNF-a�/� (28±0.6 g), and TNF-R1�/� mice (27±0.6 g), andthe mortality rate was less than 10% in all groups withoutsignificant differences.

Induction of SE

Ten days after electrode implantation, 11 mice(TNF-R1�/�, n = 6; wild type, n = 5) were subjected toelectrically induced SE (Iosif et al, 2006). Mice received1 h of suprathreshold stimulation consisting of 10 secstrains of 1 ms biphasic square wave pulses at a frequency

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1575

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

of 50 Hz. Stimulation was interrupted for 1 min every10 mins to allow for EEG recording and measurement ofafterdischarges (MacLab; AD Systems, Hastings, UK).After ending the stimulation, all mice exhibited self-sustained, continuous ictal hippocampal EEG activity andassociated motor behavioral convulsions, categorized intopartial and generalized seizures (Iosif et al, 2006).Behavioral convulsions and ictal EEG activity werearrested with pentobarbital (40 mg/kg intraperitoneal) at2 h after stimulation offset.

Bromodeoxyuridine Administration

To label mitotic cells (Dolbeare, 1995), mice wereadministered the thymidine analogue BrdU (50 mg/kg,intraperitoneal), dissolved in potassium phosphate-buffered saline (KPBS), four times with a 2 h intervaland were perfused 2 h thereafter. Animals were injectedwith BrdU at 7 days after MCAO or SE.

Immunohistochemistry

All animals received an overdose of sodium pentobarbitaland were transcardially perfused with 100 mL ofsaline followed by 100 mL of ice-cold 4% paraformalde-hyde in 0.1 mol/L KPBS. Brains were removed, postfixedovernight, and placed in 20% sucrose in 0.1 mol/Lphosphate buffer for 24 h. Coronal sections (30mm) werecut on a freezing microtome and stored in cryoprotectivesolution.

For double-label immunofluorescence with BrdU andthe neuroblast marker doublecortin (Dcx) (Brown et al,2003) or neuron-specific nuclear protein (NeuN), free-floating sections were denatured in 1 mol/L HCl for30 mins at + 651C, rinsed in KPBS, preincubated with2% donkey, 2% horse, or 2% goat serum in 0.25% TritonX-100 in KPBS (T-KPBS) for 1 h, and incubated overnightwith rat anti-BrdU antibody (1:100; Sigma, Stockholm,Sweden), goat anti-Dcx antibody (1:400; Santa CruzBiotechnology Inc., Santa Cruz, CA, USA), or mouseanti-NeuN primary antibodies (1:100; Chemicon, Temecula,CA, USA) at + 41C. Sections were then rinsed andincubated in darkness for 2 h with Cy3-conjugated donkeyanti-rat (1:200; Jackson ImmunoResearch), biotinylatedhorse anti-goat (1:200; Vector Laboratories, Peterbourough,UK), or biotinylated horse anti-mouse secondaryantibodies (1:200; Vector Laboratories), then incubatedin Streptavidin Alexa Fluor 488 (1:200; Invitrogen,Stockholm, Sweden), mounted on glass slides, andcoverslipped with glycerol-based mounting medium.

For staining of proliferating cell nuclear antigen(PCNA), which is a protein involved in DNA replication(Paunesku et al, 2001), phosphorylated histone H3 (p-H3),which labels cells in mitotic phase (Hendzel et al, 1997),and macrophage/microglia-specific protein Iba1, whichidentifies both activated and nonactivated forms ofmicroglia (Imai and Kohsaka, 2002), free-floating sectionswere quenched with 3% H2O2 and 10% methanol in KPBSfor 30 mins at room temperature, rinsed in KPBS,preincubated for 1 h with 2% horse serum in T-KPBS for

PCNA and 2% goat serum in T-KPBS for p-H3 or Iba1.Sections were then incubated overnight at + 41C withmouse anti-PCNA antibody (1:1,000; Santa Cruz Biotechno-logy Inc.), rabbit anti-p-H3 antibody (1:400; Upstate,Charlottesville, VA, USA), and rabbit anti-Iba1 antibody(1:1,000; Wako Chemicals GmbH, Neuss, Germany), rinsedand incubated with biotinylated horse anti-mouse IgGantibody (1:200; Vector Laboratories) for PCNA or bioti-nylated goat anti-rabbit IgG antibody (1:200; VectorLaboratories) for p-H3 and Iba1. Sections were incubatedwith the avidin–biotin–peroxidase complex (Elite ABCkit, Vector Laboratories) for 1.5 h, treated with diamino-benzidine (0.5 mg/mL) and 3% hydrogen peroxide, andfinally rinsed in KPBS, mounted on glass slides, anddehydrated in ethanol, before being coverslipped withPertex mounting medium.

For Fluoro-Jade staining, free-floating brain sectionswere mounted on glass slides and allowed to dryovernight. They were then rehydrated, pretreated with0.06% potassium permanganate for 15 mins, rinsed indistilled water, incubated in 0.01% Fluoro-Jade workingsolution (Histo-Chem, Jefferson, AR, USA) for 30 mins,rinsed in distilled water, immersed in xylene, and cover-slipped (Schmued et al, 1997).

For terminal deoxynucleotidyl transferase-mediatedfluorescein-dUTP nick-end labeling (TUNEL), free-float-ing sections were mounted on glass slides and allowed todry. Sections were then pretreated with 4% paraformal-dehyde for 20 mins, methanol for 30 mins, proteinase K(10mg/mL in KPBS) for 6 mins, 4% paraformaldehyde for5 mins, and ice-cold 0.1% Triton X-100 in 0.1% sodiumcitrate for 2 mins, with KPBS rinses between each step.Subsequently, sections were incubated in the dark at +371C for 60 mins in terminal deoxynucleotidyl transferasebuffer, containing 17 mL terminal deoxynucleotidyl trans-ferase enzyme solution and 150mL TUNEL label solutionwith fluorescein-conjugated dUTP (Boehringer Mannheim,Mannheim, Germany). Sections were counterstained withHoechst 33342 (Molecular Probes, Stockholm, Sweden),10mg/mL, for 10 mins in darkness (Whiteside et al, 1998).Slides were coverslipped with glycerol-based mountingmedium.

