Prophylactic Zinc and Therapeutic Selenium Administration Increases the Antioxidant Enzyme Activity in the Rat Temporoparietal Cortex and Improves Memory after a Transient Hypoxia-Ischemia

In the cerebral hypoxia-ischemia rat model, the prophylactic administration of zinc can cause either cytotoxicity or preconditioning effect, whereas the therapeutic administration of selenium decreases the ischemic damage. Herein, we aimed to explore whether supplementation of low doses of prophylactic zinc and therapeutic selenium could protect from a transient hypoxic-ischemic event. We administrated zinc (0.2 mg/kg of body weight; ip) daily for 14 days before a 10 min common carotid artery occlusion (CCAO). After CCAO, we administrated sodium selenite (6 μg/kg of body weight; ip) daily for 7 days. In the temporoparietal cerebral cortex, we determined nitrites by the Griess method and lipid peroxidation by the Gerard-Monnier assay. qPCR was used to measure mRNA of nitric oxide synthases, antioxidant enzymes, chemokines, and their receptors. We measured the enzymatic activity of SOD and GPx and protein levels of chemokines and their receptors by ELISA. We evaluated long-term memory using the Morris-Water maze test. Our results showed that prophylactic administration of zinc caused a preconditioning effect, decreasing nitrosative/oxidative stress and increasing GPx and SOD expression and activity, as well as eNOS expression. The therapeutic administration of selenium maintained this preconditioning effect up to the late phase of hypoxia-ischemia. Ccl2, Ccr2, Cxcl12, and Cxcr4 were upregulated, and long-term memory was improved. Pyknotic cells were decreased suggesting prevention of neuronal cell death. Our results show that the prophylactic zinc and therapeutic selenium administration induces effective neuroprotection in the early and late phases after CCAO.


Introduction
Zinc plays a dual role in the cerebral hypoxia-ischemia depending on its concentration in the cerebral stroke area; this concentration is known to be determined by zinc serum levels [1]. Accordingly, low serum levels of zinc have long been considered as a risk factor for stroke [2]. In contrast, the input of low concentration of zinc chloride (ZnCl 2 ) or zinc protoporphyrin (ZnPP) reduces the size of postischemic brain damage [3]. Several mechanisms can be accounted for the latter effect. For instance, the decrease in interleukin-1 (IL-1) and IL-23 expression [4], increase in chemokine and growth factor levels [5], and decrease in oxidative stress are because of the antioxidant activity of Cu and Zn superoxide dismutase (SOD1 and 3) [6]. However, excessive accumulation of zinc can also cause neuronal degeneration in the hippocampus and cerebral cortex [7]. Therefore, hypothermia by preventing the accumulation of zinc decreases the cell death [8][9][10], the degeneration of hippocampal neurons, and the loss of memory after hypoxia-ischemia [11]. The mechanism underpinning the hypothermia effect is the reduction of zinc transport from the presynaptic neurons into the postsynaptic neurons during experimental global ischemia.
Selenium has been shown to preserve mitochondrial function, stimulate mitochondrial biogenesis, and reduce infarct volume after focal cerebral ischemia [12]. Selenium treatment in a rat ischemia model decreases oxidative stress [13]. Furthermore, the oral administration of selenium improves learning and memory in an Alzheimer's disease rat model [14]. Administration of sodium selenite together with melatonin 30 min before medial cerebral artery occlusion (MCAO) and for 3 days postreperfusion decreases oxidative stress [15]. A mechanism for the antioxidant effect of selenium is the inhibition of inducible nitric oxide synthases (iNOS) and COX-2 expression through the inactivation of p38 MAPK and NF-κB [16]. Another mechanism is the incorporation of selenium into selenoproteins such as glutathione peroxidase and thioredoxin reductase, thus making the removal of peroxides more efficient [17].
