Diphenyleneiodonium and dimethylsulfoxide for treatment of reperfusion injury in cerebral ischemia of the rat

Diphenyleneiodonium (DPI) is an inhibitor of the free radical producing NAD(P)H- oxidase. We tested whether DPI shows neuroprotective properties after focal cerebral ischemia and we used dimethylsulfoxide (DMSO), a nonspecific free radical scavenger, as a solvent. In male Wistar rats middle cerebral artery occlusion (1.5 h) and subsequent reperfusion (48 h) (MCAO/R) was induced with the filament model. Immediately after reperfusion the animals received either 0.25 ml normal saline, DMSO, or a combination of DMSO and DPI; each group consisted of 10 animals. MRI was performed at different times after reperfusion. Gelatine zymography of brain tissue for MMP-2 and MMP-9 was performed. The infarct sizes and BBB damage showed a significant difference between controls and the DPI/DMSO group for almost all points in time in all sequences. The activity of MMP-2 and MMP-9 was significantly reduced by DPI/DMSO but not by DMSO alone. DMSO treatment alone resulted in a protective effect with reduced lesion sizes measured by MRI at selected points of time, consistent with its known free radical scavenger effect. The combination of DMSO with DPI partly augmented this effect, presumably due to the additional inhibition of MMP-2 and MMP-9 by DPI. Moreover, the neurological outcome in both therapeutic groups was improved compared to controls with a significant difference between the therapeutic groups in favour of DPI and DMSO. The combination of DPI and DMSO reduced the activity of MMP-2 and MMP-9, attenuated the postischemic blood–brain barrier damage and improved neurological outcome. This was most likely due to reduced oxidative stress.

1. Introduction

Several pathophysiological phenomena contribute to reperfu- sion injury, consisting of secondary release of excitatory amino acids and nitric oxide or formation of reactive oxygen species (ROS), that can enlarge the initial ischemic damage. These phenomena furthermore include upregulation of adhe- sion molecules and inflammatory cytokines which mediate the invasion and activation of leucocytes, and consecutively increase metalloproteinases (MMPs) which contribute to structural damage (del Zoppo et al., 2000; Jean et al., 1998; Maxwell and Lip, 1997; Winquist and Kerr, 1997). These mechanisms are partly interdependent and form cascades. ROS, like superoxide or peroxynitrite, play a pivotal role in the pathophysiology of ischemia when they appear at toxicologi- cally relevant or unphysiologically high concentrations (Becker et al., 1999; Hall and Braughler, 1993; Johnson and Weinberg, 1993). Consistent with this, scavengers of ROS and inhibitors of lipid peroxidation have shown neuroprotective effects after ischemia and reperfusion (Chan, 2001; Kontos, 2001). One major source of oxidative stress and ROS are leucocytes that migrate into the damaged cerebral tissue. They produce ROS mainly through the action of NAD(P)H- oxidase and other enzymes. We have evidence that the lack of the NAD(P)H-oxidase may lead to a significant reduction in ROS mediated neurotoxicity (Walder et al., 1997). Diphenyle- neiodonium (DPI) is a potent inhibitor of the NAD(P)H-oxidase which specifically and irreversibly binds to the flavicenter of the enzyme (Hampton and Winterbourn, 1995; Wang and Pang, 1993). It has shown its selective inhibitory properties in several studies (Dodd-O and Pearse, 1998; Doussiere et al., 1998). Moreover, it is also able to interact with other enzymes potentially involved in the pathophysiology of ischemia– reperfusion injury, i.e., nitric-oxide synthase, xanthine oxi- dase and NADH cytochrome P450 oxidoreductase (Doussiere and Vignais, 1992; Stuehr et al., 1991; Tew, 1993). DMSO is a simple organic compound, is highly distributive in the CNS, solubilizes many lipophilic substrates and moreover has anti- oxidative, cell membrane stabilizing and antiinflammatory properties (Kolb et al., 1967). It has already been used as a nonspecific free radical scavenger in studies of cerebral ischemia in rats and humans and proved to possess beneficial properties (Karaca et al., 2001; Shimizu et al., 1997). It is known that MMP-2 and MMP-9 are increased in cerebral ischemia cleaving extracellular matrix molecules which leads to BBB breakdown (Rosenberg et al., 1998; Wagner et al., 2003). Besides other activation pathways MMPs might also be increased by oxidative stress. The recent findings of Gursoy– Ozdemir et al. suggest that superoxide and peroxynitrite formation in and around microvessels after ischemia and reperfusion contribute to loss of selective BBB permeability by activation of MMP-9 (Gursoy-Ozdemir et al., 2004).
We therefore tested the hypothesis whether DPI protects the postischemic BBB and reduces the activity of MMPs by the postulated reduced production of free radicals. As DPI is not soluble in hydrophilic substances we used DMSO as a carrier. We quantitatively monitored the evolution of ischemic lesions and the BBB breakdown with MRI and performed quantitative zymography.

