Animal studies were performed according to the guideline for Animal Care and Health of the State Veterinary and Food Department of the Slovak Republic (approval no. 2414/06-221/3). The experiments were conducted in accordance with Directive 2010/63/EU of the European Parliament and of the Council for the protection of animals used for scientific purposes.

A total of 50 adult male Wistar rats (Velaz, Ltd.) were used. All animals were maintained on a 12/12-h light/dark cycle. Food and water were available ad libitum until the beginning of the experiments. Animal health and behaviour were monitored regularly by a doctor of veterinary medicine. Transient global cerebral ischemia was produced using the four-vessel occlusion model. Briefly, on day 1, both vertebral arteries were irreversibly occluded by coagulation through the alar foramina following anesthesia with a mixture of 2% halo-thane, 30% O₂ and 68% N₂O. On day 2, both common carotid arteries were occluded for 15 min by small clips under anesthesia with a mixture of 2% halothane, 30% O₂ and 68% N₂O. Two minutes prior to carotid occlusion, the halothane was removed from the mixture. Body temperature was maintained by means of a homeothermic blanket. Global ischemia was followed by 1, 3, 24 or 72 h of reperfusion. During a short time of reperfusion (1 or 3 h), the animals were monitored by the experimenter. Animals surviving over a longer time of reperfusion (24 or 72 h) were monitored by a doctor of veterinary medicine. Control animals underwent the same procedure, apart from the carotid occlusion. The duration of the experiment was 1-4 days, depending on the time of reperfusion. With respect to WB analysis, control animals and animals undergoing ischemia and reperfusion were first anesthetized with a mixture of 2% halothane, 30% O₂ and 68% N₂O and then sacrificed by decapitation. The cerebral cortex and both hippocampi were dissected and processed immediately. With respect to fluorescence immunohistochemistry (FIHC), control and experimental animals were anesthetized, perfused transcardially with ice-cold 0.1 mol/l phosphate-buffered saline (PBS, pH 7.4) and fixed by perfusion with ice-cold 4% paraformaldehyde in PBS. The brains were removed, post-fixed with the same solution as mentioned above for 24 h at 4°C, and cryoprotected by infiltration using 30% sucrose for the next 24 h at 4°C.

The rats were randomized into the following groups: i) Control sham-operated rats (CNT; n=5 for WB and n=3 for FIHC); ii) rats that underwent a 15-min global brain ischemia (ISCH; n=5 for WB and n=3 for FIHC); iii) rats that underwent a 15-min global brain ischemia followed by 1 h of reperfusion (I1R; n=5); iv) rats that underwent a 15-min global brain ischemia followed by 3 h of reperfusion (I3R; n=5); v) rats that underwent a 15-min global brain ischemia followed by 24 h of reperfusion (I24R; n=5); and vi) rats that underwent a 15-min global brain ischemia followed by 72 h of reperfusion (I72R; n=5 for WB and n=3 for FIHC).

Protein extracts were prepared by homogenization of either the cerebral cortex or both hippocampi in homogenization buffer (10 mM Tris-HCl, pH=7.4, 1 mM EDTA and 0.24 M sucrose) using a Potter Teflon glass homogenizer. Total cell extracts were prepared by the addition of an appropriate volume of 6X RIPA buffer [6X PBS, 6% (v/v) Nonidet P-40, 3% (w/v) sodium deoxycholate, 0.6% (w/v) sodium dodecyl sulphate (SDS)] to the homogenate.

Mitochondria from both whole hippocampi were isolated by differential centrifugation as described previously (30). Non-synaptic mitochondria from the cerebral cortex were isolated by differential centrifugation using one-step Percoll gradient (16% Percoll in 0.25 M sucrose), as described previously (31). The protein concentration was determined by the protein Dc assay kit (Bio-Rad Laboratories, Inc.) with bovine serum albumin (BSA) as a standard.

