Permissive hypercapnia and hypercapnic hypoxia inhibit signaling pathways of neuronal apoptosis in ischemic/hypoxic rats

In the present study, we aimed to test the hypothesis that hypercapnia, independently and/or in combination with hypoxia, can activate signaling pathways related to the inhibition of proapoptotic (caspase-dependent and caspase-independent) factors and the induction of antiapoptotic factors in facilitating adaptation to hypoxia/ischemia. Following exposure to permissive hypercapnia and/or normobaric hypoxia, the degree of apoptosis was evaluated in experimental ischemia models in vivo and in vitro. The percentages of caspase-3, apoptosis-inducing factor (AIF), Bax, and Bcl-2 in astrocytes and neurons derived from male Wistar rats were also calculated. In vitro, cells were subjected to various types of respiratory exposure (hypoxia and/or hypercapnia for 24 or 12 h) as well as further sublethal chemical hypoxia. The percentages of these molecules in nerve cells in the ischemic penumbra of the brain after photothrombotic injury were also calculated. The degree of apoptosis was found to decrease in ischemic penumbra, mostly due to the hypercapnic component. It was also discovered that the levels of caspase-3, AIF, and Bax decreased in this region, whereas the Bcl-2 levels increased following exposure to hypercapnia and hypercapnic hypoxia. This integrative assessment of the rate of apoptosis/necrosis in astrocyte and neuron cultures shows that the combination of hypercapnia and hypoxia resulted in the maximum neuroprotective effect. The levels of apoptosis mediators in astrocyte and neuron cultures were calculated after modeling chemical hypoxia in vitro. These results show that the exposure models where permissive hypercapnia and normobaric hypoxia were combined also had the most pronounced inhibitory effects on apoptotic signaling pathways.


Introduction
Intermittent hypoxia is an effective way to improve the ischemic tolerance of organs and tissues [1][2][3][4]. There is evidence that hypoxia is an efficient way to prevent and treat cardiovascular pathologies [5,6] and ischemic brain injuries [7]. However, the application of intermittent hypoxia in practice is limited due to the requirements for prolonged exposure (1-6 h) and lengthy treatment courses (at least seven sessions) [2,5].
We previously demonstrated that permissive hypercapnia and hypercapnic hypoxia, compared with intermittent hypoxia, improve body and brain tolerance to acute ischemia more intensively [8]. However, research on the mechanisms and signaling pathways underlying the neuroprotective efficacy of hypercapnia and its combination with hypoxia is lacking.
One of the most important mechanisms by which brain tolerance to ischemia increases is through the inhibition of apoptosis via caspase-dependent and caspase-independent pathways [9][10][11]. This protective mechanism prevents the death of partially damaged neurons during reperfusion. In their research, Cantagrel et al. [12] showed a decrease in the number of apoptotic cells in a hypoxically preconditioned brain at 24 and 48 h after the test stroke. It was also demonstrated that the neuroprotective effect of hypoxic preconditioning occurs through a shift in the ratio of Bax to Bcl-2-Bcl-xL in favor of the antiapoptotic proteins Bcl-2 and Bcl-xL [13]. It is known that the HSP70 chaperone, also activated in response to hypoxia, inhibits apoptosis by inducing the transcription factors PKR/NFкB [14]. Moreover, HSP70 inhibits reactions that contribute to an increase in mitochondrial membrane permeability and cytochrome C release, blocking Bax and increasing the expression of the antiapoptotic Bcl-2 factor [15]. In addition, apoptosisinducing factor (AIF), the principal effector of the caspaseindependent pathway, is neutralized [16]. The activation of mitoK + ATP channels prevents cytochrome C activation and, consequently, blocks the caspase-dependent apoptotic pathway [17]. It has been shown that inhibition of caspase-3 activity has a neuroprotective effect through blocking Bax in the Bcl-2/Bax signaling pathway [18].
There are data on the mechanism of apoptosis inhibition under the influence of permissive hypercapnia during reperfusion following transitory brain ischemia [19,20]. It was demonstrated that inhalation of CO 2 at moderate concentrations (Paco 2 (partial pressure of carbon dioxide) = 60-100 mm Hg) inhibits the active form of the principal effector, caspase-3, reduces the cytosolic levels of cytochrome C and proapoptotic protein Bax, and increases the concentration of the antiapoptotic protein Bcl-2 in mitochondria.
Previously, we carried out comparative assessment of the degree of apoptosis in cells of the peri-infarct region of the brain during exposure to isolated and combined hypoxia and hypercapnia prior to focal ischemic stroke [21]. This showed that maximum inhibition of the apoptotic mechanisms of cell death was achieved via combined exposure to hypoxia and hypercapnia. Exposure to hypercapnia or hypoxia alone also inhibited apoptosis, although to a lesser extent, without any significant difference between the two. Moreover, we determined that permissive hypercapnia potentiates the effect of hypoxia while increasing HSP70 expression [22] and activating mitoK + ATP channels [23]. The data provided above demonstrate the importance of inhibiting programmed cell death as a neuroprotective mechanism as well as the significant effects of both hypoxia and hypercapnia on the mentioned process. We assume that hypercapnia and/or hypoxia can facilitate ischemic/hypoxic adaptation via the activation of signaling pathways related to the inhibition of proapoptotic (caspase-dependent and caspase-independent) factors and the induction of antiapoptotic factors, as carbon dioxide impacts the adaptation systems, which in turn directly impact these factors.
The investigation of the signaling pathways involved in the inhibition of apoptosis under the intermittent effects of permissive hypercapnia and hypercapnic hypoxia may provide a theoretical basis for understanding the pathogenesis of ischemic injury to nervous tissue and the molecular mechanisms underlying the ischemic tolerance of the brain.

