Long-Lasting Regeneration After Ischemia in theCerebral Cortex

Ronen R. Leker, MD; Frank Soldner, MD; Ivan Velasco, PhD; Denise K. Gavin, PhD;Andreas Androutsellis-Theotokis, PhD; Ronald D.G. McKay, PhD

Background and Purpose—Because fibroblast growth factor 2 is a mitogen for central nervous system stem cells, weexplored whether long-term fibroblast growth factor 2 delivery to the brain can improve functional outcome and inducecortical neurogenesis after ischemia.

Methods—Rats underwent permanent distal middle cerebral artery occlusion resulting in an ischemic injury limited to thecortex. We used an adeno-associated virus transfection system to induce long-term fibroblast growth factor 2 expressionand monitored behavioral and histological changes.

Results—Treatment increased the number of proliferating cells and improved motor behavior. Neurogenesis continuedthroughout 90 days after the ischemia, and the occurrence of newly generated cells with characteristics of neuralprecursors and immature neurons was most evident 90 days after treatment.

Conclusions—Focal cortical ischemia elicits an ongoing neurogenic response that can be enhanced with fibroblast growthfactor 2 leading to improved functional outcome. (Stroke. 2007;38:153-161.)

Key Words: growth factors � neural progenitors � neural stem cells � neurogenesis � stroke

Cell therapy has been proposed in ischemia but is complexbecause it requires harvesting appropriate cells, their ex-

pansion in vitro and transplantation.1 Previous studies showedthat endogenous precursors can generate new neurons in theischemic striatum but not in the cortex, despite a large lesionburden at this site.2,3 However, newer studies show that repairmechanisms may allow cortical recovery as well,4–6 and thuscortical regeneration after ischemia remains controversial. Ourstudy was prompted by data showing a slow and growthfactor–dependent recovery from transient global ischemia.7

Fibroblast growth factor 2 (FGF2) is a major mitogen for neuralstem cells (NSC).8 An increase in infarct volume in mice lackingFGF2 and amelioration of focal deficits by FGF2 confirm theimportance of this signaling pathway in the ischemic cortex.9

We therefore used an adeno-associated viral (AAV) vector10 todeliver FGF2 over long periods of time after ischemic corticalischemia and searched for the effects on behavioral recovery andneurogenesis.

Materials and Methods

Vector PreparationAAV helper plasmids were generously provided by Drs Rabinowitzand Samulski (University of North Carolina). Recombinant AAVvectors (Figure 1) were generated and purified as previously de-scribed11 (for details please refer to supplemental Materials andMethods, available online at http://stroke.ahajournals.org).

AAV InjectionViruses were injected into 2 separate locations bordering the infarctimmediately after induction of cerebral ischemia (Figure 1) using aconvection system to enhance viral distribution. Overall, 40�109

viral particles were injected into each site.

Focal Ischemia in RatsSpontaneously hypertensive rats underwent permanent distal middlecerebral artery occlusion (PMCAO) by craniotomy and electrocoag-ulation.12 Animals were divided into 3 groups: A: shamPMCAO�AAV-FGF2 (n�3); B: PMCAO�AAV (n�28); and C:PMCAO�AAV-FGF2 (n�30). Animals were perfused 7 (n�2 inC), 21 (n�3 in B and C), 30 (n�21 in B and C, and n�3 in A) or90 days after the procedure (n�4 in B and C). All evaluation listedbelow were performed blinded to the treatment status.

BrdU InjectionsAnimals received IP injection of Bromo-deoxy-Uridine (BrdU; 50mg/kg q, 12 hours) on days 1 to 5 post-op to label dividing cells.

Ara-C InjectionsWe used ICV Cytosine Arabinosine (Ara-C) injection via mini-osmotic pumps in rats immediately after PMCAO to selectivelytarget dividing cells in the subventricular zone (SVZ) and eliminatetheir contribution to recovery.7 Rats were treated with ARA-C (2%in 0.9% NaCl) administered at a total dose of 0.5 mL into the lateralventricle via alzet mini-osmotic pumps (model 1007d). Animalswere divided into groups similarly to groups B and C describedabove and were given Ara-C or vehicle (n�3/group) for 7 days.

