Mitochondrial Events Determine the Status of Hippocampal Cells in the Post-Ischemic Period

Dear Editor,

Understanding ischemia-induced events is important for effective protective strategies in brain pathologies. Actively functioning neuronal cells, the specialization and efficiency of which were formed evolutionarily, are resistant to minor fluctuations in the extracellular environment, maintaining cell and tissue homeostasis. If the threshold for resistance is exceeded, a molecular imbalance appears, followed by a neurodegenerative process. Although the energy-supplying function of mitochondria in cellular metabolism is well known, the contribution of mitochondrial events to cell survival in extreme situations is not yet fully understood. In conditions of oxygen-glucose deficiency, the vulnerability of the brain is not identical in different cells and regions. It is known that the pyramidal neurons of the CA1 hippocampal region are most vulnerable, while interneurons and glial cells are relatively resistant to ischemia, and contribute to neuroprotection [1, 2]. The approaches in the study of mechanisms underlying neurodegeneration and/or rehabilitation are complicated by overlapping responses of functionally and structurally connected brain cells. Often, researchers focus on one type of cell in the investigation of cerebral damage. Despite numerous studies, many aspects of neuronal and glial responses to ischemia have not yet been elucidated. Deficiency of oxygen and glucose leads to systemic changes in the brain that are initially oriented towards the restoration of cellular and tissue homeostasis (adaptation), but can also trigger delayed pathological events up to irreversible neuronal degradation. In brain diseases, along with specific pathogenic mechanisms, ischemia-induced mitochondrial dysfunction is a general factor in the neurodegenerative process [3–5]. The reasons for delayed damage to hippocampal cells in a mild oxygen-glucose deficit, as well as the cellular aspects of this phenomenon, have not yet been fully elucidated. The energy supply of mammalian brain metabolism is mainly carried out by mitochondria through the aerobic oxidation of glucose. The cellular energy resource (production and buffering of ATP) is largely determined by mitochondrial activity (MA), as well as the number and morphology of mitochondria [6, 7]. This in vitro study presents an evaluation of the mitochondrial patterns of basic hippocampal cell types in the modeling of post-ischemic damage using slice cultures and transient oxygen-glucose deprivation (OGD) followed by normoxia. The hippocampal CA1 region that is most sensitive to OGD was analyzed. MA was determined based on a semi-quantitative analysis of mitochondrial membrane potential using a mitochondrial probe [MitoTracker Orange (MTO)] and confocal microscopy. Electron microscopic assessment of structural mitochondrial parameters was also carried out. In addition, given the standardized experimental conditions of our ischemic damage model in vitro, we extrapolated mitochondrial dynamics to relative parameters characterizing the state of hippocampal cells during the study period.

Double staining, namely MTO combined with immunohistochemistry, allowed us to differentiate an MTO-positive signal (MTO+) (Fig. 1, red) co-localized with specific cellular markers (Fig. 1, green). In this way, we traced the relative MA changes associated with pyramidal neurons (NeuN+), interneurons (GAD67+), astroglia (GFAP+), oligodendroglia (MBP+), and microglia (Iba-1+) for 24 h after 30-min OGD. When identifying cells, the singular cytoarchitectonics of the hippocampus, preserved in the cultured slices, as well as cell morphology (shape, size, density, and immunoreactivity) were taken into account. It is known that the pyramidal bodies are densely packed in the pyramidal zone [stratum pyramidale (SP)] and predominate there (up to 98% of the total number) [8, 9]. Given this, the widely used NeuN (neuronal nucleus) marker indicates the position of the pyramidal neural bodies in the CA1 region with a high probability [10]. Interneurons and glial cells are relatively evenly distributed in the CA1 region, mostly outside the pyramidal layer. The CA1 radial zone [stratum radiatum (SR)] is predominantly the closely intertwined appendages of hippocampal cells. In addition to MA, the mitochondrial ultrastructure was analyzed in the SP and SR using electron microscopy. The following parameters were assessed: (1) the ratio of the total mitochondrial area to the corresponding cytoplasmic area of the pyramidal soma (excluding the nucleus and processes) in the SP zone, as well as the ratio of the total mitochondrial area to the corresponding area of the processes in the SR zone (in Fig. 2, "Soma" is marked with a purple dotted line); and (2) average density and optical density of mitochondria in the pyramidal soma, as well as in the SR region.

Fig. 1.

Analysis of mitochondrial activity in the CA1 region of hippocampal slice cultures at 1, 4, and 24 h after 30 min OGD compared to control. Combined MitoTracker Orange (MTO; red)/immunohistochemical (green) staining of various hippocampal cell types: A1–A3 pyramidal neurons (NeuN+); B1–B3 interneurons (GAD67+); C1–C3 astroglia (GFAP+); D1–D3 oligodendroglia (MBP+); E1–E3 microglia (Iba-1+); co-localization of markers gives yellow. Illustrations of staining in Control (A1–E1), 1 h (A2–E2), and 24 h (A3–E3) after 30 min OGD. Scale bars, 100 µm. A4–E4 Diagrams of changes in the integrated density of MTO+ fluorescence (FL) characterizing mitochondrial activity in the respective immunopositive cells (*P <0.05 vs Control, n = 30). Data are expressed as arbitrary units (a.u.).