Our group has reported opposite effects of zinc administration in the 10 min common carotid artery occlusion (CCAO) rat model, where CCAO causes cell death by apoptosis and necrosis without producing an ischemic core [11,18]. The subacute administration of zinc (2.5 mg/kg) before CCAO exerts a neuroprotective effect by increasing the expression of CCL2/CCR2, FGF2, and IGF-1 in the temporoparietal rat cortex [5]. In contrast, chronic zinc administration at a low dose (0.5 mg/kg body weight) before CCAO decreases CCL2/CCR2, CCL3/CCR1, CCL4/CCR8, and CXCL13/CXCR5 and increases CXCL12/CXCR4, but does not prevent cell death in the late phase [19]. The antioxidant effect of selenium has been used to exert neuroprotection alone [12] or in combination with other antioxidants such as melatonin [15], Ginkgo biloba [13], and alpha-tocopherol [20]. Therefore, we propose that the combination of the prophylactic effect of zinc with the therapeutic effect of selenium can maintain the neuroprotection on neuroinflammation and neurodegeneration induced by transient CCAO.
To test that hypothesis, we administered zinc (0.2 mg/ kg body weight; ip) for 14 days before CCAO, followed by sodium selenite (6 μg/kg body weight; ip) administration for 7 days after CCAO. We determined nitrosativeoxidative stress (nitrites, lipid peroxidation, NOSs, and antioxidant enzymes), markers of neuroinflammation (chemokines and their receptors) and cell death, over time after CCAO. We also measured neuronal plasticity using the Morris-Water maze test. Because the temporoparietal cortex resulted to be more affected than the hippocampus, we only reported the findings in the temporoparietal cortex. Our results demonstrate that the combined treatment of zinc with selenium extends the effective neuroprotection against CCAO-induced hypoxia-ischemia.

Experimental Animals.
Male Wistar rats (body weight 190 to 240 g) were obtained from the vivarium of CINVES-TAV and maintained in suitable rooms with controlled conditions of temperature (22 ± 3°C) and light-dark cycles (12 h-12 h; light onset at 07 : 00). Five rats per cage (acrylic; 34 cm × 44 cm × 20 cm) were housed. Food (Laboratory Autoclavable Rodent Diet 5010, 130 ppm of zinc, and 0.47 ppm of selenium) and drinking water were provided ad libitum. All procedures were by the current Mexican legislation, NOM-062-ZOO-1999 (SAGARPA), which in turn is based on the Guide for the Care and Use of Laboratory Animals, NRC. The Institutional Animal Care and Use Committee approved the experimental procedures with the protocol number 09-102. All efforts were made to minimize animal suffering.

Zinc and Selenium Administration.
Zinc was administered as ZnCl 2 (0.2 mg/kg of body weight in water for injection, ip; Sigma-Aldrich; Saint Louis, MO, USA) every day for 14 days (chronic zinc administration). Sodium selenite (6 μg/kg of body weight in the water for injection, ip Sigma-Aldrich; Saint Louis, MO, USA) administration started 24 h after the last dose of Zn and continued every day for 7 days. Rats were grouped as follows: (1) control, healthy rats without treatment and surgery, (2) zinc, chronic zinc administration, (3) Zn + Se, chronic zinc administration followed by a single selenium administration on day 15, (4) CCAO, common carotid artery occlusion for 10 min, (5) Zn + CCAO, chronic zinc administration before CCAO, and (6) Zn + CCAO + Se, chronic zinc administration before CCAO followed by selenium administration for 7 days. The rats of groups (1), (2), and (3) were euthanized on day 15 to dissect out their brains. The brains of groups (4), (5), and (6) were obtained at different times (3,6,24, and 168 h) after reperfusion. All the variables studied were measured in the temporoparietal cortex, and all rat groups were age-matched.

Common Carotid Artery
Occlusion (CCAO). The asepsis procedures were performed in the surgical instruments and surgical area. The animals were anesthetized with a mixture of ketamine (70 mg/kg) and xylazine (6 mg/kg) at a dose of 200 μL/100 g of body weight, ip. A 0.5 cm-long midline skin incision was made in the neck area, and the left common carotid artery was carefully dissected. Then, the artery was occluded for 10 min with a clamp (Bulldog Clamps, INS6000119; Kent Scientific Corporation; Torrington, CT, USA). Upon completion of the occlusion, the reperfusion of the artery was visually verified, and the incision was sutured with a 3-0 silk thread (Atramat; Ciudad de Mexico, Mexico). The animals were kept in an individual cage under a 100-Watt, yellow light source until their complete recovery. The animals were euthanized and beheaded in the corresponding time postreperfusion using ketamine (70 mg/kg) and xylazine (6 mg/kg) at a dose of 200 μL/100 g of body weight. The ipsilateral temporoparietal cortex from the different groups was obtained for the biochemical, cellular, and molecular assays.