Proliferative History ExperimentsWe used a double-labeling strategy to determine proliferative historyin dividing cells after ischemia. PMCAO rats (n�3) were injected

Received July 23, 2006; final revision received August 23, 2006; accepted September 10, 2006.From the Laboratory of Molecular Biology (R.R.-L., R.S., I.V., D.K.G., A.A.-T., R.D.G.M.), National Institute of Neurological Disorders and Stroke,

National Institutes of Health, Bethesda, MD; and the Department of Neurology (R.R.L.), Hadassah University Medical Center, Jerusalem, Israel.Correspondence to R.R. Leker, MD, Department of Neurology, Hadassah University Medical Center, Hadassah Ein Kerem, PO Box 12000, Jerusalem

91120, Israel. E-mail leker@cc.huji.ac.il© 2006 American Heart Association, Inc.

Stroke is available at http://www.strokeaha.org DOI: 10.1161/01.STR.0000252156.65953.a9

153 by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

with Chloro-deoxy-Uridine (CldU) on days 1 to 5 and Iodo-deoxy-Uridine on days 42 to 46 as described.13 Animals were perfused 60days postischemia and their brains were evaluated.13

Motor-DisabilityAll animals were examined with a standardized disability scale(maximal deficit�10 points; no deficit�0) that incorporates motorfunction, limb placement tests and locomotion as previously de-

scribed14 (for details please refer to supplemental Materialsand Methods).

Injury SizeAnimals were killed 30 days postischemia. Brain slices 200 �m apartwere stained with Geimsa stain. Measurement of the lesioned areawas done with image-analysis software in 10 rats killed at 30 days ineach group.

Figure 1. FGF2 and GFP expression. Experi-mental Paradigm (A): Cartoon showing detailsof the experimental design. Red x’s designatepoints of sacrifice and red lines points of dis-ability evaluation. Viral Vector Structure (B):Structure of the viral vector pTR-cFGF2-IG. CellCounting (C): Cells were enumerated in prede-termined ROI including the SVZ (red), whitematter (yellow) and peri-infarct cortex (white).Note that the remains of the core region (black)are at the edge of the missing infarcted cortexthat has liquefied at 30 days postischemia.*Sites of viral injection. FGF2-GFP expressionin the AAV-FGF group (D): Photomicrograph(�100) showing expression of FGF2 (red) andGFP (green) in the brain of an AAV-FGF2–treated rat 30 days postischemia. Picturestaken from coronal brain sections at bregma—1.2 mm show the peri-infarct cortical area.Insert, Z-section (�630) through the area in theblue rectangle showing colocalization. GFP isexpressed in many cortical neurons (E): Pho-tomicrograph (�630) showing GFP (green)expression in NeuN� neurons (red) in the cor-tex. Insert, z-section through the rectangledemonstrating colocalization. Note that colocal-ization is evident in some (blue arrows) but notall (yellow arrow) neurons. Scale bars�100 �m,20 �m.

154 Stroke January 2007

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

ImmunohistochemistryBrains were frozen-sectioned at 16 �m and double- or triple-stainedfor immunohistochemistry evaluation using fate-specific antibodies(see supplemental Materials and Methods). Confocal images weretaken under Zeiss LSM-510 microscope and multiple z-sections wereobtained to ensure colocalization.

Immunopositive cells were counted in a blinded fashion using anonbiased system. Cells were enumerated on slides obtained fromhomologous coronal slices containing the infarct between �0.2 and�3.8 mm from bregma. Cells were counted in 4 regions of interest(ROI): SVZ on the infarct side, corpus callosum and white matter onboth sides and the peri-infarct cortex (Figure 1). In each slide 7high-power fields (x630) separated by 150 �m were counted in eachROI. At least 5 slides were counted for each animal.

Statistical AnalysisAnalysis was performed with the SigmaStat package (SPSS Inc).Data are presented as mean�SD and comparisons between allgroups (sham and the experimental groups) were performed usinganalysis of variance (ANOVA) with Bonferroni correction.

ResultsFGF2 Expression in the BrainAAV vectors delivered the transgene to a large region of theinjured hemisphere with prolonged expression time. FGF2expressing cells were seen throughout the infarcted hemi-sphere and were not limited to the injection sites. The resultsshow that many double-positive cells for FGF2 and greenfluorescent protein (GFP) were present in the injected hemi-sphere 30 days postischemia in the AAV-FGF group (�65%of cells) but not in the AAV controls (Figure 1A and 1B) andthat many of the infected cells were neurons (�85%; Figure1C and 1D, and supplemental Figure I, available online athttp://stroke.ahajournals.org). FGF2 was also expressed byastrocytes in the ischemic hemisphere (supplemental FigureI) but no GFP/glial fibrillary acidic protein (GFAP)� cellswere observed, suggesting that this FGF2 expression repre-sents a response to injury and does not result from AAVtransfection.

