Female Wistar rats (100 ± 5 g) were housed at the Animal Facility of the Palladin Institute of Biochemistry with free access to a standard rodent diet and drinking water ad libitum. Animals were kept in a temperature-controlled room and exposed to a daily 12/12 h light/dark cycle. The general scheme of animal experimental design and timescale are shown in Figure 1. After acclimatization for 1 week, the rats (n = 60) were randomly divided into three experimental groups: (1) the control group (n = 20); (2) the group that received the synthetic glucocorticoid prednisolone (5 mg/kg of body weight, per os, 30 days, n = 20); (3) the group that received prednisolone (5 mg/kg of body weight) and vitamin D₃ (1,000 IU/kg of body weight, per os, 30 days, n = 20). Glucocorticoid prednisolone (30 mg/ml) was purchased from Biopharma (Ukraine) and was solved in the purified water for injections for further oral administration to rats. Vitamin D₃ (cholecaciferol, Sigma-Aldrich, USA, C9756) was administered to rats as oil solution. At the end of the experimental treatment, the rats were anesthetized by intraperitoneal injection of chloral hydrate (40 mg/100 g of body weight), decapitated with a guillotine, and blood was taken from the inferior vena cava. The brain was quickly dissected and transferred to the appropriate buffer depending on the following procedures. All samples that were not used immediately were stored at −80°C.

We performed all animal procedures in accordance with the protocols approved by the Animal Care Ethics Committee of the Palladin Institute of Biochemistry (Protocol #1, 26/01/2018), adopted on the basis of national and international directives and laws relating to animal welfare: European Convention for the protection of vertebrate animals used for experimental and other scientific purposes (Strasbourg, France; 1986), Bioethical expertise of preclinical and other scientific research conducted on animals (Kyiv, Ukraine; 2006).

For histological examination, rat brains were fixed in 10% formalin solution in cold 0.1 M phosphate-buffered saline (PBS, pH 7.4), dehydrated in increasing concentrations of ethanol, and embedded in paraffin (Leica-Paraplast Regular, 39601006, Leica Biosystems Inc., USA) according to the standard procedure. Paraffin sections 5 μm thick were prepared on a Thermo Microm HM 360 Rotary microtome (Microm GmbH, Germany). Sections were deparaffined in xylene and then rehydrated in decreasing concentrations of ethanol (100%, 95%, 80% and 70%) and distilled water, followed by staining for 5 min with 0.05% toluidine blue (pH 4.5) using the Nissl technique (García-Cabezas et al., 2016) or hematoxylin and eosin (H&E; 3 min for hematoxylin and 45 s for eosin). Ten sections from hippocampus, prefrontal cortex, cerebellum, and thalamus of each animal were taken in a systematic randomized manner and covered with Histofluid coverslips (6900002, Paul Marienfeld EN, Germany). The sections were examined under a light microscope (Olympus BX 51, Japan) at ×400 magnification. Digital photographs were taken using CarlZeiss software (AxioVision SE64 Rel.4.9.1). The cells with morphological signs of cytolysis, hyperchromatosis, karyolysis, and karyopyknosis were considered affected.

Purified synaptosomes were obtained from Wistar rat brains by the differential centrifugation (Kasatkina et al., 2018) using Ficoll-400 density gradient centrifugation of crude samples. Synaptosomes were used within 2–4 h after isolation.

Synaptic vesicles (SVs) were isolated from the cerebral hemispheres according to the following method (De Lorenzo and Freedman, 1978). Crude (unpurified) synaptosomes were lysed with a buffer containing 1 mM EGTA, 10 mM Tris-HCl, pH 8.1 (3 ml/g of brain tissue) and then incubated at +4°C for 60 min. The samples were centrifuged at 20,000 g for 30 min, followed by the next supernatant centrifugation step at 55,000 g for 60 min. The last centrifugation step was carried out at 130,000 g (+4°C, 60 min) to obtain the SVs fraction (pellet) and the synaptosomal fraction of cytosolic proteins (supernatant). The SV pellet was resuspended in 10 mM Tris-HCl (pH 7.5). The purity of the SV fraction was assessed by measuring the activity of Na⁺/K⁺-ATPase, a plasma membrane marker that was not detected in the purified SV fraction.

