Evolutional Characterization of Photochemically Induced Stroke in Rats: a Multimodality Imaging and Molecular Biological Study

Seven-week-old male Sprague-Dawley rats were kept under standardized condition (12–12-h light-dark cycle, with free access to food and water). The animals received serial MRI and ¹⁸F-FDG PET/computed tomography (CT) imaging at day 1, 3, 7, and 14 after photothrombotic stroke induction. For TTC, H&E, IHC, and EB staining, rats were sacrificed at day 1, 3, 7, and 14 after stroke induction, and intact brain was carefully removed for subsequent manipulation. All animal experiments in this study were conducted according to the guidelines set by the National Laboratory Animal Center and approved by the Institutional Animal Care and Use Committee of National Yang-Ming University and Macau University of Science and Technology. Reporting of this work complies with Animal Research: Reporting in Vivo Experiments (ARRIVE) guidelines.

Rats were temporarily anesthetized with isoflurane (induction 3.0 % in air) followed by intramuscular injection of Dexdomitor and Zoletil 50 mixture (1:1) (Orion, Finland and Virbac, France, respectively) with a dose of 100 μl per 300 g body weight and placed in a stereotaxic frame (Narishige Instruments, Tokyo, Japan). After a small incision was made on the scalp, a craniotomic window (3 mm × 6 mm) was made over the somatosensory cortex with the center at the coordinate of 1 mm rostrally from the bregma and 3.5 mm lateral to the midline. A laser beam of 1.5 mm diameter and 532 nm wavelength (GPD105-M-12, Onset Electro-Optics, Taiwan) was stereotactically positioned at the middle of the craniotomic window and illuminated for 20 min. During the first 2 min of illumination, rose bengal (2 ml/kg body weight, concentration: 10 mg/ml saline) was slowly injected through the tail vein. Two control groups were performed following full procedure except for laser illumination or rose bengal injection. All rats after stroke induction were able to survive until they were sacrificed at the end point in this study.

Laser speckle contrast imaging (LSCI) is a technique in which coherent light incidence on a surface produces a reflected speckle pattern that is related to the underlying movement of optical scatters, such as red blood cells, indicating blood flow [28]. Before and after thrombosis was induced by photochemical method, regional cerebral vasculature was examined by LSCI. Briefly, a moorFLPI-2 Full-Field Laser Perfusion Imager (Moor Instruments, Axminster, UK) was placed at the center of cranial window on the somatosensory cortex where the laser beam illuminated. Before and after laser illumination, the laser speckle imaging was acquired with 25-Hz sampling frequency, 1 frame/s, 580 × 752 pixels resolution, and zoom size of 5.6 mm × 7.5 mm.

