In Vivo Evidence That the Increase in Matrix Metalloproteinase Activity Occurs Early after Cerebral Ischemia

Abstract

There is controversy over whether matrix metalloproteinases (MMPs) are activated during the early therapeutic window following ischemic stroke. Ex vivo, an increase was reported as early as 4 hours, whereas in vivo, no increase was found until 24 hours postischemia. We used fluorescence diffuse optical tomography to image MMP activity following experimental cerebral ischemia; increased MMP activity was observed in the ischemic area as early as 3 to 6 hours after ischemic onset and correlated with the volume of ischemic cerebral tissue. Therefore, MMP activation is an immediate early response to cerebral ischemia concurrent with the therapeutic window.

STROKE is a major cause of mortality and disability owing to permanent brain damage. The majority of strokes are caused by a dramatic reduction in cerebral blood flow leading to cerebral ischemia and neuronal death. Clinical evidence indicates that the restoration of cerebral blood must intervene as early as possible to limit the destruction of cerebral tissue and improve functional recovery. Multicentric studies converge on the conclusion that the benefit of acute tissue plasminogen activator (tPA) treatment is limited to a time window of less than 4.5 hours following stroke onset. ¹ Therefore, an accurate description of the complex succession of molecular and cellular events immediately following acute cerebral ischemia is important for new treatment strategies.

Matrix metalloproteinases (MMPs) are a family of zinc- and calcium-dependent proteolytic enzymes involved in the degradation and remodeling of the extracellular matrix. Clinical studies have demonstrated the relevance of MMP expression for stroke outcome: for instance, high levels of MMP-9 were related to infarct size, poor neurologic outcome, and hemorrhagic transformation complications. ² Immediately after ischemia, MMPs may disrupt the blood-brain barrier, causing hemorrhagia, edema formation, hemorrhagic transformation, and neuronal cell death. ³ In contrast, it is generally considered that MMPs' contribution at later times is beneficial to neuroplasticity and vascular and functional recovery. ⁴ MMPs are known to modulate excitotoxicity and anoikis, activate death receptors, and control neurotrophic factor availability for the brain. It remains to be elucidated how MMPs contribute to long-term tissue damage and/or repair after cerebral ischemia, a complex issue considering the diversity of MMPs, their diverse functions, and, importantly, their multiphasic expression. ⁵ In particular, a better description of MMP activation immediately after stroke is relevant to the treatment strategy.

Molecular imaging of MMP activity appears to be a relevant method to obtain well spatiotemporally localized information. Noninvasive imaging of protease activity in animal models has been reported using optical imaging techniques. ⁶ Using an optical planar imaging system and probe fluorescing in the presence of active MMPs, Klohs and colleagues showed an increase in fluorescence in the brains of mice 24 hours after induction of cerebral ischemia. ⁷ Surprisingly, however, they did not find evidence of an increase in MMP activity at 4 and 8 hours after stroke, although postmortem studies by other groups reported an increase of MMP-9 and MMP-2 activities as early as 4 hours after middle cerebral artery occlusion (MCAO).^8,9 We reasoned that this discrepancy could be explained either by (1) the limited detection sensitivity of the optical system used by Klohs and colleagues or (2) an overestimation of MMP activity in postmortem experiments, for instance, by the induction of proteases during postmortem processing of tissue.

To solve this issue of importance for future therapeutic strategies, we took advantage of a new fluorescent diffuse optical tomography (fDOT) recently validated for quantitative in vivo studies ¹⁰ and imaged MMP activation during the acute phase of cerebral ischemia in mice.

Materials and Methods

Animals and Surgery

Animal studies were approved by the animal ethics committee and conducted in accordance with the Directives of the European Union on animal ethics and welfare. Transient focal ischemia was produced in adult male C57Bl/6 mice (25 g body weight; n = 6; Charles River, Saint Germain sur l' Arbresle, France) by intraluminal occlusion of the middle cerebral artery (MCA) for 90 minutes followed by reperfusion as described elsewhere. ⁸ Briefly, mice were anesthetized with 2.5% isofluorane in 100% O₂. The tip of a 1.1 cm long monofilament nylon suture grade 8/0 that had been heat-blunted was introduced into the right external carotid artery and moved gently up to the level where the MCA branches out. Anesthesia was discontinued, and mice were allowed to recover. After 90 minutes, the animals were reanesthetized, the filament was removed, and the clip on the common carotid artery was released to allow reperfusion. Anesthesia was discontinued again, and the animals were allowed to recover.

