You are on page 1of 16

JOURNAL OF NEUROCHEMISTRY

| 2010 | 112 | 134149

doi: 10.1111/j.1471-4159.2009.06433.x

, ,

*Department of Neurology, University of New Mexico Health Sciences Center, Albuquerque, New Mexico, USA Neurovascular Research Laboratory, Neurovascular Unit, Departments of Neurology and Internal Medicine, Universitat Autonoma ` de Barcelona, Institut de Recerca, Hospital Vall dHebron, Barcelona, Spain Department of Neurosciences, University of New Mexico Health Sciences Center, Albuquerque, New Mexico, USA Departments of Cell Biology and Physiology, University of New Mexico Health Sciences Center, Albuquerque, New Mexico, USA

Abstract Increased matrix metalloproteinase (MMP) activity is implicated in proteolysis of extracellular matrix in ischemic stroke. We recently observed intranuclear MMP activity in ischemic brain neurons at early reperfusion, suggesting a possible role in nuclear matrix proteolysis. Nuclear proteins, poly-ADPribose polymerase-1 (PARP-1) and X-ray cross-complementary factor 1 (XRCC1), as well as DNA repair enzymes, are important in DNA fragmentation and cell apoptosis. We hypothesized that intranuclear MMP activity facilitates oxidative injury in neurons during early ischemic insult by cleaving PARP-1 and XRCC1, interfering with DNA repair. We induced a 90-min middle cerebral artery occlusion in rats. Increase activity of MMP-2 and -9, detected in the ischemic neuronal nuclei at 3 h, was associated with DNA fragmentation at 24 and 48 h reperfusion. The intranuclear MMPs cleaved PARP-1. Treatment of the rats with a broad-spectrum MMP

inhibitor, BB1101, signicantly attenuated ischemia-induced PARP-1 cleavage, increasing its activity. Degradation of XRCC1 caused by ischemic insult in rat brain was also signicantly attenuated by BB1101. We found elevation of oxidized DNA, apurinic/apyrimidinic sites, and 8-hydroxy-2deoxyguanosine, in ischemic brain cells at 3 h reperfusion. BB1101 markedly attenuated the early increase of oxidized DNA. Using tissue from stroke patients, we found increased intranuclear MMP expression. Our data suggest that intranuclear MMP activity cleaves PARP-1 and XRCC1, interfering with oxidative DNA repair. This novel role for MMPs could contribute to neuronal apoptosis in ischemic injuries. Keywords: nuclear gelatinolysis, oxidative DNA damage, poly-ADP-ribose polymerase-1, stroke, X-ray cross-complementary factor 1. J. Neurochem. (2010) 112, 134149.

Matrix metalloproteinases (MMPs) play fundamental roles in regulation of the extracellular matrix and have been linked to bloodbrain barrier (BBB) opening and neurodegeneration associated with ischemia and neuroinammation (Rosenberg et al. 1996; Heo et al. 1999; Planas et al. 2001; Lee et al. 2004; Rosenberg and Yang 2007; Candelario-Jalil et al. 2009). During an earlier study of the role of MMP-2 in the early reversible opening of the BBB, we unexpectedly observed gelatinase activity in the nuclei of ischemic neurons at 3 h after reperfusion (Yang et al. 2007). As nuclear matrix proteolysis is implicated in numerous activities such as apoptosis, cell cycle, and DNA fragmentation, the early increased intranuclear gelatinase activity in ischemic neurons

Received June 22, 2009; revised manuscript received September 18, 2009; accepted October 11, 2009. Address correspondence and reprint requests to Yi Yang MD, PhD, Department of Neurology, MSC11 6035, 1 University of New Mexico, Albuquerque, NM 87131-0001, USA. E-mail: yyang@salud.unm.edu Abbreviations used: 8-OHdG, 8-hydroxy-2-deoxyguanosine; aMMP, active MMP; AP site, apurinic/apyrimidinic (abasic) sites; BBB, bloodbrain barrier; BER, base excision repair; ISZ, in situ zymography; MCAO, middle cerebral artery occlusion; MMP, matrix metalloproteinase; MT1-MMP, membrane type 1-metalloproteinase; PARP-1, poly-ADP-ribose polymerase-1; PFC, piriform cortex; pMMP, proMMP; TUNEL, terminal deoxynucleotidyl transferasemediated dUTP nick end labeling; XRCC1, X-ray cross-complementary factor 1.

134

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

Intranuclear MMPs and oxidative DNA repair in focal ischemia | 135

suggested a possible role in nuclear matrix proteolysis, which may be involved in neuronal death or survival in focal ischemia with reperfusion. Poly-ADP-ribose polymerase 1 (PARP-1) is a nuclear chromatin-associated multifunctional enzyme (Hassa and Hottiger 2008) that acts as a sensor protein to detect oxidative DNA stress. DNA strand breaks induce PARP-1 activity, which plays an important role in repair of oxidized DNA and cell survival (Tanaka et al. 2005). DNA base excision repair (BER) machinery is the main mechanism in mammalian neuronal nuclei to repair various types of oxidative DNA damage. PARP-1 is required for efcient BER function by recruiting X-ray cross-complementary factor 1 (XRCC1) to oxidized DNA bases (Martin 2008). XRCC1 plays a central role in the DNA BER pathway by interacting with major DNA repair enzymes in the BER pathway (Vidal et al. 2001). Accumulation of oxidative DNA damage is associated with ischemia and, if not repaired promptly, triggers cell death. Appearance of oxidative DNA, apurinic/apyrimidinic (AP) sites and 8-hydroxy-2-deoxyguanosine (8-OHdG), and BER reduction occurs as early as 30 min after the onset of reperfusion in 2 h middle cerebral artery occlusion (MCAO) model (Lan et al. 2003; Luo et al. 2007). A recent study demonstrated that gelatinases are activated 15 min after reperfusion start in a MCAO model (Amantea et al. 2008). Expression of low level of PARP-1 partially prevented neuronal apoptosis and PARP-1 inhibition enhanced neuronal vulnerability to apoptosis (Nagayama et al. 2000; DiazHernandez et al. 2007). MMP-2 is also detected in the nucleus of human cardiac myocytes and puried MMP-2 is capable of cleaving PARP-1 in vitro (Kwan et al. 2004). Activation of MMP-2 is an early and key event in oxidative stress injury to the heart after ischemic reperfusion injury (Ali and Schulz 2009). Considering the early induction of oxidative DNA lesion and reduction of BER function, we hypothesized that the early increased activity of MMPs in the nucleus after stroke degraded PARP-1 and XRCC1, contributing to a reduction of DNA BER function and an accumulation of oxidized DNA bases in neurons, triggering their death. To test the hypothesis, we evaluated the role of intranuclear MMP activity on the cleavage of PARP-1 and XRCC1. We show for the rst time in brain that shortly after the injury MMPs in a nuclear fraction of ischemic brain tissue cleave PARP-1 and XRCC1, which facilitates accumulation of oxidized DNA at an early stage after an ischemic insult. Moreover, we show that treatment with a broad-spectrum MMP inhibitor, BB1101, signicantly attenuates ischemiainduced cleavage of PARP-1 and XRCC1 degradation and protects against oxidative DNA damage.

Middle cerebral artery occlusion with reperfusion The study was approved by the University of New Mexico Animal Care Committee and conformed to the National Institutes of Health Guidelines for use of animals in research. Male spontaneous hypertensive rats (300320 g of body weight) were subjected to 90 min of MCAO (Rosenberg et al. 2001) and reperfusion for 3, 24, and 48 h.