Lesion SizeDistal PMCAO resulted in lesions occupying the fronto-lateral cortex but left intact the vascular supply to the SVZ.15

Infarct size tended to be larger in the FGF2 group but theintergroup difference was not statistically significant (by1-way ANOVA with 2 degrees of freedom, F�2.3; P�0.31;Figure 2A).

Neurogenesis at 7 Days After IschemiaAt 7 days postischemia a robust increase in the number ofnewborn BrdU� cells was observed in the SVZ, the corpuscallosum and in the peri-infarct area. This increment waslarger in FGF2-treated rats (Figure 2B). About 10% of theseBrdU� cells colabeled with microglia markers (see below).When Ara-C was injected into the lateral ventricle for 7days,7 the proliferative response observed in the peri-infarctcortex was significantly reduced but not completely abolished(Figure 2B). Most cells that expressed BrdU in the cortex(�85%) after 7 days of treatment with ARA-C were alsopositive for the microglial markers ED1 and CD11b. Theseresults show that cell proliferation occurs in the injured cortexand suggest that most newborn cells with a nonmicroglial

phenotype in the cortex originate in the SVZ. None of the fewnonmicroglial BrdU� cells seen in the cortex after applica-tion of Ara-C expressed NeuN or GFAP and most expressednestin.

Neurogenesis at 30 to 90 Days After IschemiaIn sham-operated rats killed 30 days after the surgery, 97% ofBrdU� cells were seen in the SVZ and another 3% in thewhite matter. Extremely rare BrdU� cells were observed inthe cortex (4 cells in 60 high-power fields). The number ofSVZ BrdU� cells in sham-operated animals was signifi-cantly smaller than their number in rats with stroke (2-foldless). Most of these cells were nestin� (78%) at the SVZ.None of the BrdU� cells in these animals coexpressedneuronal or glial markers in the white matter or cortex. Inischemic rats BrdU� cells were seen in the SVZ, the whiteand the gray matter surrounding the infarct at days 30 and 90after focal ischemia in both groups (Figure 2C and 2D).Surprisingly, only a few newborn cells were present in thestriatum. A milder increase in the number of newborn cellswas also observed in the contralateral brain. A significantreduction in the number of newborn cells was observed in theperi-infarct area between day 7 and 30 postischemia (Figure2B and 2C). This cell loss suggests that most cells born earlyafter the ischemia do not survive in the intermediate-term.

The number of BrdU� cells in the SVZ remained constantbetween day 30 and 90 postischemia (Figure 2C and 2D). Inmarked contrast, the number of BrdU� cells in the peri-infarct cortex increased in all groups at 90 days after theischemia, suggesting newborn cell accumulation at this areaover time. The late peak in the number of BrdU� cells wasmore robust in the AAV-FGF group (Figure 2D).

At 7 days after ischemia, 10% of the BrdU� cells in thecortex coexpressed the macrophage/microglia markers ED1(supplemental Figure II, available online at http://stroke.ahajournals.org), cd11b or IBA-1 (data not shown). Thepercentage of BrdU� cells coexpressing microglia markersremained high for 30 days postischemia in all groups but�1% of the BrdU� cells expressed microglia antigens 90days postischemia (supplemental Figure II). Similar inflam-matory responses were observed in both groups as judged byexpression of leukocyte and microglia markers. These resultssuggest that an increase in noninflammatory BrdU� cells inthe cortex occurs only after the inflammatory responsedeclines.

Timing of NeurogenesisTo further evaluate whether cells continue to divide at theperi-infarct area or divide only when at SVZ, we used asequential labeling protocol recently described.13 Briefly,cells were labeled first with CldU on days 1 to 5 postischemiaand then with IdU 42 to 46 days after the ischemia. Theresults (Figure 3) demonstrate that most cells continue todivide at the infarct boundary as many double-labeled cellswere seen at this location. In contrast, most dividing cells atthe SVZ on day 60 were only labeled with IdU. These resultsimply that progenitors first divide at the SVZ and onmigration to the infarct zone continue to divide over a long

Leker et al Cortical Regeneration After Focal Ischemia 155

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

period of time, whereas cells labeled at the SVZ only dividethere for a limited time.

Taken together these results suggest that neurogenesis is anongoing process that continues over long periods of time andthat this process can be significantly enhanced with FGF2stimulation.

Newborn Cell IdentityThe cellular changes that follow ischemia were furtheranalyzed by use of antibodies that identify central nervoussystem stem cells, precursors and mature cells (Figure 4).