BioSpec-70/30 7T system (Bruker, Ettlingen, Germany) using a birdcage head-coil of 75 mm inner diameter for radio frequency (RF) transmission and a 20 mm diameter surface coil for reception was used for MRI experiment. The same rat at day 1, 3, 7, and 14 after thrombotic stroke was anesthetized with initial inhalation of 4 % isoflurane for 3 min and maintained with 2 % isoflurane in a mixture of 20 % oxygen and 80 % room air. Prior to imaging, rats under anesthesia were placed in the stereotaxic holder of MRI machine equipped with a heating system to maintain body temperature and a pressure detector to monitor respiration. MRI data sets consisting TWI, DWI, and PWI were acquired at corresponding time points. T₂WIs were acquired with a multislice multiecho Carr, Purcell, Meiboom, Gill (CPMG) sequence: repetition time (TR) = 2500 msec, echo time (TE) = 33 msec, 16 echoes, field of view (FOV) = 25 × 25 mm², slice thickness = 1 mm, matrix = 256 × 256. Eight coronal and eight horizontal slices were acquired covering a volume extending 10 mm in the rostrocaudal direction. Coronal slices were centered around the infarction lesion, whereas horizontal slices were aligned with the skullcap. DWIs were recorded with a multislice Stejskal-Tanner spin-echo sequence: TR = 4000 msec, echo time = 22 msec, field of view = 25 × 25 mm², slice thickness = 1 mm, matrix = 128 × 128. Data were recorded in the same coronal and horizontal slices as chosen for T₂WI. Two sets of images were acquired with two different diffusion-encoding gradient strengths (b = 30, 1000 s/mm²) in the rostrocaudal direction. PWIs were performed with pulsed arterial spin labeling (PASL) technique using a flow-sensitive alternating inversion-recovery echo planar imaging (FAIR-EPI) sequence with matrix = 96 × 96, FOV = 25 × 25 mm², inversion recovery time (TIR) = 30 to 2300 msec, number of TIR values = 22, recovery time = 10,000 msec, TE/TR > = 10/18,000 msec. The T₁ changes between slice selective inversion sequence and nonselective inversion sequence were used for CBF quantification [29]. The infarct volume and CBF ratio were measured by using PMOD software (version 3.0; PMOD Technologies). The CBF difference indexes (CBFDI) of the infarct lesion, peri-infarct region, ipsilateral remote cortex, and hippocampus were calculated as follows: CBFDI = (CBF of the mirror site of contralateral hemisphere–CBF of the selected region in the affected hemisphere)/CBF of the mirror site of contralateral hemisphere.

¹⁸F-FDG-PET/CT was performed to assess the evolutional change of cerebral glucose utilization. After a 12-h fasting, thrombotic stroke rats (n = 5) at 1, 3, 7, and 14 days were anesthetized with isoflurane and intravenously administered approximately 37 MBq (~1 mCi) of ¹⁸F-FDG. Sixty minutes later, microPET/CT images were acquired for 30 min using a FLEX X-PET and X-O small animal imaging system (GE Healthcare) with the spatial resolution of 1.6 mm and the voxel size of 0.4 mm × 0.4 mm × 1.2 mm. CT images were acquired with 256 projections over 2 min for attenuation correction and anatomy landmarks. PET data were reconstructed using 3D ordered subset expectation maximization (OSEM) method. CT images were reconstructed using a cone-beam reconstruction algorithm. PET and CT images were co-registered using commercial software (Visage Imaging) with 72-μm isotropic CT spatial resolution and 2 mm for PET imaging. For quantitative analysis, volume of interests (VOIs) were drawn on the infarct lesion, one-voxel width surrounding the lesion, and remote region on the ipsilateral cortex. VOIs were also selected at the mirror sites of the contralateral hemispheres. Percent-injected dose of ¹⁸F-FDG per c.c. of brain tissue (%ID/cm³) was obtained from each VOI. The metabolic difference index (MDI) of each VOI was calculated with the following formula: MDI = (%ID/c.c. of contralateral side VOI − %ID/c.c. of lesion VOI)/%ID/c.c. of contralateral side VOI.

Disruption of BBB was evaluated at day 1, 3, 7, and 14 after photothrombotic stroke induction. Briefly, 2 h after EB dye (Sigma, 10 mg/ml in saline, 2.5 ml/kg rat weight) was injected via tail vein, animals were sacrificed with an overdose of pentobarbital injection (RMB, Animal Health Ltd., UK). The brains were then removed from the skulls immediately after death. To detect the presence of EB, the intact brain was imaged using in vivo fluorescence and bioluminescence imaging system (IVIS 50, Perkin Elma, UK) with following steps and image acquisition settings: the brains were placed at the center of imaging field, and images were acquired for 2 s using Cy5.5 band pass filter channel (excitation/emission wavelength: 615~665 nm/695~770 nm). Region of interest (ROI) selection and quantification were performed using living image software 3.2 (IVIS Imaging System, Perkin Elma, UK).