Microscopical Analysis

All assessments were performed by an observer blind totreatment conditions. Immunostainings were examinedusing an Olympus AX-70 fluorescence light microscope.Stereological estimations of the total number of BrdU +and PCNA + cells in the SVZ were performed using theoptical fractionator method (Gundersen and Jensen, 1987;West et al, 1991). Briefly, four coronal sections throughoutthe SVZ, located 0.74, 0.50, 0.14, and 0.02 mm anterior tobregma, were analyzed ipsilaterally and contralaterallyusing an Olympus BH-2 microscope with a � 100objective, CCD-IRIS color video camera, and CAST-GRIDsoftware (Olympus, Ballerup, Denmark). For systematicsampling, the frame area was set at 1,491.7 mm2 with a steplength of 40 mm in X and Y directions, and the opticaldissector constituting a 5 mm thick fraction of the total

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1576

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

section thickness. Infarct volume was measured usingstereological equipment in four NeuN- and 4 Fluoro-Jade-stained sections throughout the striatum ( + 0.74, + 0.50,+ 0.14, and + 0.02 mm from bregma) as previouslydescribed by Thored et al (2006). The size of the SVZwas measured at the same rostrocaudal levels in BrdU/NeuN-stained sections. The density of Dcx + cells in therostral migratory stream was estimated by measuring thearea of the rostral migratory stream in a 30 mm thickcoronal section using stereological equipment. Numbersof Dcx + cells were counted using an Olympus BH-2microscope with a � 40 objective in the same section.Numbers of migrating Dcx + cells in the stroke-damagedstriatum were estimated in four BrdU/Dcx-stained sec-tions in a 150mm wide column adjacent to SVZ. Thecolumn was subdivided into three equally sized, 50 mmwide columns. Numbers of Dcx + cells in SVZ wereestimated semiquantitatively: score 1, < 20%; 2, 20% to50%; 3, > 50% of the SVZ area ipsilateral to damage iscovered by Dcx + cells. Owing to their low numbers,counts of BrdU + /Dcx + , p-H3 + , Iba1 + , and TUNEL +cells in the SVZ are given as cells per section. Colocaliza-tion of BrdU and Dcx was validated in a confocal laserscanning microscope (Leica). TUNEL + cells were countedif a clear apoptotic nuclear morphology (shrunken orfragmented nuclei), as assessed by Hoechst staining, wasidentified. Cell counts represented in Figures 1, 2, 4, and 5are from SVZ on the side ipsilateral to stroke orelectrically induced SE.

Dissection and Cell Culture

Animals were deeply anesthetized with halothane andkilled with cervical dislocation. Brains were removed,placed in ice-cold L-15 (Gibco, Stockholm, Sweden), andcut into 1 mm thick coronal slices. For gene expressionstudies, slices were immediately transferred to RNAlater(Ambion, Stockholm, Sweden). The SVZ was subdis-sected from 1 mm sections using microscissors andprocessed for cell culture, cell sorting, or gene expressionstudies. The SVZ was enzymatically dissociated in HBSSwith 0.015 mol/L HEPES, 5.4 mg/mL D-glucose, 1.33 mg/mLtrypsin (Gibco), 80 U/ml DNase, 0.7 mg/mL hyaluronidase,and 0.2 mg/mL kynurenic acid (Sigma) at + 371C for30 mins. The cell suspension was purified on a sucrosegradient (30% sucrose in 0.5� HBSS) followed bya bovine serum albumin gradient (4% bovine serumalbumin in Earle’s balanced salt solution with20 mmol/L HEPES). The SVZ tissue was dissected fromsix intact and six MCAO mice, ipsilateral and contralateralside being processed separately. For RNA isolation andcDNA synthesis, tissue from two mice was pooled, givingrise to three samples originating from each intact SVZ,ipsilateral SVZ, and contralateral SVZ after MCAO. A100 ng portion of cDNA was used from each sample andanalyzed in triplicate with quantitative PCR.

Neurospheres were grown in DMEM/F12 supplementedwith B27, 0.6% glucose (Sigma), 2 mmol/L L-glutamine,1.125% sodium bicarbonate, 15 mmol/L HEPES, and0.05 mg/mL gentamycin containing 20 ng/mL epidermal

growth factor, 10 ng/mL basic fibroblast growth factor(R&D Systems, Abingdon, UK) and 2 mg/mL heparin(Sigma) (proliferative medium). For neurosphere assays,cells were grown at 10 cells/mL in uncoated cell cultureflasks for 7 days. Neurosphere size was measured byoverlaying a scale bar on low-magnification digitizedphotographs in Adobe Photoshop. For counting, sphereswere transferred to 48-well plates with grids. In someexperiments, 20 ng/mL recombinant mouse TNF-a(R&D Systems) was added. Bulk cultures were grown at50 cells/ml. Neurospheres were fed every other day andpassaged every week using Accutase (PAA Laboratories,Pasching, Austria).

Neurospheres were passaged and cells grown adherentin proliferative medium without heparin on poly-D-lysinelaminin (Sigma)-coated 96-well plates or chamber slides(Nunc). Tumor necrosis factor-a (20 ng/mL) was added tothe medium for 24 h. To assess cell death, levels of lactatedehydrogenase were measured in culture medium using acolorimetric test (Roche, Bromma, Sweden). For TUNELstaining, chamber slides were fixed in 4% paraformalde-hyde for 15 mins at room temperature, washed in KPBS,and processed as above but without proteinase Ktreatment. The number of TUNEL + cells with clearapoptotic nuclear morphology, as assessed by Hoechststaining, was counted.

Cell Sorting, Reverse Transcription-PCR (RT-PCR),and Quantitative PCR

GFP + and GFP� cells from nestin-GFP animals weresorted on fluorescence-activated cell sorter (FACS DiVa;Becton Dickinson) directly into RLT lysis buffer (Qiagen,Solna, Sweden). An initial gate based on forward and sidescatter was set to exclude debris and cell aggregates;7AAD was used to exclude dead cells. A sorting gate wasset around the main GFP + population, at least 1 loghigher than GFP� controls.

RNA from nestin-GFP + and GFP� sorted cells, intact orstroke-damaged SVZ and striatum as well as from primaryand expanded neurospheres was isolated using RNAeasymicro (Qiagen) with DNase treatment. RNA was reversetranscribed using oligoDT primers and superscript-II(Invitrogen). Semiquantitative PCR was performed asdescribed elsewhere (Iosif et al, 2006).