Fluorescence Diffuse Optical Tomography

The MMP-activatable probe MMPsense 680 (VisEn Medical, Woburn, MA) is a substrate for various MMPs, including MMP-1, −2, −7, −9, −12, and −13. MMPsense 680 fluoresces with the spectral characteristics of Cy5.5 after activation (excitation 680 nm, emission 700 nm). Three and 21 hours after the initiation of MCA occlusion, 2 nmol of MMPsense were injected into the tail vein. Three hours after injection (6 and 24 hours after MCAO, respectively), mice were anesthetized as before and placed on a glass plate inside a contact-free fluorescence tomograph (Cyberstar, CEA-LETI, Grenoble, France). Acquisition and quantification of the infrared fluorescent signal from tissues using the transillumination geometry and image reconstruction were performed as previously described. ¹¹ Briefly, the fluorescent tomographic system was operated in the transillumination mode, with a mobile continuous wave laser at 685 nm illuminating the mouse from below the plate glass. The laser beam was guided by a two-dimensional motorized stage in X-Y positions scanning the head of the mouse. At each position, an image was recorded with a sensitive charge-coupled device camera placed above the animal and focused on the mouse surface. A long-pass filter at the 700 nm was used for the acquisition of the fluorescence emission. The complete scan covered a grid of 7 × 6 sources in steps of 2 mm and was recorded in approximately 12 minutes. For the reconstruction, the detection surface was chosen to cover an area of 15 × 13 mm², and the recorded images were mathematically processed with the aid of tomographic algorithms. The Z dimension of the reconstruction mesh depends on the particular thickness of the scanned skull area, which was recorded with a laser line scanning system and implemented inside the reconstruction algorithm. The output is a three-dimensional (3D) map of the fluorescence activity rendered to the mathematical mesh of the mouse body.

Computed Tomography

The optical images were coregistrered with anatomic images acquired on a small animal–dedicated CT scanner (FastScan 1178, SkyScan, Kontich, Belgium) as previously described. ¹⁰ This scanner incorporates a gantry with two source/camera pairs for fast scanning of the subject. Acquisition was performed using the built-in software over an axial field of view of 82 mm with a resolution of 160 mm per pixel. A total of 360 images were recorded, corresponding to 360 angle projection, for a total scanning time of approximately 5 minutes. Reconstructions were carried out using NRecon software (Skyscan), yielding 509 adjacent 0.1609 mm thick coronal slices.

Brain Damage Assessment

Mice were anesthetized and sacrificed by decapitation 24 hours after MCAO. The brain was removed and cut in 2 mm thick sections that were stained with 1% 2,3,5-triphenyltetrazolium chloride (TTC) to show functional mitochondria. The ischemic areas that appeared unstained (because of the inability of injured cells to reduce TTC to its colored form) were measured and grouped for all sections from each animal to define the total volume of the lesion.

Image Analysis

The fDOT images were coregistered to the CT images. Regions of interest (ROI) were drawn in the ipsilateral hemisphere and then copied and pasted symmetrically onto the contralateral hemisphere. The mean value and standard deviation of each ROI in the fDOT images were measured using ImageJ software.

Statistical Analyses

Statistical analyses were carried out with GraphPad Prism (GraphPad Software, La Jolla, CA). Differences between ischemic/control and different time points after reperfusion were analyzed with the Wilcoxon paired test.

Results and Discussion

MMPsense administration followed by fDOT of the cephalic area detected a local increase in the level of fluorescence in mice with unilateral occlusion of the MCA (Figure 1). Coregistration of fDOT with CT showed that the fluorescent signal coincided with the territory irrigated by the MCA (Figure 1, B and D). The appearance of fluorescence through enzymatic cleavage by MMPs was evident 3 hours after MMPsense administration, both at 6 hours (Figure 1, A and C) and 24 hours (see Figure 1, B and D) after MCAO.

Figure 1.

In vivo imaging of matrix metalloproteinase (MMP) activity after cerebral ischemia. Fluorescent diffuse optical tomographic (fDOT) images (mean intensity) were coregistered with computed tomographic (CT) images for the same animal at 6 (A, B) and 24 (C, D) hours after induction of cerebral ischemia by middle cerebral artery occlusion. The upper panels show the CT images coregistered with tridimensional reconstruction of fDOT images (A, C); the lower panels show coregistered CT + fDOT images of 1 mm thickness coronal in serial anteroposterior planes at the level of the ischemic lesion (B, D). Images at the two different time points are from the same representative animal. A threshold has been applied to the fDOT signal to eliminate the background signal corresponding to fluorescence levels in the intact brain.

Quantification of the fDOT signal (mean intensity ± SD) revealed a relative increase in fluorescence in the ipsilateral (ischemic) versus the contralateral areas of 2.5 ± 1.0-fold (15.80 ± 7.59 versus 6.94 ± 4.52, p < .05) and 10.1 ± 6.7-fold (38.07 ± 16.97 versus 5.88 ± 4.69, p < .05) at 6 and 24 hours post-MCAO, respectively (Figure 2). Comparison of the fluorescence levels between 6 and 24 hours showed a 2.4-fold increase after MCAO in the ischemic area (measured at the two time points in the same mouse, p < .05), whereas no difference was found in the contralateral area at the two time points (p > .05; see Figure 2).