2. Results

2.1. Dose finding for DPI

In our in vitro dose-finding experiment NAD(P)H-oxidase induced superoxide generation was dose-dependently decreased following addition of DPI. Approaching a concen- tration of about 100 μM DPI this inhibitory effect was saturated (possibly due to alternative sources of superoxide, non-
responsive to DPI). The results of the in vitro lucigenin assays are presented in Fig. 1. For further analysis of the in vivo capabilities of DPI in attenuating stroke damage we chose a dose in a range where DPI had a maximal inhibitory effect on the superoxide generation in vitro without affecting cardior- espiratory parameters or showing overt acute toxicity in vivo. The dose chosen, a bolus injection of 0.25 ml of a 50 mM DPI solution, would be expected to result in an effective peak blood concentration of about 0.7 mM and an estimated tissue concentration of 50–100 μM after complete distribution assuming a mean body weight of 275 g. Throughout the entire phase of the experiments including the period immediately following the bolus injection no significant alterations of mean blood pressure or respiratory parameters occurred (see Table 1) (Li and Trush, 1993).

2.2. MRI studies

All MR images displayed a progressively increasing lesion over time in all experimental groups whereas sham ani- mals showed no ischemic signs in all sequences (Fig. 2A). The combination of DPI/DMSO reduced lesion size, mea- sured as hemispheric lesion ratio (HLR) (Nagel et al., 2004), at all points in time compared to the control group ( p < 0.05). After 24 h of reperfusion DMSO alone significantly reduced the HLR in all sequences as well as after 48 h in pcT1WI ( p < 0.05). A significant additive neuroprotective effect of the combination of DPI/DMSO over DMSO alone could only be observed at 4 h in DWI and T2WI ( p < 0.05). Beyond 4 h the difference was not significant any more in these sequences; however the lesion size of BBB breakdown measured in postcontrast T1WI at 48 h was significantly reduced again by DPI/DMSO compared to DMSO ( p < 0.05). The HLR of all groups can be beheld in Fig. 2B as bar graphs and in Table 2 as numbers.

Fig. 1 – Inhibition of fMLP-triggered superoxide generation in isolated blood leucocytes by DPI. Freshly isolated rat blood leucocytes were stimulated with 100 nM of the chemotactic peptide fMLP. The maximal superoxide generation during the first 30 min was determined following 10 min of preincubation with different concentrations of DPI. Each data point represents the mean±SD of 6 independent measurements.

2.3. Physiological parameters, functional outcome and mortality

Body temperature, pH, pO2, pCO2, heart rate and mean arterial blood pressure were within the expected physiologi- cal range for all groups (see Table 1). Blood glucose levels were assessed once 1 h after treatment and were similar across all groups: control 111 ± 11, DMSO, 106 ± 14, DPI/DMSO 103 ± 12 (unit: mg/dl) (Holland et al., 1973).

Functional outcome after 48 h of reperfusion differed significantly between all three groups ( p < 0.05), rats reached the highest neuroscore in the DPI/DMSO group (15.4 ± 1.3), followed by the DMSO group (12.6 ± 0.9) and the control group (8.7 ± 1.2).

Mortality did not differ significantly between the three treatment groups and was as expected for controls (Aspey et al., 2000; Maier et al., 2001): 5/15 animals died in the control group, 1/11 in the DMSO group and 2/12 in the DPI/DMSO group. The cause of death was presumably cardiorespiratory failure in one animal of the control group and the DPI/DMSO group; the rest died of malignant infarction.