At 7 days postischemia only rare BrdU� cells in the cortexcoexpressed neuronal or glial markers (0.17% BrdU�/NeuN�, 1.8% BrdU�/GFAP�).5,16 Most of these cellsexpressed markers for immature NSC including SOX2 andnestin and some of these cells (�60%) also expressedvascular endothelial growth factor (VEGF).

At 30 days postischemia, SOX2� NSC17 were abundant inthe ventricular region and at the infarct boundary zone where

they were particularly prevalent in the subcortical whitematter (Figure 4B). VEGF is expressed in the precursor cellsof the ventricular zone and later in immature neurons andsubsequently in astrocytes.18 VEGF expression was analyzedbecause this secreted protein promotes vascular developmentand may also act as a neural survival signal.18 The majority ofSOX2� cells coexpressed VEGF (Figure 4B’). The numberof SOX2� cells was increased in the AAV-FGF groupalmost 2-fold (Figure 4C). Many neural progenitor cells(NPC) expressing the proneural transcription factors MASH1and PAX619 were seen around the infarct (peri-infarct zone;Figure 4D through 4E). Treatment with AAV-FGF2 signifi-cantly increased the numbers of NPC (Figure 4F). In AAV-FGF-treated animals, 66.8�8.6% of the BrdU� cells werealso PAX6� (versus 15.5�4.5% in AAV-controls; by 1-wayANOVA with 1 degree of freedom, F�23.07; P�0.001).Importantly, the absolute numbers of SOX2� cells at theSVZ (Figure 4A), as well as the number of PAX6� or

Figure 2. Outcomes after ischemia. Infarct Volumes (A): Infarct volumes were measured 30 days after ischemia. Differences were statis-tically nonsignificant (P�0.310 by 1-way ANOVA with 1 degree of freedom [DOF], F�1.09). Effects of Ara-C on BrdU� cell number (B):Ischemic rats received vehicle or AAV-FGF2 only, or vehicle, AAV or AAV-FGF2 in addition to Ara-C (n�3/group) and BrdU� cells wereenumerated in ROI 7 days later. P�0.001 by 1-way ANOVA with 4 DOF, F�13.9. Number of BrdU positive cells in the brain (C and D):AAV-FGF2–treated animals had significantly more BrdU� cells in all ROI at day 30 (for corpus callosum DOF�2, F�20.18, P�0.001;and for the peri-infarct area DOF�2, F�15.6, P�0.001) and 90 postischemia (for SVZ DOF�1, F�9.8, P�0.002; for corpus callosumDOF�1, F�8.5, P�0.004; and for the peri-infarct area DOF�1, F�42.08, P�0.001) by 1-way ANOVA. The late increase in the numberof BrdU� cells at 90 days was significantly more robust in AAV-FGF2 rats. HPF indicates high-power field (�630).

156 Stroke January 2007

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

MASH1� cells in the peri-infarct area (Figure 4B and 4C)were much larger than the numbers of BrdU� cells indicatingthat labeling with BrdU greatly underestimates the number ofprogenitors. At 30 days postischemia we also observed someBrdU� cells that coexpressed the migrating neuroblastmarker doublecortin16 in the peri-infarct area (Figure 4G).Moreover, 12.9�2.9% of the BrdU� cells coexpressed theimmature neuronal antigen HU16 in the AAV-FGF group(versus 4.03�1.2% in AAV-controls; P�0.027 by 1-wayANOVA with 1 degree of freedom, F�4.92; Figure 4H). At

this time coexpression of BrdU and GFAP was commonlyobserved in AAV-controls (�65% of BrdU�, Figure 4H) andless often in AAV-FGF treated rats (�21% of BrdU�;P�0.001 by 1-way ANOVA with 1 degree of freedom,F�96.4).

At 90 days after ischemia, expression of the immaturemarkers SOX2 and nestin was still present in most of theBrdU� cells in the white matter and the region of corteximmediately adjacent to the infarct border (Figure 5A and5B). In regions of cortex farther from the lesion, BrdU� cells

Figure 3. Proliferative history in neuralprogenitors after ischemia. Animals wereinjected with CldU on days 1 to 5 andIdU on days 42 to 46 after ischemia andkilled 60 days postischemia. A, Photomi-crograph (�630) through the SVZ (yellowasterisk) shows that few cells remainCldU� (green) but many other cellsincorporate the more recently adminis-tered tracer IdU (red). Insert, Z-sectionthrough the area depicted in the bluerectangle showing colabeling with bothtracers in some cells (blue arrow) or onlywith IdU (yellow arrows). Note that mostdividing cells express the NPC markerPAX6 (white). B, Photomicrograph (�630)through the peri-infarct area showingthat most cells colocalize both tracersand coexpress nestin (white).

