1. Introduction
The development of the nervous system and its subsequent functionality boasts a sequential differentiation of cellular precursors, the most prominent of which are neural stem cells (NSCs). NSCs are multipotent stem cells that give rise to neurons, astrocytes, and oligodendrocytes [
1]. Unlike the adult nervous system, in which the ratio of undifferentiated cells is negligible, the immature nervous system comprises cells in various developmental stages [
2,
3,
4].
Brain diseases, including stroke and perinatal ischemia, pose a great burden on modern society and can result in lifelong disability [
5]. From a pathophysiological perspective, the sudden lack of oxygen and glucose supply to the nervous system results in a systemic decrease in membrane potential, reduction in ATP production, and neuronal swelling [
6]. In turn, the cellular metabolism is changing from aerobic to anaerobic, leading to pH imbalance and cell death [
7,
8]. Although the occurrence of necrosis and apoptosis has been well documented in cases of nervous tissue ischemia, there is also evidence that ischemic insult could lead to mitochondrial dysfunction, followed by autophagy and mitophagy. Autophagy is a highly conserved cellular process comprising the selection and degradation of damaged intracellular organelles, or proteins, within lysosomes. Even though it could be activated by many physiological stimuli, the process of basal autophagy is generally active under normal conditions and constitutes an important homeostatic mechanism for maintaining a healthy cellular environment. In addition, autophagy has also been recognized as one of the major mechanisms underlying perinatal brain damage [
9,
10]. Even though some researchers reported that promoting autophagy after brain ischemia is beneficial, others demonstrate the opposite to be true [
11,
12]. Thus, the role of autophagy in stroke remains controversial.
Mitophagy is a subtype of autophagy that involves the lysosome-dependent clearance of damaged mitochondria. This is particularly prominent in cases of ischemic injury of cells, which involves the impaired function of mitochondria, coupled with dysregulation of calcium homeostasis and the production of reactive oxygen species (ROS) [
13]. Following such injuries, the success of subsequent improvement in mitochondrial metabolism can dictate the outcome regarding cell repair or cell death via necrosis, apoptosis, or autophagy [
14]. Accordingly, it is not surprising that mitophagy is recognized as one of the processes that might become an important target in treating perinatal brain damage [
15]. Even though some research studies suggest that increased mitophagy has beneficial effects in directing the recovery of the nervous tissue, it remains unclear how an increase or decrease in the rate of mitophagy would impact perinatal brain damage [
16].
Furthermore, even though many publications reported the occurrence of autophagy and mitophagy in the nervous system during various stages of its development, including in the neonatal neurons, a direct comparison of cells throughout these stages of differentiation is lacking [
17]. Therefore, the main goal of this work was to use our expertise in neural cell precursors and investigate the influence of oxygen and glucose deprivation on their early, middle, and late stages of differentiation. Specifically, our research focused on the differentiation and survival of immature cells of the nervous system, namely the quantification of the morphological and functional characteristics of their mitochondria upon exposure to the acute lack of oxygen and glucose, combined with a detailed overview of subsequent mitophagy and autophagy. As a result, we show that cells in various stages of maturity react differently to the lack of oxygen and glucose and that various parameters, such as the balance in the fusion/fission of mitochondria, change accordingly. Interestingly, we also note that neural cell precursors caught in the middle stage of differentiation (7 days) exhibit more significant changes in mitochondrial morphology. These findings are complemented by changes in several markers of mitophagy and autophagy. Hence, the results of this study suggest that neural cell precursors exhibit different and distinct reactions to the lack of oxygen and glucose throughout different stages of their differentiation. This finding might be important in choosing appropriate therapeutic approaches for the treatment of the immature nervous system affected by perinatal brain damage.