In situ zymography for rat and human brains Gelatinolytic activity in frozen, non-xed sections (14 lm) from ischemic rat brains was examined, using EnzCheck Collagenase Kit (Molecular Probes, Eugene, OR, USA) as previously described (Yang et al. 2007). After in situ gelatinolysis, some sections were processed for immunohistochemistry to detect the expression of proteins co-localized with active gelatinases in cells. Frozen non-xed human brain sections (12 lm) from ve deceased stroke patients were examined using EnzCheck Collagenase Kit to detect gelatinolysis. This study was approved by the Ethics Committee of the Hospital Vall dHebron and informed consent was acquired from all relatives prior to the autopsy.
Preparation of nuclear extracts Nuclear extracts from rat brains having undergone 90 min MCAO with 3 h reperfusion were prepared using the NE-PER Nuclear and Cytoplasmic Extraction Reagents and Nuclear Enrichment Kit for Tissue (Pierce, Rockford, IL, USA) following manufacturers protocol. Gelatin zymography The expression of MMP-2 and -9 in nuclear extracts from ischemic rat brains was measured using gelatin zymography as previously described (Yang et al. 2007). Immunohistochemistry Ten micrometer sections from rat brain tissues xed with 2% p-formaldehyde were used for immunohistochemical analysis. Immunoblotting Nuclear proteins extracted from tissues of ischemic and non-ischemic cortex of rats were subjected to sodium dodecyl sulfatepolyacrylamide gel electrophoresis analysis. Coomassie blue staining on the same polyvinylidene diuoride membranes was used for loading and transfer control. The results are reported as normalized band intensity for quantifying relative protein expression.

In vitro degradation assay of PARP-1 and XRCC1 In vitro degradation of PARP-1 and XRCC1 by recombinant rat MMP-2, murine MMP-9 (R&D Systems, Minneapolis, MN, USA), gelatinase extracts or total nuclear extracts was examined in an assay buffer.
Poly-ADP-ribose polymerase-1 activity assay Poly-ADP-ribose polymerase-1 activity assay in nuclear extracts of ischemic brain tissues was measured with HT Universal Colorimetric PARP Assay Kit (Trevigen, Gaithersburg, MD, USA) following manufacturers protocol. Assessment of ischemia-induced DNA damage 8-Hydroxy-2-deoxyguanosine was evaluated using an oxidative DNA ELISA kit (Cell Biolabs, San Diego, CA, USA) following the

Materials and methods


A detailed description of all methods utilized in this study is found in the Appendix S1.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

136 | Y. Yang et al.

kit instructions. AP sites in genomic DNA were detected using the OxiSelect Oxidative DNA damage quantication kit (AP sites) (Cell Biolabs). Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling assay Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assay was carried out to identify the extent of DNA fragmentation by using the NeuroTACS II kit (Trevigen). For dual immunouorescence, the slides were then exposed to the primary antibody and streptavidin uorescein and prepared as described above. Statistical analysis Statistical comparisons among groups were done using ANOVA with post hoc analysis for multiple t-test (PRISM 4.0; GraphPad Software Incorporated, San Diego, CA, USA). All data are presented as mean SEM. Statistical signicance was set at p < 0.05.
(a)

Results
Gelatinase activity in neuronal nuclei in both ischemic rat and human brains The subcellular localization of active gelatinases was determined using in situ zymography (ISZ) with a reaction buffer containing FITC-labeled DQ-gelatin (Molecular Probes). Gelatin-FITC is cleaved by gelatinases, yielding peptides whose green uorescence is representative of net proteolytic activity (Fig. 1). With ISZ, we showed previously that there was a marked increase in gelatinase activity in the ischemic hemisphere after 3 h of reperfusion, which was seen in ischemic vessels, cells, and cell nuclei (Yang et al. 2007). Intense gelatinolytic activity was seen in brain cells within 4-6-diamidino-2-phenylindole (DAPI)-positive nuclei of the ischemic cortex at 3 h after reperfusion (Fig. 1a). Confocal

(b)

(d)

(c)

Fig. 1 Gelatinase activity in the nuclei of ischemic brain cells in rat and human infarct preparations. (a) In situ zymography (ISZ) was performed in MCAO and sham-operated rat brains with 3 h post-ischemic reperfusion. Gelatinase activity is visualized with green uorescence. DAPI (blue) shows nuclear localization. Scale bars, 50 lm. (b) Z-stack confocal images show gelatinase activity is co-localized with DAPI (blue) in nucleus. Scale bars, 5 lm. (c) Z-stack confocal images show that gelatinase activity in cell nuclei is co-localized with NeuN (red), a neuronal marker. Scale bars, 5 lm. (d) Gelatinase activity in the nuclei of brain cells in human infarct preparations. 1,10-phenanthroline (1-10 Phe) eliminated proteolytic activity displayed in infarct preparations. Scale bars, 50 lm.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

Intranuclear MMPs and oxidative DNA repair in focal ischemia | 137

microscopy conrmed the strong gelatinolytic activity in the nuclei (Fig. 1b). Minimal uorescence was detected in both sham-operated rats and contralateral (non-ischemic) hemisphere (Fig. 1a). The nuclei with gelatinase activity co-localized with the neuronal marker, NEUronal Nuclei (NeuN), suggesting that the increased proteolytic activity from MMPs was in ischemic neurons (Fig. 1c). We also detected the proteolytic activity of gelatinases in human brains with an ischemic stroke within the previous range from 12 to 96 h using ISZ. Proteolytic activity was found in the cell nucleus within the infarction in brain parenchyma samples from deceased stroke patients (Fig. 1d). In contrast, nuclear proteolytic activity was rarely detected in
(a)

the contralateral hemisphere. To ensure the signal was because of MMP activity, samples from the infarct core were treated with the MMP inhibitor 1,10-phenanthroline that is known to reversibly chelate zinc ions, essential for MMP activity. Phenanthroline-treated preparations had very little signal (Fig. 1d). As ISZ cannot distinguish between MMP-2 and -9, several experiments were performed to identify the source of gelatinase activity. With co-localization of immunolabeling for MMP-2 and -9 with ISZ, we found higher levels of MMP-2 and -9 immunoreaction in the brain cells with increased gelatinase activity in ischemic brain (Fig. 2a). Confocal analysis showed that immunoreaction of both

(b)

(c)

Fig. 2 MMP-2 and -9 contribute to intranuclear gelatinase activity in rat brain at 3 h post-ischemic reperfusion. (a) Double staining of in situ zymography (ISZ) (green) and MMP-2 and -9 immunohistochemistry (red) was performed in MCAO and sham-operated rat brains with 3 h post-ischemic reperfusion. DAPI (blue) shows nuclear localization. Scale bars, 100 lm. Inset shows example of negative control for immunostaining with normal IgGs and secondary antibody conjugated with Cy-3 and no specic immunolabeling was exhibited.

(b) Representative confocal images show immunostaining for MMP-2 and -9 (red) co-localized with ISZ in nuclei in ischemic cells. Scale bars, 10 lm. (c) Pre-incubation with specic antibodies for MMP-2 and -9 (MMP-2 AB and MMP-9 AB), 1,10-phenanthroline (Phen), and MMP-2/9 inhibitor II (Inhibitor II) block the gelatinase activity by ISZ in nuclei of ischemic cells; *p < 0.05 and **p < 0.01 versus ISZ vehicle, n = 4 in each group. Fluorescent intensity of gelatinase activity was measured with IMAGEJ program. Scale bar, 50 lm.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

138 | Y. Yang et al.