2.4. Nitrites. The temporoparietal cortex (n = 5 rats in each group) was mechanically homogenized in phosphatebuffered saline solution (PBS), pH 7.4, and centrifuged at 12,500 rpm for 30 min at 4°C by using a Z 216 MK microcentrifuge (HERMLE Labortechnik; Wehingen, Germany). The production of NO was assessed through the accumulation of nitrites (NO 2 − ) in the supernatants as described elsewhere [21]. Briefly, the nitrite concentration in 100 μL of the supernatant was measured by using a colorimetric reaction generated by the addition of 100 μL of Griess reagent, composed of equal volumes of 0.1% N-(1-naphthyl)ethylenediamine dihydrochloride and 1.32% sulfanilamide in 60% acetic acid. The absorbance of the samples was determined at 540 nm with a SmartSpec 3000 spectrophotometer (Bio-Rad; Hercules, CA, USA) and interpolated by using a standard curve of NaNO 2 (1 to 10 μM) to calculate the nitrite concentration.
2.5. Lipid Peroxidation. Malondialdehyde (MDA) and 4hydroxyalkenals (4-HAD) were measured in the same supernatant where the nitrites were measured (n = 5 rats in each group), following the procedure described elsewhere [22]. The colorimetric reaction was made using 200 μL of the supernatant after the subsequent addition of 650 μL of 10.3 mM N-methyl-2phenyl-indole (Sigma-Aldrich; Saint Louis, MO, USA) diluted in a mixture of acetonitrile : methanol (3 : 1) and 150 μL of methanesulfonic acid (Sigma-Aldrich; Saint Louis, MO, USA). The reaction mixture was vortexed and incubated at 45°C for 1 h and afterward centrifuged at 3000 rpm for 10 min. The absorbance in the supernatant was read at 586 nm with a Smart-Spec 3000 spectrophotometer (Bio-Rad; Hercules, CA, USA). The absorbance values were compared to a standard curve in the concentration range of 0.5 to 5 μM of 1,1,3,3-tetramethoxypropane (10 mM stock) to calculate the content of malondialdehyde + 4-hydroxyalkenal (MDA + 4-HAD) in the samples.
2.6. Glutathione Peroxidase Activity. GPx activity was measured using the Glutathione Peroxidase Assay Kit (ab102530; Abcam; Cambridge, UK) following the manufacturer's instructions. A sample of 50 mg of the temporoparietal cortex was washed with cold PBS and homogenized in 200 μL cold assay buffer with a mechanic homogenizer on ice for 15 passes and centrifuged at 10000g for 15 min at 4°C. The supernatants were collected into a fresh corresponding tube and samples of 50 μL were distributed in the respective wells of ELISA microplates (Costar, Corning Incorporated; NY, USA). Then, 40 μL of the fresh reaction mix was added to the well, mixed, and incubated at room temperature for 15 min. Then, 10 μL of cumene hydroperoxide solution was added to each well, mixed, and read at 340 nm on a plate reader (Bio-Rad; Hercules, California, USA). Then, the plates were incubated in the dark at 25°C for 5 min and read again at 340 nm. The activity of GSH-Px in nmol/mL was calculated according to the manufacturer's instructions.
2.7. Superoxide Dismutase Activity. SOD activity was measured using the Superoxide Dismutase Activity Colorimetric Assay kit (ab65354, Abcam; Cambridge, UK) following the manufacturer's instructions. A temporoparietal cortex sample of 50 mg from each rat group was washed in cold PBS and homogenized with a mechanic homogenizer on ice for 15 passes in cold 0.1 M Tris/HCl, pH 7.4 (containing 0.5% Triton X-100, 5 mM β-ME, and 0.1 mg/mL PMSF). Then, the homogenates were centrifuged for 5 min at 4°C at 14000g. The supernatants were collected into a fresh corresponding tube, and samples of 20 μL were distributed in the respective wells of ELISA microplates (Costar, Corning Incorporated; NY, USA). Then, 200 μL of WST working solution was added into each well followed by the addition of 20 μL of Enzyme Working Solution. The plates were mixed and incubated for 20 min at 37°C before reading using a microplate reader (Bio-Rad; Hercules, California, USA) at 450 nm. The OD of each sample was normalized with untreated control (group 1).