Isolated proteins were separated by 10% SDS-PAGE. Following electrophoresis, the separated proteins were transferred onto nitrocellulose membranes using a semi-dry transfer protocol. The membranes were controlled for even load and possible transfer artefacts by staining with Ponceau Red solution. After blocking with BSA blocking buffer (50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.05% Tween-20 and 2% BSA), the membranes were first incubated for 90 min with primary mouse monoclonal antibodies against GRP75 (1:1,000, sc-133137), VDAC1 (1:1,000, sc390996), β-actin (1:2,000, sc-47778) (all from Santa Cruz Biotechnology, Inc.) and cytochrome c oxidase subunit 1 (COXI; 1 µg/ml, 459600, Invitrogen; Thermo Fisher Scientific, Inc.), rabbit polyclonal antibodies raised against Mfn2 (1:500; sc-50331), or goat polyclonal antibodies raised against DRP1 (1:500, sc-21804) (all from Santa Cruz Biotechnology, Inc.) dissolved in BSA blocking solution. Membranes incubated with primary antibodies were washed in TBS-T solution (50 mM Tris-Cl, pH 7.5, 150 mM NaCl and 0.05% Tween 20) and then incubated with secondary antibodies conjugated with horseradish peroxidase (1:5,000, Santa Cruz Biotechnology, Inc). After extensive washes with TBS-T solution (4 times, 15 min), the membranes were incubated in SuperSignal West Pico Chemiluminescent Substrate solution (Thermo Fisher Scientific, Inc.) for 3 min. Following exposure of the membranes to Chemidoc XRS (Bio-Rad Laboratories, Inc.), the intensities of the corresponding bands were quantified using Quantity One software (BioRad Laboratories, Inc.). The intensities of the bands of interest were normalized to the corresponding band intensities of either β-actin or COXI.

The brains from control and experimental rats were frozen and cut with a cryostat into 30-µm sections; the sections were mounted on Superfrost Plus glass slides (Thermo Fisher Scientific, Inc.). Mounted brain sections were permeabilized with a permeabilization solution (0.1% Triton X-100 with 10% BSA) for 1 h. Mouse monoclonal antibody against β3 tubulin (1:50; sc-80005; Santa Cruz Biotechnology, Inc.), as a specific marker of neuronal cell body cytoplasm and axon guidance, was used as a primary antibody. Mfn2 rabbit polyclonal antibodies (1:50; sc-50331; Santa Cruz Biotechnology, Inc.) were used to detect Mfn2. The tissue sections were incubated at 4°C overnight in primary antibodies diluted in permeabilization solution. Alexa Fluor 488 goat-anti-rabbit IgG (1:50; cat. no. 4412, Cell Signaling Technology, Inc.) was applied as a secondary antibody for Mfn2, and Alexa Fluor 594 goat-anti-mouse IgG (1:100, cat. no. 8890, Cell Signaling Technology, Inc.) was applied as a secondary antibody for β3 tubulin. Finally, the brain sections were cover-slipped with Fluoromount-G^® medium containing 4′,6-diamidino-2-phenylindole (DAPI, CA 0100-20, SouthernBiotech). In the absence of primary antibody, no immunoreactivity was observed. The slides were examined under an Olympus FluoView FV10i confocal laser scanning microscope (Olympus Corporation) equipped with an objective of ×10 with a zoom up to a magnification of ×40 and filters for fluorescein isothiocyanate for Alexa Fluor 488 (excitation: 499 nm; emission: 520 nm) and Texas Red (excitation: 590 nm; emission: 618 nm). Images were captured using Olympus Fluoview FV10-ASW software, version 02.01 (Olympus Corporation) and Quick Photo Micro software, version 2.3 (Promicra, s.r.o.) and further processed in Adobe Photoshop CS3 Extended, version 10.0 for Windows (Adobe Systems, Inc.).

The brightness and contrast of each image file were uniformly calibrated using Adobe Photoshop CS3 Extended, version 10.0 for Windows (Adobe Systems, Inc.). Values of background staining were obtained and subtracted from the immunoreactive intensities.

All statistical analyses were performed using GraphPad InStat V2.04a (GraphPad Software, Inc.). For the comparison of the ischemia-induced changes among all groups, one-way analysis of variance was first performed to determine any differences among all experimental groups. Additionally, an unpaired Tukey's test was used to determine differences between individual groups. The significance level was set at P<0.05.

