Molecular and Pharmacological Modulation of CALHM1 Promote Neuroprotection against Oxygen and Glucose Deprivation in a Model of Hippocampal Slices

Calcium homeostasis modulator 1 (CALHM1) is a calcium channel involved in the regulation of cytosolic Ca2+ levels. From a physiological point of view, the open state of CALHM1 depends not only on voltage but also on the extracellular concentration of calcium ([Ca2+]) ions. At low [Ca2+]e or depolarization, the channel is opened, allowing Ca2+ influx; however, high extracellular [Ca2+]e or hyperpolarization promote its resting state. The unique Ca2+ permeation of CALHM1 relates to the molecular events that take place in brain ischemia, such as depolarization and extracellular changes in [Ca2+]e, particularly during the reperfusion phase after the ischemic insult. In this study, we attempted to understand its role in an in vitro model of ischemia, namely oxygen and glucose deprivation, followed by reoxygenation (OGD/Reox). To this end, hippocampal slices from wild-type Calhm1+/+, Calhm1+/−, and Calhm1−/− mice were subjected to OGD/Reox. Our results point out to a neuroprotective effect when CALHM1 is partially or totally absent. Pharmacological manipulation of CALHM1 with CGP37157 reduced cell death in Calhm1+/+ slices but not in that of Calhm1−/− mice after exposure to the OGD/Reox protocol. This ionic protection was also verified by measuring reactive oxygen species production upon OGD/Reox in Calhm1+/+ and Calhm1−/− mice, resulting in a downregulation of ROS production in Calhm1−/− hippocampal slices and increased expression of HIF-1α. Taken together, we can conclude that genetic or pharmacological inhibition of CALHM1 results in a neuroprotective effect against ischemia, due to an attenuation of the neuronal calcium overload and downregulation of oxygen reactive species production.


Introduction
Ischemic stroke is the most frequent cause of cerebrovascular disease [1]. It is the third leading cause of death and the main cause of adult disability worldwide. Despite this high unmet medical need, there is no current effective neuroprotective treatment [2]. To date, thrombolytic or mechanical removal of the occlusion are the only therapeutic options. Mechanical thrombectomy complications occur in 11% of the patients and the intravenous thrombolytic treatment, using recombinant tissue exchanger located in the mitochondrial membrane and partially blocks voltage-gated Na + and L-type voltage-dependent calcium channels. This compound, which is able to cross the BBB, has been shown to exert neuroprotective properties in different oxidative stress models [22].
In this context, we aimed to identify by which molecular mechanisms the absence or blockade of CALMH1 exerts a protective effect under brain ischemia conditions by using hippocampal slices of Calhm1 +/+ , Calhm1 +/− , and Calhm1 −/− mice.

Animal Usage and Care
Native male and female mice (C57BL/6J) for CALHM1, Calhm1 +/− and Calhm1 −/− were used in the experiments following The Guide for the Care and Use of Laboratory Animals and were previously approved by the Institutional Ethics Committee of the Universidad Autónoma de Madrid, Spain. Mice were housed under controlled lighting and temperature conditions, and water and food were administered ad libitum. All efforts were made to minimize the number of animals used and their suffering.

Isolation and Preparation of Mouse Hippocampal Slices
Adult mice (2-4 months) were used to obtain hippocampal slices (HSs). Mice were quickly decapitated by cervical dislocation and brains were removed from the skull and placed into cold Krebs bicarbonate dissection buffer (pH 7.4), containing: KCl 2 mM, NaCl 120 mM, NaHCO 3 26 mM, CaCl 2 0.5 mM, KH 2 PO 4 1.18 mM, MgSO 4 10 mM, sucrose 200 mM, and glucose 11 mM. The hippocampus was placed in a McIlwain tissue chopper and was cut up in 300-µm-thick hippocampal slices. After that, slices were introduced into a chamber at 34 • C and exposed to 95% O 2 /5% CO 2 mixture for 45 min stabilization period before the beginning of the ischemic procedure.