Quantitative RT-PCR was performed with TaqManuniversal master mix and TaqMan Gene expression assayswere performed for GAPDH (Mm99999915_g1), TNF-a(Mm00443258_m1), and TNF-R1 (Mm00441889_m1).100 ng of cDNA was used for each quantitative PCR andall experiments were run in triplicate. cDNA input wasnormalized to GAPDH. Relative gene expression wascalculated using the DDCT method.

Statistical Analysis

Comparisons between animal groups were performedusing unpaired Student’s t-test. Side differences,ipsilateral vs contralateral to the ischemic damagewere assessed with paired Student’s t-test.

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1577

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

Figure 1 Deletion of TNF-R1 does not affect basal progenitor proliferation in SVZ and deletion of TNF-a does not alter progenitorproliferation in SVZ under basal conditions or after stroke. (A–C) Number of cells in the intact SVZ expressing the proliferationmarkers BrdU (A), p-H3 (B), and PCNA (C) in wild-type and TNF-R1�/� mice. Bromodeoxyuridine was injected four times with a 2 hinterval and animals perfused 2 h thereafter. Values denote means±s.e.m. n = 5 and 4 for wild type and TNF-R1�/�, respectively.No significant differences were observed, P > 0.05, unpaired Student’s t-test. (D) Photomicrographs showing the distribution ofBrdU + cells in the intact SVZ of wild-type and TNF-R1�/� mice. Scale bar: 100 mm. LV, lateral ventricle. (E–G) Number of cells inthe SVZ expressing the proliferation markers BrdU (E), p-H3 (F), and PCNA (G) in intact mice and at 1 week after 40 minsMCAO. Bromodeoxyuridine was injected four times with a 2 h interval and animals were perfused 2 h thereafter. Valuesdenote means±s.e.m.; n = 6 and 7 for intact wild-type and TNF-a�/� mice, and n = 6 and 5 for stroke-affected wild-typeand TNF-a�/� mice, respectively. *P < 0.05 compared with intact; no significant differences were observed between TNF-a�/� andwild-type mice, P > 0.05, unpaired Student’s t-test.

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1578

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

Comparisons between neurosphere numbers and sizewere performed using one-way analysis of variance withBonferroni–Dunn post hoc test. Comparisons of semi-quantitative scores were analyzed with nonparametricMann–Whitney U-test. Data are presented as means±s.e.m., and differences are considered significantat P < 0.05.

Results

Deletion of TNF-R1 and TNF-a Does Not Affect BasalProgenitor Proliferation in SVZ

We first studied whether absence of TNF-R1influenced SVZ cell proliferation in the intact

Figure 2 Deletion of TNF-R1 enhances the increased progenitor proliferation in SVZ after stroke. Photomicrographs showing thedistribution of BrdU + (A), p-H3 + (C), and PCNA + cells (E), and number of cells expressing BrdU (B), p-H3 (D), and PCNA (F) inwild-type and TNF-R1�/� mice in SVZ at 1 week after 40 mins MCAO. Bromodeoxyuridine was injected four times with a 2 h intervaland animals perfused 2 h thereafter. Values denote means±s.e.m.; n = 5 and 8 for wild-type and TNF-R1�/� mice, respectively.*P < 0.05 compared with wild type, unpaired Student’s t-test. Scale bar: 100 mm. LV, lateral ventricle.

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1579

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

Figure 3 Deletion of TNF-R1 enhances the formation of neuroblasts in the SVZ after stroke but the size of ischemic injury isunchanged. (A) Confocal images of a BrdU + /Dcx + cell showing BrdU and Dcx immunoreactivity separately and as a mergedimage. Orthogonal reconstructions from confocal z-series are represented as viewed in x–z (bottom) and y–z (right) planes.(B) Number of BrdU + /Dcx + cells in the SVZ ipsilateral and contralateral to the striatal injury at 1 week after 40 mins MCAO.(C) Confocal images showing Dcx + cells in wild-type and TNF-R1�/� mice in the SVZ at 1 week after 40 mins MCAO.(D–G) Photomicrographs showing NeuN staining in intact wild-type mice (D), and in wild-type (E), TNF-a�/� (F), and TNF-R1�/�

(G) mice at 1 week after 40 mins MCAO. Infarct volume was calculated in wild-type, TNF-a�/�, and TNF-R1�/� mice. The damage isoutlined with a broken black line. Values are means±s.e.m.; n = 4 and 6 for wild-type and TNF-R1�/�, respectively. *P < 0.05compared with wild type, unpaired Student’s t-test #P < 0.05 compared with contralateral SVZ, paired t-test. Scale bar: (A) 2.5 mm,(C) 50 and 12mm, (G) 200 mm. LV, lateral ventricle.

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1580

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

brain. Three different proliferation markers wereanalyzed using immunohistochemistry: BrdU, p-H3,and PCNA. No differences in the numbers ofBrdU + , p-H3 + , or PCNA + cells in SVZ weredetected between TNF-R1�/� and wild-type mice(Figures 1A–1C).

We also determined whether deletion of TNF-acaused alterations of cell proliferation in SVZ. Wedid not find any significant differences betweenTNF-a�/� and wild-type mice in the number of BrdU+ ,p-H3 + , or PCNA + cells under basal conditions(Figures 1E–1G).

Deletion of TNF-R1 but not TNF-a Enhances theIncreased Progenitor Proliferation in the SVZ AfterStroke

We wanted to explore whether after stroke, causingdamage to the adjacent striatum and microgliaactivation, signaling through TNF-R1 regulates cellproliferation in SVZ. Wild-type and TNF-R1�/� micewere subjected to 40 mins MCAO and cell prolifera-tion was analyzed 1 week thereafter. Stroke gave riseto higher numbers of BrdU + cells in the SVZbilaterally in both animal groups (Figure 2B)compared with intact animals (Figure 1A). Interest-ingly, after stroke, mice lacking TNF-R1 had ele-vated numbers of BrdU + , p-H3 + , and PCNA +cells in the SVZ compared with wild-type mice(Figure 2). There was no difference betweenwild-type and TNF-R1�/� mice in SVZ size at 1week after MCAO (data not shown).

