Alzheimer's disease (AD) is one of the most common neurodegenerative diseases associated with cognitive impairment and progressive neuronal loss. Currently, there is no doubt that mitochondria are involved in the pathogenesis of AD, and some hypotheses place mitochondrial dysfunction in the leading role in the onset of AD [1, 2]. Decreased respiratory capacity, increased mitochondrial fragmentation, and fractures in the mitochondrial cristae structure occur in the brain in AD, and abnormalities in mitochondria appear before the deposition of pathologic Aβ plaques [3]. Consistent with the observation that impaired energy metabolism invariably precedes the clinical onset of AD, mitochondrial dysfunction has been established as an early and prominent feature of the disease [4]. Cell functionality is highly dependent on the state of the mitochondria, which are essential for ATP production (general support of cell metabolism) and maintaining calcium homeostasis (regulation of neurotransmitters and communication with other cells). There are different fractions of mitochondria in different parts of the cell at the same time, performing different functions. These fractions must be in constant turnover to maintain the normal physiological state of the cell. This form of mitochondrial dynamics is known as mitochondrial fission and fusion [5]. Under normal conditions, mitochondrial fusion and fission occur continuously. A constant balance is maintained between mitochondria in the fusion-fission process, which deviates from one side to the other in response to the energy and metabolic needs of the cell.
Improper mitochondrial function leads to the formation of lipofuscin (Figure 1). It is a substance composed of oxidized lipids, covalently cross-linked proteins, oligosaccharides, and metals. The composition of lipofuscin is largely dependent on the intracellular site of action of reactive oxygen species on proteins and lipids [6]. Lipofuscin accumulation is directly proportional to pathologies of various genesis and also increases under conditions of oxidative stress and reactive oxygen species produced by damaged mitochondria. Although lipofuscin is considered a marker of senescent cells, recent work shows its close association with pathologic tau protein and Aβ in AD [7, 8]. Moreover, a recent study shows that the α-structured protein lipofuscin is a toxic component of β-structured amyloid plaques [9]. Since lipofuscin has the property of autofluorescence, it can be used as a marker of disorders even before the first morphologic manifestations. That said, one of the most common descriptions of lipofuscin sounds like "intracellular trash". This term denotes one of the interesting properties of lipofuscin—extreme resistance to cellular proteolysis, which is explained by the "cross-linking" of the aldehyde and amino groups of the peptide to form stable polymeric structures.

The search for a model of sporadic AD is incomplete, mainly due to the inability to understand and reproduce the mechanism of disease development. This leads researchers to the fact that one obvious imbalance process is reproduced in models (disturbance of the balance in the systems of generation and detoxification of reactive oxygen species in rats of the OXYS line [10], mice with surgically removed olfactory bulbs (olfactory bulbectomized mice) [11] or direct introduction of pathological tau protein or beta-amyloid into the culture of neurons [12, 13].
In this work, we sought to investigate whether disruption of mitochondrial dynamics in healthy cultures could lead to a process similar to the pathology manifested in the 5xFAD mouse line. We directly manipulate the mitochondrial fusion-division system in primary mixed neuroastrocyte hippocampal cultures of mice by inducing shifts toward a prevalence of fusion or fission. We used the primary hippocampal cell culture of 5xFAD transgenic mice for comparison. Cells in the transgenic culture gradually accumulate high levels of the neurotoxic agent betaamyloid from the beginning of their lives, causing neurodegeneration similar to AD [14]. In addition to the formation of amyloid plaques, it is characterized by marked astrogliosis and loss of neuronal and synaptic density, as well as increased levels of lipofuscin, an early marker of cellular pathology [15]. Altering the balance of fusiondivision processes in native primary cell cultures of the hippocampus creates conditions in which an already adult neuronal culture without pathologies encounters a forced disturbance of cellular homeostasis. In 5xFAD mice, it is amyloid overproduction that triggers a cascade of reactions affecting cellular and subcellular morphology, mimicking the processes typically seen in AD.