Oxygen and Glucose Deprivation in Mouse Hippocampal Slices
In order to perform the ischemic protocol, one group of hippocampal slices were incubated in a Krebs solution named as control, composed by: NaCl 120 mM, NaHCO 3 26 mM, MgSO 4 1.19 mM, KCl 2 mM, KH 2 PO 4 1.18 mM, CaCl 2 2 mM, and glucose 11 mM. This mixture was bubbled with 95% O 2 /5% CO 2 . On the other hand, another pool of slices were subjected to oxygen-glucose deprivation (OGD) for 15 min. For this, slices were incubated in a glucose-free Krebs solution, where glucose was replaced by 2-deoxyglucose and equilibrated with a 95% N 2 /5% CO 2 gas mixture. Thereafter, the OGD solution was removed, and oxygenated Krebs solution with glucose (reoxygenation period) was added and maintained for 2 h. When CGP37157 (10 and 30 µM) was used, it was added during the OGD and Reox periods.

Glutamate Excitotoxicity in Mouse Hippocampal Slices
Apart from hypoxia, Ca 2+ overload-induced by excessive glutamate release is one of the most important molecular events that takes place during an ischemic insult. With the aim of mimicking this process in vitro, we subjected the slices to an excitotoxicity protocol well described in the literature [23]. First, the hippocampal slices were subjected to a stabilization period in a solution composed of MgSO 4 1.19 mM, CaCl 2 2 mM, NaCl 120 mM, KH 2 PO 4 1.18 mM, KCl 2 mM, glucose 11 mM and NaHCO 3 26 mM, with 95% O 2 /5% CO 2 . After that, we immersed the slices in a culture media containing: 50% of Dulbecco's Modified Eagle's Medium (DMEM), 50% of Krebs-Ringer Bicarbonate Buffer, 100 units/mL penicillin and 100 mg/mL streptomycin, HEPES 20 mM and 1 mM glutamate for 4 h in a CO 2 atmosphere at 37 • C. CGP37157 (10 µM) was applied as a CALHM1 antagonist and co-incubated for 4 h in the presence of glutamate.

Determination of Hippocampal Viability by the MTT Assay
Cell viability was measured by adding the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) to the hippocampal slices. This method is based on the ability of functional mitochondrial dehydrogenases to reduce the tetrazolium ring of MTT in an insoluble salt called formazan; being the amount of salt precipitate an indicator of cell viability. After incubation of hippocampal slices for 30 min in the presence of MTT (0.5 mg/mL) in Krebs solution at 37 • C, the formazan salt was solubilized with 200 µL of dimethyl sulfoxide (DMSO) for another 30 min, resulting in a purple-colored substrate. Finally, we measured the absorbance at 540 nm wavelength in an ELISA microplate reader. Absorbance values of control slices from wild-type mice were considered as 100% viability.

Quantification of Reactive Oxygen Species Production
To determine the production of reactive oxygen species (ROS) in hippocampal slices, we used the fluorescent dye 2',7-Dichlorofluorescein diacetate (H 2 DCFDA) purchased from Molecular Probes (Invitrogen, Madrid, Spain). This compound crosses the plasma membrane and is cleaved by cytosolic esterases from the non-fluorescent product dichlorodihydrofluorescein, which binds to intracellular H 2 O 2 generating dichlorofluorescein (DCF), a green fluorescent probe [24].
During the reoxygenation period, slices were loaded for 45 min with 10 µM H 2 DCFDA in Krebs solution and to stain nuclei, Hoechst33342 at 1 µg/mL was added for 5 min. Then, hippocampal slices were washed with the control Krebs solution. The fluorescence records were determined in a fluorescence inverted NIKON eclipse T2000-U microscope. The excitation and emission wavelengths were 485 and 520 nm, respectively and analysis was performed using the Metamorph program version 7.0 The area of interest was the CA1 region of the hippocampus, and images were obtained at magnifications of 10×. ROS production was calculated as the ratio of the mean DCFDA intensity divided by the mean Hoechst fluorescence. Data from control slices of wild-type were taken as 100%.