Figure 2.

Matrix metalloproteinase (MMP) activity increases early after cerebral ischemia. A, The mean fluorescence intensity was quantified in two region of interest areas defined for each animal over the ipsilateral (containing the infarcted area) and the contralateral hemisphere. Fluorescent diffuse optical tomography (fDOT) at 6 and 24 hours was conducted in the same animal. MMP activity (mean intensity ± SD) shows a significant increase in the ipsilateral (ischemic) side versus the contralateral side at 6 hours (*p < .05) and 24 hours (^&p < .05) after ischemia. There is also a significant increase in the ipsilateral side from 6 to 24 hours (^#p < .05), whereas no significant difference was found in the contralateral side between the two time points. Statistical analyses were carried out by Wilcoxon paired test. B, Correlation between the fDOT measurements (mean fluorescence intensity) at 24 hours after cerebral ischemia and the infarct volume obtained by TTC staining in the same animals. Linear regression analysis shows a significant correlation (p < .001). The goodness of fit was assessed by r ² . Each point represents an individual animal.

Protease-activated near-infrared fluorescent probes are stable and retain their capacity to detect protease activity in blood several hours after injection in mice. ⁶ In contrast to our results, Klohs and colleagues did not find any significant increase in the MMPsense signal until 24 hours after inducing ischemia in a similar MCAO model of cerebral ischemia in mice. The discrepancy observed with our results could be explained by differences in the protocols used in the two studies, specifically (1) the duration of the occlusion period: the animals used in our study were subjected to 90 minutes of MCAO, whereas Klohs and colleagues used 60 minutes, ⁷ as well as by (2) differences in the dose and time of administration of the MMP probe between both studies and/or by (3) the use of fDOT (our study) instead of a planar system (Klohs and colleagues ⁷ ) for optical imaging.

fDOT is a volumetric (3D) imaging technique that takes into account the diffusive propagation of photons in tissue and can quantify picomoles of fluorochromes in whole animals at depths reaching several centimeters of tissue. ⁶ We recently calibrated the fDOT system used in the present study and showed that it allows accurate quantification of fluorescent probes in the near-infrared range at concentrations ranging from 3 nm to 1 μm in a deep-seated organ, with a spatial resolution of 0.7 mm in the x and y axes and 2.0 mm in the z axis. ¹⁰ Moreover, recent progress in the accuracy of the measurements made with fDOT ¹⁰ allows intersubject and intrasubject comparisons of independently acquired experiments.

Coregistration of CT scans of the mouse body and fDOT provided 3D fluorescent maps fused with anatomic information and localized MMP activation in the predicted ischemic area (see Figure 1). This was confirmed by ex vivo histologic TTC staining of ischemic brain showing the extent of brain damaged in the MCA territory. Interestingly, the fDOT signal at 24 hours after cerebral ischemia correlated with the infarct volume obtained after TTC staining (r ² = .90, p < .001; see Figure 2B).

Infarction size shows variability among animals; therefore, a 90-minute MCAO duration was applied to limit as much as possible the variability of the infarction size and the number of animals involved in the study. It should be noted that the importance of the resulting infarct is linked to the duration of the MCAO and, consequently, to the importance of metalloproteinase activation. In the present study, MMP activity detected by fDOT showed stroke volume dependence 24 hours after reperfusion. Nevertheless, fDOT failed to detect MMP activity in animals presenting with small (10–20 mm³) lesions mainly affecting part of the striatum at early time points (6 hours) after reperfusion. Thus, fDOT evidenced MMP activity successfully at early stages after cerebral ischemia in relatively large (ie, corticostriatal) cerebral infarctions. Given that the volume of ischemia-induced infarcted brain is highly correlated with the clinical outcome poststroke, ¹² fDOT imaging of MMP activation may represent a convenient means to assess the severity of the consequences of an ischemic insult in animal models of stroke.

The present results indicate that an increase in MMP activity rapidly follows cerebral ischemia in mice. They are consistent with in vitro studies reporting the activation of MMP-9 4 hours after cerebral ischemia ⁹ and with the suggestion that in the acute phase immediately after cerebral ischemia, MMPs contribute to pathologic neuro-vascular dysfunction through brain barrier disruption and vasogenic edema formation. ¹³ Given that it has been reported that the level of MMP activity is correlated with clinical outcome in stroke patients, ¹⁴ our observation that MMPs are activated immediately or very soon after ischemia is of interest for the testing of early therapy intervention, such as the use of broad-spectrum MMP inhibitors reported to significantly decrease the degree of brain edema and improve neurologic outcome after the onset of ischemia. ¹⁵

In summary, fDOT noninvasively detects MMP activation in vivo during the early phase of acute cerebral ischemia in mice and should prove useful for monitoring strategies modulating MMP activities aimed at promoting recovery after stroke.