Animal models

The work was performed on animals that are a model of AD: transgenic mice of the 5XFAD (TG6799) line and healthy littermates. Cells in the transgenic culture gradually accumulate high levels of the neurotoxic agent betaamyloid, which causes neurodegeneration in AD, from the beginning of their lives. Mice of the 5xFAD line were obtained on a congenic SJL/C57B16 background to minimize concerns related to allelic segregation and high variability in the original hybrid background. The 5xFAD transgenic mice overexpress the following five FAD mutations: the APP (695) transgene contains the Swedish (K670N, M671L), Florid (I716V), and London (V7171) mutations, and the PSEN1 transgene contains the M146L and L286V mutations. Expression of both transgenes is regulated by murine Thy1 promoter elements to drive their overexpression specifically in neurons. The transgenes are inserted into a single locus Chr3:6297836 where they do not affect any known genes. Animals were kept in a specialized vivarium with free access to water and standardized feed at 22-24°C and natural light. Laboratory animals were treated in accordance with the European Convention for the Protection of Vertebrates used for Experimental and Other Purposes (Strasbourg, 1986) and the principles of the Helsinki Declaration (2000). All animal procedures performed with mice were approved by the Committee on Biosafety and Bioethics (Institute of Cell Biophysics—Pushchino Scientific Center for Biological Research of the Russian Academy of Sciences) in accordance with Directive 2010/63/EU of the European Parliament.
To establish groups of animals for breeding, mice were genotyped by classical PCR using DNA isolated from ear biopsy specimens. The presence of a Tg cassette of 377 bp in length was detected using primers 5'-AGG ACT GAC CAC TCG ACC AG-3' and 5'-CGG GGG TCT AGT TCT GCA T-3', followed by electrophoretic visualization. Newborn mice are subjected to a second genotyping to validate their transgenicity.

Primary culture of brain hippocampus

Mice (aged 0-1 day) were decapitated using a special guillotine for instant and humane slaughter of laboratory rodents. The hippocampus was mechanically crushed and treated with trypsin-EDTA solution (Gibco, USA). The cell suspension obtained by enzymatic and mechanical dissociation was then added to the wells of a 12-well plate coated with poly-D-lysine support substrate (Gibco, USA) or slides, cells were added to 2 or 1 mL Neurobasal Medium (Gibco, USA) containing 2% B-27 supplement (Gibco, USA) and 1% penicillin-streptomycin-glutamine (Gibco, USA), respectively. Half of the medium was replaced with fresh medium every five days. Cells were cultured in a CO₂ incubator at 37°C and 5% CO₂ until the required time. Thus, we obtained both control transgenic (Tg) and non-transgenic (nTg) cultures from the hippocampi of littermate animals. Cultures at 14 days of age were used in the experiments. At this term transgenic cultures already reproduce such characteristic features of AD as a decrease in neuronal density on the background of an increase in astrocyte density, a decrease in the number of synaptic contacts, as well as a significant increase in the concentration of beta-amyloid.

Regulation of mitochondrial fusion and fission processes

Two substances with similar functions but different pathways were selected for mitochondrial fusion shifts. The first substance, Mitochondrial Division Inhibitor 1 (Sigma-Aldrich 475856) (mdivi-1), is a cell-permeable quinazolinone compound that reversibly inhibits dynamin-like proteins (Drp1) (dynamin-related GTPases) responsible for mitochondrial fission. The second substance: Mitochondrial Fusion Promoter M1 (Sigma-Aldrich SML0629) (MFP), is a cell-permeable hydrazone that enhances mitochondrial fusion and does not affect the morphology of the endoplasmic network and lysosomes. Preparation of solutions of substances and application into wells was performed according to the manufacturer's recommendations. In our experiments, we used short (1 day) and long (5 days) observations after Mdivi-1 and MFP administration to evaluate the response and dynamics of changes in the structure of the primary hippocampal culture over time.
Forcing mitochondrial fission seemed to be a much more difficult task. Currently, the commercial market is unable to provide compounds capable of directly affecting mitochondrial fission processes. In our experiments, we used the introduction of zinc chloride solution at high (1 μM) and low (0.1 μM) concentrations into the cultures. Since "overloading" the culture cells with zinc causes too much cellular and mitochondrial stress, which directly affects both mitochondrial and culture morphology, this experiment was performed within a single day. The concentration we chose allowed for effects on mitochondrial morphology without causing immediate cytotoxicity, which is primarily reflected in the integrity of the neurite structure. This experiment reflects a short (1 day) and a long (5 days) follow-up.