We then determined whether the enhancement ofstroke-induced cell proliferation in the SVZ, causedby deletion of TNF-R1, also led to increased

neuroblast generation. As a neuroblast marker weused Dcx. Bromodeoxyuridine had been injected 8,6, 4, and 2 h before the animals were perfused at 1week after MCAO. The number of BrdU + /Dcx +neuroblasts was significantly higher in the SVZbilaterally in TNF-R1�/� as compared with wild-type mice (Figure 3B). In both TNF-R1�/� and wild-type mice, the number of BrdU + /Dcx + cells washigher in the SVZ on the stroke-damaged side. Weobtained no evidence that the higher number ofproliferating cells in the SVZ of TNF-R1�/� mice wasbecause of decreased apoptotic cell death. Numbersof TUNEL + cells with an apoptotic morphology(shrunken or fragmented nuclei) in SVZ did notdiffer between groups (2.9±0.6 cells in TNF-R1�/�

and 1.8±0.5 cells in wild-type mice).Hypothetically, the increased number of newly

proliferated neuroblasts in the SVZ of TNF-R1�/�

mice could be because of decreased migration.Tumor necrosis factor-a can induce migration ofneutrophils (Oliveira et al, 2008) and smooth musclecells (Rajesh et al, 2008). Moreover, administrationof a TNF-a-converting enzyme proteolytic inhibitorto SVZ cultures reduced neuroblast migration(Katakowski et al, 2007). We did not find any BrdU + /Dcx + cells in the injured striatum, rostralmigratory stream, or olfactory bulb, probably be-cause the time point was too early for them to haveleft SVZ. No differences in the density of Dcx + cellsin rostral migratory stream were observed betweenwild-type and TNF-R1�/� mice, either ipsilateral(2.5� 105±4.9� 104 cells/mm3 in wild-type and3� 105±2.4� 104 cells/mm3 in TNF-R1�/� mice) orcontralateral to the MCAO (2.5� 105±3.7�104 cells/mm3 in wild-type and 2.6� 105±1.7�104 cells/mm3 in TNF-R1�/� mice). Similarly, we

Figure 4 Deletion of TNF-R1 does not affect the increased progenitor proliferation in SVZ after SE. (A–C) Number of cells expressingthe proliferation markers BrdU (A), p-H3, (B) and PCNA (C) in the SVZ of wild-type and TNF-R1�/� mice after SE and in theirrespective, electrode-implanted but nonstimulated controls. Bromodeoxyuridine was injected four times with a 2 h interval on day 7after SE and animals perfused 2 h thereafter. Values are means±s.e.m.; n = 5 and 4 for wild-type SE and wild-type electrode-implanted controls, and n = 6 and 4 for TNF-R1�/� SE and TNF-R1�/� electrode-implanted controls, respectively. *P < 0.05compared with nonstimulated, unpaired Student’s t-test. No significant differences were observed between wild-type andTNF-R1�/� mice, P > 0.05.

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1581

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

found no differences in the total numbers ofDcx + cells in the striatum between wild-type andTNF-R1�/� mice ipsilateral (29±0.7 cells inwild-type and 31±1 cells in TNF-R1�/� mice) orcontralateral to the ischemic damage (17±0.7 cellsin wild-type and 18±0.6 cells in TNF-R1�/� mice).Also, the distance the new neuroblasts migrated intothe damaged striatum did not differ between groups(data not shown). In line with more efficientneuroblast generation, we observed higher numbersof Dcx + cells in the SVZ ipsilateral to the damage inTNF-R1�/� than in wild-type mice (SupplementaryFigure 1).

Figure 5 Stroke but not SE increases microglia numbers in SVZ.(A) Numbers of Iba1 + cells in the SVZ of TNF-a�/� and TNF-R1�/� mice and their respective wild-type controls in the intactbrain and at 1 week after 40 mins MCAO or SE. Both ‘Intact’and ‘MCAO’ show results from two separate experiments.(B) Photomicrographs showing the distribution of Iba + cells inthe SVZ in intact wild-type mice and in wild-type, TNF-a�/�,and TNF-R1�/� mice after MCAO. The SVZ is outlined with abroken white line. Means±s.e.m. n = 4 to 7 in each group.*P < 0.05 compared with corresponding intact, unpairedStudent’s t-test. No significant differences were observedbetween wild-type and TNF-R1�/� or TNF-a�/� mice,P > 0.05. Scale bar: 20mm. LV, lateral ventricle; SVZ,subventricular zone.

Figure 6 TNF receptor-1 is expressed in SVZ neural progenitorcells and upregulated together with TNF-a after stroke. (A) TNFreceptor-1 mRNA expression, as assessed with RT-PCR, inintact SVZ tissue, primary and expanded neurospheres, andin sorted nestin-GFP + and GFP� cells from intact SVZ.(B) Individual values (one dot for each sample) and mean (bar)values of TNF-a and TNF-R1 gene expression levels in striatumand SVZ at 1 week after 40 mins MCAO measured withquantitative real-time PCR on the side ipsilateral and contral-ateral to the damage. NS, neurospheres; P, passage.

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1582

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

Deletion of TNF-R1 has been reported to amelio-rate ischemic damage (Hallenbeck, 2002) and wehave previously shown (Thored et al, 2006) that theextent of striatal injury influences the magnitude ofthe neurogenic response after stroke. We wanted toexclude the possibility that the increased SVZ cellproliferation in TNF-R1�/� mice was an indirecteffect, caused by alterations in the size of theischemic injury. In NeuN- and Fluoro-Jade- (datanot shown) stained sections, we found no differ-ences in striatal infarct volume between wild-type(15.1±1.0 mm3; n = 6), TNF-a�/� (13.4±1.0 mm3;n = 5), and TNF-R1�/� (14.4±2.0 mm3; n = 8) miceafter stroke, amounting to 79%, 80%, and 81% ofthe total volume, respectively (Figures 3D–3G).

We finally analyzed the effect of deletion of TNF-aon cell proliferation 1 week after MCAO. Also in thisexperiment, we observed a significant increase in

the number of proliferating SVZ cells comparedwith intact animals using all three markers.However, no differences in the number of BrdU + ,p-H3 + , and PCNA + cells in SVZ, or the size ofstriatal injury, were detected between wild-type andTNF-a�/� mice (Figures 1E–1G and 3E and 3F).