Western Blot Analysis
Slices of the different experimental groups were lysed in 100 µL cold lysis buffer (1% Nonidet P-40, 1 mM PMSF, 20 mM Tris-HCl pH 7.5, 10% glycerol, 20 mM NaF, 137 mM NaCl, 1 µg·mL−1 leupeptin, 1 mM Na 3 VO 4 and 1 mM sodium pyrophosphate) and disaggregated with Sonicator Ultrasound. The resulting tissue mixture was centrifuged at 13,000 rpm for 5 min at 4 • C, and the supernatant was collected. From each supernatant, 30 µg of protein were resolved by SDS-PAGE and transferred to Immobilon-P membranes (Millipore Corp.GE). Membranes were incubated with anti-HIF-1α at 1:500 (Cayman Chemical, Michigan, MI, USA). Seahorse peroxidase-conjugated secondary antibodies (1:10,000) were used to detect proteins by chemiluminescence. Protein bands density was analyzed using the ImageJ analysis software (National Institutes of Health, USA).

Statistical Analysis
Results are represented in bars as mean ± S.E.M. Statistical differences were determined by one-way ANOVA test followed by Newman Keuls post hoc and paired t-test. The program used for the analysis was GraphPad Prism 5.00. The threshold of statistical significance was established at p < 0.05.

Partial or Total Molecular Ablation of CALHM1 Results in a Protective Effect in Hippocampal Slices Exposed to OGD/Reox
Hippocampal slices from Calhm1 +/+ , Calhm1 +/− and Calhm1 −/− mice were subjected to the OGD/Reox protocol previously described. Considering the high Ca 2+ permeability of CALHM1, and the excessive release of this cation to the synaptic cleft during an ischemic insult, it is presumable that the lack of the channel exerts a beneficial effect against ischemic damage. The OGD/Reox protocol Cells 2020, 9, 664 5 of 13 was carried out and cell viability was determined by the MTT method in hippocampal slices from each group. Figure 1A, illustrates the cell viability of hippocampal slices from Calhm1 +/+ and Calhm1 +/− mice. Slices from animals with deletion of one allele of CALHM1 showed significantly higher viability (74.2% ± 6.2) compared to wild-type (58.4% ± 2.7) when subjected to OGD/Reox condition. Following the same experimental approach, we also observed that slices from animals with total ablation of Calhm1 −/− showed less toxicity in comparison to those from Calhm1 +/+ mice (60.4% ± 2.3 vs. 72.5% ± 4, respectively) ( Figure 1B). These results indicate that either partial or total genetic deletion of CALHM1 exerts a neuroprotective effect against OGD/Reox damage in our ex-vivo model of brain ischemia.

HIF-1α is Overexpressed under OGD/Reox Conditions when CALHM1 is Absent
To further understand the molecular mechanism by which the lack of CALHM1 induces neuroprotection upon OGD/Reox conditions, new experiments were carried out, but this time we searched for a Ca 2+ dependent protein such as HIF-1α. Under hypoxic conditions, HIF-1α accumulates and dimerizes with HIF-1β, and translocates into the nucleus where it binds to the hypoxic-response elements (HREs), inducing the expression of genes that have been linked with the adaptation to hypoxia, such as antioxidant genes, as well as genes involved in angiogenesis and glycolytic energy metabolism. This metabolic adaptation decreases the formation of ROS and mitochondrial toxicity [26]. As illustrated in Figure 4, Calhm1 −/− slices showed a reduction in the production of free radicals, in consequence, some antioxidant or pro-survival pathways could be activated. Due to its role as a master regulator in hypoxia, we decided to study HIF-1α levels in Calhm1 +/+ and Calhm1 −/− mice. As shown in Figure 5, HIF-1α expression was significantly higher in Calhm1 −/− hippocampal slices upon OGD/Reox compared to Calhm1 +/+ slices.  (A) Represents cell viability measured by MTT reduction in hippocampal slices from Calhm1 +/+ and Calhm1 +/− mice under basal conditions or exposed to the OGD/Reox protocol. OGD/Reox-induced cytotoxicity is prevented in Calhm1 +/− hippocampal slices. Data are the mean ± S.E.M of seven different animals; all variables were run in parallel. *** p < 0.001 and ** p < 0.01 in comparison to basal; # p < 0.05 respect to Calhm1 +/+ OGD. (B) Hippocampal slices from Calhm1 −/− mice subjected to OGD/Reox show higher cell viability when compared to those obtained from Calhm1 +/+ mice. Data represent the mean ± S.E.M of 7 different animals; all variables were run in parallel. *** p < 0.001 versus basal; # p < 0.05 respect to Calhm1 +/+ OGD. One-way ANOVA followed by Newman-Keuls was performed. Data were normalized to Calhm1 +/+ , which was considered as 100% of viability.