2.9. Retrotranscription. Total RNA was isolated from 100 mg of temporoparietal cortex using 1 mL of TRIzol (Invitrogen Corporation; Carlsbad, CA, USA) and then RNA-treated with RNase-free DNase I and quantified using a NanoDrop Spectrophotometer (Thermo Scientific NanoDrop Technologies; Wilmington, DE, USA). cDNA was obtained from 5 μg of total RNA using 1 μL of SuperScript III reverse transcriptase kit (Catalog 18080093, Invitrogen; Carlsbad, CA, USA),1 μL of Oligo dT 50 μM, 1 μL of dNTP mix 10 mM, and water grade molecular biology to 13 μL. Retrotranscription conditions were denaturation at 70°C for 10 min, hybridization at 42°C for 5 min, synthesis of cDNA at 55°C for 50 min and then 70°C for 15 min, and removal of RNA at 37°C for 20 min. Finally, 1 μL of RNase H (Invitrogen; Carlsbad, CA, USA) was added and samples were incubated at 37°C for 20 min.
2.10. qPCR. Fresh cDNA was used to amplify each gene using TaqMan probes (Table 1)  Rat β-actin was used as internal control and for normalization. The amplification assays were made using a 7900HT Fast Real-Time PCR System (Applied Biosystems; Foster City, CA, USA). The 2 −ΔΔCt analyses were used to calculate the relative transcript levels expressed as fold change for gene expression.
2.11. Spatial Reference Learning and Memory. The Morris water maze was used to measure the spatial reference memory. The measurements were conducted in a round tank, 150 cm in diameter and 80 cm deep, filled with water, and divided into four imaginary quadrants. Water was maintained at a temperature of 23 ± 2°C. Several distal visual cues were placed on both walls of the Morris water maze and the room in which it had been installed. This evaluation consisted of five test days of four consecutive trials per day. During the trial, each animal was placed in the tank facing the wall and allowed to swim freely to an escape platform (40 cm in height and 15 cm in diameter), which was submerged by 2 cm under the water surface and conserved to the center of the southeast (SE) quadrant of the tank. If the animals did not find the platform during a period of 60 s in the first trial of each test day, they were gently guided to it, allowed to remain on the platform for 30 s, and then removed from the tank. This procedure was used to ensure that the animals retained the visual-spatial information of the maze online during the execution of the swimming task [24]. Long-term memory was evaluated in the absence of the platform on day 7 after learning. The latency to reach the platform and the number of times that rats pass by the platform location were measured.
2.12. Histopathological Study. The animals were anesthetized with ketamine (70 mg/kg) and xylazine (6 mg/kg) at a dose of 200 μL/100 g of body weight, ip, and intracardially perfused with 200 mL of PBS followed by 100 mL of 4% paraformaldehyde. The brains were obtained and kept in 4% paraformaldehyde at 4°C for 24 h. The brains were embedded in paraffin using a Histokinette (Leica Microsystems; Wetzlar, Germany). The steps of the tissue processing were consecutive dehydration in different ethanol concentrations (80% for 1 h, 96% for 3 h, and 100% for 3), clearance in pure xylol for 2 h, and inclusion in paraffin at 56°C for 2 h. The tissues were placed in blocks using metallic cassettes. The histological sections of 3 μm were made in a rotary microtome-type Minot (Leica RM2135; Wetzlar, Germany) and placed on slides recovered with poly-L-lysine and finally fixed with heat at 60°C for 30 min. The slices were deparaffinized in an oven at 60°C and placed in xylene 2 times for 15 min. The hydration of slices was made by two consecutive passages in decreasing concentration of ethanol (100%, 96%, and 80%) and finally in tap water for 5 min. Hematoxylin staining was performed for 5 min or less until sections look blue. After washing with tap water, the differentiation was carried out with 1% acid alcohol (1% HCl in 70% alcohol) for 1 dip. The slices were washed in running tap water and dipped again in an alkaline solution (i.e., a saturated solution of lithium carbonate) followed by a tap water washing. The staining in 1% eosin Y was done for 10 minutes. The slices were Table 1: List of chemokines and receptors assessed in qPCR: TaqMan probe used was obtained from Thermo Fisher Scientific Inc.