Visualization of mitochondria

To visualize mitochondria in cells, we used transduction/transfection of cell cultures with the CellLight BacMam 2.0 Red System. This ready-to-use construct is transduced/transfected into cells using BacMam 2.0 technology, where it expresses a fluorescent protein fused to the E1 alpha-pyruvate dehydrogenase leader sequence. An appropriate volume of reagent is injected into a cell culture of desired density (reagent calculation is based on the number of cells in the culture according to the manufacturer's recommendations) and gently mixed. The cells are returned to the culture incubator overnight and after approximately 16 hours the culture is ready for experiments. Mitochondrial morphology was examined using a Leica DM IL LED microscope (Leica, Germany) with a ×100 oil immersion objective. We found that the fixation of nerve cells disrupts mitochondrial morphology and triggers an immediate cascade of mitoptosis reactions with the displacement of mitochondrial material along with fluorescent agents from the cell, which is especially noticeable when working with astrocytes. Therefore, all mitochondrial morphology data were obtained from live, unfixed cell cultures.

Detection of intracellular lipofuscin

We also used spectrofluorimetric detection of lipofuscin, an autofluorescent aging pigment, as a marker of changes in cellular metabolism. For this purpose, culture plates containing primary hippocampal cultures were washed three times with PBS and poured into a solution of 5% sodium dodecyl sulfate in water for 5 minutes with simultaneous stirring on a shaker equipped with a refrigerant, which allowed to completely destroy cell cultures and transfer them into solution. The resulting suspensions were frozen and used for further experiments. Sodium dodecyl sulfate disintegrates protein complexes and destroys cell membranes, but does not destroy lipofuscin due to its N-substituted imines-Schiff bases. The addition of detergent in our experiments, in addition to chemical dissociation, reduced light scattering and eliminated the effect of Tyndall-Rayleigh hypochromism, directly affecting the registration of spectral features.
The protein concentration in the obtained suspensions was determined and equilibrated by the Lowry method and the UV-express method by the optical density of tryptophan protein in the ultraviolet at 286 nm [16]. UV spectroscopy methods are based on the ability of solutions to absorb light at about 280 nm due to the presence of the amino acids tryptophan, tyrosine, and phenylalanine in proteins. Detection results are directly proportional to the content of these amino acids in proteins. To obtain reliable and accurate results, the optical density values of the solutions must meet the requirements of linearity over the range of protein concentrations to be determined.
Lipofuscin formation was detected by a characteristic fluorescence band at about 450 nm under excitation at a wavelength of 360 nm. Lipofuscin content in homogenates was determined using a Perkin Elmer MPF-44B spectrofluorimeter (USA) in multiway mirror microcuvettes for measuring the fluorescence of weakly absorbing solutions. The cuvettes provide multiple enhancements of fluorescence intensity due to the increased optical path length of the excitation light and the added light collection of the emission. The excitation light enters through a narrow window in the front mirror wall of the cuvette, passes through the solution to the opposite side of the mirror, is deflected, and undergoes two or three reflections within the cuvette. The radiation is collected at right angles. Additional light collection of radiation is provided by a side wall of the mirror that directs the fluorescence into the registration channel. When used, light loss and artifact polarization are minimal compared to a conventional cuvette placed near concave mirrors.