Genetic Ablation of CALHM1 Promotes Neuroprotection in the Glutamate Excitotoxicity Model
Due to the fact that the absence of CALHM1 proved to be neuroprotective in the OGD/Reox model, new experiments were performed, but this time focusing on the neurotransmitter glutamate, known to be cytotoxic during brain ischemia. As previously described, glutamate induces neurotoxic effects due to activation of NMDA and metabotropic receptors, leading to massive Ca 2+ entry and cell death.
As previously described in our laboratory [21], CGP37157 reduces Ca 2+ influx through CALHM1 in Hela cells. Hela cells lack L-type VDCCs, and when transiently transfected with CALHM1, the only Ca 2+ -permeant channel expressed in the plasma membrane is CALHM1. So, according to our previous results, CGP37157 blocks CALHM1 since in control Hela cells, the contribution of the Na + /Ca 2+ exchanger upon Ca 2+ add-back protocol seems to be insignificant compared with the Ca 2+ entry in those cells overexpressing CALHM1. Therefore, we carried out experiments to determine the Cells 2020, 9, 664 6 of 13 protective role of this compound in our models. The addition of the CGP37157 (10 µM) to the medium during the glutamate incubation period increased cell viability significantly in hippocampal slices from Calhm1 +/+ mice, but it did not provide significant protection in Calhm1 −/− slices. The increase in cell survival in Calhm1 +/+ slices may be related to the blocking action on NMDAR, VDCCs and sodium calcium exchanger (NCX) reported for CGP37157, preventing the Ca 2+ overload upon exposure to high glutamate concentrations [22]. This effect is higher in Calhm1 +/+ slices because CGP37157 is also blocking Ca 2+ influx through CALHM1, which is not present in Calhm1 −/− slices ( Figure 2). These findings highlight the idea that in hippocampal slices from Calhm1 −/− mice, CGP37157 is not exerting neuroprotection mediated by blockade of CALMH1 channels. Taken together, these results support that the genetic deletion or the pharmacological blockade of CALHM1 attenuates the massive release of glutamate and Ca 2+ overload to provide neuroprotection. However, the effects of CGP37157 on voltage dependent calcium channels (VDCCs) in CALHM1 mouse models have still not been reported, but further investigations are required to solve this crucial point. shows a schematic representation of the OGD/Reox protocol used, as described in materials and methods. (A) Represents cell viability measured by MTT reduction in hippocampal slices from Calhm1 +/+ and Calhm1 +/− mice under basal conditions or exposed to the OGD/Reox protocol. OGD/Reox-induced cytotoxicity is prevented in Calhm1 +/− hippocampal slices. Data are the mean ± S.E.M of seven different animals; all variables were run in parallel. *** p < 0.001 and ** p < 0.01 in comparison to basal; # p < 0.05 respect to Calhm1 +/+ OGD. (B) Hippocampal slices from Calhm1 −/− mice subjected to OGD/Reox show higher cell viability when compared to those obtained from Calhm1 +/+ mice. Data represent the mean ± S.E.M of 7 different animals; all variables were run in parallel. *** p < 0.001 versus basal; # p < 0.05 respect to Calhm1 +/+ OGD. One-way ANOVA followed by Newman-Keuls was performed. Data were normalized to Calhm1 +/+ , which was considered as 100% of viability. (1 mM, 4 h) in hippocampal slices was measured by the MTT method. As represented, slices from Calhm1 −/− mice showed significantly less cell damage than Calhm1 +/+ slices. Moreover, the neuroprotective profile described previously for CGP37157 (10 μM) in the glutamate model of ischemia was partially lost when CALHM1 was not present. Data correspond to the mean ± S.E.M of five different mice; all conditions were performed in parallel. *** p < 0.001 and ** p < 0.01 compared to basal; # p < 0.05 versus Calhm1 +/+ glutamate; $ p < 0.05 respect to Calhm1 +/+ glutamate. Paired T-Test was performed. Data were normalized to Calhm1 +/+ , considering it as 100% of cell viability. As represented, slices from Calhm1 −/− mice showed significantly less cell damage than Calhm1 +/+ slices. Moreover, the neuroprotective profile described previously for CGP37157 (10 µM) in the glutamate model of ischemia was partially lost when CALHM1 was not present. Data correspond to the mean ± S.E.M of five different mice; all conditions were performed in parallel. *** p < 0.001 and ** p < 0.01 compared to basal; # p < 0.05 versus Calhm1 +/+ glutamate; $ p < 0.05 respect to Calhm1 +/+ glutamate. Paired T-Test was performed. Data were normalized to Calhm1 +/+ , considering it as 100% of cell viability.