2.4. Zymography

We could not detect any processed forms of MMP-2 and MMP- 9 in all groups. As no clear bands could be identified, MMP-9 was below detection limits by zymography in sham animals and in the contralateral hemispheres of the experimental groups. MMP-2 in sham animals and contralateral hemi- spheres was consistently lower than in both experimental groups ( p < 0.05). MMP-9 showed a trend towards reduction following DMSO treatment alone. However, it was signifi- cantly reduced by the combination treatment of DPI/DMSO ( p < 0.05) compared to the control group. In addition MMP-2 was significantly reduced by DPI/DMSO and not by DMSO alone ( p < 0.05). Activities are displayed in figure as bar graphs of integrated optical density (see Fig. 3).

3. Discussion

DMSO treatment resulted in an improved outcome after transient cerebral ischemia measured by MRI and by a functional neuroscore. The combination of DPI and DMSO was significantly superior to DMSO treatment alone with reference to functional outcome, to MMP-2 and -9 inhibition and to selected MRI measurements, i.e., lesion size at 4 h and volume of BBB breakdown after 48 h of reperfusion.

DMSO has already proved its neuroprotective effects in rats in different doses when administered before and after permanent ischemia of 24 h and MRI observations showed a persistent diffusion/perfusion mismatch (Bardutzky et al., 2005; Shimizu et al., 1997). Now we can confirm these results after transient ischemia and moreover show an attenuation of the BBB damage which was confirmed with pc T1WI. There is controversy as to whether DMSO itself enhances the perme- ability of the BBB (Greig et al., 1985; Kleindienst et al., 2006; Neuwelt et al., 1983). With a molecular weight of approxi- mately 90 kDa Magnevist® exceeds the upper limit for the size of substances delivered across the BBB by DMSO that has been determined by Broadwell et al. (1982). Nevertheless we cannot exclude an effect of DMSO on the contrast media permeability over the BBB. However, an attenuation of focal enhancement is clearly related to its neuroprotective properties and cannot be seen as a confounding factor as DMSO was used in both therapeutic groups.

An additional effect of DPI over DMSO measured in lesion sizes in DWI and T2WI could only be observed after 4 h of reperfusion. This is possibly due to the decrease of the first burst of superoxide and peroxynitrite (resulting from the interaction of superoxide and NO) and the consecutive reduction of excitotoxicity by inhibition of the circulating, marginalized and transmigrated leucocytes (Afshar-Kharghan and Thiagarajan, 2006; Ishikawa et al., 2004; Lossinsky and Shivers, 2004). This action, as well as potential additional effects on the generation of ROS from nonphagocytic sources (glial or neuronal cells) is likely to mediate a substantial attenuation of oxidative stress in the infarcted tissue and surrounding tissue at risk. In accordance with this assumption it has recently been demonstrated in vitro that DPI inhibits the inflammatory microglial response after lipopolysaccharide stimulation (Wang et al., 2004). Consistent with previous data on the inhibitory effect of DPI on granulocyte NAD(P)H- oxidase, we measured decreased LGLC of leukocytes after DPI exposure, providing a mechanism for reduced superoxide production (see Fig. 1). An exacerbation of ROS production after DPI as described by Riganti et al. (2004) could not be observed.

We were not able to show an additional effect of DPI to DMSO treatment on the BBB breakdown after 4 h of reperfu- sion. It is likely that this observation time was too short to display any effects on BBB breakdown as contrast enhance- ment was too weak and lesion volumes were too small. As DPI mediated inhibition of NAD(P)H-oxidase is due to an irrever- sible interaction, we assume that ongoing protective effects far beyond the initial stages of ischemia–reperfusion would take place. This would be highly desirable keeping in mind the delayed cellular responses to ischemia. After 24 h of reperfu- sion the lesion volumes in DWI and T2WI just showed a trend towards smaller lesions in the combination group but no significant difference. The BBB damage after 48 h however was significantly reduced by the combination therapy. We suppose this was due to the inhibited activity of both MMP-9 and MMP-2. DMSO alone reduced activity of neither MMP-2 nor MMP-9. DPI and DMSO provided inhibition of MMP-2 activity as well as inhibition of MMP-9. MMP-2 and MMP-9 have been clearly identified as a major cause of BBB disruption in cerebral ischemia due to their degradation of components of the basal lamina, and therefore constitute a highly attractive target for innovative pharmacological interventions (Romanic et al., 1998; Rosenberg et al., 1998; Wagner et al., 2003). MMPs are secreted as inactive precursors and require activation by other proteases and free radicals (Mun-Bryce et al., 2002). Gasche et al. already demonstrated that oxidative stress mediates BBB disruption through metalloproteinase activation in mice lacking copper/zinc-superoxide dismutase-1 (Gasche et al., 2001). Moreover, the association between superoxide and MMP-2 and MMP-9 activation has also been proved for other tissues like rat cardiocytes or bovine smooth muscle cells (Egi et al., 2004; Mandal et al., 2004).