Deletion of TNF-R1 Does Not Affect the IncreasedProgenitor Proliferation in the SVZ After SE

We then explored if the enhancement of cellproliferation caused by deletion of TNF-R1 wasspecific for stroke or also occurred in other patho-logic situations with increased proliferation in SVZ.Animals were subjected to SE for 2 h and during thisperiod showed partial or generalized convulsivebehavior. We previously observed no differences

Figure 7 Tumor necrosis factor-a inhibits neural progenitor proliferation without affecting survival in vitro. (A) Number and size(diameter in mm) of neurospheres formed from the SVZ of wild-type and TNF-R1�/� mice with and without the addition of TNF-a tothe medium. Means±s.e.m., n = 5. *P < 0.05, one-way analysis of variance with Bonferroni–Dunn post hoc test. Levels ofcytotoxicity (B), as assessed with lactate dehydrogenase release, and apoptosis (C), as assessed with Hoechst/TUNEL staining (% oftotal Hoechst + cells), in neurosphere cells from wild-type mice and TNF-R1�/� mice with and without the addition of TNF-a to themedium. No significant differences were observed between cultures obtained from wild-type and TNF-R1�/� mice, P > 0.05. Valuesare means±s.e.m., n = 8 and 4, respectively. (D) Number of spheres formed from wild-type and TNF-R1�/� mice at 1 week afterstroke. Means±s.e.m., n = 5, P = 0.08, one-way analysis of variance with Bonferroni–Dunn post hoc test.

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1583

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

between TNF-R1�/� and wild-type mice in anyseizure parameter, that is, percentage of partial andgeneralized convulsions, or time to developcontinuous ictal hippocampal EEG activity (Iosifet al, 2006).

We analyzed cell proliferation in the SVZ at 1week after SE. The epileptic condition gave rise tosignificant increases (25% to 30%) in the number ofproliferating BrdU + , p-H3 + , and PCNA + cells inboth wild-type and TNF-R1�/� mice as comparedwith their electrode-implanted, nonstimulated con-trols (Figure 4). We found no differences in SVZ inany of the cell proliferation markers between thetwo groups of mice after SE (Figure 4).

Stroke but not SE Increases Microglia Numbers in theSVZ: No Effect of TNF-a or TNF-R1 Deletion

Because microglia is a major source of TNF-a, weexplored if the numbers of these cells in SVZ aredifferentially regulated after stroke and SE, andinfluenced by deletion of TNF-R1 or its ligand.Tumor necrosis factor-a can induce microglia pro-liferation by means of TNF-R1 (Dopp et al, 1997) andeffects of TNF-R1 deletion on SVZ cell proliferationcould hypothetically be caused by changes inmicroglia numbers. As a marker for microglia, weused Iba1. No changes in SVZ Iba1 + cell numberswere observed in wild-type mice subjected to SE(Figure 5). In contrast, we found that MCAO gaverise to major increases in the numbers of Iba1 + cellsbilaterally in the SVZ (Figure 5A). Numerous Iba1 +cells were detected not only in the SVZ but also inthe striatum ipsilateral to MCAO, whereas highnumbers of these cells on the contralateral side werefound only in SVZ. Neither in intact mice nor afterSE or MCAO did deletion of TNF-R1 or TNF-a causeany changes in the number of Iba1 + cells in the SVZcompared with wild type (Figure 5).

TNF-R1 is Expressed by Neural Stem/Progenitor Cellsin SVZ and Upregulated Together with TNF-a AfterStroke

To explore if the observed effect of TNF-R1 signalingon SVZ cell proliferation could be mediated directlyon the progenitor cells themselves, we analyzed thepresence of TNF-R1 using RT-PCR. The mRNA ofTNF-R1 was expressed in SVZ tissue and in primaryand expanded neurospheres (Figure 6A). We thensorted GFP + and GFP� cells from the SVZ ofnestin-GFP mice. In these animals, GFP is expressedunder the nestin promoter and, therefore, labelsprogenitor cells in the SVZ (Mignone et al, 2004).We found expression of TNF-R1 in GFP + andinterestingly also in GFP� cells (Figure 6A). Levelsof TNF-a were very low or undetectable by RT-PCRunder basal conditions.

Both TNF-a and TNF-R1 are upregulated in thestroke-damaged brain (Lambertsen et al, 2005) but

whether any changes occur in SVZ is unknown. Wesubdissected SVZ and striatum 1 week after strokeand analyzed gene expression for TNF-a and TNF-R1 using quantitative real-time PCR. We foundmassive upregulation of TNF-a (296-fold) andTNF-R1 mRNA (8-fold) in the stroke-injured stria-tum, and also a 50- and 6-fold upregulation of TNF-aand TNF-R1 mRNA, respectively, in the tissuedissected from the ipsilateral SVZ (Figure 6B).Levels of expression of ligand and receptor incontralateral SVZ were not changed.

Tumor Necrosis Factor-a Decreases Proliferation ofNeural Stem/Progenitor Cells Through TNF-R1 in SVZNeurosphere Cultures

To further establish if deletion of TNF-R1 influencesSVZ progenitor proliferation, we isolated cells fromthe SVZ of wild-type and TNF-R1�/� mice and usedthe neurosphere formation assay. In agreement within vivo data, we found no differences between wild-type and TNF-R1�/� mice in numbers or size ofprimary neurospheres under basal conditions(Figure 7A). However, when recombinant mouseTNF-a was added, there was a 30% and 25%reduction in the numbers and size, respectively, ofneurospheres formed from wild-type SVZ. In con-trast, in cultures from TNF-R1�/� mice, addition ofTNF-a had no effect on sphere numbers or size(Figure 7A). As signaling through TNF-R1 caninduce cell death (Fontaine et al, 2002), we exploredwhether the effects of TNF-a addition were becauseof decreased survival. The lactate dehydrogenasetest was used to assess general cell death (Figure 7Band Supplementary Figure 2) and TUNEL stainingto show apoptosis (Figure 7C). However, we did notdetect any differences between wild-type and TNF-R1�/� mice under basal conditions or after theaddition of TNF-a to the cultures.

We compared neurosphere formation from theSVZ of wild-type and TNF-R1�/� mice at 1 weekafter stroke, that is, at the time point when we hadfound upregulation of TNF-a and TNF-R1 geneexpression. There was a trend (P = 0.08) towardincreased number of neurospheres formed in thetissue from TNF-R1�/� mice (Figure 7D). The mostlikely explanation to this modest change is that thelevels of endogenously produced TNF-a in the stem/progenitor cells in vitro were too low to significantlysuppress neurosphere formation.