CGP37157 Requires CALHM1 to Exert Neuroprotection under OGD/Reox Conditions
As mentioned above, previous findings from our laboratory [21] demonstrated that CGP37157 modulated massive Ca 2+ entry through CALHM1 by blocking it. Due to the properties of CGP37157, new experiments were performed in Calhm1 +/+ , Calhm1 +/− , and Calhm1 −/− mice hippocampal slices exposed to OGD/Reox in the presence or absence of CGP37157 at 10 µM. The OGD/Reox protocol induced 51.9% ± 4.3 cell death in Calhm1 +/+ hippocampal slices, while the incubation of CGP37157 (10 µM) reverted the cytotoxic effect to 76.3% ± 5.3 ( Figure 3A). Incubation with 10 µM of CGP37157 of Calhm1 +/− or Calhm1 −/− hippocampal slices per se did not induce an observable neuroprotective effect. Despite not observing a significant effect, cell viability was increased in the presence of CGP37157 Cells 2020, 9, 664 7 of 13 (10 µM) compared to non-treated hippocampal slices of Calhm1 +/− and Calhm1 −/− mice (Figs. 3B and 3C). This increase in cell survival could be attributed to the blocking action of CGP37157 on NMDAR, VDCCs, and NCX, preventing the Ca 2+ overload upon exposure to ischemic conditions [22]. Moreover, CGP37157 (30 µM) was neuroprotective in the OGD/Reox protocol in Calhm1 +/+ slices ( Figure 3D), but not in Calhm1 −/− mice ( Figure 3E). These results indicate that the pharmacological modulation of CALHM1 could be a novel neuroprotective mechanism under ischemic brain conditions.  all variables were run in parallel. *** p < 0.001, ** p < 0.01 and * p < 0.05 compared to basal; ## p < 0.01 and # p < 0.05 respect to OGD. One-way ANOVA followed by Newman-Keuls was performed. Data were normalized to the basal condition which was considered as 100% of viability.

The Production of Free Radicals Induced by OGD/Reox is Prevented when CALHM1 Is Absent
It is well described that under hypoxic conditions, such as ischemia, there is intracellular Ca 2+ overload and ROS production. Those pathological conditions induce overactivation of astrocytes and microglia. These two cell types in brain parenchyma under pathological situations lead to an increase in the production and release of ROS and inflammatory mediators, worsening neuronal damage [25]. Once demonstrated that CALHM1 ablation has a neuroprotective profile, we decided to go further by measuring ROS production in hippocampal slices from each experimental group. Figure 4A illustrates the production of ROS in the CA1 region of hippocampal slices of both, Calhm1 +/+ and Calhm1 −/− mice measured by the fluorescent probe DCFDA (green). In basal conditions Calhm1 +/+ and Calhm1 −/− showed similar ROS release as represented in the Figure 4B; however, after exposure to OGD/Reox, Cells 2020, 9, 664 8 of 13 slices from Calhm1 +/+ mice showed significantly higher levels of free radical production in comparison with those from Calhm1 −/− (178.1% ± 17.01 vs. 132.6% ± 1.97, respectively). These findings support the idea that the absence of CALHM1 not only decreases [Ca2+ ] c entry but, in addition, contributes to lowering ROS production, opening a new neuroprotective pathway for CALHM1. Figure 3. CALHM1 is required for the protective effect of CGP37157 in the model of oxygen and glucose deprivation followed by reoxygenation (OGD/Reox). Representation of the OGD/Reox protocol used on the top of figures A-E. (A-C) Cell viability in hippocampal slices of Calhm1 +/+ , Calhm1 +/− and Calhm1 −/− mice was determined through the MTT method. Treatment with CGP37157 (10 μM) during the OGD/Reox protocol only provided neuroprotection in Calhm1 +/+ slices (A), while this effect was not observed in Calhm1 +/− (B) or Calhm1 −/− (C) genotypes. (D-E) CGP37157 (30 μM) only exerted benefits in Calhm1 +/+ slices (D), without having an effect on Calhm1 −/− samples (E). Data represent the mean ± S.E.M of 3-5 different animals; all variables were run in parallel. *** p < 0.001, ** p < 0.01 and * p < 0.05 compared to basal; ## p < 0.01 and # p < 0.05 respect to OGD. One-way ANOVA followed by Newman-Keuls was performed. Data were normalized to the basal condition which was considered as 100% of viability.  group; all variables were performed in parallel. ** p < 0.01 versus basal Calhm1 +/+ ; ## p < 0.01 compared to OGD Calhm1 +/+ . One-way ANOVA followed by Newman-Keuls was made. Data were normalized to Calhm1 +/+ , which was considered as 100% of viability.