Animals
The research was conducted on 40 male Wistar rats (Institute of Cytology and Genetics of SB RAS, Novosibirsk, Russia), with an average weight of 250 to 300 g, using in vivo and in vitro methodologies. Animal testing was approved by the bioethics commission of the local ethics committee of Krasnoyarsk State Medical University and was carried out in conformity with the principles of the European Convention for the Protection of Vertebrate Animals. The animals were randomized using SPSS 11.5 (SPSS Inc, USA). The rats were kept in cages at a controlled room temperature (~ 23 °C) and under natural lighting, and had free access to food and water. The animals were weighed before and after the tests.

Groups and experimental design
The research consisted of three series of experiments (Fig. 1). The first series was conducted on an in vivo model and included four animal groups, which differed in terms of the partial pressure of oxygen (Po 2 ) and partial pressure of carbon dioxide (Pco 2 ): In the first series of experiments (n = 40), the rats in vivo were subjected to respiratory actions in a special chamber for 30 min per day for 15 days. The following day, after the course's completion, a focal cerebral ischemia was simulated in all the animals by photochemical thrombosis [24,25]. Seventy-two hours after the operation, we carried out perfusion fixation, followed by decapitation and brain extraction. The level of apoptosis was evaluated by the ISOL method for a fraction of the cerebral section of each animal from the first series; for another part of the section, immunohistochemical assessment and Western Blot analysis of the modulators of apoptosis (caspase-3, AIF, Bax, and Bcl-2) was carried out. In the second and third series of experiments in vitro, we used astrocytes and neurons, differentiated from primary tissue cultures of progenitor cells of the central nervous system and a primary culture of rat neurospheres [26]. After differentiation, the astrocytes and neurons were kept in a special hypoxic chamber under the conditions of normobaric hypoxia, permissive hypercapnia, and a combination of hypoxia and hypercapnia for 24 or 12 h or with 12-h intervals of hypoxia and reoxygenation, depending on the group assignment. Six hours after exposure, a part of each cell culture from the second series was subjected to chemical hypoxia. After chemical hypoxia, the cell cultures from the second series were subjected to immunocytochemical analysis of the levels of apoptotic and necrosed cells, whereas the cell cultures from the third series were subjected to immunocytochemical evaluation of the levels of the modulators of apoptosis (caspase-3, AIF, Bax, and Bcl-2). The cells that were kept under standard conditions in the CO 2 incubator were used as a reference for those series: temperature, 37 °C; relative humidity, 80%; Po 2 ≈150 mm Hg; and Pco 2 ≈35 mm Hg; the rest-N 2 .

Respiratory action delivery
The flow chamber described previously was used for the respiratory actions [23]. The test groups of rats inhaled a gas mixture whose composition depended on the group assignment. The control group was placed into the chamber under similar conditions; however, instead of the gas mixture, the compressor delivered atmospheric air. The gas mixture in the chamber was controlled by a gas analyzer (Microlux O 2 + CO 2 , Microlux Ltd, Russia).