The isolated synaptic terminals (synaptosomes) were preincubated at +37°C for 10 min and fluorescence (F) was measured on a Hitachi 650-10S spectrofluorimeter (Japan) at excitation and emission wavelengths of 490 and 530 nm (2 nm slit), respectively. High K⁺ (35 mM) depolarization of the plasma membrane was achieved in Ca²⁺-containing medium at the steady-state of acridine orange (AO) fluorescence. The traces were normalized to the similar data in the absence of synaptosomes:

where F₀ is the fluorescence intensity measured upon AO quenching in the solution, and F_t is the fluorescence intensity of AO in the presence of synaptosomes.

We employed the octadecyl rhodamine B chloride (R18) technique to detect SV fusion with target synaptic plasma membranes. The fusion was recorded on a Hitachi 650-10S spectrofluorimeter (Japan) in a suspension of R18-labeled SVs (1 nmol/240 μg of synaptosomal protein) and unlabeled plasma membranes in a ratio of 1:8 (according to protein concentration) after addition of Ca²⁺ (Trikash et al., 2010). An increase in the fluorescence signal was recorded as a result of membrane fusion and R18 dequenching. To quantify the probe dilution during the fusion of vesicles with target membranes, the initial fluorescence signal of R18-labeled SVs (I_o) was measured. The time-dependent dequenching of the fluorescence signal (I) was recorded within 4 min after the start of testing. At the end of each experiment, C₁₂E₈ detergent was added to the reaction mixture at a final concentration of 0.1%, at which the maximum value (Id) of the dequenched fluorophore signal was obtained. The fusion rate at any given time is proportional to the percentage of fluorescence dequenching (%FD) and thus can be calculated as:

Total RNA was isolated from rat brain using the innuPREP RNA Mini Kit (Analytik Jena AG, Germany). The mRNA concentration was determined on a DS-11 Spectrophotometer/Fluorometer (DeNovix, USA). The Maxima H Minus First Strand cDNA Synthesis Kit (Thermo Fisher Scientific Inc., USA) was used to synthesize cDNA samples for subsequent RT-PCR on a Standard real-time PCR Thermal Cycler (Analytik Jena AG, Germany). Specific primer sequences for Vdr, Vdbp, Cyp27b1, Cyp24a1, Nf-κb, Iκb-alpha, and the glyceraldehyde 3-phosphate dehydrogenase (Gapdh) reference gene were designed using Primer BLAST software and used at a working concentration of 10 μM:

Target genes were amplified for 60 cycles using Maxima SYBER Green/ROX qPCR Master Mix (Thermo Fisher Scientific Inc., USA). Relative mRNA expression calculations were performed according to the 2^–ΔΔCt comparison method. The expression level of each gene was normalized for GAPDH in the same samples and then calculated as a fold change compared to the control.

Using Western blotting, we measured the following protein levels in total brain tissue lysates. First, VDR was examined to evaluate the effect of chronic administration of prednisolone on the main component of the vitamin D auto/paracrine system. 3-Nitrotyrosine, carbonylated and poly(ADP)-ribosylated proteins (PAR), as well as PARP-1 attracted our attention as reliable markers of oxidative-nitrosative and genotoxic stress in brain tissue. Finally, the glucocorticoid receptor (GR), NF-κB p65 phosphorylated at Ser311, and IκB were measured to characterize the relationship of VDR and GR with NF-κB transcriptional activity. Total protein extracts were prepared from frozen brain tissue according to a standard protocol using RIPA buffer (20 mM Tris-HCl, pH 7.5; 150 mM NaCl; 1% Triton X-100; 1 mM EGTA; 0.1% SDS; 1% sodium deoxycholate; 10 mM sodium pyrophosphate) and a protease inhibitor cocktail (PIC, Sigma, USA), and then the samples were sonicated and centrifuged. The supernatants were stored at −80°C until required. Lysate samples containing 60 μg of protein were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using a 10–15% gel followed by protein transfer to a nitrocellulose membrane. The membrane was then blocked for 1 h with 5% skimmed milk diluted in Tris-buffered saline with Tween (TBST, 150 mM NaCl, 10 mM Tris-HCl, and 0.1% Tween; pH 7.5) and incubated overnight at +4°C with one of the following primary antibodies (VDR: 1:200, sc-13133, Santa Cruz; 3-nitrotyrosine: 1:2,500, #05-233, Merck Millipore; carbonylated proteins: 1:1,000, ab-178020, Abcam; poly(ADP)-ribosylated proteins: 1:1,000, 4335-MC-100_AC, Trevigen; PARP-1: 1:1,000, #9542, Cell Signaling; GR: 1:250, sc-1004, Santa Cruz; NF-κB p65 phosphorylated at Ser 311: 1:200, sc-33039, Santa Cruz; and IκB-α: 1:1,000, sc-847, Santa Cruz). Subsequently, the membrane was thoroughly rinsed and incubated for 1 h with secondary antibodies: anti-rabbit IgG (H + L)-HRP conjugate (1:4,000, #1721019, Bio-Rad) or anti-mouse IgG (Fab Specific)-Peroxidase (1:2,500, A9917, Sigma) at room temperature. Thereafter, the membrane was developed with chemiluminescent agents: p-coumaric acid (Sigma, USA) and luminol (AppliChem GmbH, Germany). Target protein immunoreactive signals were quantified by densitometry and values were adjusted for the corresponding β-actin level (1:10,000, A3854, Sigma). Immunoreactive bands were quantified using Gel-Pro Analyzer v3.1 software.