HIF-1α is Overexpressed under OGD/Reox Conditions when CALHM1 is Absent
To further understand the molecular mechanism by which the lack of CALHM1 induces neuroprotection upon OGD/Reox conditions, new experiments were carried out, but this time we searched for a Ca 2+ dependent protein such as HIF-1α. Under hypoxic conditions, HIF-1α accumulates and dimerizes with HIF-1β, and translocates into the nucleus where it binds to the hypoxic-response elements (HREs), inducing the expression of genes that have been linked with the adaptation to hypoxia, such as antioxidant genes, as well as genes involved in angiogenesis and glycolytic energy metabolism. This metabolic adaptation decreases the formation of ROS and mitochondrial toxicity [26]. As illustrated in Figure 4, Calhm1 −/− slices showed a reduction in the production of free radicals, in consequence, some antioxidant or pro-survival pathways could be activated. Due to its role as a master regulator in hypoxia, we decided to study HIF-1α levels in Calhm1 +/+ and Calhm1 −/− mice. As shown in Figure 5, HIF-1α expression was significantly higher in Calhm1 −/− hippocampal slices upon OGD/Reox compared to Calhm1 +/+ slices. Cells 2020, 9, 664 9 of 13 p < 0.01 versus basal Calhm1 +/+ ; ## p < 0.01 compared to OGD Calhm1 +/+ . One-way ANOVA followed by Newman-Keuls was made. Data were normalized to Calhm1 +/+ , which was considered as 100% of viability.

Discussion
In the present work, we have shown that the partial or total absence of CALHM1 results neuroprotective against neuronal death in an ex vivo model of brain ischemia, mice hippocampal slices subjected to OGD/reox. We hypothesized that the Ca 2+ channel CALHM1 could worsen cell damage upon ischemia-reperfusion due to its high Ca 2+ permeability [17]. Recently, it has been established that the CALHM1 channel is implicated in ischemia brain damage [20]. These previous findings were focused on Calhm1 −/− mice without explaining the effects of hypoxia in Calhm1 +/− mice. In cortical neurons transfected with an shRNA-CALHM1 [20] showed a 13% reduction in lactate dehydrogenase release, indicating a mild contribution of CALHM1 to brain-ischemia neurotoxicity. Considering that presumably, CALHM1 is exerting its cytotoxic effect during the reperfusion phase of brain ischemia, due to its activation properties (Ca 2+ add-back protocol), we hypothesize that the absence of one allele of CALHM1 should be sufficient to reduce a large amount of Ca 2+ influx into the neurons during the reoxygenation period, exerting a similar neuroprotective effect to the total ablation of CALHM1 gene. Also, during a hypoxic insult in the brain, the depolarization wave and the excessive amount of glutamate released to the synaptic cleft, activates VDCCs and NMDAR, respectively, being the effects of these two types of channels predominant on CALHM1 participation [7][8][9][10]. As an example, heterozygous mice of the aryl hydrocarbon receptor (AhR +/− mice) show a significant reduction in infarct volume as well as an improvement of neurological deficits after a middle cerebral artery occlusion (MCAO) in comparison to wild-type mice [27]. These results indicate that the deletion of only one allele of AhR results in neuroprotectivity against brain ischemia, similar to our observations with Calhm1 +/− mice.
In view of the physiological role of CALHM1 in the regulation of AMPA and NMDA receptor expression and its implication in memory consolidation in the hippocampus [19], as well as, the participation of CALHM1 in the maintenance of neuronal excitability [17], we decided to perform our experiments in brain hippocampal slices from Calhm1 +/+ , Calhm1 +/− and Calhm1 −/− mice. In this study, we have performed experiments with Calhm1 +/− mice, showing that the deletion of just one allele of CALHM1 results neuroprotective in the OGD/Reox protocol. In addition, we have seen that