Surgical manipulations and photochemical thrombosis
Before the focal ischemia simulation, all the animals were deprived of food but had free access to water. Each rat received anesthesia by the intraperitoneal injection of chloral hydrate (400 mg/kg). A sterile incision was performed in the left groin. A solution of 4% rose Bengal dye (Sigma-Aldrich, Germany) in 0.9% NaCl was injected into the left femoral vein at a dose of 40 mg/kg. Ischemic cortical injury was carried out by the method of transcranial photochemical thrombosis [24,25]. A 20 mW green laser light (532 nm) was applied to scalped bones of the skull for 10 min. The light was applied to an area of the right parietal bone, 2 mm in diameter, positioned midway between the bregma and lambdoid suture and 2 mm lateral of the sagittal suture.
Before decapitation, the animals were subjected to transcardial perfusion fixation (stream infusion of 500 mL of phosphate-buffered saline (PBS) followed by 250 mL of 4% PBS-buffered paraformaldehyde). After the perfusion fixation, the brain was extracted and kept in a 4% paraformaldehyde solution in PBS for 2 days and then in a 30% sucrose solution for a further 2 days. After the brain was dissected using a vibratome (into 60 µm-thick frontal slices), the sectioned slices were individually placed into 30% sucrose solution in 24-well plates for safekeeping and analysis.

Obtaining the primary culture of neurospheres and the primary cultures of neurons and astrocytes
The hippocampi were separated from the brains of intact rats (Wistar males, 5-6 months old) in 2% glucose solution in PBS and reduced to fragments. The obtained fragments were placed into a fresh 2% glucose solution in PBS, followed by the removal of the supernatant after passive deposition. Following the deposition, the tissues were resuspended in 1 mL of medium from the NeuroCult Proliferation Kit (Mouse & Rat; StemCell, Canada) and triturated until a homogeneous cell suspension was obtained. After this, the supernatant obtained after deposition was centrifuged at 150 g for 5 min.
The progenitor cells extracted from the brain, with a density of 6 to 12 mln viable cells/mL, were seeded into T25 culture flasks with 10 mL of medium from the Neuro-Cult Proliferation Kit (Mouse & Rat; StemCell, Canada). Later, the cells were incubated in a CO 2 incubator at 5% CO 2 and a temperature of 37 °C.

Differentiation of neurospheres
After 24 h of incubation, free-floating neurospheres started to form (clusters of cells that proliferated to form spheroids). Neurospheres were used for differentiation to astrocytes 4 to 5 days after cultivation.
To differentiate the neurospheres, we added differentiating factors to the culture medium. We used commercial medium from the NeuroCult Proliferation Kit (Mouse & Rat; StemCell, Canada) as a growth medium. Seven days later, we carried out the immunocytochemical assessment of the purity of the obtained cultures. Fluorescent labeling was applied to astrocytes (glial fibrillary acidic protein) and neurons (neuron-specific enolase) according to the standard protocols for indirect immunocytochemistry.
Upon the completion of the differentiation of neurospheres and after accumulating the necessary cell mass, astrocytes and neurons were seeded in 24-well plates for further experiments. 100 ml of cell suspension (neurons or astrocytes) and 100 ml of culture medium were added to the wells of the plate. The total cell content is 3х10 4 in each well.

Modeling chemical hypoxia in vitro
Astrocytes and neurons were incubated in a CO 2 incubator with sodium iodoacetate for 30 min at 37 °C. The concentration of sodium iodoacetate in the medium was 50 µM. After the incubation, the cells were washed, and the culture medium was changed entirely; after this, the cells were further cultivated in a CO 2 incubator under standard conditions for 24 h until the start of the immunocytochemical analysis.

Modeling hypoxia and hypercapnia in vitro
For progenitor-cell-derived astrocyte and neuron cultures, hypoxic and hypercapnic-hypoxic exposure was simulated in a special chamber, HypoxyLab (Oxford Optronix Ltd, Oxford, UK), which automatically maintained the preset atmosphere composition (Po 2 ≈ 35 mm Hg for hypoxia; Pco 2 ≈ 50 mm Hg for hypercapnia), a temperature of 37 °C and a humidity of 80% for 12 or 24 h, or in two to three 12-h intervals of hypoxia and hypercapnia with alternate 12-h periods of normoxia.

