All animal procedures were approved by the Ethics Committee of the Catholic University of Korea and were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80-23). Chronic cerebral hypoperfusion was induced in male Wistar-ST rats (Joongang Lab, Seoul, Korea) and male SD rats (Koatech, Kyungki-do, Korea) weighing 250~270 g, as previously described with minor modifications [5]. Briefly, after induction with 4% halothane, rats were anesthetized with 1.5% halothane in a 70% nitrous oxide and 30% oxygen mixture using a face mask; anesthesia was continued throughout the surgical procedure. After making a midline incision, both common carotid arteries were carefully exposed without damaging the vagus nerve. The carotid arteries were double ligated using silk sutures, trying to minimize injuries of the vessels and adjacent tissues. With the exception of occlusion of the carotid arteries, surgical procedures in sham-operated animals were the same as those in the BCCAo-operated animals. During surgery, rectal temperature was maintained at 37.0~37.9℃ with a heating pad. After the operation, all animals were returned to their cages with free access to food and water. On the 7^th, 14^th, and 21^st days after the surgery (BCCAo or sham operation), the animals were euthanized with 15% chloral hydrate.

Reference memory was assessed using the five-radial arm maze with different tactile cues. To encourage reward-seeking behavior, food was gradually restricted to 85% of the initial weight. SD and Wistar rats were first placed in the maze where small pieces of food pellets were scattered in all five arms with different tactile cues to become familiarized with the radial arm maze for 30 min. On subsequent days, food was placed only at the ends of the 2^nd and 4^th arms, and the animals were given two training sessions per day for 5 consecutive days. One rat at a time was placed in the central zone of the maze. After 1 min, all guillotine doors were opened, and the animal was allowed to enter the arms for 10 min to receive a food reward. To eliminate scent trails, the maze was wiped with 70% alcohol after each trial session. The total number of entries into the arms without a food pellet and entries into the arms with food but no ingestion was divided by the total number of arms visited to give the error rate. Only rats showing error rates lower than 5% were selected and further subjected to sham or BCCAo operations. After 7 days of recovery following BCCAo, the animal was placed for 5 min in the maze where a food reward was located in the 2^nd and 4^th arms, which was exactly the same as the training sessions. For 10 consecutive days, the number of incorrect arm choices, i.e., entering the arms without a reward or entering the arms with a reward but no intake, was evaluated to test the reference memory of SD or Wistar rats that underwent sham- and BCCAo-manipulation, as BCCAo does not induce motor deficits in general [912].

To monitor cortical blood flow in the whole brain, we used a Periscan PIM 3 Laser Doppler Perfusion Imager (LDPI; PIM3, Perimed, Järfälla, Sweden). The LDPI system uses a low-power laser light that is scattered in the tissue and detects changes in wavelength when the light hits moving blood cells in the cortical blood vessels. For efficient transfer of the laser beam to the cortex, the skull surface was thinned using a hand drill. The distance between the LDPI scanner head and the skull was 6.5 cm, producing high-resolution LDPI images with a 99×85 pixel matrix. Cortical microperfusion was repetitively scanned before and at 30 min, 7 days, 14 days, and 21 days after BCCAo and was presented as color-coded images. The blood flow values of the middle cerebral artery (MCA) and the posterior cerebral artery (PCA) territories were numerically determined by the concentration of moving blood cells and mean velocity of these blood cells, which was calculated from the magnitude of the Doppler signal and the frequency shift. Finally, data were presented as % CBF based on the value collected immediately before BCCAo or sham manipulation.

At each time point after sham or BCCAo, animals were transcardially perfused with saline, followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The brain vasculature was visualized with the latex-perfusion technique for cerebral vessels [10]. The whole brain was photographed with a digital camera (EOS 300D; Canon, Tokyo, Japan), followed by post-fixation in 4% paraformaldehyde for 4 h. The degree of plasticity of the PcomA and PCA was evaluated by comparing the diameter with that of the basilar artery using Image-Pro Plus software version 5.1 (MediaCybernetics, Silver Spring, MD, USA), represented as a relative plasticity of PcomA and PCA.

White matter lesions were evaluated by Luxol fast blue staining in regions located −2.64 to −3.12 mm from bregma [5]. After post-fixation, the brains were dehydrated with 30% sucrose and embedded in Tissue-Tek (Sakura Finetechnical, Tokyo, Japan). Tissue sections (20 µm) were mounted on the slide and incubated with Luxol fast blue (Sigma, St. Louis, MO, USA) at 56℃ overnight. Next, the slides were sequentially soaked with 0.05% lithium carbonate solution, distilled water, and 70% ethanol. Finally, the slides were dehydrated in 100% ethanol, cleared in xylene, and coverslipped. White matter damage was evaluated in a double-blind manner, and three sections per animal were analyzed. The severity of white matter rarefaction was graded as normal (grade 0), disarrangement of nerve fibers (grade 1), formation of marked vacuoles (grade 2), and disappearance of myelinated fibers (grade 3).