Discussion
In the present work, we have shown that the partial or total absence of CALHM1 results neuroprotective against neuronal death in an ex vivo model of brain ischemia, mice hippocampal slices subjected to OGD/reox. We hypothesized that the Ca 2+ channel CALHM1 could worsen cell damage upon ischemia-reperfusion due to its high Ca 2+ permeability [17]. Recently, it has been established that the CALHM1 channel is implicated in ischemia brain damage [20]. These previous findings were focused on Calhm1 −/− mice without explaining the effects of hypoxia in Calhm1 +/− mice. In cortical neurons transfected with an shRNA-CALHM1 [20] showed a 13% reduction in lactate dehydrogenase release, indicating a mild contribution of CALHM1 to brain-ischemia neurotoxicity. Considering that presumably, CALHM1 is exerting its cytotoxic effect during the reperfusion phase of brain ischemia, due to its activation properties (Ca 2+ add-back protocol), we hypothesize that the absence of one allele of CALHM1 should be sufficient to reduce a large amount of Ca 2+ influx into the neurons during the reoxygenation period, exerting a similar neuroprotective effect to the total ablation of CALHM1 gene. Also, during a hypoxic insult in the brain, the depolarization wave and the excessive amount of glutamate released to the synaptic cleft, activates VDCCs and NMDAR, respectively, being the effects of these two types of channels predominant on CALHM1 participation [7][8][9][10]. As an example, heterozygous mice of the aryl hydrocarbon receptor (AhR +/− mice) show a significant reduction in infarct volume as well as an improvement of neurological deficits after a middle cerebral artery occlusion (MCAO) in comparison to wild-type mice [27]. These results indicate that the deletion of only one allele of AhR results in neuroprotectivity against brain ischemia, similar to our observations with Calhm1 +/− mice.
In view of the physiological role of CALHM1 in the regulation of AMPA and NMDA receptor expression and its implication in memory consolidation in the hippocampus [19], as well as, the participation of CALHM1 in the maintenance of neuronal excitability [17], we decided to perform our experiments in brain hippocampal slices from Calhm1 +/+ , Calhm1 +/− and Calhm1 −/− mice. In this study, we have performed experiments with Calhm1 +/− mice, showing that the deletion of just one allele of CALHM1 results neuroprotective in the OGD/Reox protocol. In addition, we have seen that pharmacological treatment with CGP37157 resulted neuroprotective upon OGD/Reox conditions in Calhm1 +/+ mice, whereas this effect was reduced in Calhm1 +/− or Calhm1 −/− hippocampal slices; however, protection elicited by CGP37157 in Calhm1 +/+ mice did not reach basal condition values. These finding could be related to the reported lack of protective effects of CGP37157 against oxidative stress cell damage; for example, in SHSY5Y neuroblastoma cells subjected to oxidative stress stimulus such exposure to Oligomycin/Rotenone, CGP37157 was not neuroprotective [22]. OGD/Reox increases oxidative stress, therefore, under these conditions, it would be expected that CGP37157 would not be able to provide significant neuroprotection. These results indicate that the expression of CALHM1 seems to be necessary to induce pharmacological neuroprotection with CGP37157 ( Figure 3). As mentioned earlier [21], it has been reported that CGP37157 can block Ca 2+ flux through the mitochondrial sodium/calcium exchanger (mNCX), L-type VDCCs, the plasmalemmal sodium/calcium exchanger (pNCX), and voltage-gated sodium channels, being more selective at mNCX [28]. However, there is currently no experimental evidence in the literature that describes any alteration in the expression or functionality of mNCX, voltage-gated sodium channels, and L-type VDCC when CALHM1 is absent, so it should not affect the expression or function of these channels. CALHM1 controls [Ca 2+ ] c [15][16][17], mitochondrial [Ca 2+ ] [29] and ER [Ca 2+ ] [30]. Now we have more evidence that the opening of CALHM1 upon ischemia in the OGD/Reox model or glutamate excitotoxicity, induces large amounts of Ca 2+ entry and plasma membrane depolarization [15], but its pharmacological blockade with