MMP-2 and -9 was co-localized with the proteolytic activity in nuclei (Fig. 2b). To validate the nding, brain sections were pre-incubated with MMP-2 and -9 antibodies, broadspectrum MMP inhibitor 1,10-phenanthroline and selective MMP-2/9 inhibitor II, prior to an ISZ analysis. Pre-incubation with MMP-2 and -9 antibodies and MMP inhibitors reduced the uorescence intensity in ischemic brain tissue, suggesting that both MMP-2 and -9 were involved in the gelatinase activity (Fig. 2c). For quantitation of the gelatinase activity, uorescence intensity in elds of ischemic cortex, piriform cortex, and striatum in each section derived from bregma area were measured using IMAGEJ software (http://www.rsbweb.nih.gov/ij). The data strongly suggest that MMP-2 and -9 contribute to the intranuclear gelatinase activity in ischemic neurons. We next performed gelatin zymography, a common method for examining and quantifying MMP-2/-9, with ischemic rat brains. Supernatants of human brosarcoma HT-1080 cells were used as gelatinase standards for MMP-2 and -9 (Fig. S1). Zymography showed increased expression of MMP-2 and -9 in nuclear extracts from ischemic rat brains compared with sham-operated controls and non-ischemic brains (p < 0.001 and 0.01, respectively) (Fig. 3a and b). A statistically significant increase in active bands of MMP-2 (aMMP; p < 0.05) was seen in the ischemic hemisphere. We used an inhibitor of MMP activity, BB1101, to determine the role of intranuclear

MMP activity on the nuclear matrix proteolysis. Treatment with BB1101 signicantly reduced proMMP-2 and MMP-9 (pMMP-2 and -9) in the ischemic brains compared with ischemic vehicle-treated control (p < 0.05) (Fig. 3a and b). Zymography revealed aMMP-2 bands in the nucleus. To conrm the nuclear presence of proteolytic activity of MMP2, we analyzed the expression of membrane type 1-MMP (MT1-MMP), a major activator of MMP-2, in ischemic rat brains. With immunohistochemistry of MT1-MMP and ISZ, co-localization of MT1-MMP and gelatinolytic activity was seen in the ischemic cells (Fig. 4a). Z-stack and 3D confocal images showed that the expression of MT1-MMP was detected in both cellular membrane and nucleus (Fig. 4a and b), with greater MT1-MMP staining seen in the nuclei. The nuclear presence of MT1-MMP was conrmed by western blot analysis in nuclear extracts (Fig. 4c). We also detected the expression and location of furin, an activator of MT1MMP, using immunohistochemistry and confocal microscopy (Fig. 4d). The co-localization of MT1-MMP and furin was seen in ischemic nuclei. PARP-1 expression and degradation by intranuclear MMPs in vivo and in vitro To investigate whether the increase in nuclear MMP-2 and -9 is involved in delayed cell death, we compared MMP-2 and -9 immunohistochemistry with the spatial patterns of DNA

(a)

(b)

Fig. 3 ProMMP-9, MMP-2, and active MMP-2 (pMMP-9, pMMP-2, and aMMP-2) in nuclear preparations of rat brains at 3 h post-ischemic reperfusion with and without BB1101 treatment. (a) Gelatin zymography analysis was performed in nuclear preparations of rat brains. Human HT-1080 was used as a standard for MMP-2 and -9; pMMP-9, MMP-2 and aMMP-2 were detected in MCAO brains. (b) Relative intensity of pMMP-9 (95 kDa), pMMP-2 (68 kDa) and aMMP-2 (62 kDa). 3 h R, ischemic hemispheres; 3 h L, non-ischemic

hemispheres; BB1101 R, ischemic hemisphere with BB1101 treatment; BB1101 L, non-ischemic hemisphere with BB1101 treatment; Sham, sham-operated brains. Sham group n = 4, 3 h group n = 8, BB1101 group n = 5; pMMP-9 level: ***p < 0.001 and *p < 0.05 versus 3 h L, BB1101 L, and sham, respectively; #p < 0.05 versus 3 h R; pMMP-2 level: **p < 0.01 versus 3 h L, BB1101 L, and sham; *p < 0.05 versus sham; #p < 0.05 versus 3 h R; aMMP-2 level: *p < 0.05 versus 3 h L, BB1101 L, and sham.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

Intranuclear MMPs and oxidative DNA repair in focal ischemia | 139

(a)

(b)

(b) (i)

(d) (i)

5 m

5 m

5 m

5 m

5 m

(ii) (ii)

5 m

5 m

5 m

(iii)

Fig. 4 Expression and location of MT1-MMP and furin in nuclei at 3 h post-ischemic reperfusion. DAPI shows nucleus localization. (a) Colocalization of MT1-MMP immunohistochemistry with in situ zymography (ISZ) in ischemic (top panel) and non-ischemic (bottom panel) brain sections. Scale bars, 5 lm. (b) Z-stack confocal analysis for the co-localization of MT1-MMP and gelatinase activity in nucleus (i). (ii) Image from top panel corresponding to z-axis consecutive images (16 sections). In all sections MT1-MMP co-localizes with ISZ in the nucleus. A three-dimensional image shows the reconstruction from the

stack of 16 optical sections. Scale bars, 5 lm. (c) Western blot analysis for MT1-MMP in nuclear extract preparations; *p < 0.05 versus non-ischemic hemispheres, n = 4 in each group. (d) Immunostaining for furin is co-localized with MT1-MMP (i). Scale bar, 20 lm. Z-stack confocal images show furin is co-localized with MT1-MMP in nucleus (ii). (iii) Image from top panel corresponding to z-axis consecutive images (10 sections). A three-dimensional image with a 90 rotation shows the reconstruction from the stack of 10 optical sections. Scale bars, 5 lm.

fragmentation (TUNEL expression) in brain cells. At 24 h, we detected scattered TUNEL-positive cells in the ischemic hemispheres that co-localized with MMP-2 (Fig. 5a). By 48 h, we observed extensive TUNEL-positive cells in the entire lesion area of the ischemic hemisphere but not in the non-ischemic side. Most TUNEL-positive cells expressed MMP-2 (Fig. 5a) and co-localized with NEUronal Nuclei

(NeuN) (Fig. 5c). We found MMP-9 expression is co-localized with TUNEL expression in some brain cells in the lesion area of the ischemic hemisphere at 24 and 48 h. However, compared with MMP-2, the association between the MMP-9 expression and DNA fragmentation is complicated at 48 h. We detected some TUNEL-positive cells without presence of MMP-9 expression. In contrast, some

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

140 | Y. Yang et al.

(a)

(b)

(c)

(d)

(e)

Fig. 5 Cell apoptosis and expression of MMP-2 and PARP-1. (a and b) Representative double staining of MMP-2 or MMP-9 with TUNEL in rat brain at 24 and 48 h post-ischemic reperfusion. (a) Arrows indicate few TUNEL-positive cells with the expression of MMP-2 at 24 h. Scale bar, 20 lm. At 48 h reperfusion, extensive colocalization of TUNELpositive cells with MMP-2 expression was seen in ischemic brain section. Arrows and inset: TUNEL-positive cells show very weak DAPI staining that labels nuclei. Scale bars, 50 lm. (b) Arrows indicate TUNEL-positive cell with the expression of MMP-9 at 24 h. Scale bar, 20 lm. Expression of MMP-9 and TUNEL in ischemic and nonischemic brain cells at 48 h reperfusion (middle and bottom panels). Scale bar, 50 lm. Arrows: TUNEL-positive cells with MMP-9 expression. Arrowheads: TUNEL-negative cells with MMP-9 expression.