All data are presented as mean±SEM. SPSS software (version 21.0, IBM SPSS Corp.) was used for statistical comparison. Differences in reference memory and CBF among four different groups were determined by repeated measures analysis of variance (ANOVA) with Duncan's post hoc mean comparison. PCA and PcomA plasticity were analyzed using one-way ANOVA, followed by Duncan's post hoc test. Fold change of relative PCA plasticity based on the sham controls was determined by two-tailed Student's t test with assumption of normal distribution. Finally, white matter damage was analyzed by Pearson Chi-square test. p<0.05 was considered statistically significant.

To test reference memory impairment after BCCAo, SD and Wistar rats were pre-trained for 5 days to remember the arms with a food reward. For 10 consecutive days from the 8^th day after BCCAo, we placed the rats in the five-radial arm maze once a day and evaluated whether they could acquire the reward (Figs. 1A, 1C). Because of white matter damage in the optic tract, different tactile cues were provided at each arm to help blinded rats find the correct arms (Fig. 1B). Entry into the arms without a food pellet (wrong arm) and entry into the food-containing arms but leaving the food uneaten were defined as incorrect choices (Fig. 1C). SD rats made a similar number of errors at all times assessed regardless of hypoperfusion injury, demonstrating that SD animals in both sham and BCCAo groups maintained reward-related reference memory (Fig. 1D). However, Wistar rats showed a higher number of incorrect choices after BCCAo compared with sham-treated animals, suggesting reference memory deficits induced by BCCAo (Fig. 1D).

To explain why the BCCAo procedure resulted in cognitive dysfunction in the Wistar strain but not in SD rats, we monitored cortical microperfusion under the entire skull for 3 weeks. Tracking temporal CBF in the same animal could provide integrated views of CBF alterations after BCCAo (Fig. 2A). Forebrain perfusion examined at the MCA territory was markedly reduced immediately after BCCAo in both SD and Wistar rats (Fig. 2B). Moreover, over the subsequent 21 days, the cortical blood flow of Wistar and SD rats was lower after BCCAo than in sham-manipulated animals, as analyzed by repeated measures of ANOVA. However, in PCA territory, Wistar rats remained hypoperfused after BCCAo, while SD rats recovered CBF similar to that of sham-operated animals (Fig. 2C). Compared with the SD strain, Wistar rats showed significantly lower CBF in both MCA and PCA territories after BCCAo (Figs. 2B, 2C). In summary, in MCA territory, BCCAo induced hypoperfusion in both SD and Wistar rats, with more profound CBF reduction in Wistar rats. In contrast, in PCA territory, only Wistar rats showed low CBF. These data suggest that posterior circulation is a critical factor in the difference in BCCAo-induced hypoperfusion between SD and Wistar rats.

Considering that Wistar rats demonstrated memory deficits and hypoperfusion in PCA territory after BCCAo, which were not observed in SD rats, we decided to examine brain vasculature. The latex-perfusion technique for cerebral vessels was applied at 7, 14, and 21 days after BCCAo or sham-manipulation. In sham-operated animals, significant differences in the plasticity of the PcomA and the PCA were found between SD and Wistar rats (Figs. 3A, 3B). The ratio of the diameter of the PcomA to that of the basilar artery in SD and Wistar rats was 0.84±0.05 and 0.64±0.11, respectively (Fig. 3B). In addition, relative PCA diameter in Wistar rats was clearly smaller than that of SD rats (Fig. 3B). Although sham manipulation may affect the plasticity of these vessels, however, as we tried to minimize unnecessary dissections of the common carotid arteries and the adjacent tissues, it is likely that SD and Wistar rats have a basal difference in the diameter of PcomA and PCA. After BCCAo, no significant difference was observed in plasticity of the PcomA in SD or Wistar rats during the course of the study (Figs. 3A, 3B). However, the relative diameter of PCA was markedly increased in both SD and Wistar rats at all time points after BCCAo (Figs. 3A, 3B). Notably, the fold change in PCA plasticity based on the sham value was significantly greater in Wistar than SD rats (Fig. 3C), suggesting that more reactive changes occurred in the vascular system of Wistar rats.

To determine whether chronic cerebral hypoperfusion and vascular plasticity can be translated into histological changes, we performed Klüver-Barrera luxol fast blue myelin staining (Fig. 4). After 21 days, sham-operated SD and Wistar rats demonstrated an intact optic tract, which was chosen as one of the most vulnerable regions to white matter damage (Fig. 4A). Interestingly, prominent vacuolation and serpentine white matter fibers were observed in Wistar rats after BCCAo, whereas SD rats showed almost intact myelin fibers (Fig. 4A). Moreover, Chi-square statistical analysis revealed that BCCAo significantly induced white matter lesions in Wistar rats compared with SD rats (Fig. 4B).