In order to study the impact of global brain ischemia and ischemia followed by reperfusion on the levels of proteins involved in mitochondrial dynamics and MAM, WB analysis of total cell extracts from the cerebral cortex and hippocampus of control and experimental animals was performed. In total cell extracts prepared form the cerebral cortex, significantly decreased total levels of Mfn2 were observed after ischemia followed by 3 h (69.5% of control, P<0.05), 24 h (58.3% of control, P<0.01), and 72 h (65.9% of control, P<0.05) of reperfusion. In addition to Mfn2, the total level of VDAC1 was significantly increased after 72 h of reperfusion (417.7% of control, P<0.05). The levels of DRP1 were increased, whereas those of GRP75 were decreased after ischemia and after ischemia followed by reperfusion, but the observed changes were not statistically significant (Fig. 1). In hippocampal total cell extracts, ischemia followed by reperfusion led to an increase in the levels of the Mfn2 protein, although these changes were not statistically significant (Fig. 2). A statistically significantly increased level of VDAC1 was documented after 3 h of reperfusion (189.8% of control, P<0.01). The levels of other investigated proteins in hippocampal total cell extracts were unaltered after ischemia and after ischemia followed by reperfusion (Fig. 2).

In addition to the analysis of total cell extracts, the levels of selected proteins were determined in mitochondria isolated from the cortex and hippocampus of control and experimental animals (Figs. 3 and 4). We observed that ischemia and ischemia followed by reperfusion led to a decrease in the levels of Mfn2 in mitochondria isolated from the cerebral cortex; this decrease was significant after ischemia (60.8% of control, P<0.001) and after 1 h (54.8% of control, P<0.001), 3 h (65.2% of control, P<0.01), 24 h (60.6% of control, P<0.001), and 72 h (66.7% of control, P<0.01) of reperfusion (Fig. 3). We also observed that ischemia followed by reperfusion led to a decrease in the levels of the VDAC1 protein in cortical mitochondria, although the changes were not statistically significant (Fig. 3). The levels of the other investigated proteins in the cortical mitochondria were not significantly altered after ischemia and after ischemia followed by reperfusion (Fig. 3). In hippocampal mitochondria, the levels of Mfn2, VDAC1 and GRP75 were increased after ischemia followed by reperfusion, but the changes were not statistically significant. The levels of DRP1 in hippocampal mitochondria were unaltered after ischemia and after ischemia followed by reperfusion (Fig. 4).

To further confirm the WB results, immunofluorescence was used to detect any immunoreactivity of Mfn2 and β3 tubulin, as a neuronal cytoskeletal marker, in the hippocampal area and in the M1 region of the rat brain cortex. Representative images from control rats (CNT), rats that were subjected to global brain ischemia for 15 min (ISCH), and from rats that underwent reperfusion with a duration of 72 h after 15 min of global brain ischemia (I72R) are shown in Figs. 5 and 6. Signals corresponding to Mfn2 were predominantly located within the perikaryal cytoplasm, in the processes of histologically normal tissue in the M1 region of the cortex (Fig. 5B), and in the CA1 layer of the hippo-campus (Fig. 6B). In the M1 region of ischemic rats (Fig. 5E), the intensity of the Mfn2 signal was slightly decreased. The localization of Mfn2 was shifted, and Mfn2 and β3 tubulin signals were co-localized in some neurons (Fig. 5F, arrows). In the CA1 layer of the hippocampus, the signal for Mfn2 was reduced compared with that of the control, and the localization of Mfn2 was transferred to the neuronal processes and the neurophil (Fig. 6E). In the M1 region of the cortex, a 15-min ischemia followed by 72 h of reperfusion resulted in a diminution of the Mfn2 signal, a finding that was consistent with the data from the WB analysis. The localization of Mfn2 was limited to the neuronal processes and the neurophil (Fig. 5H). In the CA1 layer of the hippocampus, Mfn2 immunoreactivity was decreased after 15 min of ischemia followed by 72 h of reperfusion (Fig. 6H), and was relocated to the periphery of the perikaryon, with minimal intersection with the neuronal processes or neurophil. The cytoskeleton of neurons in CA1 was highly disintegrated with no specific morphology in this group of rats (Fig. 6G).