Stars: TUNEL-positive cells without MMP-9 expression. V: vessel. (c) Representative confocal images of double staining for TUNEL with NeuN immunoreaction reveal that most of TUNEL-positive cells are neurons in ischemic hemisphere at 48 h reperfusion. Scale bars, 50 lm. (d) Representative confocal analysis of expression of PARP-1 and ISZ in rat brain at 3 h post-ischemic reperfusion. The z-stack image shows the co-localization of PARP-1 and gelatinase activity in nucleus. A three-dimensional image with a 90 rotation shows the same cells with ISZ and PARP-1 in nuclei. Scale bars, 20 lm. (e) Representative co-localization of MMP-2 and PARP-1 in cell nuclei of rat brain at 3 h post-ischemic reperfusion. DAPI shows nucleus localization. Scale bars, 20 lm.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

Intranuclear MMPs and oxidative DNA repair in focal ischemia | 141

cells that expressed MMP-9 were TUNEL negative. In addition, we also detected the expression of MMP-9 in vessels in the lesion area of the ischemic hemisphere at 48 h (Fig. 5b). We next examined the relationship of PARP-1 expression in cells to gelatinase activity using an antibody that recognizes both full-length (116 kDa) and cleaved PARP-1. Compared with the non-ischemic tissue, higher PARP-1 immunostaining was seen in the ISZ-positive cells in the ischemic tissue (Fig. 5d). This co-localization was conrmed by double staining of MMP-2 and PARP-1 (Fig. 5e). To determine whether the increase in PARP-1 staining reected either increased accumulation of PARP-1 or differential cleavage in the lesions, or possibly both, western blot analyses and densitometry were carried out. Nuclear extracts from ischemic rat brains with or without BB1101 treatment were used to determine the involvement of MMPs in changes of PARP-1 expression and degradation. After 3 h of reperfusion, the full-length form of PARP-1 (116 kDa) was detected in nuclear extracts from both ischemic and nonischemic hemispheres (Fig. 6a). However, in ischemic nuclear extracts, cleavage products of PARP-1 yielded an 89 kDa band and a 43 kDa band that were signicantly increased compared with the non-ischemic side (Fig. 6a). More importantly, even though not all rats showed the increased 89 kDa band, all animals had an additional 43 kDa band. To determine whether the cleavage of PARP-1 was a consequence of the early nuclear gelatinase proteolysis, we examined PARP-1 cleavage in ischemic nuclear extracts prepared from rats treated with BB1101. Treatment with BB1101 signicantly reduced the degradation of PARP-1 as shown by a signicant reduction in the bands migrating at 89 and 43 kDa (Fig. 6a). It is well known that PARP-1 cleavage by caspase 3 yields 89- and 24-kDa fragments during apoptosis. However, very low immunostaining of cleaved caspase 3 was detected in the cells that showed active gelatinase at 3 h reperfusion after ischemia. Increased cleaved caspase 3 was detected at 72 h reperfusion when excessive apoptosis appears in the ischemic area (Fig. 6b). In in vitro studies, incubation of puried bovine PARP-1 with recombinant rat MMP-2 and -9 or with gelatinase extracts prepared from nuclear fractions of rat brain using gelatin beads also showed the loss of full-length PARP-1 (116 kDa) by western blot (Fig. 6c, d, and e). After 1 h of incubation with MMP-2 and -9, there was appearance of a cleaved product of PARP-1 around 43 kDa (Fig. 6c and d). Co-incubation with BB1101 inhibited the PARP-1 cleavage caused by both MMPs and gelatinase extracts. MMP inhibitors GM6001 and 1,10-phenanthroline were also tested and these compounds signicantly inhibited PARP-1 cleavage caused by both rat MMP-2 and gelatinase extracts (data not shown). Western blot did not detect any cleaved band of PARP-1 after incubation with gelatinase extracts (Fig. 6e). This could be because of the action of other MMPs or

proteases that may be present in the bead extracts. These proteases could degrade PARP-1 into fragments that are not recognized by the PARP-1 antibody. To further demonstrate the cleavage of PARP-1 by nuclear MMPs, the pure bovine PARP-1 was incubated with total nuclear extracts from ischemic brains in assay buffer (Fig. 6f). In addition to pure PARP-1, we added total nuclear extracts alone as another control since the total nuclear extracts contain PARP-1, XRCC1, and other proteases. Immunoblotting showed a signicant loss of full-length PARP-1 (116 kDa) and a signicant increase in the 43-kDa cleavage product of PARP-1 after incubation with the ischemic nuclear extracts. Co-incubation with selective MMP-2/9 inhibitor II attenuated PARP-1 cleavage, suggesting the involvement of nuclear MMP-2 and -9 in this process. We observed the cleaved product of PARP-1 at 89 kDa (Fig. 6f), but treatment with MMP-2/9 inhibitor II failed to inhibit its appearance. Poly-ADP-ribose polymerase-1 is activated by DNA strand breaks caused by oxidative stress. We measured PARP activity in nuclear extracts obtained from ischemic rat brains to determine whether the cleavage of PARP-1 by early active gelatinases may affect PARP-1 activity. There was a signicant reduction in PARP-1 activity in ischemic brains compared with those treated with BB1101 (Fig. 6g). Reduction of XRCC1 in ischemic brain by nuclear gelatinase activity The nuclear protein XRCC1 plays a central role in the DNA BER pathway. We therefore investigated the expression of XRCC1 in ischemic rat brain with 3 h reperfusion using immunohistochemistry and western blot. Immunohistochemistry showed the constitutive expression in the nucleus of XRCC1 in the entire region of the normal rat brain (data not shown) and it was mainly seen in the nucleus (Fig. 7a). At 3 h after reperfusion, reduction of XRCC1 was observed in the lesion areas, including the ischemic piriform cortex (PFC) when compared with the non-ischemic PFC. Treatment with BB1101 reversed the reduction of XRCC1 in the ischemic PFC. Double staining with XRCC1 and ISZ showed the spatial relationship between XRCC1 loss and nuclear gelatinase activity (Fig. 7a, inset). Using an antibody against XRCC1 that recognizes the full-length XRCC1, western blot detected two bands around 95.74 kDa in the nuclear extracts from rat brain and in the positive control XRCC1, which represent the full-length XRCC1, as well as several bands around 43 kDa (Fig. 7b). All bands of XRCC1 detected were signicantly decreased in the nuclear extracts of ischemic brains compared with the non-ischemic brain. The decrease of XRCC1 in the ischemic side was signicantly reversed by BB1101 treatment (Fig. 7b). These results from both immunohistochemistry and western blot strongly suggest that the early nuclear gelatinase

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

142 | Y. Yang et al.

proteolysis is involved in the early reduction of XRCC1 induced by transient focal cerebral ischemia. To provide additional evidence in support of this conclusion, we next performed a study of XRCC1 degradation in vitro by recombinant rat MMP-2, murine MMP-9 or with gelatinase extracts prepared from nuclear fractions of rat brain using gelatin beads. Western blots showed that incubation of recombinant XRCC1 with synthetic rat MMP-2 or with
(a) (c)

gelatinase extracts signicantly reduced the level of XRCC1 protein (Fig. 7c and e). Co-incubation with BB1101 inhibited the XRCC1 cleavage caused by both MMP-2 and gelatinase extracts. Matrix metalloproteinase-9 produced a signicant reduction in full-length XRCC1 (96 kDa) as well as in the bands migrating in the 43-55 kDa range (Fig. 7d). Compared with MMP-2 or gelatinase extracts, MMP-9-mediated cleavage of

(d)

(b)

(e)

(f)

(g)