Therefore, a neuroprotective effect of DPI in vivo after focal cerebral ischemia through the reduction of oxidative stress and the consecutive inhibition of MMP-2 and MMP-9 can be suggested (Jian Liu and Rosenberg, 2005). However, part of the measured effect in reduction of lesion sizes is rather attributable to DMSO. Additionally the inconsistency of the significant protective effect on the BBB and the inability to reduce lesion sizes significantly at 48 h of reperfusion in DWI and T2WI imply also a possible direct inhibitory mechanism of DPI on MMP-2 and -9. Most interestingly we found an improvement of functional outcome after DPI/DMSO which was superior to DMSO alone, correlating with the attenuation of the BBB and displaying a neuroprotective effect of DPI beyond reduction of lesion volumes.

In summary we proved after transient focal cerebral ischemia the neuroprotective potential of DMSO and showed for the first time neuroprotective properties of DPI through MMP inhibition and attenuation of the ischemic BBB and improvement of functional outcome. Further studies with DPI or other inhibitors of the NAD(P)H-oxidase, especially with repetitive applications of the drugs and longer observation periods, are warranted to confirm these results and to make the desired transfer form bench to bedside. Recently neuroprotec- tive properties of another selective NAD(P)H-oxidase inhibitor have been demonstrated and the use of another free radical scavenger showed a neuroprotective effect in a large clinical trial (SAINT) (Lees et al., 2006; Wang et al., 2006). The reduction in symptomatic and asymptomatic hemorrhagic transforma- tion, besides indifferent infarct sizes, documented in the SAINT trial might also be mediated by reduced MMP activity.

heating pad in all animals. A femoral artery was cannulated for the continuous monitoring of arterial blood pressure and for measurement of arterial blood gases before occlusion, during occlusion and during reperfusion. A femoral vein was cannulated for delivering the drugs and the contrast agent for MRI. During surgery and up to 4.5 h after reperfusion the animals were maintained in anesthe- sia with 1–2% halothane, 75% N2O and 23% O2. Immediately after the induction of reperfusion the animals received randomly either a bolus of 0.25 ml isotonic NaCl (N), DMSO or DMSO and 50 mM DPI (n = 10 in each group). Animals lost from any group during the period of the experiment were replenished. Three animals were replaced due to technical problems with the MRI machine. After the initial anesthesia.

4. Experimental procedures

4.1. Dose finding for DPI

In order to establish an optimal dose for the treatment with DPI and to assess toxicity we performed a range of in vitro experiments. Polymorphonuclear phagocytes were prepared from fresh, heparinized rat blood by Ficoll-Hypaque gradient centrifugation. The cells were rinsed in PBS several times. Absence of endotoxin in any reagent or material was verified by Limulus amebocyte lysate (LAL) assays. The resulting cell pellet was repeatedly resuspended in PBS+ by gentle aspiration through an 18-G needle. Trypan blue exclusion tests were performed to verify the integrity of cells. Aliquots of the cell suspension were transferred into the test tube and allowed to settle on the bottom of the cavities to minimize light scattering. After addition of lucigenin (bis-N-methylacridi- niumnitrate, 100 μM final concentration) the test tube was inserted into the measurement chamber of the luminometer for 15, further to achieve reagent-uptake in the dark. After background correction the basic light output and the reaction towards pharmacological interventions were continuously recorded by measuring the photon emission. The specificity of Lucigenin-enhanced chemiluminescence (LGCL) for stimu- lated O− release and the relative contribution of intra- and extracellular ROS sources were demonstrated by adding superoxide dismutase (SOD) or cell-permeable SOD mimics MnTBAP (manganese[III]tetrakis[4-benzoic acid]porphyrin) or MnTMPyP (manganese[III]tetrakis[1-methyl-4-pyridyl]por- phyrin), or the low molecular weight O− scavenger tiron (4,5- dihydroxy-1,3-benzene-disulfonic acid) (Fig. 1).