Fig. 2.

Electron microscopic analysis of mitochondria in the CA1 region in hippocampal slice cultures. A Representative electron micrographs: control (A1), 1 h (A2), and 24 h (A3) after OGD. Images illustrate the transverse profile of a part of a pyramidal neuronal soma (purple dashed line) and the region of the cell processes [(stratum radiatum (SR)]. Arrows indicate: Mt, mitochondria; V1, synaptic vesicles; V2, other cellular vesicles; Inv, invaginations. Scale bars, 500 nm. B1 Ratio of the mitochondrial area/area of the neuronal cytoplasm (marked on the graph as “Soma”) and mitochondrial area/area of the measured SR region in the Control, 1 h, 4 h, and 24 h after 30 min OGD (*P < 0.05 vs Control, n = 60). B2 Changes in the optical density of mitochondria in the neuronal cytoplasm and the SR region (*P < 0.05 vs Control, n = 60). B3 Scatter plot showing a correlation between mitochondrial optical density and integrated density of MTO+ fluorescence (FL) in CA1 pyramidal neurons; n = 90, r = 0.62. C Images illustrating exocytosis/endocytosis involving: pyramidal soma and SR cell processes (C1), SR cell processes (C2), and pyramidal soma and astroglia (C3). The pyramidal soma is marked with a purple dashed line. Scale bars, 200 nm.

In the hippocampus, excitable pyramidal neurons are functionally dominant, but they are also the most vulnerable. The molecular mechanisms of OGD-induced pyramidal neuron neurodegeneration have been well studied. They are mainly associated with disruption of ion transport pumps resulting from ATP deficiency and glutamate excitotoxicity, as well as the accumulation of reactive oxygen species (ROS) and subsequent oxidative stress [3, 4]. The imbalance of transmembrane ionic gradients involving glutamate/calcium-mediated cascades leads to hyperactivation of excitable pyramidal neurons [7], which can initiate cell damage. There is a fine line between normal functioning and the beginning of neurodegeneration. Using an ischemic damage model, we investigated changes in some mitochondrial features and then compared the studied parameters with the status of hippocampal cells. The post-OGD effects can be differentiated into initial (reactive stage), intermediate (adaptive stage), and delayed (destructive stage) [11], which is conditionally comparable to 1, 4, or 24 h of normoxia after 30-min OGD in our experimental model (Fig. S1).

The experimental protocol we used did not produce rapid disturbances, but had signs of delayed damage that simulates post-ischemic injury events. The reactive stage (1 h after OGD) was characterized by an increase in the integrated density of MTO fluorescence compared to the control in all types of cells (Fig. 1A1–E1, A2–E2, A4–E4; n = 30, P < 0.05), which may indicate the activation of their metabolism. The parameters that characterize the status of various hippocampal cell types (namely, relative immunoreactivity and number) remained at the control level (Fig. S2B, C; n = 15, P < 0.05). These data indicate the adequacy of the available mitochondrial resources to maintain cellular homeostasis during this period. In the reactive stage after temporary OGD, electron microscopy showed statistically significant increases in total mitochondrial area (Fig. 2B1; n = 60, P < 0.05) and density (i.e., average number of mitochondria per unit area, Table S1) just in CA1 pyramidal soma, although their shape (Fig. 2A1, A2) and the average optical density (Fig. 2B2; n = 60, P < 0.05) remained without significant changes. Thus, in the early period after OGD, along with mitochondrial activation (based on MTO staining), the number of mitochondria increases in CA1 pyramidal somata.

There are data that neuronal cells can exchange their contents. Neurons are able to release and accept mitochondria, when astrocytes can form mitochondrial fragments that can be transported to neurons through exocytosis/endocytosis [12, 13]. Delivery of neuroprotective factors and removal of reactive molecules can be carried out in the same way [14]. In this study, electron microscopy allowed us to detect morphological signs of intercellular communication in the border zone between the SP and SR of the CA1 region. CA1 pyramidal neurons were identified based on their spatial localization and morphology (characteristic shape, size, density, and features of the internal ultrastructure). In addition to regular synaptic contacts, we have found many examples of interaction through exocytosis/endocytosis with the formation of membrane invaginations and/or subcellular vesicles (Fig. 2A2, 2C), suggesting the possibility of direct exchange of cellular contents. Due to the formation of invaginations, an increase in the surface area of contacts between the pyramidal soma and inhibitory interneurons, which dominate in the zone of interest, was observed. As a result of exocytosis/endocytosis, intracellular vesicles of various sizes were formed, carrying the contents of neighboring cells, not excluding small organelles (namely, mitochondria). The GIF images well illustrate the phenomena described above (Fig. S3). The described form of intercellular communication can be one of the mechanisms for the normalization of the state of pyramidal neurons during their hyperactivation in the reactive stage after OGD.