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

Intranuclear MMPs and oxidative DNA repair in focal ischemia | 143

XRCC1 was unique in that the lowest size XRCC1 band was increased compared with the positive control and BB1101treated samples. MMP inhibitors GM6001 and 1,10-phenanthroline were also tested and these compounds signicantly inhibited XRCC1 cleavage caused by both human MMP-2/9 and gelatinase extracts (data not shown). Similarly to PARP-1, incubation of XRCC1 with total nuclear extracts produced a signicant degradation of all XRCC1 bands. This was signicantly inhibited by the selective MMP-2/-9 inhibitor II (Fig. 7f). This strongly suggests that gelatinases present in the nuclear extracts of the ischemic brain are involved in the degradation of XRCC1. Nuclear gelatinase activity enhances the induction of oxidative DNA damage Based on the observation that the early nuclear gelatinase activity was involved in the proteolysis of PARP-1 and XRCC1, which mediate repair of oxidized DNA, we next measured the formation of AP sites, one of the prevalent lesions of oxidative DNA damage, and 8-OHdG, a ubiquitous marker of oxidative stress, in ischemic rat brains at 3 h after reperfusion. An antibody against 8-OHdG was used to detect whether there was an induction of oxidative DNA damage at 3 h after reperfusion. Double staining with 8-OHdG and ISZ showed that a strong signal of 8-OHdG was seen in nuclei of the cells that expressed active gelatinases (Fig. 8a). Figure 8b and c illustrate the proles of AP sites and 8-OHdG induction, respectively, in the ischemic rat brains at 3 h after reperfusion. Both AP sites and 8-OHdG were signicantly increased in the ischemic brains compared with the sham-operated animal and non-ischemic side. In contrast, in BB1101-treated animals the number of AP sites was signicantly decreased compared with ischemic animals receiving the vehicle. We did not nd a signicant difference of 8-OHdG between the ischemic animals and BB1101Fig. 6 PARP-1 cleavage by MMPs in vivo and in vitro. (a) Western blot analyses for PARP-1 cleavage in nuclear preparations of rat brain at 3 h post-ischemic reperfusion. 3 h R, ischemic hemispheres; 3 h L, non-ischemic hemispheres; BB1101 R, ischemic hemisphere with BB1101 treatment; BB1101 L, non-ischemic hemisphere with BB1101 treatment; n = 5 in each group. Full-length PARP-1 was visualized as a 116 kDa band. Cleavage of PARP-1 yielded two bands at 89 and 43 kDa 89 kDa band: *p < 0.05, 3 h R versus 3 h L and BB1101 L and #p < 0.01 BB1101 R versus 3 h R 43 kDa band: ***p < 0.001 3 h R versus 3 h L and BB1101 L and ###p < 0.001 BB1101 R versus 3 h R. (b) Representative confocal analysis of expression of cleaved caspase 3 and ISZ in rat brain at 3 h post-ischemic reperfusion. Bottom panel shows the positive immunostanining for cleaved caspase 3 in ISZ-positive cells in ischemic brain section at 72 h reperfusion. DAPI shows nucleus localization. Scale bars, 5 lm. (c, d, e, and f) In vitro degradation of PARP-1 by recombinant rat MMP-2 (c), MMP-9 (d), gelatinase extracts prepared from ischemic nuclear extracts (e and f) ischemic nuclear extracts. (c) n = 4, **p < 0.01, PARP-1

treated animals using absolute levels of 8-OHdG in each group. However, the 8-OHdG ratio of ischemic/non-ischemic showed BB1101 treatment signicantly decreased the 8-OHdG levels when compared with ischemic control. 3,3diaminobenzidine (DAB) immunohistochemical staining showed fewer 8-OHdG-positive cells in areas of piriform cortex in BB1101-treated animals (Fig. 8d). These results suggested that the early nuclear gelatinase activity enhanced the nuclear accumulation of oxidative DNA damage in the ischemic rat brain with reperfusion injury.

Discussion
We demonstrated an increase in intranuclear gelatinase activation in neurons in rat brain at 3 h of post-ischemic reperfusion. This increase of nuclear gelatinolytic activity is also observed in human brain after stroke. We showed that both MMP-2 and -9 contribute to the early nuclear gelatinase activity, which produced cleavage of PARP-1 in ischemic rat brain leading to a decrease of PARP-1 activity. Moreover, we demonstrated that the nuclear gelatinase proteolysis is involved in the early reduction of XRCC1 induced by the ischemic insult. Finally, we report that the early oxidative DNA damage caused by transient focal cerebral ischemia was attenuated in animals treated with the MMP inhibitor, BB1101. Our results provide the rst experimental evidence that early intranuclear aMMPs promote neuronal accumulation of oxidative DNA after an ischemic injury by cleaving PARP-1 and BER enzymes, such as XRCC1, reducing their ability to mediate DNA repair. Although best known for their role in the proteolysis of extracellular protein targets, recent studies have revealed that MMPs are also localized to various intracellular sites including nucleus (Si-Tayeb et al. 2006; Schulz 2007; Ali and Schulz 2009). Nuclear gelatinolytic activity in transient focal ischemic models has been previously observed (Gasche
(116 kDa) positive control versus PARP-1 incubated with MMP-2; *p < 0.05, PARP-1 incubated with MMP-2 versus BB1101 inhibition. (d) n = 4, **p < 0.01, PARP-1 (116 kDa) positive control versus PARP-1 incubated with MMP-9. (e) n = 5, **p < 0.01 PARP-1 (116 kDa) positive control versus PARP-1 incubated with gelatinase extracts; *p < 0.05, PARP-1 incubated with gelatinase extracts versus BB1101 inhibition. (f) n = 6, **p < 0.01, PARP-1 (116 kDa) positive control versus PARP-1 incubated with nuclear extracts (NE); *p < 0.05, PARP-1 (116 kDa) incubated with NE versus inhibition of MMP-2/9 inhibitor II (Inhibitor II); *p < 0.05, PARP-1 (43 kDa) positive control versus PARP-1 incubated with NE; #p < 0.05, PARP-1 (43 kDa) incubated with NE versus inhibition of MMP-2/9 inhibitor II. (g) PARP activity assay in nuclear extract preparations at 3 h postischemic reperfusion. Treatment with BB1101 signicantly increased the PARP activity in ischemic nuclear extracts: *p < 0.05 versus 3 h R, ## p < 0.01 versus 3 h L, BB1101 L, and sham. Sham group n = 3, 3 h group n = 6, and BB1101 group n = 5.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

144 | Y. Yang et al.

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 7 XRCC1 degradation by MMPs in vivo and in vitro. (a) XRCC1 immunostaining in ISZ-positive cells of non-ischemic, ischemic brains and ischemic brains with BB1101 treatment at 3 h post-ischemic reperfusion. The bottom panel from the inset shows less XRCC1 in the cells that expressed high gelatinase activity (arrows). DAPI shows nucleus localization. Scale bar, 50 lm. (b) Western blot analysis for degradation of XRCC1 in nuclear extract preparation. XRCC1 immunoblotting detected a band 95.74 kDa in the positive control XRCC1 (Abnova) and two bands around 96 kDa in the nuclear extracts, which represent the full-length XRCC1, as well as several bands around 43 kDa. All bands of XRCC1 detected were quantied, n = 5 in each group, *p < 0.05 versus 3 h L and BB1101 L; **p < 0.01 versus 3 h R.

(c, d, e, and f) In vitro degradation of XRCC1 by recombinant rat MMP2 (c), MMP-9 (d), gelatinase extracts prepared from ischemic nuclear extracts (e), and (f) ischemic nuclear extracts. All bands of XRCC1 detected were quantied unless indicated. (c) n = 4 in each group, **p < 0.01 versus positive control; *p < 0.05 versus XRCC1 incubated with MMP-2. (d) n = 5 in each group, *p < 0.05 versus XRCC1 positive control. (e) n = 5 in each group, ***p < 0.001, versus XRCC1 positive control; ##p < 0.02 versus XRCC1 positive control; ###p < 0.001 versus XRCC1 incubated with gelatinase extracts. (f) n = 4 in each group, ***p < 0.001 versus XRCC1 positive control; ##p < 0.01 versus XRCC1 positive control; #p < 0.05 versus XRCC1 incubated with nuclear extracts (NE). Inhibitor II, MMP-2/9 inhibitor II.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

Intranuclear MMPs and oxidative DNA repair in focal ischemia | 145

(a)

(b)

(d)

(c)