Leker et al Cortical Regeneration After Focal Ischemia 157

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

Figure 4. Newborn cells identity 30 days postischemia. Schematic representation of the antibodies used to identify newborn cells (A). NB indicatesneuroblasts; DCX, doublecortin; NEU, Neuron. Neural stem cell (B and C): Photomicrograph (�100) showing the SVZ of an AAV-FGF2 animal. Cellsexpressing the NSC marker SOX2 (B) are evident. Number of SOX2� cells is higher than that of BrdU� (B). Some SOX2� cells coexpress VEGF(B). Their number was significantly higher in animals treated with AAV-FGF (C). Neural progenitor cells (D through F): Z-series (�630) from the peri-infarct cortex of an AAV-FGF2 animal. MASH1� (red, D) and BrdU� (green) cells showing colocalization. PAX6� cells (red, E–E) were also presentin the peri-infarct area. Many of these cells were BrdU� (blue arrows depict BrdU�/PAX6� cells, yellow arrows depict BrdU�/PAX6- cells). Pro-genitors were more common in AAV-FGF2 animals (F and not shown, by 1-way ANOVA with 1 degree of freedom, F�23.07, P�0.001). Neuro-blasts (G): Photomicrograph (�250) showing the peri-infarct area from an AAV-FGF2 animal (G). Cells expressing DCX (red) frequently coexpressBrdU (green). G - z-series (�630) from the peri-infarct area showing the features of doublecortin� cells (red). Young Neurons (H): Z-series from theperi-infarct cortex of an animal treated with AAV-FGF2. About 12% of cortical newborn cells (BrdU�; green) coexpress the immature neuronalmarker Hu (red, thin light blue arrow). Relative percentage of mature cells (I): Bar graph showing the percentage of newborn cortical neurons to besignificantly higher (by 1-way ANOVA with 1 degree of freedom, F�4.92, P�0.027) and that of GFAP� lower (by 1-way ANOVA with 1 degree offreedom, F�96.4, P�0.001) in the AAV-FGF2 rats compared with the AAV-controls as shown in the bar graph (I). LV indicates lateral ventricle; PL,peri-infarct area; C, infarct cavity. Scale bars�100 �m (A and E), 20 �m (D and G), 10 �m (H).

158 Stroke January 2007

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

Figure 5. Differentiation of newborn cells 90 days postischemia. NSC: Photomicrograph (�100) of the peri-infarct area from anAAV-FGF2 rat (A) showing that many BrdU� cells (green) present in the peri-infarct white matter coexpress SOX2� (red). Notethat most BrdU� cells remain in the white matter but some migrate into the surrounding peri-infarct cortex where many BrdU�are negative for SOX2. Z-section (�630) from the area depicted in the blue rectangle demonstrating colocalization of BrdU andSOX2 (A). Z-section (B) obtained from the peri-infarcted cortex of an AAV-FGF2 showing colocalization of SOX2 (red) and nestin(green). Neuroblasts: Z-section (�630) obtained from the peri-infarcted cortex showing colocalization of BrdU (green) and dou-blecortin (red; blue arrows; C). Some cells are only positive for doublecortin (yellow arrow). Neurons: Photomicrograph (�630) ofthe peri-infarct cortex of an animal treated with AAV-FGF2 (D). Some BrdU� cells (green) also expressed the pan-neuronalmarker NeuN (red; blue arrow). D’, Z-section (�630) obtained from the area in the blue rectangle demonstrating colocalization.However, about 70% of the BrdU� cells in this area were negative for NeuN (yellow arrows). Note that numbers of BrdU� cellscoexpressing NeuN are significantly higher with FGF2 (E; by 1-way ANOVA with 1 degree of freedom, F�12.87, P�0.001) andGFAP/BrdU� cells are more prevalent in the AAV group (E; by 1-way ANOVA with 1 degree of freedom, F�13.05, P�0.001).Sensory-motor impairment (F): Deficits were evaluated with a standardized disability scale (maximal disability�10, no disability�0). Animalshad similar degree of disability at 1 day postischemia but AAV-FGF2 rats had significantly better improvement rates with better out-comes observed at 20 (by 1-way ANOVA with 1 degree of freedom, F�4.5, P�0.039), 30 (by 1-way ANOVA with 1 degree of freedom,F�4.19, P�0.047) and 45 days postinjury (by 1-way ANOVA with 1 degree of freedom, F�10.75, P�0.017). WM indicates white matter;*PL, perilesioned cortex; C, infarct cavity. Scale bars�100 �m (A), 20 �m (A, C, D and D).