2. Materials and Methods
2.1. Isolation and Differentiation of Neural Stem Cells
NSCs were isolated from the telencephalic wall of 14.5-day-old C57/BL6 albino mice embryos. They were cultivated in a proliferation medium made from 1% N2 (Gibco, 17502-048, New York, NY, USA), 1% Pen/Strep (penicillin/streptomycin, 5000 U/mL, Gibco, 15070063), 2% B27 (Gibco, 17502), 20 ng/mL EGF (epidermal growth factor, Gibco, PMG8041), 10 ng/mL bFGF (fibroblast growth factor basic, Gibco, PMG0035), and 5 mM HEPES (Sigma-Aldrich, H0887, Burlington, MA, USA), where they formed structures called neurospheres. To initiate the process of their differentiation, the cells were cultivated in pretreated plates containing 50 μg/mL of poly-D-lysin (PDL, Sigma-Aldrich, P6407) and 10 μg/mL of laminin (Sigma-Aldrich, L2020). Herewith, the complete proliferation medium was replaced with the differentiation medium comprising 1% N2, 1% Pen/Strep, 1% FBS (fetal bovine serum, Gibco, 15070063), 2% B27+ (Gibco, A3582801), and 5 mM of HEPES. Half of the old medium was replaced with a new differentiation medium every 4 days.
2.2. Oxygen–Glucose Deprivation Model
To induce ischemic injury in 3 specific timepoints, we utilized the OGD (oxygen–glucose deprivation) method. Briefly, this included the replacement of the complete differentiation medium with Dulbecco’s no-glucose medium (Gibco, 11966025) on differentiation days 1 (D1), 7 (D7), and 14 (D14). After this, the cells were placed in a low-oxygen (1% O2) incubator for 24 h, after which the samples were taken. The samples from the control group were collected after the same period, albeit from the cells that were placed in a separate incubator and were grown under normal conditions.
2.3. Immunocytochemistry
The cells were grown on coverslips and fixed in 4% PFA for 10 min, followed by 3 washes with PBS. Permeabilization was performed in 0.2% Triton in PBS for 15 min, followed by another 3 washes with PBS. The cells were then blocked in filtered 3% goat serum in PBS at room temperature for 2 h. The primary antibodies were directly added onto the coverslips and incubated overnight at +4 °C. The following day, the cells were washed in PBS 3 times, after which a secondary antibody was added and incubated at room temperature for 1 h. Here, we used 1:1000 goat anti-mouse 488, 1:1000 goat anti-chicken 546, and 1:500 goat anti-rabbit 633. The cells were, once again, washed 3 times in PBS. The counterstain procedure was carried out using 1: 20,000 DAPI (4′,6-Diamidine-2′-phenylindole dihydrochloride, Roche, 10236276001)
2.4. Western Blot
Protein lysates were prepared in a RIPA lysis buffer using ice and a 27 G needle with the addition of protease (Roche, 11836170001, Indianapolis, IN, USA) and phosphatase (Roche 4906837001) inhibitors. Quantification was performed using a detergent-compatible Bradford reagent (Thermo Scientific, 1863028, Waltham, MA, USA). The proteins were loaded into a 12% stain-free gel, and blocking was performed using 3% low-fat milk for 2 h. The membrane was incubated overnight at +4 °C with the following antibodies: anti-PINK1 (Abcam, ab23707, Waltham, MA, USA) 1:1000, anti-LC3 (Cell Signalling, #3868) 1:1000, anti-FUNDC1 (Novus Biologicals, NBP1-81063, Centennial, CO, USA) 1:2500, and anti-p62 1:15,000 (Abcam, ab109012). The membranes were then washed 3 times in TBST and incubated with secondary antibodies (1: 200,000) for 1 h. Subsequent detections were accomplished using the Femto SuperSignal chemiluminescent reagent (Thermo Scientific, 34095). Because of the low protein yield on D1, Western blot was only performed on D7 and D14 of differentiation. The results obtained with the Western blot were normalized to the total protein amount. All images of the Western blots can be found in the
supplementary data.