Fig. 8 Analysis of oxidative DNA lesion in rat brain at 3 h postischemic reperfusion. (a) Representative confocal analysis of 8-OHdG (red) induction in ISZ-positive cells of ischemic brain section. Scale bar, 10 lm. (b) Induction of AP sites in nuclear DNA of ischemic brain cells. 3 h R, ischemic hemispheres; 3 h L, non-ischemic hemispheres; BB1101 R, ischemic hemisphere with BB1101 treatment; ***p < 0.001 and *p < 0.05 versus sham R and L, 3 h L, and BB1101 L, respectively; ##p < 0.01 versus 3 h R. Sham group n = 4, 3 h and BB1101

groups n = 5. (c) Induction of 8-OHdG in nuclear DNA of ischemic brain cells. The level of 8-OHdG is presented as ratio of ischemic/nonischemic 8-OHdG levels. **p < 0.01 versus sham; *p < 0.05 3 h postischemic reperfusion. Sham n = 4, 3 h group n = 7, BB1101 group n = 5. (d) Representative immunohistochemistry for 8-OHdG induction (brown) in the treated and non-treated piriform cortex in rat brain at 48 h post-ischemic reperfusion. Scale bar, 50 lm.

et al. 2001; Yang et al. 2007; Amantea et al. 2008). Up-regulation of aMMP-13 in the nuclei of both rat and human ischemic brains was recently reported and the intranuclear MMP-13 activity is detectable as early as

30 min after the occlusion onset (Cuadrado et al. 2009). Our present ISZ results conrmed the early activation of gelatinases in neuronal nuclei in rat, as well as in human brain cells after an ischemic insult.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

146 | Y. Yang et al.

In experimental focal cerebral ischemia, MMP-2 increases initially followed by a later increase in MMP-9 expression (Rosenberg et al. 1996; Heo et al. 1999; Chang et al. 2003; Yang et al. 2007). In the present study, gelatin zymography revealed increases in pMMP-9 and aMMP-2 in the nucleus; pMMP-2 is activated by a molecular cascade that involves a trimolecular complex made up of MMP-2, tissue inhibitor of metalloproteinase-2 (TIMP-2), and MT1-MMP (Ratnikov et al. 2002). We observed that MT1-MMP was induced not only along the cytosolic membrane but also in the nuclei of ischemic brain cells, which was co-localized with active gelatinases, as well as with furin, an activator of MT1-MMP (McMahon et al. 2005). Immediate up-regulation of the activation system for pMMP-2 has been demonstrated in a focal cerebral ischemic model (Chang et al. 2003). This atypical location of MT1-MMP that co-localized with MMP2 in the nucleus is also observed in hepatocellular carcinoma (Ip et al. 2007). The nding that the enzymes involved in the activation of MMP-2 are present in the nucleus along with MMP-2 supports the presence of nuclear MMP-2 activity. Matrix metalloproteinases cleave most components of the extracellular matrix including bronectin, laminin, proteoglycans, type IV collagen, and tight junction proteins (Sternlicht and Werb 2001; Rosenberg 2002; Yang et al. 2007). The intranuclear location of MMP activity after stroke suggests a novel role in nuclear matrix proteolysis. Proteolysis of nuclear matrix is implicated in numerous processes such as apoptosis, cell cycle, and DNA fragmentation. Oxidative stress generated during stroke is a critical event leading to BBB disruption and apoptosis. The BER pathway is a critical mechanism for repair of oxidative DNA lesions in the brain after ischemia (Fujimura et al. 1999; Dutra et al. 2006; Li et al. 2007). The active PARP-1 triggers the recruitment of XRCC1 (Schreiber et al. 2002). During BER and single-strand break repair, a functional XRCC1 is critical for the accurate repair of damaged bases, abasic (AP) sites (Campalans et al. 2005; Nazarkina et al. 2007). XRCC1 is involved in all the steps of the repair of oxidized bases by interacting with the DNA repair enzymes (Marsin et al. 2003; Wiederhold et al. 2004). Oxidative DNA damage as shown by AP sites and 8-OHdG, and BER reduction occurred as early as 1530 min after the onset of reperfusion in a 2 h MCAO model (Lan et al. 2003; Luo et al. 2007). Strong 8-OHdG was observed 36 h after reperfusion (Ohtaki et al. 2007). Considering the compatible timing of when the oxidative DNA lesion and BER reduction occurred and when the intranuclear aMMPs were rst seen after ischemia and reperfusion, we reasoned that early intranuclear MMPs may cleave nuclear enzymes like PARP-1 and XRCC1 and interfere with oxidized DNA repair, which could promote apoptosis after an ischemic injury. We found MMP-2 associated with TUNEL-positive neurons, suggesting a possible involvement of MMP-2 in neuronal apoptosis. In vivo and in vitro studies showed that

nuclear MMP activity cleaves PARP-1, reducing PARP-1 activity in ischemic cell nuclei of rat brain. We used the MMP inhibitors to conrm the role of MMPs on the nuclear matrix cleavage. Both the PARP-1 cleavage (in vivo and in vitro) and the reduction of PARP-1 activity could be reversed by MMP inhibitor treatment. The PARP-1 cleavage by nuclear MMP-2 produced a 43 kDa band, which was also seen in a previous in vitro study of PARP-1 degradation by puried human MMP-2 (Kwan et al. 2004). We detected very low level of active caspase 3 in cells with gelatinase activity, suggesting that PARP-1 cleavage was mainly because of MMPs. In addition to PARP-1 cleavage, we found a signicant reduction of XRCC1 in the nuclear extracts from ischemic rat brain at 3 h reperfusion, which was reversed by MMP inhibitors in vivo and in vitro. We also detected an increase in MMP-9 in the ischemic nuclei. In vitro, 4-aminophenyl mercuric acetate (APMA)activated MMP-9 acts similarly to MMP-2 in cleavage of PARP-1 and XRCC1, and it could have contributed to nuclear proteolysis. The co-localization of MMP-9 in TUNEL-positive cells in ischemic brain suggested the involvement in cell apoptosis. However, the association of MMP-9 expression and DNA fragmentation in ischemic brain cells is different from the pattern of MMP-2 involvement in the DNA fragmentation at 48 h reperfusion. We detected some TUNEL-positive cells without presence of MMP-9 expression and some cells that express MMP-9 were TUNEL negative. It seems that there is a delayed involvement of MMP-9 in cell death compared with MMP-2 after ischemia. It has been reported that MMP-9 co-localized with neuronal nitric oxide synthase in ischemic brain. During oxidative stress, MMP-9 is activated by peroxynitriteinduced S-nitrosylation. This may trigger proteolytic cascades to disrupt the extracellular matrix leading to apoptotic neuronal death (Okamoto et al. 2001; Gu et al. 2002). On the other hand, studies in myocardial ischemiareperfusion injury showed intracellular location of MMP-2 that was rapidly activated by peroxynitrite and that MMP-2 may rapidly act on intracellular substrates on a minute timescale (Schulz 2007; Ali and Schulz 2009). These suggest a possibility that MMP-2 and -9 induce neuronal death via different pathways. We demonstrated that the oxidative DNA damage as assessed by 8-OHdG levels and AP sites was markedly induced in the brain cells during the early stage of postischemic reperfusion. More importantly, we found that treatment with BB1101 efciently reduced ischemia-induced oxidative DNA damage, suggesting that more oxidized DNA was repaired. To our knowledge, this is the rst report of MMP proteolytic activity involvement in brain cell oxidative DNA damage in an ischemia model. Our results indicated that inhibition of the early intranuclear MMP activity could reduce neuronal DNA fragmentation and death at a later stage after ischemic insult.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

Intranuclear MMPs and oxidative DNA repair in focal ischemia | 147

Fig. 9 Schematic drawing of hypothesis on how intranuclear MMPs facilitate the oxidative DNA damage in neurons after ischemic insult.