Immunohistochemical analysis
Immunohistochemical analysis was performed according to the protocol using the free-floating section method. Three hundred microliters of blocking solution with bovine serum albumin (BSA) was added to the wells for 60 min, with shaking and two further washes (500 µL of wash solution + Triton-X100 × 2). Then, the solution of primary antibodies was applied to the sections (300 µL), which were further transferred to the incubator shaker and left overnight at a temperature of + 4 °C. After this, the sections were rinsed three times, the solution of secondary antibodies was applied, and the sections were transferred to the shaker and incubated in the dark for 2 h at room temperature.
Following incubation, the sections were rinsed three times and transferred to object plates; after this, DAPI (4′,6-diamidino-2-phenylindole) Fluoromount was added, and a coverslip was applied. The number of immunopositive cells expressing target antigens in the peri-infarct zone was calculated in at least five fields of view under a confocal microscope, FV10i-W (Olympus, Japan). The microphotographs were processed with the program ImageJ 1.41 (Scion Inc, USA).

Immunocytochemical analysis
Immunocytochemical assessment of the levels of caspase-3, AIF, Bax, and Bcl-2 in astrocyte and neuron cultures was conducted by indirect fluorescence double staining. Prior to the assessment, the cells were subjected to prefixation using 2% paraformaldehyde solution in the culture medium for 5 min. The following cell fixation was carried out using 4% paraformaldehyde solution for 15 min.
After fixation, the cells were permeabilized with 0.1% Triton X-100 (10 min at room temperature), and the nonspecific antigens were blocked with a solution of 10% BSA in PBS (30 min at 37 °C). The bovine serum albumin was then removed, and the cells were washed with PBS three times; then, a solution of primary antibodies was added to the astrocytes for 2 h at 37 °C. After the primary antibodies were removed and the cells were washed with PBS three times, a solution of secondary antibodies was added to the astrocytes for 1 h at 37 °C. The nuclei were counterstained with DAPI.
The primary and secondary antibodies used in this assessment were similar to those used for immunohistochemical assessment at similar dilutions (Section "Immunohistochemical analysis").
Fluorescent microscopy was carried out using the ZOE Cell Imaging System (Bio-Rad, USA), counting the number of immunopositive cells containing target antigens relative to the total number of cells. The microphotographs were processed with the program ImageJ 1.41 (Scion Inc, USA).

The assessment of the intensity of apoptosis in the cerebral ischemic penumbra in rats
A fraction of each section in the first series of experiments was analyzed according to the double-fluorescent labeling protocol for the assessment of apoptosis using the ApopTag® ISOL Dual Fluorescence Apoptosis Kit (Sigma-Aldrich, cat. no. APT1000, Germany), following the manufacturer's instruction manual for the focal microscope FV10i-W (Olympus, Japan). The method is based on the use of an oligonucleotide probe containing two labels: carboxyfluorescein (FAM) and Cal Fluor Red 590 (CR590). The probe labeled with CR590 detects DNA breaks induced by DNase I, whereas the probe labeled with FAM detects DNA breaks induced by DNase II. The chosen method is more selective than conventional ones, such as ISEL (the Klenow Fragment), TUNEL (terminal deoxynucleotidyl transferase, TdT), and ISNT (DNA polymerase I), and allows the precise differentiation of apoptotic cells from reversibly damaged ones by the presence of both labels in the nucleus and cytoplasm.

Assessment of the intensity of apoptosis and necrosis in astrocyte and neuron cultures
Upon the completion of the courses of exposure to hypoxia and/or hypercapnia in vitro, we assessed the intensity of apoptosis and necrosis using an Apoptosis/Necrosis Detection Kit (blue, green, and red) (Abcam, cat. no. ab176749, USA), following the manufacturer's instruction manual. Fluorescent microscopy was carried out using the ZOE Cell Imaging System (Bio-Rad, USA), counting the number of immunopositive cells containing target antigens relative to the total number of cells. The count was conducted in at least five fields of view containing 1000 cells in a sample. Microphotographs were processed with the program ImageJ 1.41 (Scion Inc, USA).