Using a commercial ELISA kit ≪25-OH-Vitamin D₃≫ (Immunodiagnostics, Germany), in accordance with the manufacturer’s instructions, the VD₃ status was assessed by determining the concentration of 25-hydroxyvitamin D₃ in the serum, cerebrospinal fluid and brain homogenate of rats. Cerebrospinal fluid was obtained by puncture of the atlanto-occipital space. Brain homogenates were prepared from frozen brain tissue (100 mg) that was ground in a porcelain mortar with liquid nitrogen, lysed in PBS with 1% Tween-20 (pH 7.4), then sonicated and incubated on ice for 2 h. Next, the samples were centrifuged at 14,000 g for 20 min at +4°C. The supernatant was carefully collected and aliquots were stored at −20°C avoiding thawing/freezing.

The results of all experiments were expressed as the mean ± standard deviation for at least seven rats per group. The hypothesis about the normality of data distribution was tested using the Shapiro–Wilk test. Differences between means were estimated using the ANOVA followed by Tukey’s post hoc test. To compare difference between groups in neuronal morphometric parameters, statistical significance was determined using a two-tailed unpaired Student’s t-test. For behavioral data, the non-parametric Kruskal–Wallis test was used, if necessary, followed by Dunn’s post hoc test. The significance criterion was p ≤ 0.05. Statistical analysis was performed using Origin Pro 8.5 software (OriginLab Corporation, Northampton, MA, USA). For behavioral datasets, statistical significance was analyzed using the Origin 9 and Graph Pad Prism 6 software.

The brain-to-body weight ratio is commonly used in neurotoxicity studies to estimate the presence of general morphopathological changes in the brain. As can be seen from Table 1, the administration of prednisolone slowed down the growth of rats and reduced the average body weight by 70.65% compared with the control group (Tukey’s test, p = 0.000051), while there was no significant difference in the average brain mass between the two groups (Table 1, Tukey’s test, p > 0.05). An assessment of the brain-to-body weight ratio revealed that GC slightly increased it by 29.04% compared with the control group (Tukey’s test, p = 0.0031). Treatment with VD₃ enhanced mean body weight by 15.97% compared with the prednisolone group (Tukey’s test, p = 0.033), however, it did not lead to any ameliorative effect on the brain-to-body weight ratio compared with the prednisolone group (Tukey’s test, p > 0.05).

In the context of the essential role of vitamin D₃ for the CNS functioning, the question arose whether chronic GC treatment could affect the circulating VD₃ pool and alter the content of the main components of the VD₃-auto/paracrine system. First, we measured the level of 25-hydroxyvitamin D₃ as a reliable marker of VD₃ bioavailability in serum, cerebrospinal fluid, and brain homogenates. Figure 2A shows a pronounced drop in 25OHD₃ in the serum of rats supplemented with GC (by 3.24-fold, up to 30.00 ± 3.04 nmol/L) compared with the control animals (97.23 ± 7.01 nmol/L, Tukey’s test, p = 0.000001). In the CSF, we observed half the content of 25OHD₃ in the prednisolone group (16.00 ± 2.05 nmol/L) than in the control group (33.00 ± 2.50 nmol/L, Figure 2A). Finally, depletion of the circulating VD₃ pool was correlated with a significant decrease in the 25OHD₃ level in the brain tissue extracts, indicating impaired bioavailability of the VD₃-prohormone in the nervous tissue. Figure 2B demonstrates that the 25OHD₃ level after prednisolone supplementation was 8.64 ng/g brain tissue compared to 18.60 ng/g brain tissue in the control animals, representing a 2.15-fold decrease (Tukey’s test, p = 0.022).