Subsequent events (adaptive stage) showed that CA1 hippocampal cells remained viable for 4 h after OGD, but MA differed significantly in the considered cell types. Variations in MA parameters were as follows: the average MA level stabilized to control in pyramidal neurons and oligodendrocytes; and remained elevated or increased in interneurons, astroglia, and microglia (Fig. 1A4–E4; n = 30, P < 0.05). The fact that MA in pyramidal neurons and oligodendrocytes decreased (compared to OGD + 1 h) indicates a probable decrease in the adaptive potential and resistance of these cells. The ultrastructural parameters of mitochondria in the SP and SR remained stable for 4 h after OGD stress. Thus, the data indicate that in our experimental model at the adaptive stage there are sufficient resources to maintain cellular homeostasis, but the energy status is approaching a critical level.

In further studies, delayed destructive post-OGD effects became apparent and were in direct correspondence with MA dynamics. It was revealed that 24 h after OGD, the average MA level was lower than the control in pyramidal neurons and oligodendrocytes, while it significantly increased in astroglia and microglia (Fig. 1A4–E4; n = 30, P < 0.05). The described MA changes were correlated with the degradation of pyramidal neurons and oligodendrocytes and the hyperactivation of astroglia and microglia, respectively (Fig. S2). MA parameters of interneurons were similar to those in control. These data demonstrate the relatively greater resistance of the last mentioned three types of cells in the post-OGD period.

It should be noted that the basal MA level and mitochondrial morphology were largely comparable in different cell types of control hippocampal cultures. It is clear that the functional potential of mitochondria has limits. It can be assumed that when energy resources for functioning and maintaining homeostasis are depleted, cells degrade. Vulnerability/resistance varies both for different hippocampal cells and within the CA1 pyramidal cell population as well. In our experimental model, morphological abnormalities of mitochondria (membrane deformation, amorphousness, absence of cristae) were observed even in the somata of CA1 pyramidal neurons, which did not yet have pronounced signs of cellular degradation at the destructive stage, as shown in Fig. 2A3. At this stage, we evaluated the parameters of the mitochondrial ultrastructure without taking into account extremely damaged cells, when mitochondria could not be marked. It was in the pyramidal soma that a significant decrease in the total area, density, and optical density of mitochondria was found 24 h after OGD (Fig. 2B1, B2; n = 60, P < 0.05; Table S1). It should be noted that the parameters of the optical density of mitochondria in the pyramidal soma correlated with the integrated fluorescence density characterizing MA in the SP region (Fig. 2B3; n = 90, r = 0.62).

Based on the analysis of mitochondrial and cellular changes in the post-OGD period, we presume that glutamatergic pyramidal neurons most actively spend their energy resources on glutamate/calcium-mediated events at the reactive stage (OGD + 1 h). Subsequently, to restore homeostasis at the adaptive stage (OGD + 4 h), they also consume mitochondrial resources more actively than cells of other types. At the destructive stage (OGD + 24 h), when the resistance threshold is reached, pyramidal neurons are more likely to degenerate. Since oligodendrocytes involved in signaling are in close contact with excitatory glutamatergic synapses and express glutamate receptors [15], they are very sensitive to the excitotoxicity of glutamate, as well as to other OGD-induced factors. Our data demonstrate an MA decrease and cell degradation in the oligodendrocytic population, similar to pyramidal neurons. During the post-OGD period studied, the results showed a stable state and/or activation of interneurons, as well as astroglial and microglial cells, which were more resistant and able to modulate the state of pyramidal neurons, at least in part.

The higher vulnerability of hippocampal pyramidal neurons in cerebral ischemia, which has been widely described in both in vivo and in vitro studies, can be largely due to the specific cytoarchitectonics of the hippocampus. That is, the evolutionarily formed hippocampal patterns are functionally expedient, but they are also limiting. Apparently, processes similar to those described can also occur in any brain regions in the context of the energy resources of various cellular types.

Thus, mitochondrial events largely determine the homeostatic threshold of hippocampal cells in OGD, and thus their survival. A comprehensive study of cellular and mitochondrial patterns in the hippocampus expands the understanding of the causes of the high vulnerability of CA1 pyramidal neurons in cerebral ischemia and the phenomenon of delayed neuronal death after transient cerebral ischemia. We believe that in brain diseases caused by ischemia, it is advisable to take into account both the cell specificity and the dynamics of destructive processes, which can expand the therapeutic window, as well as increase the effectiveness of neuroprotective strategies that also target mitochondria.

Supplementary Information

Below is the link to the electronic supplementary material.

Conflict of interest

All authors claim that there are no conflicts of interest.