2.5. Quantitative Polymerase Chain Reaction (qPCR)
To isolate RNA, the samples were first stored in an RLT buffer. The RNeasy kit (Qiagen, 74104, Hilden, Germany) was used for RNA isolation, and the concentration was measured using the Nanodrop device. Equal amounts of RNA (25 ng/µL) were later transcribed to cDNA using a high-capacity RNA to cDNA kit (Applied Biosystems, 4374966, Waltham, MA, USA). qPCR was performed by adding the same amount of the specific TaqMan assay and the sample of interest. mRNA was then analyzed using specific TaqMan probes for nestin, GFAP, MAP2, and FUNDC1. β-Actin was used as a housekeeping probe. Finally, the qPCR results were depicted as 2–∆∆Ct, which indicates the fold difference.
2.6. Lactate Dehydrogenase (LDH) Detection Kit
For the purposes of evaluating cellular damage, LDH-GloTM Cytotoxicity Assay (Promega, J2380, Madison, WI, USA) was used. Firstly, 5 µL of the medium was mixed with 95 µL of a storage buffer (200 mM Tris-HCl (pH 7.3), 10% Glycerol, 1% BSA). The samples were then combined with an LDH detection enzyme mix (detection enzyme and reductase) in a 1:1 ratio. The incubation time was 30 min, and the luminescence was detected with a 0.9 integration time. The amount of released LDH was proportional to the luminescence signal. The percentage of cytotoxicity was calculated according to the following formula: 100 × (cellular LDH release − medium background)/(maximum LDH release − medium background), where cellular LDH release denotes the level of LDH in specific conditions. On the other hand, the maximum LDH release denotes the amount of released LDH following a 15 min treatment with Triton X-100, which caused complete damage to the cells.
2.7. MitoSOX
In order to detect the mitochondrial oxygen radicals, we used a MitoSOX red detection kit (ThermoFisher, M36008, Waltham, MA, USA). Once it enters the live cells, it is oxidized by superoxide, resulting in the appearance of a fluorescent signal. Briefly, a 500 nM working concentration was directly added to the cells and incubated for 15 min. Cells were then washed 2× with preheated HBSS, after which the signal was detected at 580 nm. The oxygen radicals were observed using live-cell imaging with a glass bottom Petri dish and an Olympus FV3000 confocal microscope. Images were analyzed in Imaris 9.9 (Oxford Instruments, Oxfordshire, UK).
2.8. TMRE
To evaluate the mitochondrial membrane potential (MMP), we used a TMRE detection kit (Abcam, 113852) that is designed to quantify changes in the MMP of live cells. It consists of a positively charged dye that selectively binds to negatively charged active mitochondria. The dye was added to the cells at a final concentration of 200 nM for 20 min. The cells were then washed with preheated PBS twice, and a signal was detected at 575 nm. The mitochondrial membrane potential was, once again, observed using live-cell imaging.
2.9. LC3-GFP/Lysotracker Colocalization
As a way of analyzing colocalization, we used a Premo Autophagy Assay (ThermoFisher, P36235) and LysoTracker Deep Red (ThermoFisher, L12492). The Premo Autophagy sensor is based on non-replicative BacMaM technology for the stable transduction of LC3B-GFP chimera. The multiplicity of infection (MOI) was 20, and the incubation time was 24 h. LysoTracker Deep Red was used in a final concentration of 50 nM. The kit was incubated for 30 min, after which the medium was replaced with a new preheated medium. Images were analyzed in Imaris 9.9 (Oxford Instruments).
2.10. Mitochondrial Morphology
The morphology of mitochondria was analyzed using the Tomm20 marker. All the images were analyzed with an ImageJ/Fiji macro developed by our group (Šimunić et al., under review).
2.11. Statistical Analyses
Statistical analyses were conducted in GraphPad Prism version 8.00 (GraphPad Software, San Diego, CA, USA). All the experiments were repeated 3 times, and their significance was quantified using the t-test. A p value of p < 0.05 was considered statistically significant (*), p < 0.01 was considered highly statistically significant (**) and p > 0.05 was considered nonsignificant (ns). The data are presented as mean ± SEM (Standard error of mean).