Immunocytochemistry

For immunocytochemical staining for neuronal and astrocyte markers, cell cultures were fixed with 4% paraformaldehyde for 10 minutes. Membrane permeability was then increased with 0.2% Triton X-100 solution, and non-specific binding to antigens was blocked for 1 hour at room temperature in PBST (PBS + 0.1% Tween 20) + 1% BSA solution with 10% normalized donkey serum (Abcam, ab7475) and 5% normalized goat serum (Invitrogen, 31872). The cultures were left overnight at 4°C for incubation with primary antibodies. After incubation with primary antibodies, binding was performed with appropriate secondary antibodies conjugated to fluorescent tags for 2 hours at room temperature. After each treatment, the cells were washed 3 times with PBS solution (pH 7.4) for 5 minutes each. The following primary and secondary antibodies were used to stain neurons, anti-MAP2 antibody (ab32454, 1:200) and corresponding Alexa594nm (ab150080, 1:400); astrocytes, anti-GFAP antibody (ab4674, 1:800); Alexa488nm (ab150169, 1:1000).

Statistical analysis

Statistical analysis of the results was performed using Sigma Plot 12.5 software. Data are presented as mean and standard error of the mean for visual convenience. The hypothesis of normal distribution was tested using the Shapiro-Wilk test. Two-sample Student's t-test and MannWhitney test were used for pairwise comparisons of groups. One-way analysis of variance (ANOVA) with post hoc comparison of groups using Dunn's and Bonferroni's methods was used to compare differences between groups and to determine statistical significance.

Differences between 5xFAD hippocampal cultures and healthy native cultures

We found that astrocyte mitochondria in 5xFAD cultures are more prone to form fused mitochondrial networks (Figure 2A & B), whereas mitochondria in 5xFAD neurons conversely show a tendency to accumulate smaller mitochondria (Figure 2C & D). Primary hippocampal cultures from 5xFAD animals show marked astrogliosis—an overgrowth of the astrocytic network (68.6 ± 1.2 in 5xFAD, 52.3 ± 0.9 in native culture) with a concomitant decrease in neuronal density (37.3 ± 1.0 in 5xFAD, 47.4 ± 1.0 in native culture) as shown in Figure 2E-G. We also found that the intensity level of lipofuscin fluorescence (Figure 2H & I) was significantly higher in 5xFAD cultures than in healthy non-transgenic cultures (571 ± 3 in 5xFAD, 345 ± 3 in native culture).

Effects of MFP on primary neuronal culture

Utilization of the mitochondrial fusion promoter M1 (MFP) in astrocytes converts the mitochondrial network into a pool of fused and hyperfused mitochondria (Figure 3A). A similar pattern is observed in neurons, where fused mitochondria become like elongated sticks or large fused mitochondria (Figure 3B). Figure 3C-E shows a tendency for neuronal density to fall by day 1 of exposure and it falls significantly relative to the control native culture level by day 5 (47.4 ± 1 at native level, 36.7 ± 1.4 at day 5). There was a decrease in astrocytic density by day 1 (52.3 ± 1 in control, 48.6 ± 1.6 at day 1) and a further progresive decrease in density by day 5 (37.9 ± 1.5 at day 5). We also observe an interesting effect of decreasing lipofuscin fluorescence by day 1 (345 ± 2 in native levels, 277 ± 2 at day 1) and day 5 (131 ± 1.5) of fusion activator exposure in Figure 3F & G. Overall, prolonged exposure to MFP resembles a depletion of internal culture resources.