Leker et al Cortical Regeneration After Focal Ischemia 159

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

expressing doublecortin (Figure 5C) and the mature neuronalmarker NeuN were present (Figure 5D–D). Immature andmature newborn neurons were more frequent in the AAV-FGF group (NeuN 29.78�3.4% and Hu 22.1�5.8% versus12.05�3.5% and 1.6�0.6%, respectively; P�0.001 for bothby 1-way ANOVA with 1 degree of freedom, F�12.87;Figure 5E). In contrast, newborn cells coexpressing GFAPwere significantly less frequent in the AAV-FGF group(P�0.001 by 1-way ANOVA with 1 degree of freedom,F�13.05; Figure 5E).

Collectively, these results show that by 30 and 90 daysafter the ischemia, FGF2 treatment results in a robust increasein the number of NSC and NPC at the SVZ and white matterand of NPC and immature neurons in the peri-infarct cortex.Importantly, newborn immature neurons could only be ob-served in the peri-infarct cortex located a few cell layers awayfrom the immediate infarct border where only more primitiveSOX2�/nestin� cells were seen.

Impact of Newborn Cells on Behavioral ChangesAnimals treated with AAV-FGF2 had significantly betterimprovement in behavior (Figure 5F). Differences betweengroups became apparent starting 20 days postischemia andbecame more robust with time as animals treated with FGF2continued to improve.

DiscussionThe data presented here suggest that: (1) partial behavioralrecovery can occur after a focal ischemic injury restricted tothe cortex; (2) newborn cells are first found in the SVZ,subsequently migrate through white matter, and accumulatein the cortex and subcortical white matter areas surroundingthe infarct forming a regenerative zone with many primitiveneural precursor cells immediately adjacent to the injuredtissue; (3) this accumulation is an ongoing process that occursover long periods of time; and (4) that AAV-FGF2 promotesthe expression of genes characteristic of telencephalic pre-cursors and immature neurons in the cortical regions sur-rounding the infarct. The data suggest that therapy based onendogenous repair would target distinct cells with features ofneural precursors or immature neurons at specific times afterischemia.

Reports of an endogenous neurogenic response restrictedto the striatum after a large focal ischemia in the forebrain3,6

and of a slow endogenous repair process in the hippocampusafter global ischemia7 have been published before. Cellularregeneration in the striatum and hippocampus might be aconsequence of their known proximity to proliferative zonesthat are known to generate neurons in the adult. However, thepresence of cortical regeneration is still debatable.2 Althoughcortical neurogenesis is controversial, very localized lesionsof neurons in this area have been shown to generate cellreplacement with newly generated neurons surviving for longperiods and extending axons to appropriate distant targetsites.4 Occluding the MCA at the pial surface results inischemic injury restricted to the cerebral cortex, rapid activa-tion of proliferation in the SVZ and migration of newborncells into the parenchyma.5,15 Newborn cells were recentlyshown to continuously form after ischemic injury involving

the striatum.6 In that set of experiments BrdU was adminis-tered either within the first 2 weeks after ischemic onset or atweeks 7 to 8 after ischemia. BrdU/DCX double-positive cellswere found to be formed throughout the experiment, and theSVZ was found to be of increased size over time and tocontinuously produce DCX� cells. The data presented hereextends these reports by suggesting that persistent stimulationwith FGF2 promotes long-lasting precursors in the ventricu-lar region that migrate to the peri-infarcted area where theycontinue to divide over time, resulting in a continuousaccumulation of neural precursors in the cortex. Our study islimited because we did not perform a direct lineage analysisto ensure that all progenitors emerge from the SVZ. Futureexperiments with labeling of NPC at the SVZ with retrovi-ruses expressing GFP and tracking these cells at intervalsuntil at the cortex should give the complete answer to whetherall NPCs emerge from the SVZ or whether at least someemerge from other locations including the cortex, where theymay reside in a quiescent form. In agreement with 2 recentarticles that could not detect newborn neurons at the imme-diate infarct border,5,20 our results show that newborn pro-genitors are constantly present in the area immediatelybordering the lesion and that a second domain of newlygenerated cells with the properties of immature neurons existsonly a few cell-layers deeper to the lesion border. The originof these 2 distinct zones is unknown. It is possible that factorssuch as inflammation, metabolic stress or local inhibitorycomponents within the white matter surrounding the infarctprevent terminal differentiation of newborn cells into neu-rons. Importantly, newborn cells expressing neuronal anti-gens could be seen far more often in the group treated withFGF2 compared with AAV-controls.