Gene
Gene name Assay Ccl2 Nitric oxide synthase 1, neuronal Rn00583793_m1 Nos2 Nitric oxide synthase 2, inducible Rn00561646_m1 Nos3 Nitric oxide synthase 3, endothelial Rn02132634_s1 Sod1 Superoxide dismutase 1, cytosolic Rn006566938_m1 Sod2 Glutathione peroxidase 4 Rn00820188_g1 washed again with tap water for 1 min. The dehydration of slices was made by ten consecutive dippings in increasing concentration of alcohols (80%, 96%, and 100%) and clearing with xylene. Finally, the slices were mounted on glass slides using Entellan (Merck KGaA; Darmstadt, Germany) and protected with coverslips. The slides were then examined with a light microscope equipped with 10x objective (Leica Microsystems; Wetzlar, Germany). The count of pyknotic cells in 20 fields of 4 brains in layer 5 of the temporoparietal cortex was made using ImageJ software (RRID:SCR_003070, National Institutes of Health).

Experimental Design and Statistical Analysis.
All values were expressed as mean ± SEM from 5 independent experiments including the controls. The values of each variable studied were normalized concerning group 1 (untreated control), except the qPCR values that were expressed as a fold change (2 −ΔΔCt ). One-way ANOVA and Dunnett's post hoc test were used to compare all groups with the untreated control group. Student's t-test was used to compare the treated groups concerning the CCAO group. All statistical analyses were performed using the GraphPad Prism 6 software. The results of learning memory were analyzed with Kruskal-Wallis one-way analysis of variance for comparison of multiple groups, and the Mann-Whitney U test was used for the statistical analysis of two groups. P values <0.05 were considered statistically significant. All statistical analyses were performed using data analysis software (GraphPad Prism, RRID:SCR_0158070).
The histopathology studies showed the presence of pyknotic cells at 168 h after CCAO (Figure 7(b)) as compared with the control group (Figure 7(a)). Zinc (Figure 7(c)) or Zn + Se (Figure 7(e)) did not modify the histological morphology as compared with the untreated control. On the contrary, CCAO caused a significant increase in the number of pyknotic cells (indicative of apoptosis) by 1460 ± 188% (P = 0 0001) at 168 h postreperfusion as compared with the untreated control (Figures 7(b) and 7(h)). Zn + CCAO reduced the number of pyknotic cells by 80% ± 3% (P = 0 001) at 168 h, as compared with the CCAO group (Figures 7(d) and 7(h)), but there was a statistical difference as compared with the untreated control group (Figure 7(g)). Zn + CCAO + Se significantly decreased the number of pyknotic cells by 90 ± 2% (P = 0 0001) as compared with CCAO (Figure 7(f) and 7(h)) and reached the basal values of the untreated control (Figure 7(g)).
The functional recovery from CCAO was assessed through the learning and memory test using the Morris water maze. There was not any statistical difference in information acquisition among the groups (Figure 8(a)). In memory evaluation, 7 days after the training, CCAO increased the escape latency by 53.9 ± 17% (P = 0 0238; Figure 8(b)) and decreased the crossing by the platform location by 37 ± 10% (P = 0 0272; Figure 8(c)) as compared with the controls. On the contrary, Zn + CCAO or Zn + CCAO + Se significantly decreased the latency time to remember the localization of the escape platform by 44 ± 11% (P = 0 0197) and 56 ± 7% (P = 0 0046), respectively, as compared with CCAO, suggesting improvement of consolidation of information. Only the Zn + CCAO + Se group showed a crossing by the platform location (65 ± 14% P = 0 0348) higher than CCAO, thus confirming that this treatment favors the neuronal functionality and exerts effective neuroprotection against hypoxia-ischemia.