Effects of mdivi-1 on primary neuronal culture

Morphologically, the use of the mitochondrial fission inhibitor mdivi-1 resulted in larger mitochondria being found in neuron bodies and neurites. At the same time, in astrocytes there was fragmentation of the mitochondrial network and instead fused mitochondria were found in greater numbers. We also observed a decrease in the pool of smallest mitochondria, which almost completely disappeared in astrocytes and remained only in neurons (Figure 4A & B). After one day of mitochondrial fission inhibition, neuronal density decreased relative to the control group (47.4 ± 1 at native level, 35 ± 2 at day 1) and did not decrease further by day 5. There was an increase in astrocytic density by day 1, and by day 5 (52.3 ± 1.2 at native level, 58.2 ± 1.7 at day 1, 68.1 ± 1.8 at day 5) as shown in Figure 4C-E. In addition, there was a significant increase in lipofuscin fluorescence correlating with the time of mdivi-1 exposure (Figure 4F & G). Moreover, by day 5 of mitochondrial fission inhibition, the intensity of lipofuscin fluorescence increased such that it was not significantly different from lipofuscin fluorescence in transgenic culture (571 ± 3 in Tg culture, 571 ± 2 at day 5). Thus, inhibition of mitochondrial fission in healthy cultures leads to a cascade of events reminiscent of those in transgenic culture.

Effects of exogenous zinc on primary neuronal culture

We used exogenous zinc chloride solution to modulate mitochondrial fission. In this case, zinc overload at high, rather uncomfortable concentrations for nerve cells, induces cellular and mitochondrial stress affecting mitochondrial morphology, which is consistent with the results of [17]. Primary cultured neurons are the first to suffer from zinc overload at the same time in culture it was observed compensatory astrogliosis (Figure 5C). Treatment of cultures with excessive zinc resulted in neurite destruction and a drop in neuronal density relative to control cultures (47.4 ± 1 in native, 29 ± 1.5 under toxic influence), while there was a strong compensatory astrocyte overgrowth (52.38 ± 1 in native, 62.7 ± 2.2 under toxic influence) and an increase in lipofuscin fluorescence intensity (Figure 5H) (344.9 ± 2.7 in native, 440 ± 1.2 under toxic influence). Using lower concentrations of zinc chloride solution allowed us to achieve the same morphologically detectable mitochondrial fission (Figure 5A & B) in which neurites do not receive such high zinc overload (Figure 5D-F). Low zinc concentrations caused no significant change in neuronal and astrocytic density by day 1, but reduced lipofuscin fluorescence intensity (265 ± 1 at day 1). However, on day 5 (Figure 5F) of the experiment, a fall in neuronal density was already again observed (47.4 ± 1 at native, 27.7 ± 1.3 at day 5) with a concomitant drop in astrocyte density (52.3 ± 1 at native, 35.4 ± 1.8 at day 5), and the cluster morphology of the culture was impaired, neurons became less likely to assemble into clustered structures, and began to show altered immunoreactivity for glial fibrillary acidic protein, and pathologic aggregation of GFAP increased, but lipofuscin fluorescence intensity was still reduced (190 ± 3 at day 5) as shown in Figure 5G & H.