Western blot
After isolation of the cortex and selective extraction of cytoplasmic and nuclear proteins using the Mammalian Nuclear and Cytoplasmic Protein Extraction Kit (SERVA, Cat. No. 39243.01), the nuclear protein extract was stored until the day of the analysis at − 80 °C. Prior to electrophoresis, the samples were thawed at room temperature and equivalently diluted with Tris/Glycine/SDS with addition of mercaptoethanol, and then heated for 5 min at 95 °C on a solid-state thermostat.
Protein electrophoresis was carried out in a commercially available 8% TG PRiME acrylamide gel (SERVA, Cat. No. 43260.01). To measure the total protein content in the samples and determine the volume for application, the Lowry micromethod was used on a CM2203 spectrofluorometer (Solar, Belarus). The electrophoresis procedure was carried out in a miniVE Amersham Biosinces chamber (GE Healthcare, USA) for 90 min at 20 mA and 230 V. To transfer proteins from the gel to the PVDF membranes, a set of Semi-Dry Blotting Buffers was used (SERVA, Cat. No. 42559.01). Semi-dry blotting was carried out on a TE77 Amersham system (GE Healthcare, USA) at 65 mA for 60 min.

Statistical analysis
The total sample size and the sample size of each group was calculated based on the results of our previous studies on a similar model using the quantitative scale method [27]. Statistical analysis was performed using SPSS 11.5 software (SPSS Inc, Chicago, IL). The Shapiro-Wilk test was used for normality testing.
Comparisons of normally distributed data were performed using Student's t-test. Comparisons of non-normally distributed data were performed using non-parametric methods. Variables in independent groups were compared using the Kruskal-Wallis one-way analysis.
Normally distributed quantitative data are presented as mean (M) ± standard deviation (σ); non-normally distributed data are shown in the form of the median (Me), the lower quartile (25%) and the upper quartile (75%). The differences were considered significant if the level of significance (p value) was less than 0.05. Data are presented as the median ± 25/75 percentiles. The quantitative data are presented as the median (Me), the lower quartile (25%), and the upper quartile (75%). The differences for which the error probability (P) was less than 0.05 were considered statistically significant. The values are presented as medians ± 25th/75th percentiles.

Apoptosis in the ischemic penumbra in the cortex of rats
Following preliminary respiratory exposure to intermittent hypercapnia and/or hypoxia, the level of apoptosis in rats in the ischemic penumbra of the brain mainly decreases under the influence of the hypercapnic component (Fig. 2). This is proven by the lower level of apoptotic nerve cells in the group of permissive hypercapnia as compared with the control group (five times lower); at the same time, there was no significant difference between this index and the reference figures in the group of normobaric hypoxia. Given this, the combined effect of hypercapnia and hypoxia shows more pronounced efficacy in inhibiting apoptosis: 20.8 times more as compared with the control group and 5.8 and 4.3 times more as compared with the groups of normobaric hypoxia and permissive hypercapnia, respectively.

Levels of apoptosis mediators in nerve cells of the ischemic penumbra in the cerebral cortex of the rat
The results for immunohistochemical analysis of apoptosisrelated factors in ischemic penumbra tissues of the rat cortex following preliminary exposure to intermittent hypercapnia and/or hypoxia show that all respiratory actions inhibit the principal effector of apoptosis, caspase-3, but to varying extents (Fig. 3A). At the same time, the caspase-independent (mitochondrial) signaling pathway of apoptosis, whose principal effector is AIF, is mainly inhibited following preexposure to permissive hypercapnia with or without hypoxia. Hypercapnic exposure also affects the Bax inductor of the caspase-independent pathway of apoptosis. However, the concentration of the antiapoptotic factor Bcl-2 in cells of the ischemic penumbra equally increases after pre-exposure to isolated and combined hypoxia and hypercapnia.
The results of Western blot analysis generally reflect the trends that were found in immunohistochemical studies of the ischemic penumbra zone in the cerebral cortex (Fig. 3B). At the same time, data from a quantitative study of proapoptotic mediators of apoptosis (AIF and Bax) in ischemic tissue showed lower levels after exposure to permissive hypercapnia than after exposure to hypercapnic hypoxia. At the same time, the level of anti-apoptotic factor Bcl-2 in these groups was higher than after exposure to normobaric hypoxia, and the maximum increase was observed when there was permissive hypercapnia.