Discussion

Here, we show that mice with loss of TNF-R1function respond to stroke with enhanced SVZ cellproliferation. In contrast, deletion of TNF-R1 did notinfluence basal cell proliferation in SVZ. However,we have previously found increased number ofproliferating cells in the SGZ in intact TNF-R1�/�

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1584

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

mice (Iosif et al, 2006), indicating that under basalconditions, TNF-R1 signaling acts to suppressprogenitor proliferation in SGZ but not in SVZ.

Several lines of evidence indicate that the highernumber of newly formed cells in the SVZ of TNF-R1�/� mice after stroke was because of increasedproliferation of neural stem/progenitor cells. First,the number of proliferated BrdU + /Dcx + was higherin TNF-R1�/� as compared with wild-type mice.Second, deletion of TNF-R1 did not alter theelevated microglia numbers in the SVZ. Third, wefound expression of TNF-R1 in the GFP + progenitorcells sorted from SVZ tissue of nestin-GFP mice.Because we also detected TNF-R1 expression in theGFP� fraction, the effect of TNF-R1 signaling onprogenitor proliferation could also be indirect andmediated through other cells in the SVZ, such asmicroglia or astrocytes, which are known to expressTNF-R1 (Dopp et al, 1997). Fourth, in the in vitroexperiments, addition of TNF-a decreased neuro-sphere size and numbers through a TNF-R1-depen-dent mechanism without affecting survival. Takentogether, our results provide the first evidence thatTNF-R1 signaling mediates a suppressant effect on theproliferation of progenitor cells in the SVZ after stroke.

Previous work has shown a similar role of TNF-R1signaling in other stem/progenitor cell systems.Activation of TNF-R1 inhibited proliferation ofhematopoietic stem cells (Dybedal et al, 2001).Also, TNF-a suppressed the proliferation ofTNF-R1-expressing neural progenitors derived fromneonatal rat striatum (Ben-Hur et al, 2003). We haverecently obtained evidence indicating that in SGZ,TNF-R1 is a negative regulator of progenitor pro-liferation both under physiologic conditions andafter SE (Iosif et al, 2006). Seemingly in contrast toour present data, Katakowski et al (2007) observedthat proliferation in the SVZ was promoted by theprotease activity of TNF-a-converting enzyme. Thisdiscrepancy is most likely explained by the factthat besides being a convertase for TNF-a,TNF-a-converting enzyme activates many othersubstrates, for example, growth factors andcytokines, which could promote proliferation(Katakowski et al, 2007). The results presented hereare also at variance with those of Widera et al (2006),who reported that TNF-a triggered proliferation ofneural stem cells derived from the adult SVZ.However, Widera et al (2006) studied cultured cellsderived from rats, whereas we have analyzed cellproliferation in the SVZ of mice both in vitro andin vivo. Also, we specifically investigated signalingthrough TNF-R1 in knockout mice, whereasWidera et al (2006) used human recombinantTNF-a, supposedly acting only through TNF-R1.Crossreactivity of human TNF-a on TNF-R2or species differences in signaling through TNFreceptors cannot be excluded.

In contrast to the effects of deletion of TNF-R1 onthe SVZ response after stroke, we observed nochanges in proliferation in mice with loss of TNF-a

function. Similarly, basal proliferation in SGZ didnot differ between TNF-a�/� and wild-type mice (REIosif et al, unpublished observations) even if therewere alterations in mice with deletion of thereceptors (Iosif et al, 2006). Also, retinal ischemia,which caused reduced and increased neuronal lossin TNF-R1�/� and TNF-R2�/� mice, respectively, didnot induce any changes in TNF-a�/� mice (Fontaineet al, 2002). Taken together, these findings indicatethe possible existence of a TNF-a-independentmechanism operating, for example, through theaction of another ligand such as lymphotoxin a,which binds to TNF-R1 (Wajant et al, 2003). Thismechanism could, hypothetically, be compensatoryand a consequence of the ablation of TNF-a alreadyduring embryonic development.

Loss of TNF-R1 function had differential effectson SVZ cell proliferation following the two patho-logic conditions stroke and SE. Although bothinsults gave rise to increased cell proliferation inSVZ, the dampening action of TNF-R1 on prolifera-tion was revealed only after stroke. It is interestingto note that at the time of increased cell prolifera-tion, that is, 1 week after stroke or SE, we detectedhigh numbers of microglia in the SVZ of stroke-damaged brains, but not in mice subjected to SE.The elevated number of microglia in the SVZ,possibly associated with increased release of TNF-a,could attenuate the stroke-induced proliferativeresponse through TNF-R1 signaling. It should bepointed out, though, that at 1 week after MCAOin wild-type animals, we detected elevated TNF-aand TNF-R1 mRNA levels only in SVZ tissue fromthe side ipsilateral to the damage. In contrast, theenhancement of progenitor proliferation in the SVZof TNF-R1�/� mice was observed bilaterally. Thisdiscrepancy could indicate that the TNF-R1-mediated suppressant effect on progenitor prolifera-tion is exerted by bilateral release of TNF-a oranother ligand, possibly from microglia, alreadybefore the 1-week time point. The high TNF-amRNA level in ipsilateral SVZ despite bilaterallyincreased microglia numbers suggests that themicroglia in SVZ on the side of the damage at 1week after stroke had adopted a more cytotoxicphenotype as compared with contralateral SVZ(Schwartz et al, 2006). The observed ipsilateraland contralateral effects on proliferation could alsobe because of elevated levels of TNF-a in serum andcerebrospinal fluid, which have been reported inpatients after stroke (Zaremba and Losy, 2001) andwould not be detected with gene expression analy-sis. Finally, it cannot be excluded that the elevatedmRNA expression in ipsilateral SVZ samples wasbecause of inclusion of some tissue from thedamaged striatum, in which we found TNF-a andTNF-R1 mRNA levels to be much higher.

In this study, we have identified the first negativeregulator of SVZ progenitor proliferation afterstroke. Many compounds and treatments have beenreported to increase SVZ cell proliferation after

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1585

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

stroke (for references, see Lindvall and Kokaia,2008). Much less is known about endogenousmechanisms regulating progenitor proliferation inthe SVZ. Notch1 and its naturally occurring activa-tor, Jagged1, are expressed in the adult SVZ (Stumpet al, 2002) and administration of Notch ligandsincreases the number of proliferating cells in theSVZ after stroke (Androutsellis-Theotokis et al,2006). Insulin-like growth factor-1 has also beenproposed to be a mediator of the increased SVZprogenitor proliferation after stroke (Yan et al, 2006).