Poly-ADP-ribose polymerase-1 is a repair enzyme and is involved in maintenance of nuclear homeostasis, cell survival and death. However, over-activation of PARP-1 caused by massive DNA damage may result in cell necrosis by ATP depletion, and DNA fragmentation and cell apoptosis induced by translocation of apoptosis-inducing factor (Fig. 9) (Yu et al. 2002; Kauppinen and Swanson 2007; Pacher and Szabo 2008). It has been shown that inhibition of PARP activity is neuroprotective in brain ischemia (Skaper 2003; Chiarugi 2005a,b; Tanaka et al. 2005). On the other hand, PARP-1 could act as a survival factor through its capacity to efciently repair damaged DNA by binding to DNA and interacting with BER factors including XRCC1 (Hassa and Hottiger 2008). Activation of PARP-1 by mild genotoxic stimuli may facilitate DNA repair and cell survival. Inhibition of PARP-1 enhanced the vulnerability of neurons to apoptosis. Thus, in mild progressive damage that occurs in neurodegenerative diseases, PARP-1 activation plays a neuroprotective role and may contribute to cellular recovery following sublethal transient global ischemia (Nagayama et al. 2000; Diaz-Hernandez et al. 2007). Based on these reports and our present results, we propose that PARP-1 activity might be benecial for neuronal survival at an early stage during ischemia/reperfusion injury when tissue has not yet become excessively damaged and PARP-1 has not yet become over-activated. We showed that ischemiareperfusion triggers oxidative DNA damage, which leads to PARP-1 activation, initiating the DNA/BER mechanism for repair of oxidized DNA. When repair is successful, the neuron may be protected from apoptosis. However, excessive nuclear gelatinase proteolysis in neurons impairs the protective action of the DNA/BER pathway by cleaving PARP-1 and XRCC1, leading to

neuronal DNA fragmentation and apoptosis (Fig. 9). Our results indicate that early increase in MMP activity in ischemia contributes to oxidative DNA damage in neurons, which can be reduced by MMP inhibitors. As MMPs are important in angiogenesis and neurogenesis, the benecial effects of MMP inhibitors in the early stages of an injury need to be balanced with the later interference with recovery (Lee et al. 2006; Zhao et al. 2006). There are some limitations in this study. Based on our data, it is not possible to discern the specic role of PARP-1 in early ischemic neuronal death. PARP-1 could act as a benecial factor or have detrimental effects in cell survival. Further experiments will be needed to provide direct evidence of whether PARP-1 has a benecial or a detrimental role in early neuronal death in ischemia, and how MMPs could affect this balance. In summary, the present study demonstrated a novel role for MMPs in nuclear DNA damage. We found increased MMP-2 and -9-mediated proteolysis in nuclei of neurons during the early stage of post-ischemic reperfusion both in human stroke and in an animal model. We propose that gelatinase proteolysis in the nucleus plays a role in oxidative DNA damage by cleaving nuclear matrix proteins, PARP-1 and XRCC1, reducing their ability to repair DNA damage caused by oxidative stress after stroke. More importantly, an inhibitor to MMPs protected neurons from the gelatinasemediated oxidative DNA damage after stroke.

Acknowledgements
This work was supported by an American Heart Association beginning grant-in aid (0765473Z), a COBRE pilot grant and a RAC grant from University of New Mexico to YY, and by a National Institutes of Health grant (5RO1 NS04547) to GAR.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

148 | Y. Yang et al.

Confocal images in this paper were generated in the University of New Mexico Cancer Center Fluorescence Microscopy Facility: http://www.hsc.unm.edu/crtc/microscopy.

Supporting information
Additional Supporting information may be found in the online version of this article: Appendix S1 Detailed Materials and methods. Fig. S1 The supernatant of HT-1080 human brosarcoma cell line contains MMP-9 and MMP-2 (pro and active forms). As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing les) should be addressed to the authors.

References
Ali M. A. and Schulz R. (2009) Activation of MMP-2 as a key event in oxidative stress injury to the heart. Front Biosci. 14, 699716. Amantea D., Corasaniti M. T., Mercuri N. B., Bernardi G. and Bagetta G. (2008) Brain regional and cellular localization of gelatinase activity in rat that have undergone transient middle cerebral artery occlusion. Neuroscience 152, 817. Campalans A., Marsin S., Nakabeppu Y., OConnor T. R., Boiteux S. and Radicella J. P. (2005) XRCC1 interactions with multiple DNA glycosylases: a model for its recruitment to base excision repair. DNA Repair (Amst.) 4, 826835. Candelario-Jalil E., Yang Y. and Rosenberg G. A. (2009) Diverse roles of matrix metalloproteinases and tissue inhibitors of metalloproteinases in neuroinammation and cerebral ischemia. Neuroscience 158, 983994. Chang D. I., Hosomi N., Lucero J., Heo J. H., Abumiya T., Mazar A. P. and del Zoppo G. J. (2003) Activation systems for latent matrix metalloproteinase-2 are upregulated immediately after focal cerebral ischemia. J. Cereb. Blood Flow Metab. 23, 14081419. Chiarugi A. (2005a) Intrinsic mechanisms of poly(ADP-ribose) neurotoxicity: three hypotheses. Neurotoxicology 26, 847855. Chiarugi A. (2005b) Poly(ADP-ribosyl)ation and stroke. Pharmacol. Res. 52, 1524. Cuadrado E., Rosell A., Borrell-Pages M., Garcia-Bonilla L., Hernandez-Guillamon M., Ortega-Aznar A. and Montaner J. (2009) Matrix metalloproteinase-13 is activated and is found in the nucleus of neural cells after cerebral ischemia. J. Cereb. Blood Flow Metab. 29, 398410. Diaz-Hernandez J. I., Moncada S., Bolanos J. P. and Almeida A. (2007) Poly(ADP-ribose) polymerase-1 protects neurons against apoptosis induced by oxidative stress. Cell Death Differ. 14, 12111221. Dutra A. V., Lin H. F., Juo S. H., Mohrenweiser H., Sen S. and Grewal R. P. (2006) Analysis of the XRCC1 gene as a modier of the cerebral response in ischemic stroke. BMC Med. Genet. 7, 78. Fujimura M., Gasche Y., Morita-Fujimura Y., Massengale J., Kawase M. and Chan P. H. (1999) Early appearance of activated matrix metalloproteinase-9 and blood-brain barrier disruption in mice after focal cerebral ischemia and reperfusion. Brain Res. 842, 92 100. Gasche Y., Copin J. C., Sugawara T., Fujimura M. and Chan P. H. (2001) Matrix metalloproteinase inhibition prevents oxidative stressassociated blood-brain barrier disruption after transient focal cerebral ischemia. J. Cereb. Blood Flow Metab. 21, 13931400.