Photochemical induction of embolic stroke was completed by 532-nm LASER illumination at somatosensory cortex through the cranial window upon rose bengal infusion. Vasoconstriction and blockade of blood flow were examined by LSCI before and after stroke induction. As shown in Fig. 1, the surface vessel network was clearly appeared before induction in each group. Thirty minutes after stroke induction, remarkably vanishing of blood vessel network was observed in rats injected with rose bengal through the cranial window where the LASER beam illuminated at, while hyperemic blood vessel network was presented in the group of only illuminated by laser beam and there was no change of blood flow in the group of sole injection of rose bengal. The rapid and massive coagulation of vessels within the illumination area features the pathomechanism of this method which is not similar to that in clinical ischemic stroke usually caused by an embolus or two.

Tissue viability and morphologic change of the photochemically induced stroke were evaluated by TTC and H&E stains at day 1, 3, 7, and 14 post-induction. Results of TTC and H&E staining at each time point were shown in Fig. 2. TTC-negative region which represents the necrotic and non-viable tissue was well demonstrated on the first day after stroke induction, and gradually decreased in size afterward. At day 3, in addition to the shrinkage of ischemic lesion, the boundary between TTC-positive and TTC-negative regions became blurred, as compared to a sharp, clear dividing line at day 1, indicating an increasing number of cells which were migrating and infiltrating to the ischemic margin from outside the region. The volume of the infarct lesions at day 3 to day 14 reduced gradually and remarkably compared to that of the initial infarct volume at day 1. TTC-negative region became hardly observed at day 14 (Fig. 2a). Microscopic examination of the H&E-stained tissues showed a typical stroke-induced liquefactive necrosis at day 1, dilated vessels at day 3, infiltration of cells into the surrounding area of ischemia at days 3 and 7, and neovascular formation at day 14 (Fig. 2b, c).

As a marker of macrophage and microglia, CD68 expression represents the evolution of inflammatory response post-stroke [30]. Figure 3a shows the presence of CD68-positive cells in the border zone of infarction lesion at day 7, and the cell number increased in the lesion at day 14. IHC staining using antibody against GFAP showed the infiltration of astrocytes into infarct boundary at day 7, and interdigitating compact astrocyte scar surrounding the lesion core was observed at day 14 (Fig. 3b). In addition, NeuN-expressing mature neuronal cells were absent in stroke region throughout 14 days after induction (Fig. 3c). These results indicated that the cellular evolutional changes in the lesion were mainly the inflammatory response and astrocyte scar formation, instead of neurogenesis.

As shown in Fig. 4, there were scattered vWF (a marker of endothelial cell) positive vessels in infarct boundary at day 3, and the number of capillaries and αSMA (a marker of vascular smooth muscle cells) positive vessels significantly increased along the infarct margin at day 7. Further, neovascularization was detected in the core of infarct lesion at day 14. These results suggested that angiogenesis and vasculogenesis occurred in the regions of infarct and peri-infarct lesion within 2 weeks after photochemically induced stroke.

To monitor the evolutional brain edema and hemodynamic change of photothrombotic stroke, T₂WI, DWI, and PWI were performed on rats at day 1, 3, 7, and 14 after stroke induction. The ischemic lesion showed a hyperintensity in T₂WI with sharp margin at day 1, and the signals gradually declined afterward (Fig. 5a). Quantitative analysis indicated that the lesion volume declined from 11.4 mm³ at day 1 to 0.7 mm³ at day 14 (Fig. 5b). In DWI, restriction of water diffusion was observed at the margin of stroke shown as an intense halo at day 1 and day 3, but not in the core of the lesion which showed a hypointense signal instead. At day 7, the halo was getting smaller. The hypointense lesion almost completely disappeared at day 14 (Fig. 5a). As T₂WI and DWI have been the reliable methods for investigation of vasogenic edema and cytotoxic edema, respectively [20, 21, 23], gradual reduction of hypointense signal of T₂WI and DWI suggested the improvement of brain edema, which contributed to decrease of infarct volume. PWI at day 1 after photochemical induction demonstrated a region of compromised CBF on arterial spin labeling (ASL) corresponding to the photothrombotic infarct lesion (6 ± 51 ml/100 g/min vs 200 ± 95 ml/100 g/min of contralateral cortex), and slightly decreased CBF in the remote cortex of ipsilateral hemisphere. The CBF compromised zone was getting smaller with time. At day 14, a tiny infarct core was still noted, suggesting a persisted damage of brain tissue. CBF of the ischemic region recovered to 195 ± 55 ml/100 g/min (Fig. 5c). CBFDI of the infarct region (zone A) and the peri-infarct region (zone B) declined rapidly at day 7. CBFDIs of the infarct region and the peri-infarct region showed no significant difference with the contralateral cortex at day 7 and day 14, respectively. CBFDI of ipsilateral remote cortex (zone C) and hippocampus (zone D) closed to zero means CBF almost identical to that of the contralateral hemisphere throughout the whole study of 14 days (Fig. 5d). In summary, the MRI results revealed progressive shrinkage of stroke volume at the periphery of the infarct lesion early after stroke and almost complete recovery of CBF within 2 weeks after stroke induction. Moreover, the remarkable improvement of CBF in infarct lesion and peri-infarct region at day 7 might be related to the synchronous enhancement of angiogenesis and vasculogenesis.

