The methods used for animal experiments were in accordance with the international guiding principles for biomedical research involving animals recommended by the World Health Organization. Permission was granted by Canakkale Onsekiz Mart University Animal Experiments Local Ethics Committee (Protocol number : 2014/03-02, date of approval : 21/03/2014). This study was conducted in Canakkale Onsekiz Mart University Experimental Research Center.

In this study, we used 24 male Sprague-Dawley rats weighing 250-350 g. All animals were fed ad libitum with 7-8 mm pellet rat food (Bil-Yem Ltd., Ankara, Turkey) and tap water. To provide 12 hour darkness, 12 hour light environment photoperiodic white fluorescent light was used and the temperature and humidity were held at 21±2℃ and 55-60%. Before the operation no rat was found to have any abnormalities.

Rats were randomly divided into three equal groups (consisting of eight rats each) : 1) sham operated control group. Laparotomy and infrarenal abdominal aorta dissection was completed but occlusion was not performed; 2) SCIR (ischemia) group fed on a standard diet without kefir pretreatment; and 3) kefir group fed on a standard diet plus kefir (NBL Probiotic ATP, Cell Biotech Co., Ltd., Seoul, Korea) (10 cc/kg/day; 2×10⁹ cfu/kg/day) by gastric gavage during 1 month before SCIR.

The dosage to reveal the beneficial effects of kefir was determined as 10 cc/kg/day body weight (2×10⁹ cfu/kg/day) based on preliminary studies with various doses (3×10⁸, 9×10⁹, 1.8×10¹⁰ cfu/kg bw)^14,23).

SCIR was induced as reported by Lafci et al.¹⁶⁾. A Biopac MP36 (BIOPAC Systems, Inc., Goleta, CA, USA) device was used as a monitor. Mean arterial pulse was 375 per minute during surgery. Body temperature was monitored with a rectal probe and was adjusted to 37.1 to 37.4℃ with a heating pad during surgery. Rats were given premedication with intraperitoneal ketamine (50 mg/kg) and xylazine (5 mg/kg). Anesthesia was continued with ketamine injections at intervals without intubation or mechanical ventilation. Surgical approach was supine position. After the operating field was prepared in sterile fashion, laparotomy was performed with a standard midline incision. After retracting the intestines laterally, the retroperitoneum was opened and was reached to the abdominal aorta. The spinal cord ischemia was induced by cross-clamping the aorta with mini aneurysm clip between just below the left renal artery and just proximal to the aortic bifurcation. Loss of aortic pulse was confirmed by palpation. The duration of ischemia was set at 45 minutes and later the cross clamps were removed and distal reperfusion was observed visually. At the end of the procedure, the abdominal wall was closed with 5/0 prolene sutures. Animals in sham group underwent a surgical procedure similar to the other groups but the aorta was not clamped. This group of animals was used for eliciting the effects of anesthesia and operation on results and also determining the biochemical parameters studied in the normal spinal cord tissue. Animals fed on a standard diet and water ad libitum in their cages after surgery. At 24th hour all animals were anesthetized with penthobarbital (20 mg/kg) and sacrificed. The lumbar spinal cord was harvested immediately via posterior approach. Each spinal cord was longitudinally divided into 2 equal parts with a fine scalpel. Half of the specimen taken for histopathological investigation and it was fixed in formalin for 7 days. The other half was stored in a freezer at -80℃ for biochemical estimations.

Neurological status of animals was assessed blindly by a neurologist at 1, 12, and 24th hour after SCIR. To assess the motor function of rats after SCIR, rats were scored by a modified Tarlov's scale³⁾ as follows : 0, no lower extremity movement; 1, lower extremity motion without gravity; 2, lower extremity motion against gravity; 3, able to stand with assistance; 4, able to walk with assistance; 5, normal.

For biochemical investigation, MDA levels, CAT, and SOD activities from each supernatant were measured in duplicate with highly sensitive ELISA spectrophotometry, respectively. The protein concentrations were determined by the Lowry method¹⁹⁾ using commercial protein standards (Sigma Aldrich, Inc., St. Louis, MO, USA; Total protein kit-TP0300-1KT).

Tissue MDA levels were determined according to Buege's method⁵⁾. MDA levels were analyzed for lipid peroxidation products in spinal cord tissue using the thiobarbituric acid reactive substance test (Cell Biolabs, Inc., San Diego, CA, USA; STA-330 MDA Assay Kit). The results were expressed as nmol/mg protein.

Tissue CAT activity was determined according to Aebi's method¹⁾. Decomposition of H2O2 in the presence of CAT was followed at 540 nm (Cell Biolabs, Inc., San Diego, CA, USA; STA-341 Catalase Activity Assay Kit). CAT activity was reported as U/mg protein.