4.2. Experimental stroke model in the rat

The experimental protocols used in this study were approved by the ethics committee for animal research, Karlsruhe, Germany.After randomization occlusion of the middle cerebral artery for 90 min with subsequent reperfusion for 48 h (MCAO/R) was induced in 45 (including 4 sham animals) adult male Wistar rats (275 g± 10%) according to a modification of the method of Zea Longa et al. (1989). Each experimental group consisted of 10 animals. Experimental occlusion of the MCA has been described in detail elsewhere (Garcia et al., 1993, 1995).Body core temperature of 37 °C was controlled by a rectal thermometer connected to a feedback controlled the rats were allowed to wake up with free access to food and water and were reanaesthetized for MRI measurements shortly thereafter. After each experiment rats were evaluated with a func- tional neuroscore for neurological outcome on a scale from 3 (dead) to 18 (no deficit) (Garcia et al., 1995).

4.3. Tissue processing

After each experiment transcardiac perfusion with chilled (4 °C) isotonic NaCL solution was performed; thereafter brains were removed from the skull. Specimens were frozen in isopentane and stored at − 80 °C until postproces- sing. With the brains of each group including the sham animals zymography was performed. We used the ipsilat- eral hemisphere after mechanical and chemical homogenization according to a common protocol for gelatine zymography (Wagner et al., 2003); the contralateral hemi- sphere served as control.

4.4. Zymography

Gelatinases were bound to gelatin-coated sephadex beads washed with elution buffer (50 mM Tris, 1 M NaCl, 10 mM CaCl2, 7.5% DMSO, 0.16% Triton X-100) and subsequently loaded on the gel for zymography. Protease activity was visualized as clear bands. Gels were scanned and bands quantitated by volume and density integration using the NIH Image program (Wayne Rasband, NIH, Bethesda, USA http:// rsb.info.nih.gov/ij). as previously described (Wagner et al., 2003).

4.5. MRI protocol

The animals were examined in a 2.35 T scanner (Biospec 24/40, BRUKER Medizintechnik, Ettlingen, Germany). An actively shielded gradient coil with 120-mm inner diameter was used. This coil was driven by the standard 150V/100 A gradient power supply. In this configuration, 180 mT/m could be reached in 180 μs. For the RF-coil we used a home- built birdcage resonator with 40-mm inner diameter. Serial MR examination started 4 h after initiation of reperfusion and further examinations were done after 24 h and 48 h. In each animal we performed diffusion-weighted MR imaging using a spin-echo echo-planar-imaging (EPI) sequence (TR= 3 s, TE= 45 ms, NA= 4, 6 different b-values from 200 to 700 s/mm2, diffusion time Δ= 25 ms, duration of diffusion gradient δ= 5, FOV= 4 cm× 4 cm, matrix= 128 × 64, 6 slices, thk= 2 mm), T2-weighted MR imaging using a multi-spin- echo sequence (TR= 2 s, TE= 8, 16, 24, … 96 ms, NA= 1, FOV= 4 cm× 4 cm, matrix= 128 × 96, 6 slices, thk= 2 mm), and T1-weighted MR imaging using a spin-echo sequence (TR = 350 ms, TE = 15 ms, NA = 4, FOV = 4 cm × 4 cm, matrix= 256 × 192, 6 slices, thk= 2 mm). T1-weighted MR imag- ing was performed 5 min after injection of 0.5 mmol/kg bw. Gd-DTPA (Magnevist®, Schering AG, Berlin, Germany). During all scans the body temperature was maintained at 37 °C by a heating fan.Image data were transferred to a SUN Sparcstation 10 (SUN Microsystems, USA). Measurement of infract volumes in DWI, T2WI and postcontrast T1WI was done by a blinded investi- gator as previously described (Nagel et al., 2004; Wagner et al., 2003).

4.6. Statistical analysis

All data are presented as the mean±standard deviation (SD). For all intergroup comparisons the Kruskall–Wallis test and the Mann–Whitney test (with Bonferroni adjustment) when appropriate were applied. To test significance in mortality rates between the treatment groups we used the Fisher’s Exact Test.