CGP37157 reduced cell death in Calhm1 +/+ , but not in Calhm1 +/− and Calhm1 −/− mice. In other words, in hippocampal slices from Calhm1 +/+ mice, OGD/Reox-induced cell death can be pharmacologically prevented by CGP37157 by a mechanism that implicates inhibition of CALMH1 channels. These results demonstrate that the Ca 2+ driven through CALHM1 plays a neurotoxic effect and its blockade results neuroprotective. To further elucidate the neuroprotective properties of CALHM1, we performed experiments with Calhm1 +/− mice. Upon OGD/Reox, the partial lack of CALHM1 exerts neuroprotective properties, pointing out that the possibility of only the deletion of one allele of CALHM1 is enough to induce neuroprotection against brain ischemia. However, in the model of glutamate toxicity, hippocampal slices from Calhm1 +/− genotype did not show the cytoprotective effect present in those from Calhm1 −/− mice (data not shown). To go further, we decided to investigate the molecular pathway by which the absence of CALHM1 resulted in neuroprotection. Mitochondria efficiently buffer Ca 2+ driven through CALHM1 channels [29]. Ca 2+ entering through the mitochondrial uniporter is translated to more hyperpolarized mitochondrial membrane potential and to more ATP production, but much more ROS production [31]. In fact, this was the case in our experimental conditions; Calhm1 +/+ hippocampal slices exposed to OGD/Reox released higher amount of ROS than the Calhm1 −/− ones. These findings highlight that the deletion of CALHM1 results neuroprotective not only to the reduction of Ca 2+ entry driven by the channel but also to lower ROS production. Focusing on our results, CALHM1 absence results in high cell viability upon exposure of hippocampal slices to ischemia-related protocols and less ROS production in the CA1 hippocampal region.
Another key marker activated after exposure to OGD/Reox is HIF-1α. This protein is inducible upon hypoxia and cytosolic Ca 2+ elevation, triggering cell survival under low oxygen concentrations [32]. We found a strong elevation of HIF-1α in the hippocampus of Calhm1 −/− mice, subjected to OGD/Reox, compared to Calhm1 +/+ . An increase in HIF-1α levels might be beneficial because it could further induce antioxidant protein expression [33]. The finding that HIF-1α is upregulated in Calhm1 −/− samples is conflicting. However, intracellular Ca 2+ increase is not the only molecular mechanism that induces HIF-1α expression and stabilization. HIF-1α stabilization in the cytosol promotes metabolic adaptation to hypoxia, proliferation, survival, and resistance to apoptosis, although the lack of CALHM1 leads to less Ca 2+ entry into the cell, it is no surprising that the expression levels of HIF-1α protein are increased in knock-out mice. Also, HIF-1α regulates the expression of some Ca 2+ permeant channels, such as the transient receptor potential (TRP) C1 (TRPC1), stromal interaction molecule-1 (STIM1), or the Golgi Ca 2+ pump secretory pathway Ca 2+ -ATPase 2 (SPCA2). Based on the latter observations, we can hypothesize that the deletion of CALHM1 could trigger an elevation in HIF-1α levels as a compensatory mechanism in order to restore Ca 2+ signaling through other Ca 2+ channels [34] Although different CALHM1 antagonists have been described in the literature such as Ni 2+ and Co 2+ , their high toxicity and non-specificity make them inadequate for use in the clinic. Something similar occurs with 2-APB and Ruthenium Red [17], they block CALMH1 channels but at high concentrations that could also be toxic. The organic properties of CGP37157 and its blocking action on CALMH1 channels at the micromolar range, contrary to other CALMH1 blockers previously described [17], indicate that it could be a druggable compound for stroke.

Conclusions
Genetic or pharmacological inhibition of CALHM1 in hippocampal slices from mice, was neuroprotective in the ODG/Reox protocol, suggesting that Ca 2+ entry through CALHM1 contributes to cell death upon ischemic conditions. The neuroprotective molecular pathway results from lowering ROS production and activating of HIF-1α. Taken together, these results open a novel neuroprotective pathway for ischemic stroke, which includes the inhibition of CALMH1.