Vitamin D₃ supplementation caused a partial restoration of the 25OHD₃ content. To a greater extent, the normalizing effect concerned the serum, where the 25OHD₃ level rose to 111.43 ± 13.17 nmol/L and there was no significant difference with the control group. In the CSF, we also observed a tendency to restore the level of 25OHD₃, the value of which reached 22 ± 3.1 nmol/L. This is 37.5% higher than in the prednisolone group. The 25OHD₃ level also increased in the brain extract to 12.88 ± 0.9 ng/g of tissue, which was 49% higher than in the prednisolone group (Tukey’s test, p = 0.042), but remained still lower (by 44.4%) than in the control. Collectively, these data indicate a profound GC-induced VD₃ deficiency and depletion of the circulating VD₃ pool under the influence of prednisolone, however, vitamin D₃ treatment may partially restore the circulating VD₃ pool, suggesting that this abnormality is reversible.

Since the hormonally active form of VD₃ in different cell types exerts its biological action through specific receptors for 1,25(OH)₂D₃ – VDR, we also examined the VDR expression in the brain at the transcriptional and translational levels (Figures 2C, D). It was shown that after prednisolone administration, the level of Vdr mRNA in the brain increased by 6.45 times (Tukey’s test, p = 0.000003) (Figure 2D). We further confirmed an increase in VDR protein by western blot (Figures 2C, D), which was 41% higher than in the control group (Tukey’s test, p = 0.005). In turn, a strong reducing effect of VD₃ on the expression of Vdr mRNA (by 2.14-fold, Tukey’s test, p = 0.000001) and protein (by 3.44-fold, Tukey’s test, p = 0.000002) was found compared with the prednisolone group.

Despite the observed disturbances in the VDR expression, the next question arises whether prednisolone is able to affect the VD₃ transport. VDBP is considered to be the major plasma transporter for all VD₃ metabolites; however, its function in the brain is not fully understood. Therefore, we examined the level of Vdbp mRNA and found that prednisolone caused a 42%-lowering effect (Tukey’s test, p = 0.002) compared to the control, while VD₃ supplementation restored its level to the control value (Figure 2E).

To complete this part of the study, we focused on enzymes involved in VD₃ metabolism and considered important factors that contribute to the proper functioning of the VD₃-auto/paracrine system. As presented in Figure 2F, prednisolone induced a significant (5.46-fold, Tukey’s test, p = 0.00002) increase in Cyp27b1 mRNA expression in the brain compared with the control animals, which may reflect a compensatory response of the VD₃-auto/paracrine system to the lack of circulatory and brain levels of 25OHD₃. Vitamin D₃ treatment reduced Cyp27b1 gene expression to a level that was 9.72-fold lower than in the prednisolone group (Tukey’s test, p = 0.00001) and 1.79-fold lower than in the control group (Tukey’s test, p = 0.0029).

Another component of the VD₃-auto/paracrine system that we examined was CYP24A1, that catalyzes the inactivation of both 25OHD₃ and 1,25(OH)₂D₃ (Hamamoto et al., 2006). In contrast to the effect of prednisolone on Cyp27b1, this glucocorticoid caused a 5.41-fold decrease in Cyp24a1 mRNA level compared with the control group (Tukey’s test, p = 0.0005), while VD₃ did not significantly affect the level of Cyp24A1 mRNA compared with the prednisolone group (Tukey’s test, p > 0.05, Figure 2F).