4. Discussion
The deregulation of circulation, which leads to the lack of oxygen and glucose within the tissue, is the backbone of many diseases that cause life-long impairment, including stroke and perinatal brain damage. Although perinatal brain damage can be separated into five distinct entities, namely hypoxic–ischemic encephalopathy, intraventricular hemorrhage, periventricular leukomalacia, and perinatal stroke, they have one pathophysiological element in common: oxygen and glucose deprivation. With one-third of neonates affected by perinatal brain damage, those who do not die develop severe seizures, and motor, cognitive, and memory impairments, alongside cerebral palsy. Since this represents approximately 30–40% of survivors, it poses a great medical burden against which our therapeutic options are still insufficient and, therefore, lacking [
19]. With the goal to elucidate how the sudden lack of oxygen and glucose influences the immature cells of the nervous system, we exposed an in vitro model of neural precursors to OGD during three stages of their development to OGD. The first step was to fine-tune the levels of OGD exposure needed to trigger measurable cell death with an LDH test, including all the major markers of differentiation present. This was followed by a direct comparison of their capability for differentiation. With the exception of the earliest precursors, which expressed more GFAP, the precursors exposed to OGD exhibited decreased nestin, MAP2, and GFAP gene expression in all differentiation stages.
Considering that one of the goals of this research was to quantify the influence of oxygen–glucose deprivation on the mitochondria and mitochondria-related processes such as mitophagy, we simultaneously developed a protocol for semi-automatic quantification of the parameters needed in such an investigation, namely Lusca. Lusca is an ImageJ/Fiji macro designed for detailed and accurate recognition of circular, fibrillar, and punctate structures in digital images of tissue or cells (Šimunić et al., under review). This macro allowed us to quantify the total area, the number of mitochondrial branches, and the total length of mitochondrial filaments in developing cells of the nervous system using Tomm20, a marker of the mitochondrial membrane. Interestingly, our results demonstrate that the cells significantly increased the ratio of the cytoplasm volume normalized to the total cell volume filled with mitochondria through the increasing stages of differentiation. One potential reason behind this is the fact that, as cells mature, the mitochondria gradually become the more important organelles in cells of the nervous tissue. When exposed to OGD, the cells within all three differentiation timepoints reacted by decreasing the total volume, length, and branching of the mitochondrial membranes. This effect was most prominent in cells on D7.
In order to quantify these changes in mitochondrial parameters and morphology, we used Lusca to separate the observed mitochondria into three distinct forms—tubular, intermediate, and punctate—with tubular mitochondria denoting the healthy and punctate denoting damaged mitochondria undergoing fission [
18,
20]. Since the mitochondrial network and its metabolism are a very dynamic system, undergoing constant fusion/fission cycles, the balance between the main fusion proteins Opa1 (optic atrophy 1) and Mitofusin 1 and 2, and the fission protein Drp1 (dynamin-related protein), is an important mechanism underlying the regulation of mitochondrial architecture [
21]. In an undifferentiated state, immature cells, such as NSCs, maintain a low metabolic activity characterized by a low oxygen consumption rate, fragmentation, and glycolysis [
22,
23]. This hypothesis is supported by the results of our study, wherein, during the first day of differentiation, almost exclusively SOX2- and Nestin-positive cells were observed, characterized by a small number of Tomm20-positive regions. Additionally, cell differentiation is also initiated and, therefore, regulated through changes in ATP production, namely its increase. This is accomplished by an increase in the mitochondrial membrane potential and changes in the electron transport chain located in the inner mitochondrial membrane [
23]. Similar phenomena can also be seen within our study, wherein the growth of the mitochondrial network was observed on D7 and D14, depicted as an increase in the Tomm20-positive area and filament length.