The levels of apoptosis/necrosis in astrocyte and neuron cultures after respiratory actions in vitro
The assessment of the number of cells involved in apoptosis or necrosis in astrocyte and neuron cultures after prior exposure to hypoxia and hypercapnia showed a clear dependence on the duration of the hypoxic/hypercapnic mode (Fig. 4). Thus, astrocytes manifested increased apoptosis under physiological conditions in the groups of hypoxic exposure (NbH24 and NbH12), in the group of 24-h permissive hypercapnia (PermH24), and in the groups of intermittent hypercapnic hypoxia (IHHx3 and IHHx2). At the same time, 12-h exposure to hypercapnic hypoxia had an apoptosisinhibiting effect upon astrocytes, whereas groups PermH12 Fig. 2 The number of apoptotic cells in the ischemic penumbra in rats (A). NbH normobaric hypoxia, PermH permissive hypercapnia, HyperH hypercapnic hypoxia, Con control group. *Differences found to be significant as compared with control (P < 0.05). # Differences found to be significant as compared with NbH group (P < 0.05). & Differences found to be significant as compared with the PermH group and HyperH24 did not manifest any significant differences from the reference values. Given this, in all the test groups except the groups of combined hypercapnia and hypoxia (HyperH24 and HyperH12), the numbers of necrosed cells were higher than those in the control group.
In neuron culture, a similar tendency regarding the induction of apoptosis was observed in the groups of hypoxic exposure (NbH24 and NbH12) and the group of 24-h permissive hypercapnia (PermH24). However, together with 12-h exposure to hypercapnic hypoxia, the respiratory actions in groups PermH12 and HyperH24 also inhibited apoptosis, whereas group IHHx2 did not show any differences from the control. It is worth noting that none of the test groups manifested elevated levels of necrosed neurons relative to the control; moreover, in the groups of combined hypercapnia and hypoxia (both continuous and intermittent), there was a decrease.
In a fraction of the test groups, the results of the assessment of the apoptosis/necrosis intensity in nerve cell cultures after simulating chemical hypoxia presented a situation similar to that in physiological conditions (Fig. 5). For instance, the induction of astrocyte apoptosis was detected after 24-h exposure to permissive hypercapnia and normobaric hypoxia; in addition, there were no differences between the control group and the PermH12 and HyperH24 groups. However, the inhibition of astrocyte apoptosis under the condition of chemical hypoxia occurred after prior exposure to 12-h hypercapnic hypoxia and intermittent hypercapnic hypoxia (HyperH12, IHHx3, and IHHx2). After chemical hypoxia, as well as under physiological conditions, the necrosis among astrocytes was elevated after 12-and 24-h exposure to permissive hypercapnia. Nevertheless, in contrast to physiological conditions, the number of necrosed cells decreased after 12-h exposure to normobaric hypoxia and hypercapnic hypoxia, as well as after two-cycle intermittent hypercapnic hypoxia.
In contrast to in physiological conditions, chemical hypoxia did not produce any increase in the intensity of apoptosis in neuron culture in the test groups; however, the given index did not differ from the reference level in the groups subjected to 24-h normobaric hypoxia and permissive hypercapnia. The other test groups demonstrated apoptosis-inhibiting efficacy. There was no increase in the number of necrosed neurons detected in the test groups, or under physiological conditions, whereas a decrease in this index was noted in the groups of combined hypercapnia and hypoxia (both continuous and intermittent).
For an objective interpretation of the data on the effects of various modes of hypercapnic and hypoxic exposure upon the mechanisms of apoptosis and necrosis activation, the effects were expressed on the integrative assessment scale (Table 1). On this scale, the inhibition of apoptosis/necrosis was marked as + 1 point, whereas their induction was − 1 point.

The level of apoptotic mediators in astrocytes and neurons in vitro
The results of the immunocytochemical analysis of pro-and antiapoptotic factors in astrocyte/neuron cultures after prior exposure to hypercapnia and hypoxia showed that 24-h exposure to normobaric hypoxia and permissive hypercapnia, as well as 12-h exposure to hypercapnic hypoxia, induced the expression of caspase-3 in astrocytes (Fig. 6A). At the same time, 12-h exposure to permissive hypercapnia and threetime exposure to intermittent hypercapnic hypoxia, on the contrary, inhibited the studied mediator of apoptosis in the mentioned cell type. No caspase-3 induction was noted in the neuron culture after exposure to permissive hypercapnia and normobaric hypoxia (Fig. 6B), whereas 24-h exposure to hypercapnic hypoxia and both modes of intermittent hypercapnic hypoxia inhibited caspase-3 activity.
The antiapoptotic mediator Bcl-2 was inhibited in astrocyte cultures after 12-h exposure to normobaric hypoxia and hypercapnic hypoxia, as well as after 24-h permissive Fig. 3 The containing mediators of apoptosis in the ischemic penumbra of the rat cortex: immunohistochemistry (A) and western blot analysis (B). Bottom panels (C) show AIF, caspase-3, Bax, Bcl-2 protein bands and corresponding β-actin bands representative for each experimental group. Data are presented as the median ± 25/75 percentiles (A) and mean ± standard deviation (B). NbH normobaric hypoxia, PermH permissive hypercapnia, HyperH hypercapnic hypoxia, Con control group. *Comparison versus control at P < 0.05. **Comparison versus control at P < 0.001. # Comparison versus NbH at P < 0.01. ## Comparison versus NbH at P < 0.001. & Comparison versus PermH at P < 0.05. && Comparison versus PermH at P < 0.001 ◂ 1 3 hypercapnia (Fig. 7A). At the same time, only 12-h permissive hypercapnia induced it in astrocytes. It is noteworthy that all the isolated and combined modes of exposure to hypoxia and hypercapnia, except the intermittent ones, induced Bcl-2 in neuron cultures (Fig. 7B).
As an extension of the analysis of the data on the effects of hypercapnia and hypoxia upon the mechanisms of apoptosis and necrosis initiation ("The levels of apoptosis/necrosis in astrocyte and neuron cultures after respiratory actions in vitro" section), we converted the data on the effects of