Normal mitochondrial fusion and fission dynamics are the basis for maintaining a healthy mitochondrial pool also through mitophagy, the disruption of which leads to the degradation of mitochondria into lipofuscin granules [18]. First, we would like to talk about lipofuscin. We see that the accumulation of lipofuscin in our experiments occurs in two cases. The first is the toxic effect of high concentrations of exogenous zinc, which is well explained by the activation of peroxidation and, in general, a quite common pathway of lipofuscin increase. The second is the inhibition of mitochondrial fission. Thus, by day 5 of fission inhibition, the fluorescence intensity of lipofuscin in the experimental groups is compared to control 5xFAD transgenic animals. According to [19], the normal mitochondrial fission cycle involves the formation of two daughter mitochondria, one of which has a higher membrane potential and goes on to the fusion-fission cycle, and the other, with a more depolarized membrane, remains separated until the membrane potential is restored or until elimination by autophagy. We can assume that the increase in lipofuscin in this case is a process of accumulation of intracellular debris associated with insufficient mitochondrial division and clearance. Therefore, activation of fission by non-toxic concentrations of zinc, although leading to deleterious effects in culture, decreased the level of lipofuscin. At the same time, activation of mitochondrial fusion, although leading to the degradation of the culture, did not lead to the formation of lipofuscin. On the contrary, it decreased over time. This can be taken to mean that by acting on the fusion promoter, we are not blocking fission, and mitochondria can still maintain fission dynamics. Also, given the close relationship between lipofuscin and peroxidation, we can assume that when mitochondria fuse, cells are not subjected to oxidative stress.
It has been reported that 5xFAD transgenic mice and a reproducible neuro-astrocytic culture from them, which we used as a control, already have a shift in the mitochondrial balance toward excessive fission [20]. Loss of the central mitochondrial fission protein dynamin-related mitochondrial fission protein 1 (Drp1) also increases the toxicity of mutant APP in vivo [21]. In the 5xFAD mouse model, all mutations target the hyperproduction of human Aβ1-42 [22]; it is the overproduction of toxic agents that causes the dysfunction in this model. We assume that the increased mitochondrial fission in this model is secondary and related to the fact that the cells are trying to get rid of toxic amyloid in this way. In our experiments, inhibition of mitochondrial fission resulted in a cascade of reactions characteristic of the 5xFAD culture. In addition to a pathologic increase in lipofuscin, we observed a decrease in neuronal density and a compensatory overgrowth of astroglia, which can be considered a manifestation of astrogliosis in this model. That is, cellular toxin accumulation and mitochondrial fission mediate each other. At the same time activation of cell division did not lead to such consequences. However, we can assume that zinc is a toxic agent in general and its mechanism of operation is a reflection of how the introduction of an external toxic agent leads to an attempt of mitochondria to clean themselves and the cell of contaminants. In doing so, the prolonged action of the fusion promoter resembles the depletion of internal resources in a healthy culture. This process of slow degeneration is broadly similar to one pathway of cell culture degeneration. It loses clustering and shrinks (although there is generally another pathway of culture degradation, abrupt neurosphere clustering, in which thin neuro-astrocytic strands stretch between densely packed neurospheres, linking them to other neurospheres; this is characteristic of primary cultures with high initial stem cell content or their increased level of activity). Mitochondrial fusion is thought to result in enhanced ATP energy supply to the cells, as they have more cristae, increased levels of dimerization and ATP synthase activity, and support ATP synthesis [23, 24]. However, excess ATP leads to neuronal dysfunction and death [25] and is a prerequisite for the realization of the cellular apoptosis cascade [26]. In this case, the mitochondria resulting from fusion may be too large for mitochondrial transport. We find it difficult to say what the mechanism of culture degeneration is in this case. However, the absence of oxidative stress and the drop in lipofuscin levels shows that this process is more likely a controlled extinction of activity rather than an abrupt pathology. It can be stated unequivocally that only harmonious functioning of this system can ensure normal physiological functions of mitochondria, and any shift to one side of the dynamic process leads to disorders.

Forced alteration of mitochondrial fusion and fission leads to morphologic changes in primary neuro-astrocytic cultures of mouse hippocampus. Activation of mitochondrial fusion leads to slow depletion of the culture in the absence of visible degenerative processes and oxidative stress. Inhibition of mitochondrial fission leads to a cascade of events reminiscent of those occurring in a culture of 5xFAD animals (a genetic model of AD)—astrogliosis, decreased neuronal density, and increased lipofuscin levels. Exogenous zinc induces mitochondrial fission and provokes compensatory astrogliosis at toxic concentrations, and at lower concentrations leads to a slow decrease in culture clustering, and changes in astrocyte immunoreactivity to GFAP and its aggregation, whereas lipofuscin fluorescence was still decreased. We suggest that mitochondrial fission in models of toxic mitochondrial damage is a compensatory process in which mitochondria strive to clear the cell of toxic agents, but which may lead to a cascade of reactions that provoke further development of pathologies.