The present data provide evidence for a suppres-sant role of TNF-R1 signaling in progenitor prolif-eration during stroke-induced neurogenesis. TNFreceptor-2 had no significant influence on cellproliferation in the SGZ (Iosif et al, 2006) and thisreceptor was, therefore, not studied here. However,TNF-a and its receptors most likely affect also othersteps of neurogenesis after stroke. Consistent with arole for the survival of the new neurons, infusion of anantibody against TNF-a reduces the number of striatalneuroblasts generated after stroke, possibly by inhibit-ing a neuroprotective action of TNF-a mediated byTNF-R2 (Heldmann et al, 2005). Hypothetically,TNF-a could also influence the functional synapticintegration of the new neurons, for example, bymodulating excitatory synaptic transmission (Picker-ing et al, 2005). Obviously, the actions of TNF-a and itsreceptors are complex and various consequences forstroke-induced neurogenesis could be envisaged. Thepresent data suggest that suppression of TNF-R1signaling might be a novel strategy to promotethe proliferative response in the SVZ after stroke.However, it will be important to establish how such anapproach will affect other steps of neurogenesis andthe functional outcome after stroke.

Acknowledgements

We thank M Lundahl and U Sparrhult-Bjork fortechnical assistance. This work was supported bythe Swedish Research Council, Juvenile DiabetesResearch Foundation, Swedish Diabetes Founda-tion, EU project LSHB-CT-2006-037526 (STEM-STROKE), and the Soderberg, Crafoord, Kock, andKing Gustav V and Queen Victoria Foundations. TheLund Stem Cell Center is supported by a Center ofExcellence grant in Life Sciences from the SwedishFoundation for Strategic Research.

Disclosure/conflict of interest

None.

References

Androutsellis-Theotokis A, Leker R, Soldner F, Hoeppner D,Ravin R, Poser S, Rueger M, Bae S, Kittappa R, McKay R(2006) Notch signalling regulates stem cell numbersin vitro and in vivo. Nature 442:823–6

Arvidsson A, Collin T, Kirik D, Kokaia Z, Lindvall O(2002) Neuronal replacement from endogenous precur-sors in the adult brain after stroke. Nat Med 8:963–70

Bederson JB, Pitts LH, Tsuji M, Nishimura MC, Davis RL,Bartkowski H (1986) Rat middle cerebral artery occlu-sion: evaluation of the model and development of aneurologic examination. Stroke 17:472–6

Ben-Hur T, Ben-Menachem O, Furer V, Einstein O,Mizrachi-Kol R, Grigoriadis N (2003) Effects of proin-flammatory cytokines on the growth, fate, and motilityof multipotential neural precursor cells. Mol CellNeurosci 24:623–31

Brown JP, Couillard-Despres S, Cooper-Kuhn CM, Winkler J,Aigner L, Kuhn HG (2003) Transient expression ofdoublecortin during adult neurogenesis. J Comp Neurol467:1–10

De Ryck M, Van Reempts J, Borgers M, Wauquier A,Janssen PA (1989) Photochemical stroke model: flunar-izine prevents sensorimotor deficits after neocorticalinfarcts in rats. Stroke 20:1383–90

Dolbeare F (1995) Bromodeoxyuridine: a diagnostic tool inbiology and medicine, Part I: historical perspectives,histochemical methods and cell kinetics. Histochem J27:339–69

Dopp J, Mackenzie-Graham A, Otero G, Merrill J(1997) Differential expression, cytokine modulation,and specific functions of type-1 and type-2 tumornecrosis factor receptors in rat glia. J Neuroimmunol75:104–12

Dybedal I, Bryder D, Fossum A, Rusten L, Jacobsen S(2001) Tumor necrosis factor (TNF)-mediated activa-tion of the p55 TNF receptor negatively regulatesmaintenance of cycling reconstituting human hemato-poietic stem cells. Blood 98:1782–91

Fontaine V, Mohand-Said S, Hanoteau N, Fuchs C,Pfizenmaier K, Eisel U (2002) Neurodegenerative andneuroprotective effects of tumor Necrosis factor (TNF)in retinal ischemia: opposite roles of TNF receptor 1and TNF receptor 2. J Neurosci 22:RC216

Gebicke-Haerter P (2001) Microglia in neurodegeneration:molecular aspects. Microsc Res Tech 54:47–58

Gundersen HJ, Jensen EB (1987) The efficiency ofsystematic sampling in stereology and its prediction.J Microsc 147(Part 3):229–63

Hallenbeck J (2002) The many faces of tumor necrosisfactor in stroke. Nat Med 8:1363–8

Hara H, Huang P, Panahian N, Fishman M, Moskowitz M(1996) Reduced brain edema and infarction volume inmice lacking the neuronal isoform of nitric oxidesynthase after transient MCA occlusion. J Cereb BloodFlow Metab 16:605–11

Heldmann U, Thored P, Claasen J, Arvidsson A, Kokaia Z,Lindvall O (2005) TNF-alpha antibody infusion impairssurvival of stroke-generated neuroblasts in adult ratbrain. Exp Neurol 196:204–8

Hendzel MJ, Wei Y, Mancini MA, Van Hooser A, Ranalli T,Brinkley BR, Bazett-Jones DP, Allis CD (1997) Mitosis-specific phosphorylation of histone H3 initiatesprimarily within pericentromeric heterochromatinduring G2 and spreads in an ordered fashioncoincident with mitotic chromosome condensation.Chromosoma 106:348–60

Imai Y, Kohsaka S (2002) Intracellular signaling in M-CSF-induced microglia activation: role of Iba1. Glia 40:164–174

Iosif R, Ekdahl C, Ahlenius H, Pronk C, Bonde S, Kokaia Z,Jacobsen S, Lindvall O (2006) Tumor necrosis factor

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1586

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587

receptor 1 is a negative regulator of progenitorproliferation in adult hippocampal neurogenesis.J Neurosci 26:9703–12