Discussion
Apoptosis in the ischemic penumbra was assessed in a fraction of the cerebral section using the ApopTag® ISOL detection method. This method produces more precise and reliable results than the more common terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) method. The results show that permissive hypercapnia, including its combination with hypoxia, inhibits apoptosis in the ischemic penumbra. Similar findings for permissive hypercapnia were obtained by Zhou et al. [19]; they showed that moderate hypercapnia was efficacious in inhibiting apoptosis during cerebral ischemia-reperfusion injury. We obtained similar results regarding the neuroprotective effect of permissive hypercapnia and hypercapnic hypoxia in our previous research, where we assessed the number of TUNEL-positive cells in the cerebral ischemic penumbra [21].
The active form of caspase-3 is the principal caspase effector in the cytoplasm that initiates apoptosis. It can be activated as part of the intrinsic mitochondrial pathway as well as through extrinsic metabolic pathways via death receptors [28,29]. Flavoprotein AIF released from the mitochondrial intermembrane space acts via pathways independent of caspase. It then enters the nucleus, induces chromatin condensation, and activates endonucleases, which participate in DNA fragmentation [30]. Bax, in turn, together with porin, creates a mitochondrial channel in the outer membrane; cytochrome C and AIF are released into the cytoplasm through this channel, whereas Bcl-2 directly or indirectly prevents this process [31,32]. It is known that proapoptotic and antiapoptotic proteins are overexpressed during the first 24 h after a stroke in the ischemic penumbra, which is why the survival of certain cells in this area is determined by the balance between them [11]. Given this, the inhibition of apoptosis in the ischemic penumbra is regarded as a potential approach in antistroke treatment.
Assessment of the levels of proapoptotic factors (caspase-3 and Bax) in the cells in the ischemic penumbra showed that they were inhibited by permissive hypercapnia and hypercapnic hypoxia (Fig. 8). The level of the antiapoptotic factor Bcl-2 increased after all modes of exposure. These data are comparable with the results obtained by Tao et al. [20] regarding permissive hypercapnia. According to their findings, moderate hypercapnia inhibits caspase-3 and cytochrome C in the cytosol; it also decreases Bax levels and elevates Bcl-2 levels in mitochondria. Such effects of hypercapnia, mainly occurring on the mitochondrial apoptotic pathway, could be caused by its antioxidative effects [32][33][34][35][36] as well as by stabilization of the NAD + /NADH (oxidized and reduced forms of nicotinamide adenine dinucleotide, respectively) ratio and the buffering effect of bicarbonate on Ca 2+ [10,37]. The apoptosis-inhibiting impact of hypoxia in this case is probably related to the activation of the phosphatidylinositol 3-kinase system (PI3K) [38,39] and the antiapoptotic effect of hypoxia-inducible factor 1α transcription factor [40,41].