Gu Z., Kaul M., Yan B., Kridel S. J., Cui J., Strongin A., Smith J. W., Liddington R. C. and Lipton S. A. (2002) S-nitrosylation of matrix metalloproteinases: signaling pathway to neuronal cell death. Science 297, 11861190. Hassa P. O. and Hottiger M. O. (2008) The diverse biological roles of mammalian PARPS, a small but powerful family of poly-ADPribose polymerases. Front Biosci. 13, 30463082. Heo J. H., Lucero J., Abumiya T., Koziol J. A., Copeland B. R. and del Zoppo G. J. (1999) Matrix metalloproteinases increase very early during experimental focal cerebral ischemia. J. Cereb. Blood Flow Metab. 19, 624633. Ip Y. C., Cheung S. T. and Fan S. T. (2007) Atypical localization of membrane type 1-matrix metalloproteinase in the nucleus is associated with aggressive features of hepatocellular carcinoma. Mol. Carcinog. 46, 225230. Kauppinen T. M. and Swanson R. A. (2007) The role of poly(ADPribose) polymerase-1 in CNS disease. Neuroscience 145, 1267 1272. Kwan J. A., Schulze C. J., Wang W. et al. (2004) Matrix metalloproteinase-2 (MMP-2) is present in the nucleus of cardiac myocytes and is capable of cleaving poly (ADP-ribose) polymerase (PARP) in vitro. FASEB J. 18, 690692. Lan J., Li W., Zhang F., Sun F. Y., Nagayama T., OHoro C. and Chen J. (2003) Inducible repair of oxidative DNA lesions in the rat brain after transient focal ischemia and reperfusion. J. Cereb. Blood Flow Metab. 23, 13241339. Lee S. R., Tsuji K. and Lo E. H. (2004) Role of matrix metalloproteinases in delayed neuronal damage after transient global cerebral ischemia. J. Neurosci. 24, 671678. Lee S. R., Kim H. Y., Rogowska J., Zhao B. Q., Bhide P., Parent J. M. and Lo E. H. (2006) Involvement of matrix metalloproteinase in neuroblast cell migration from the subventricular zone after stroke. J. Neurosci. 26, 34913495. Li N., Wu H., Yang S. and Chen D. (2007) Ischemic preconditioning induces XRCC1, DNA polymerase-beta, and DNA ligase III and correlates with enhanced base excision repair. DNA Repair (Amst.) 6, 12971306. Luo Y., Ji X., Ling F., Li W., Zhang F., Cao G. and Chen J. (2007) Impaired DNA repair via the base-excision repair pathway after focal ischemic brain injury: a protein phosphorylation-dependent mechanism reversed by hypothermic neuroprotection. Front Biosci. 12, 18521862. Marsin S., Vidal A. E., Sossou M., Menissier-de Murcia J., Le Page F., Boiteux S., de Murcia G. and Radicella J. P. (2003) Role of XRCC1 in the coordination and stimulation of oxidative DNA damage repair initiated by the DNA glycosylase hOGG1. J. Biol. Chem. 278, 4406844074. Martin L. J. (2008) DNA damage and repair: relevance to mechanisms of neurodegeneration. J. Neuropathol. Exp. Neurol. 67, 377387. McMahon S., Grondin F., McDonald P. P., Richard D. E. and Dubois C. M. (2005) Hypoxia-enhanced expression of the proprotein convertase furin is mediated by hypoxia-inducible factor-1: impact on the bioactivation of proproteins. J. Biol. Chem. 280, 65616569. Nagayama T., Simon R. P., Chen D., Henshall D. C., Pei W., Stetler R. A. and Chen J. (2000) Activation of poly(ADP-ribose) polymerase in the rat hippocampus may contribute to cellular recovery following sublethal transient global ischemia. J. Neurochem. 74, 16361645. Nazarkina Z. K., Khodyreva S. N., Marsin S., Lavrik O. I. and Radicella J. P. (2007) XRCC1 interactions with base excision repair DNA intermediates. DNA Repair (Amst.) 6, 254264. Ohtaki H., Takeda T., Dohi K. et al. (2007) Increased mitochondrial DNA oxidative damage after transient middle cerebral artery occlusion in mice. Neurosci. Res. 58, 349355.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

Intranuclear MMPs and oxidative DNA repair in focal ischemia | 149

Okamoto T., Akaike T., Sawa T., Miyamoto Y., van der Vliet A. and Maeda H. (2001) Activation of matrix metalloproteinases by peroxynitrite-induced protein S-glutathiolation via disulde S-oxide formation. J. Biol. Chem. 276, 2959629602. Pacher P. and Szabo C. (2008) Role of the peroxynitrite-poly(ADP-ribose) polymerase pathway in human disease. Am. J. Pathol. 173, 213. Planas A. M., Sole S. and Justicia C. (2001) Expression and activation of matrix metalloproteinase-2 and -9 in rat brain after transient focal cerebral ischemia. Neurobiol. Dis. 8, 834846. Ratnikov B. I., Deryugina E. I. and Strongin A. Y. (2002) Gelatin zymography and substrate cleavage assays of matrix metalloproteinase-2 in breast carcinoma cells overexpressing membrane type-1 matrix metalloproteinase. Lab. Invest. 82, 15831590. Rosenberg G. A. (2002) Matrix metalloproteinases in neuroinammation. Glia 39, 279291. Rosenberg G. A. and Yang Y. (2007) Vasogenic edema due to tight junction disruption by matrix metalloproteinases in cerebral ischemia. Neurosurg. Focus 22, E4. Rosenberg G. A., Navratil M., Barone F. and Feuerstein G. (1996) Proteolytic cascade enzymes increase in focal cerebral ischemia in rat. J. Cereb. Blood Flow Metab. 16, 360366. Rosenberg G. A., Cunningham L. A., Wallace J., Alexander S., Estrada E. Y., Grossetete M., Razhagi A., Miller K. and Gearing A. (2001) Immunohistochemistry of matrix metalloproteinases in reperfusion injury to rat brain: activation of MMP-9 linked to stromelysin-1 and microglia in cell cultures. Brain Res. 893, 104112. Schreiber V., Ame J. C., Dolle P., Schultz I., Rinaldi B., Fraulob V., Menissier-de Murcia J. and de Murcia G. (2002) Poly(ADP-ribose) polymerase-2 (PARP-2) is required for efcient base excision DNA repair in association with PARP-1 and XRCC1. J. Biol. Chem. 277, 2302823036. Schulz R. (2007) Intracellular targets of matrix metalloproteinase-2 in cardiac disease: rationale and therapeutic approaches. Annu. Rev. Pharmacol. Toxicol. 47, 211242.

Si-Tayeb K., Monvoisin A., Mazzocco C. et al. (2006) Matrix metalloproteinase 3 is present in the cell nucleus and is involved in apoptosis. Am. J. Pathol. 169, 13901401. Skaper S. D. (2003) Poly(ADP-ribosyl)ation enzyme-1 as a target for neuroprotection in acute central nervous system injury. Curr. Drug Targets CNS Neurol. Disord. 2, 279291. Sternlicht M. D. and Werb Z. (2001) How matrix metalloproteinases regulate cell behavior. Annu. Rev. Cell Dev. Biol. 17, 463 516. Tanaka S., Takehashi M., Iida S., Kitajima T., Kamanaka Y., Stedeford T., Banasik M. and Ueda K. (2005) Mitochondrial impairment induced by poly(ADP-ribose) polymerase-1 activation in cortical neurons after oxygen and glucose deprivation. J. Neurochem. 95, 179190. Vidal A. E., Boiteux S., Hickson I. D. and Radicella J. P. (2001) XRCC1 coordinates the initial and late stages of DNA abasic site repair through protein-protein interactions. EMBO J. 20, 65306539. Wiederhold L., Leppard J. B., Kedar P. et al. (2004) AP endonucleaseindependent DNA base excision repair in human cells. Mol. Cell 15, 209220. Yang Y., Estrada E. Y., Thompson J. F., Liu W. and Rosenberg G. A. (2007) Matrix metalloproteinase-mediated disruption of tight junction proteins in cerebral vessels is reversed by synthetic matrix metalloproteinase inhibitor in focal ischemia in rat. J. Cereb. Blood Flow Metab. 27, 697709. Yu S. W., Wang H., Poitras M. F., Coombs C., Bowers W. J., Federoff H. J., Poirier G. G., Dawson T. M. and Dawson V. L. (2002) Mediation of poly(ADP-ribose) polymerase-1-dependent cell death by apoptosis-inducing factor. Science 297, 259263. Zhao B. Q., Wang S., Kim H. Y., Storrie H., Rosen B. R., Mooney D. J., Wang X. and Lo E. H. (2006) Role of matrix metalloproteinases in delayed cortical responses after stroke. Nat. Med. 12, 441445.

2009 The Authors Journal Compilation 2009 International Society for Neurochemistry, J. Neurochem. (2010) 112, 134149

You might also like