It is also important to note that newborn progenitors onlybegan to accumulate in significant numbers at timepointsafter the major microglial inflammatory response declined.Thus, early inflammatory changes may hamper initial effortsfor neurogenesis after ischemia and probably negativelyinfluence functional outcome.

The lack of spontaneous behavioral recovery in the AAVgroup is somewhat surprising because some degree of recov-ery is expected in the stroke model used and was indeed seenwhen we compared the results to historical controls at ourlaboratory. However, this may be attributable to the proin-flammatory effects of AAV (however minor) which mayhave compounded on the ischemic damage. At least part ofthis proinflammatory damage which was similar in the AAV-FGF group was probably reversed by excess neurogenesisobserved in the latter group.

The exact mechanisms responsible for behavioral improve-ments observed in AAV-FGF animals remain unclear be-cause the generation of newborn neurons appears to be alengthy process and newborn neurons were only observed insignificant numbers relatively late after lesioning. FGF2 mayexert neuroprotective effects in ischemia,9 but the lack ofsignificant differences in infarct volumes, as well as the lateimprovement in functional deficits, argues against a protec-tive effect of FGF2 in our experimental settings. Our resultsindicate that the mere increase in the numbers of newborncells is associated with better functional outcomes after

160 Stroke January 2007

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

cortical ischemia. The results also show that young, undiffer-entiated SOX2�/nestin� cells secrete trophic factors such asVEGF which has been shown to salvage motor neurons.21

Furthermore, VEGF-induced angiogenesis in the peri-infarctarea,18 as well as the presence of young cells that are capableof metabolizing and neutralizing excitatory amino acids andfree radicals, may also contribute to improved metabolismand function of surviving neurons.

Importantly, FGF2 treatment resulted in increased numbersof progenitors in the cortex only in the presence of ischemia.This implies that factors related to the presence of ischemicinjury (eg, stromal growth factor 1�) are necessary to furtherincrease neurogenesis and induce migration of newborn cellsto the cortex.

Taken together, our results suggest that in ischemic lesionsrestricted to the cortex, long-term stimulation with FGF2 canstimulate regenerative processes that induce partial behav-ioral recovery. These preliminary data encourage studies todetermine the regenerative consequences of these long-termcellular responses to further define appropriate ligands thatcan promote recovery.

DisclosuresNone.

References1. Savitz SI, Rosenbaum DM, Dinsmore JH, Wechsler LR, Caplan LR. Cell

transplantation for stroke. Ann Neurol. 2002;52:266–275.2. Arvidsson A, Collin T, Kirik D, Kokaia Z, Lindvall O. Neuronal

replacement from endogenous precursors in the adult brain after stroke.Nat Med. 2002;8:963–970.

3. Parent JM, Vexler ZS, Gong C, Derugin N, Ferriero DM. Rat forebrainneurogenesis and striatal neuron replacement after focal stroke. AnnNeurol. 2002;52:802–813.

4. Magavi SS, Leavitt BR, Macklis JD. Induction of neurogenesis in theneocortex of adult mice. Nature. 2000;405:951–955.

5. Gotts JE, Chesselet MF. Migration and fate of newly born cells after focalcortical ischemia in adult rats. J Neurosci Res. 2005;80:160–171.

6. Thored P, Arvidsson A, Cacci E, Ahlenius H, Kallur T, Darsalia V,Ekdahl CT, Kokaia Z, Lindvall O. Persistent production of neurons fromadult brain stem cells during recovery after stroke. Stem Cells. 2006;24:739–747.

7. Nakatomi H, Kuriu T, Okabe S, Yamamoto S, Hatano O, Kawahara N,Tamura A, Kirino T, Nakafuku M. Regeneration of hippocampal pyram-

idal neurons after ischemic brain injury by recruitment of endogenousneural progenitors. Cell. 2002;110:429–441.

8. Johe KK, Hazel TG, Muller T, Dugich-Djordjevic MM, McKay RD.Single factors direct the differentiation of stem cells from the fetal andadult central nervous system. Genes Dev. 1996;10:3129–3140.