Jin K, Minami M, Lan J, Mao X, Batteur S, Simon R,Greenberg D (2001) Neurogenesis in dentate subgranu-lar zone and rostral subventricular zone after focalcerebral ischemia in the rat. Proc Natl Acad Sci USA98:4710–5

Jin K, Wang X, Xie L, Mao X, Zhu W, Wang Y, Shen J, Mao Y,Banwait S, Greenberg D (2006) Evidence for stroke-induced neurogenesis in the human brain. Proc NatlAcad Sci USA 103:13198–202

Katakowski M, Chen J, Zhang ZG, Santra M, Wang Y,Chopp M (2007) Stroke-induced subventricular zoneproliferation is promoted by tumor necrosis factor-alpha-converting enzyme protease activity. J CerebBlood Flow Metab 27:669–78

Lambertsen K, Clausen B, Fenger C, Wulf H, Owens T,Dagnaes-Hansen F, Meldgaard M, Finsen B (2007)Microglia and macrophages express tumor necrosisfactor receptor p75 following middle cerebral arteryocclusion in mice. Neuroscience 144:934–49

Lambertsen K, Meldgaard M, Ladeby R, Finsen B (2005) Aquantitative study of microglial-macrophage synthesisof tumor necrosis factor during acute and late focalcerebral ischemia in mice. J Cereb Blood Flow Metab25:119–35

Lindvall O, Kokaia Z (2008) Neurogenesis following strokeaffecting the adult brain. In: Adult neurogenesis (GageFH, Kempermann G, Song H, eds), New York: ColdSpring Harbor Laboratory Press, 549–70

Marchetti L, Klein M, Schlett K, Pfizenmaier K, Eisel U(2004) Tumor necrosis factor (TNF)-mediated neuro-protection against glutamate-induced excitotoxicity isenhanced by N-methyl-D-aspartate receptor activation.Essential role of a TNF receptor 2-mediated phospha-tidylinositol 3-kinase-dependent NF-kappa B pathway.J Biol Chem 279:32869–81

Mignone J, Kukekov V, Chiang A, Steindler D, Enikolopov G(2004) Neural stem and progenitor cells in nestin-GFPtransgenic mice. J Comp Neurol 469:311–24

Oliveira SH, Canetti C, Ribeiro RA, Cunha FQ (2008)Neutrophil migration induced by IL-1beta dependsupon LTB(4) released by macrophages and uponTNF-alpha and IL-1beta released by mast cells.Inflammation 31: 36–46

Parent J, Vexler Z, Gong C, Derugin N, Ferriero D(2002) Rat forebrain neurogenesis and striatal neuronreplacement after focal stroke. Ann Neurol 52:802–13

Pasparakis M, Alexopoulou L, Episkopou V, Kollias G(1996) Immune and inflammatory responses in TNFalpha-deficient mice: a critical requirement for TNFalpha in the formation of primary B cell follicles,follicular dendritic cell networks and germinal centers,and in the maturation of the humoral immuneresponse. J Exp Med 184:1397–411

Paunesku T, Mittal S, Protic M, Oryhon J, Korolev SV,Joachimiak A, Woloschak GE (2001) Proliferating cell

nuclear antigen (PCNA): ringmaster of the genome.Int J Radiat Biol 77:1007–21

Peschon JJ, Torrance DS, Stocking KL, Glaccum MB, Otten C,Willis CR, Charrier K, Morrissey PJ, Ware CB, MohlerKM (1998) TNF receptor-deficient mice reveal diver-gent roles for p55 and p75 in several models ofinflammation. J Immunol 160:943–52

Pickering M, Cumiskey D, O’Connor J (2005) Actions ofTNF-alpha on glutamatergic synaptic transmission inthe central nervous system. Exp Physiol 90:663–70

Rajesh M, Mukhopadhyay P, Hasko G, Huffman JW,Mackie K, Pacher P (2008) CB(2) cannabinoid receptoragonists attenuate TNF-alpha-induced human vascularsmooth muscle cell proliferation and migration.Br J Pharmacol 153:347–57

Schmued LC, Albertson C, Slikker W, Jr (1997) Fluoro-Jade: a novel fluorochrome for the sensitive and reliablehistochemical localization of neuronal degeneration.Brain Res 751:37–46

Schwartz M, Butovsky O, Bruck W, Hanisch UK (2006)Microglial phenotype: is the commitment reversible?Trends Neurosci 29:68–74

Stump G, Durrer A, Klein A, Lutolf S, Suter U, Taylor V(2002) Notch1 and its ligands Delta-like and Jagged areexpressed and active in distinct cell populations in thepostnatal mouse brain. Mech Dev 114:153–9

Thored P, Arvidsson A, Cacci E, Ahlenius H, Kallur T,Darsalia V, Ekdahl C, Kokaia Z, Lindvall O (2006)Persistent production of neurons from adult brain stemcells during recovery after stroke. Stem Cells 24:739–747

Wajant H, Pfizenmaier K, Scheurich P (2003)Tumor necrosis factor signaling. Cell Death Differ10:45–65

West MJ, Slomianka L, Gundersen HJ (1991) Unbiasedstereological estimation of the total number of neuronsin the subdivisions of the rat hippocampus using theoptical fractionator. Anat Rec 231:482–97

Whiteside G, Cougnon N, Hunt S, Munglani R (1998) Animproved method for detection of apoptosis in tissuesections and cell culture, using the TUNEL techniquecombined with Hoechst stain. Brain Res Brain ResProtoc 2:160–4

Widera D, Mikenberg I, Elvers M, Kaltschmidt C,Kaltschmidt B (2006) Tumor necrosis factor alphatriggers proliferation of adult neural stem cells viaIKK/NF-kappaB signaling. BMC Neurosci 7:64

Wu J, Kuo J, Liu Y, Tzeng S (2000) Tumor necrosis factor-alpha modulates the proliferation of neural progenitorsin the subventricular/ventricular zone of adult ratbrain. Neurosci Lett 292:203–6

Yan Y, Sailor K, Vemuganti R, Dempsey R (2006) Insulin-like growth factor-1 is an endogenous mediator of focalischemia-induced neural progenitor proliferation. Eur JNeurosci 24:45–54

Zaremba J, Losy J (2001) Early TNF-alpha levels correlatewith ischaemic stroke severity. Acta Neurol Scand104:288–95

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

Stroke-induced progenitor proliferation and TNF-R1RE Iosif et al

1587

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 1574–1587