Discussion
Our results show that the combined prophylactic of zinc and therapeutic of selenium administration had better effective protection against a transient hypoxic-ischemic event in the temporoparietal cortex, unlike other strategies we have tested such as the prophylactic administration of Se alone or combined with Zn (data not shown). This neuroprotection can be mainly explained by the increase in transcription and enzymatic activity of GPx and SOD, which prevented lipid peroxidation, and the significant decrease in neuronal cell death that is shown by the improvement of long-term memory.
Several studies have shown that chronic prophylactic administration of zinc shows a preconditioning effect [25][26][27]. This preconditioning effect can be explained by the induction of antioxidant enzymes, chemokines, and DNA methylases through zinc finger proteins [28][29][30]. Selenium has also been involved in the epigenetic regulation at least of antioxidant enzymes and DNA methylases [31]. Accordingly, our results show that the administration of those elements caused an upregulation of Nos3, Gpx4, and Sod and the chemokines Ccl2, Cxcl12/Cxcr4, and Cxcl13/Cxcr5; the translation of NOS3; and the increase in the enzymatic activity of GPx and SOD. However, the protein levels of these chemokines and their receptors were not modified by zinc or selenium administration in the period studied. Therefore, the major contributor to the preconditioning effect in the transient hypoxia-ischemia model was the antioxidant effect and the preservation of NO bioavailability through NOS3 expression. As reported in a similar hypoxia-ischemia model, NOS3 is essential in the preservation and maintenance of microcirculation, inhibiting platelet aggregation, leukocyte adhesion, and migration and decreasing the inflammatory response [32]. The increased expression of NOS3 induced by zinc might be associated with the zinc finger protein ZFP580 [33]. Also, zinc stabilizes the dimerization of NOS3, which can prevent the production of superoxide anion and promote the increase in nitric oxide (NO) in the early phase of the hypoxia-ischemia process [34,35]. Interestingly, selenium enhanced NOS3 expression similarly to only zinc administration at the early phase of CCAO, thus recovering the endothelial function as shown in an endothelial dysfunction model [36].
Our results showed that the prophylactic chronic zinc administration in the transient hypoxia-ischemia model increased the enzymatic activity of SOD and the levels of the transcripts of sod1, sod2, and sod3. These enzymes are known to play a major role in protecting from intracellular, extracellular, and mitochondrial oxidative stress, as reported in the cerebral cortex and hippocampus [37][38][39]. A mechanism that accounts for the increase in SOD1 and SOD3 activity is their stabilization by zinc [40]. Also, the antioxidant effect of zinc might be due to the induction of metallothioneins, which are involved in the homeostasis of zinc and ROS [41]. Our results also show that the therapeutic administration of selenium maintains the level of enzymatic activity and transcription of SOD2 in the early phase, SOD1 in the late phase, and SOD3 in the complete period of the study post-CCAO. These three enzymes could have provided an effect of resistance/tolerance to ischemia in the early and late phases of cerebral hypoxia-ischemia, as reported in transgenic mice [42][43][44][45]. Furthermore, the effects of SOD on preventing the disruption of the bloodbrain barrier [46], decreasing karyorrhexis, and attenuating the activation of NF-κB [47] might also explain the neuroprotection induced by the combined treatment with zinc and selenium. Results in sod1 [48], sod2, and sod3 knockout models [37,49,50] also confirmed the neuroprotective effect of SOD. Accordingly, the deficiency of SOD enzymatic activity in the late phase of ischemia has been associated with increases in the size of the infarction, the release of cytochrome c, and the production of mitochondrial superoxide radicals [51]. Of the three SOD isoforms, SOD2 is thought to be the primary contributor to the protective effect in both transient and permanent occlusion [52,53]. Our results support this proposal. We found that the combined treatment with zinc and selenium also causes upregulation of Gpx4 in the early and late phases of CCAO. This result suggests that GPx4 also contributes to the neuroprotective effect of zinc and selenium, removing peroxides from cell membranes and macromolecules such as lipids, proteins, and DNA [54,55]. Another mechanism of neuroprotection by GPx4 is to prevent apoptosis, counteracting the activity of lipoxygenase (LOX) [56] and promoting survival and proliferation [57]. In agreement with the antiapoptotic effect, Gpx knockout mice develop an increased volume of myocardial infarction in a hypoxia-ischemia event [58]. Furthermore, selenium can inhibit TRPM2 and TRPV1 receptors (activated by increasing H 2 O 2 ), thus preventing the entry of calcium into the cell that is known to detonate oxidative stress and inflammation [59].