Since appropriate mitochondrial functioning is a crucial factor in determining the proper development of the nervous system, our research also included investigations into the reaction of neural precursors to OGD. As expected, these effects depend on the cells’ developmental stage. Interestingly, significant changes in the cell mitochondria, namely an increased incidence of punctate elements, were observed only on D7. Based on these observations, it appears that cells in the middle stage of differentiation react to OGD with a significant reduction in the tubular form of mitochondria, which are also prone to mitochondrial fission. These findings are in accordance with other research within the field, wherein evidence of mitochondrial fragmentation following ischemia, including an increase in mitophagy, was observed [
16]. Alongside fragmentation, we noticed that the affected mitochondria are more spherical, which possibly correlates with its swollen and rounded ultrastructural appearance analyzed using electron microscopy [
21]. Such an increase in the ratio of punctate mitochondria could be a mechanism through which cells accelerate the removal of damaged mitochondria and improve their chances of survival [
24].
As the polarization of the mitochondrial membrane plays a significant role in energy transformation during oxidation, and therefore ATP production, we also evaluated ROS production and the corresponding MMP. Any imbalance in the oxidative status of the system results in misfolding of proteins, lipid peroxidation, genome instability, general cellular toxicity, and opening of mPTPs (mitochondrial permeability transition pores) [
20,
25]. Even though an increase in ROS production in all three stages of differentiation following OGD was detected, no significant difference was observed on D7. This can possibly be explained by the fact that the largest number of neurons are present on D7 under normal conditions. Once they are exposed to OGD, they die. The fact that neurons produce more oxygen radicals than astrocytes explains why our results depicted a lower presence of ROS on D7 [
26].
Similarly, TMRE revealed a hyperpolarization of the mitochondrial membrane in all three stages of cell differentiation. This hyperpolarization may be explained as a consequence of cellular shrinkage and subsequent mitochondrial accumulation. Such findings are supported by those of Agarwal et al., who suggested that cellular heat-shock response—usually induced by an increase in temperature, ischemia, ROS production, and other stressors—leads to perinuclear mitochondrial clustering and an increase in overall mitochondrial intensity [
27]. Alternatively, since we also detected excessive ROS production, the hyperpolarization of the mitochondrial membrane could be due to the phenomenon called RIRR (ROS-induced ROS release) that opens mPTP as a part of the cellular adaptive mechanism of releasing ROS to the environment [
20]. The research by Ward et al. suggested that delayed apoptotic injury of neurons is first characterized by a depolarization period, followed by the hyperpolarization of the mitochondrial membrane [
28]. Under normal conditions, ROS increase is accompanied by a decrease in the mitochondrial membrane potential. On the other hand, Zorov et al. reported that, under certain conditions, ROS overproduction could also lead to the occurrence of a hyperpolarization period [
20]. Similarly, Korenić et al. described the same phenomenon in astrocytes exposed to OGD [
29]. The fact that our cell culture boasts predominant numbers of astrocytes over neurons, as is the case in the human brain, might explain why our results depict clear cases of hyperpolarization of the mitochondrial membrane, as opposed to depolarization noted in experiments conducted on predominantly neuronal cultures [
30].
Further analyses performed on markers of mitophagy, namely PINK1 and FUNDC1, suggest that cells already undergo this process on D7 and D14. PINK1 is a mitochondrial serine–threonine protein kinase that acts as an important mitochondrial quality control system. In physiological conditions, it is effectively imported to the inner mitochondrial membrane, where it is subsequently cleaved and degraded. On the other hand, PINK1 is also known to accumulate on the outer membrane of damaged mitochondria, where it can induce the ubiquitin-dependent removal of mitochondria [
31]. Similar to what was observed with PINK1, a significant decrease in FUNDC1, a novel mitochondrial receptor, was observed on D7. This mirrors the results of Liu et al., reporting the hypoxia-induced degradation of mitochondria [
32]. To quantify the extent of autophagy, we also evaluated the levels of LC3-II, in combination with its colocalization with a lysosome. The results revealed that the levels of LCR-II were significantly increased during both the middle and late stages of differentiation.