To investigate the metabolic change of the photochemically induced brain stroke with time, ¹⁸F–FDG PET imaging was performed at day 1, 3, 7, and 14 after induction of stroke. As shown in Fig. 6a, a remarkably reduced ¹⁸F-FDG uptake was noted at the site of infarct lesion and moderately reduced uptake in the remote region of ipsilateral hemisphere at day 1 following induction of stroke. At day 3, the cortical metabolic defect partially recovered in the outer region of the ischemic zone and almost completely recovered in the remote cortex. At day 14, the ¹⁸F-FDG uptake at the infarct site almost completely recovered. MDI was calculated to monitor the change of glucose utilization in the infarct lesion (zone A), in the region at one-pixel width surrounding the lesion (zone B), in the remote region of the ipsilateral cortex (zone C), as well as in the hippocampus (zone D). MDIs of zones A and B were decreasing with time, indicating that the glucose consumption was gradually restored at both the infarct region and the margin (Fig. 6b). There was significant decrease in MDIs with time in zone A and zone B in 14 days (ANOVA, p < 0.05). From day 7 to day 14, MDI of zone A decreased by 10 % compared to 20 % decrease in zone B, indicating that the recovery of glucose metabolism at the margin of the infarct is better than that in the infarct core. There was no significant statistical difference of MDI observed for 14 days in zones C and D (ANOVA, p > 0.05), suggesting no obvious compromise of glucose metabolism observed in the ipsilateral remote cortex and hippocampus. Using photothrombotic method, infarct induction at somatosensory cortex did not interfere the glucose utilization in the remote ipsilateral cortex and hippocampus and contralateral cortex as well.

Evidence indicates that BBB disruption occurs after ischemic stroke and leads to the leakage of normally excluded substances into interstitial fluid of the brain. To evaluate the extent of BBB disruption and its restoration, we performed the EB staining and ex vivo fluorescence imaging at day 1, 3, 7, and 14 post-stroke. Figure 7a demonstrated an infarction core surrounded by an EB-stained zone in the somatosensory cortex at day 1 after photochemical induction. Subsequently, the leakage of BBB gradually increased at day 3 and began to restore at day 7; a small EB-stained lesion still remained at day 14. The ex vivo fluorescence imaging showed a high intensity fluorescence corresponding to the EB-stained ischemic lesion. The fluorescence signals were much less intense and less extensive at day 14. Quantitative analysis of EB fluorescence at the ROIs of ischemic region showed a markedly reduced signal intensity at day 14 compared to the highest intensity at day 3 (Fig. 7b, p < 0.05). Fluorescent signal at peri-infarct region was much lower than that at infarct core throughout 7 days and greatly reduced to a background level while the infarct core still showed a high signal at day 14. These results suggest a favorable self-repair of BBB in 14 days after photothrombotic stroke.