Our results show that CCAO upregulated the mRNA for Cxcl12/Cxcr4 and Cxcl13/Cxcr5 without modifying their protein levels, although it decreased CXCL13 protein levels. In contrast with our transient CCAO model for 10 min, the permanent occlusion of the middle cerebral artery (MCAO) increases CCL2, CXCL2, and CXCL13 levels after 2 days, gradually decreasing after 7 days of MCAO [60,61]. Therefore, the transient effect of hypoxia-ischemia might be insufficient to alter the protein levels of chemokines and receptors in the time points we have studied. Moreover, the lack of the effect in protein levels can be explained by posttranscriptional regulation of miRNAs [62,63] or a posttranslational regulation at the level of  Figure 8: The administration of the combined treatment with zinc and selenium improves information consolidation in the hypoxiaischemia model in rats. Control = untreated rats; Zn = zinc; CCAO = common carotid artery occlusion; Zn + CCAO = chronic zinc administration before CCAO; Zn + CCAO + Se = chronic zinc administration before CCAO followed by selenium (Se) administration. Each value represents mean ± SEM of 5 independent experiments made in triplicate. * P < 0 05, one-way Kruskal-Wallis analysis of variance with post hoc Dunnett's test when compared with the untreated control group; † P < 0 05, Mann-Whitney U test when compared with the CCAO group.
degradation after receptor-ligand desensitization [64]. In this latter case, those chemokines could have exerted their function before their degradation. Then, the upregulated Ccl2/Ccr2 by the combined zinc and selenium administration in the late phase of hypoxia-ischemia might be neuroprotective because they are known to decrease cell death and improve memory [65]. Furthermore, CCL2 also stimulates the migration of neuronal precursor cells to the damaged area [66]. We have previously reported that high levels of CCL2 by a subacute prophylactic administration of zinc are associated with a preconditioning process [5]. However, the combined treatment with zinc and selenium did not maintain the preconditioning effect of zinc but exerted the therapeutic effect of selenium in the late phase. This effect is reflected by a decrease in cell death and recovery of long-term memory. CXCL12 and CXCL13 have been associated with a deleterious role during cerebral ischemia [67] attracting lymphocytes [68]. Nevertheless, CXCL12 and CXL13 can also attract neuronal precursor cells mainly through interaction with CXCR4 or CXCR5, respectively [69][70][71][72]. CXCL12 and CXCL13 promote the migration of neuroblasts from the subventricular zone in neonatal mice [73], although the main promoter of neuroblast migration is CXCL12 [74]. In this study, we found that the combined treatment with zinc and selenium upregulated Cxcl12 and Cxcl13 and increased CXCL13 protein levels. Therefore, these chemokines can be associated with the neuroprotective effect of the combined treatment with zinc and selenium.
The effect of selenium on memory consolidation has been shown in other models different from hypoxia-ischemia like Alzheimer's disease [75]. The facilitation of learning and improvement of cognitive development have been associated to the neuroprotective effect of zinc, which decreases free radicals produced by cerebral ischemia [76]. Our group showed similar results in the preconditioning effect of subacute zinc administration in the CCAO rat model [5]. However, the prophylactic chronic zinc administration (0.2 mg/kg of body weight/days) exerted a partial effect because there was no memory consolidation as reported previously with a higher dose of zinc [19]. In contrast, the therapeutic administration of selenium improved the long-term memory consolidation, which is consistent with the significant decrease in neuronal cell death induced by cerebral ischemia in the temporoparietal cortex.
In summary, the combination of the chronic prophylactic zinc administration with the therapeutic selenium administration exerts effective neuroprotection against transient hypoxia-ischemia. In this effect, GPx and SOD seem to be the key players in reducing oxidative stress and cell death, suggesting possible participation in neuroregeneration. The perspective of this work consists of challenging the present therapeutic strategy with a longer time of common carotid artery occlusion as it happens in humans.

Data Availability
The data used to support the findings of this study are included within the article.

Conflicts of Interest
The authors declare no competing financial interest, personal or other relationships with other people or organizations within five years of beginning the submitted work.