9. Kiprianova I, Schindowski K, von Bohlen und Halbach O, Krause S,Dono R, Schwaninger M, Unsicker K. Enlarged infarct volume and lossof BDNF mRNA induction following brain ischemia in mice lackingFGF-2. Exp Neurol. 2004;189:252–260.

10. Kaspar BK, Vissel B, Bengoechea T, Crone S, Randolph-Moore L,Muller R, Brandon EP, Schaffer D, Verma IM, Lee KF, Heinemann SF,Gage FH. Adeno-associated virus effectively mediates conditional genemodification in the brain. Proc Natl Acad Sci U S A. 2002;99:2320–2325.

11. Haberman R, Lux GK, McCown T, Samulski RJ. Production of recom-binant adeno-associated viral vectors and use in in vitro and in vivoadministration. In: Crawley J, Gerfen C, Rogawski M, Sibley D, SkolnickP, Wray S, eds. Current Protocols in Neuroscience. New York, NY: JohnWiley & Sons, Inc;1999:4.17.11–14.17.25.

12. Leker RR, Teichner A, Grigoriadis N, Ovadia H, Brenneman DE, FridkinM, Giladi E, Romano J, Gozes I. Nap, a femtomolar-acting peptide,protects the brain against ischemic injury by reducing apoptotic death.Stroke. 2002;33:1085–1092.

13. Vega CJ, Peterson DA. Stem cell proliferative history in tissue revealedby temporal halogenated thymidine analog discrimination. Nat Methods.2005;2:167–169.

14. Leker RR, Gai N, Mechoulam R, Ovadia H. Drug-induced hypothermiareduces ischemic damage: effects of the cannabinoid hu-210. Stroke.2003;34:2000–2006.

15. Gotts JE, Chesselet MF. Vascular changes in the subventricular zone afterdistal cortical lesions. Exp Neurol. 2005;194:139–150.

16. Couillard-Despres S, Winner B, Schaubeck S, Aigner R, Vroemen M,Weidner N, Bogdahn U, Winkler J, Kuhn HG, Aigner L. Doublecortinexpression levels in adult brain reflect neurogenesis. Eur J Neurosci.2005;21:1–14.

17. Komitova M, Eriksson PS. Sox-2 is expressed by neural progenitors andastroglia in the adult rat brain. Neurosci Lett. 2004;369:24–27.

18. Ogunshola OO, Stewart WB, Mihalcik V, Solli T, Madri JA, Ment LR.Neuronal VEGF expression correlates with angiogenesis in postnataldeveloping rat brain. Brain Res Dev Brain Res. 2000;119:139–153.

19. Murray RC, Navi D, Fesenko J, Lander AD, Calof AL. Widespreaddefects in the primary olfactory pathway caused by loss of mash1function. J Neurosci. 2003;23:1769–1780.

20. Komitova M, Mattsson B, Johansson BB, Eriksson PS. Enriched envi-ronment increases neural stem/progenitor cell proliferation and neuro-genesis in the subventricular zone of stroke-lesioned adult rats. Stroke.2005;36:1278–1282.

21. Storkebaum E, Lambrechts D, Dewerchin M, Moreno-Murciano MP,Appelmans S, Oh H, Van Damme P, Rutten B, Man WY, De Mol M,Wyns S, Manka D, Vermeulen K, Van Den Bosch L, Mertens N, SchmitzC, Robberecht W, Conway EM, Collen D, Moons L, Carmeliet P.Treatment of motoneuron degeneration by intracerebroventriculardelivery of VEGF in a rat model of ALS. Nat Neurosci. 2005;8:85–92.

Leker et al Cortical Regeneration After Focal Ischemia 161

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from

Androutsellis-Theotokis and Ronald D.G. McKayRonen R. Leker, Frank Soldner, Ivan Velasco, Denise K. Gavin, Andreas

Long-Lasting Regeneration After Ischemia in the Cerebral Cortex

Print ISSN: 0039-2499. Online ISSN: 1524-4628 Copyright © 2006 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Stroke doi: 10.1161/01.STR.0000252156.65953.a9

2007;38:153-161; originally published online November 22, 2006;Stroke. 

http://stroke.ahajournals.org/content/38/1/153World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://stroke.ahajournals.org//subscriptions/

is online at: Stroke Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer process is available in the

Request Permissions in the middle column of the Web page under Services. Further information about thisOnce the online version of the published article for which permission is being requested is located, click

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Strokein Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on February 17, 2015http://stroke.ahajournals.org/Downloaded from