Tissue SOD activity was measured with a modified spectrophotometric method at 560 nm as described by Sun et al.²⁹⁾. SOD activity was reported as U/mg protein.

The spinal cord samples were fixed with 10% neutral formalin and then histologically processed and embedded in paraffin. Five micron-thick sections were mounted onto slides for histochemistry. Toluidine-blue staining was used to evaluate general histological examination. In the evaluation of toluidine blue-stained sections the dead neurons which were characterized by karyolitic, karyorectic nucleus and vacuolated cytoplasm were counted in 6 different random areas under ×20 objective magnification. All the sections were evaluated under light microscope (Eclipse E-600, Nikon, Tokyo, Japan). The data were statistically analyzed.

The spinal cord samples were fixed with 10% neutral formalin and then histologically processed and embedded in paraffin. Five micron-thick sections were mounted onto poly-l-lysine coated slides for immunohistochemistry. Tissue samples were deparaffinized and hydrated and immunohistochemically stained with HIF1-α and caspase 3 primary antibodies. Citrate buffer (pH=6.0) was used for antigen retrieval. Endogenous peroxidase activity was blocked by 3% hydrogen peroxidase in methanol. Then primary antibodies were dropped and incubated overnight for both in dilution 1 : 50. After incubation, HRP secondary antibody kit was used as a secondary antibody and AEC kit for chromogen. Finally, all the slides were counter-stained with Mayers hematoxylin and mounted with water-based mounting medium. All the chemicals were purchased from Labvision corp (Fremont, CA, USA). HIF1-α and caspase 3 primary antibodies were used to label and evaluate these proteins immunohistochemically. The immunopositive cells in slides were counted in 6 different random areas under ×20 objective magnification. Immunopositive cells were counted by Image Analysis Software (NIS Elements, Nikon, Tokyo, Japan). The data were statistically analyzed.

Results were subjected to one-way analysis of variance using the Statistical Package for the Social Sciences (SPSS 19.0, SPSS Inc., Chicago, IL, USA) software. Differences among the groups were obtained using the Bonferonni's test option and were considered to be significant at p<0.05. All data was expressed as mean±standard deviation in each group. Tarlov scale results were analyzed by the Kruskal-Wallis followed by the Mann-Whitney U test. Difference was considered significant when p<0.05.

Fourty-five minutes of ischemia resulted in severe motor deficit in the hind limbs of ischemia groups, while all sham group animals maintained normal motor behavior (score of 5), as assessed by the motor deficit score. Pre-treatment with kefir did not prevent the development of paraparesis. Most of animals in the kefir group exhibited score 3 or score 4 motor function at 24 h. The scores in kefir group were significantly higher than ischemia group at 24 h (3.25±0.37 vs. 1.37±0.46; p<0.05) (Table 1).

MDA levels were found to be significantly higher in the ischemia group compared to sham group (p<0.05). Also it was determined that spinal cord SOD and CAT activities decreased in the ischemia group (p<0.05). When the kefir group was compared with ischemia group, a significant decrease in MDA levels of spinal cord tissues was observed (p<0.05). CAT and SOD levels in spinal cord tissues of the kefir group were significantly higher than ischemia group (p<0.05) (Table 2).

Histopathological examination of toluidine blue staining in the sham group showed normal morphology. Ischemia group showed many dead and degenerated neurons with karyolitic, karyorectic nucleus and vacuolated cytoplasm. Comparing the ischemia group with sham group, there was a significant increase in the number of dead and degenerated neurons in ischemia group (p<0.05). Comparing the ischemia and kefir group with the sham group, the number of the dead and degenerated neurons were significantly higher than sham group (p<0.05). When the kefir group is compared with ischemia group there was a significant decrease in numbers of dead and degenerated neurons in kefir group (p<0.05) (Fig. 1).

Immunohistochemical results showed that cytoplasmic positivity was found in neuron cells with caspase 3 immunostaining. Sham group showed few caspase 3 immunopositive neurons. Ischemia group showed many caspase 3 immunopositive neurons compared with the sham group and caspase 3 immunopositive neurons significantly increased in the ischemia group (p<0.05). Kefir group showed fewer caspase 3 immunopositive neurons than ischemia group and compared with ischemia group caspase 3 immunopositive neurons were significantly decreased in the kefir group (p<0.05) (Fig. 2).

In staining with HIF1-α, positive staining in cytoplasm and cytoplasmic membrane was observed. Sham group showed few immunopositive neurons. Ischemia group showed many HIF1-α immunopositive neurons and HIF1-α immunopositive neurons were significantly increased compared with the sham group (p<0.05). The kefir group showed fewer HIF1-α immunopositive neurons than ischemia group and compared with the ischemia group the HIF1-α immunopositive neurons were significantly decreased in the kefir group (p<0.05) (Table 3, Fig. 3).