3
The integrative assessment of the intensity of apoptosis/ necrosis following various modes of respiratory action in in vitro models in the second series demonstrates that combined exposure to hypercapnia and hypoxia provides the maximum protective effect, manifesting a distinct dependence on the duration and number of exposure cycles. Moreover, continuous exposure is more effective than intermittent exposure, and 12-h exposure is more effective than 24-h exposure. The two-cycle  12-h hypercapnic hypoxia, IHHx3 3-cycle intermittent hypercapnic hypoxia, IHHx2 2-cycle intermittent hypercapnic hypoxia, Con control group. *Comparison versus control at P < 0.05. **Comparison versus control at P < 0.01 during chemical hypoxia. It is worth noting that 12-h exposure to normobaric hypoxia and intermittent hypercapnic hypoxia has a negative effect upon astrocyte and neuron cultures under physiological conditions, whereas it has a neuroprotective effect during chemical hypoxia. This could be connected to a pronounced subdamaging stimulus that appears in these modes and causes the death of this fraction of astrocytes and neurons, which are the least resistant to hypoxia, thus inducing the maximum ischemic tolerance for the remaining cells. It can also be pointed out that the numbers of apoptotic cells differ following exposure to permissive hypercapnia for in vitro and in vivo models. Thus, the intensity of apoptosis in the ischemic penumbra decreases after prior exposure to permissive hypercapnia; at the same time, this factor has no such effect upon astrocyte and neuron cultures. This could be caused by the absence of glial-neuronal interactions in pure neuron and astrocyte cultures as well as by the excess stress effect of prolonged exposure. Moreover, the direct identification of the effects is limited in different models, because of the various modes of in vivo exposure applied through intermittent sessions in rats and through the continuous prolonged exposure of cells in vitro.
Evaluation of the levels of proapoptotic (caspase-3, Bax, and AIF) and antiapoptotic (Bcl-2) factors in astrocytes and neurons after the in vivo simulation of chemical hypoxia, carried out in the third test series, showed results comparable with those obtained in the second series. A pronounced inhibitory effect on apoptotic signaling pathways also occurred for exposure modes combining permissive hypercapnia and normobaric hypoxia. Moreover, intermittent exposure was found to have positive effects on both astrocytes and neurons, whereas continuous exposure only affected neuron cultures. However, in contrast to the second series, 12-h exposure to permissive hypercapnia had a pronounced suppressive effect on proapoptotic mediators and activated Bcl-2 in all types of cells, whereas 12-h exposure to normobaric hypoxia had a negative effect, inducing apoptosis. The above results indicate that intermittent modes of moderate hypercapnia and hypoxia have more pronounced neuroprotective effects, impacting both neurons and glial cells, which allows high ischemic tolerance to be achieved [13,20]. Table 2 Integrative assessment of hypoxic and/or hypercapnic impact upon astrocytes and neurons in vitro NbH24 24-h normobaric hypoxia, NbH12 12-h normobaric hypoxia, PermH24 24-h permissive hypercapnia, PermH12 12-h permissive hypercapnia, HyperH24 24-h hypercapnic hypoxia, HyperH12 12-h hypercapnic hypoxia, IHHx3 3-cycle intermittent hypercapnic hypoxia, IHHx2 2-cycle intermittent hypercapnic hypoxia, Con control group. ↓Induction of proapoptotic mediators (caspase-3, AIF, and Bax)/inhibition of antiapoptotic mediator (Bcl-2); ↑Inhibition of proapoptotic mediators (caspase-3, AIF, and Bax)/induction of antiapoptotic mediator (Bcl-2) It is worth specifying that, even though the apoptosisinhibiting effect is associated with reduced caspase-3 activity, there are some data presupposing that preconditioning can contribute to the survival of neurons and the maintenance of DNA integrity in an ischemic brain, despite the activation of caspase-3 [42]. The authors explain this by the fact that some stages of caspase-3 activation are necessary for the maximum development of ischemic tolerance. In the present research, elevated caspase-3 was observed, particularly in astrocyte cultures after 12-h exposure to hypercapnic hypoxia, accompanied by the inhibition of Bax and Bcl-2. At the same time, a similar exposure mode had a pronounced antiapoptotic effect on neuron cultures, which suggests the possible existence of the mentioned mechanism.
It is also important to note that the research described here has certain limitations; the findings on the levels of the mediators of apoptosis in cell cultures were not confirmed by Western blotting or gene expression analysis based on qPCR analysis. Consequently, the obtained results should be interpreted with caution, and a detailed understanding of the mechanism underlying the hypercapnic and hypoxic effects on caspase-dependent and caspase-independent signaling pathways of apoptosis requires further study.

Conclusions
Preliminary exposure to permissive hypercapnia and hypercapnic hypoxia can activate the Bcl-2 antiapoptotic factor and inhibit caspase-dependent and caspase-independent apoptotic signaling pathways during cerebral ischemic injury.