In general, our results suggest that, despite the development of VD₃-deficiency, prednisolone elevated, most likely compensatory, the expression of such important components of the VD₃-auto/paracrine system as CYP27B1 and VDR, while the levels of mRNA transcripts of CYP24A1 and VDBP genes decreased in the brain, thus indicating an impairment of VD₃ transport and catabolism. Administration of cholecalciferol restored the circulating 25OHD₃ pool and normalized, at least in part, the levels of key components of the VD₃-auto/paracrine system in the brain, although VD₃ 24-hydroxylase remained unchanged.

To support the concept that VD₃ deficiency may be associated with behavioral abnormalities (Schoenrock and Tarantino, 2016), we further estimated prednisolone-induced behavioral changes in animals that we found to exhibit significantly reduced VD₃ bioavailability. The open-field test allows to examine motor function and exploratory behavior by measuring spontaneous activity in the OFT box. Animals of all groups showed no significant difference in walking distance (p > 0.05) and in time spent exploring 9 inner squares and 16 outer squares (p > 0.05), as presented in Figures 3A–C respectively.

The forced swim test has been used to assess the presence of depression-related behaviors (Slattery and Cryan, 2012). Prednisolone-administered rats demonstrated an increase in immobility time (220 ± 15 s, Figure 3D) and a decrease in active behavior (swimming 30 ± 3 s, Figure 3E, climbing 50 ± 6 s, Figure 3F) compared with the control rats (160 ± 12 s, 60 ± 5 s, 80 ± 9 s, respectively), indicating depressive-like changes. VD₃ exerted antidepressive-like action, diminishing immobility time (150 ± 11 s) and stimulating active behavior (swimming 65 ± 5 s, climbing 85 ± 8 s) compared with the GC action.

The next task was to determine the level of anxiety in the elevated plus maze test by measuring the amount of time spent in open/closed arms. Behavior in performing this task (activity in the open arms) reflects a conflict between rodent’s preference for protected areas (closed arms) and their innate motivation to explore novel environments. In the EPM test, we observed a tendency to an increased percentage of entries into the open arms (38.33 ± 5.9%, Figure 4A) and time spent in the open arms (14.92 ± 3.16%, Figure 4B) after prednisolone administration compared with the control group (21.03 ± 3.9% and 7.3 ± 1.27%, respectively). Unexpectedly, these results showed that prednisolone did not cause the anxiety-related behavioral changes. Increased time spent in the open arms and percentage of entries are conventionally interpreted as decreased anxiety-like behavior (Shoji and Miyakawa, 2021), however, another possible interpretation can be considered: an open arm study may reflect heightened anxiety and panic response to a novel situation under certain conditions. VD₃ administration reduced the percentage of entries into the open arms (13.88 ± 3.4%, p = 0.0272) and time spent in the open arms (5.92 ± 2.18%, p = 0.0306) compared to the prednisolone-administered group. In addition, there was no meaningful difference when comparing the EPM parameters of VD₃-treated rats with the control group and the prednisolone group. Interestingly, time spent in the open arms of the EPM was positively correlated with immobility time in the FST (r = 0.296, p < 0.05) and negatively correlated with active behavior (r = −0.382, p < 0.01 for swimming).

After completion of the OFT, FST, and EPM tests, the rats were subjected to a contextual/cued fear conditioning test to assess an associative fear learning and memory. As shown in Figures 4C, D, there was no statistically significant difference in spontaneous freezing between all groups during the training day in the novel context (p = 0.65, Figure 4C) and after an auditory tone cue (p = 0.99, Figure 4D). We also did not observe significant differences between groups in the level of spontaneous freezing during the day of testing in both contextual and cued conditioning. However, we noted a tendency that the time of contextual conditioning freezing was significantly shorter in prednisolone-administered rats (2.63 ± 0.64 s vs. 4.44 ± 1.33 s in the control). A separate analysis of freezing time between each group revealed the statistically significant differences between training and testing days for three groups in contextual conditioning (Figures 4C, p = 0.049). However, no valuable difference was observed in cued conditioning freezing between all 3 groups (Figures 4D, p = 0.068).

Thus, we did not find any significant effects on locomotor activity or memory parameters (based on OFT and contextual/cued conditioning), while FST and EPM tests reflect the behavioral changes following prednisolone administration. In particular, FST demonstrated an increase in immobility time and a decrease in active behavior in glucocorticoid-treated rats, indicative of depressive-like changes. At the same time, the EPM results may point to the stimulation of anti-anxiety behavior induced by prednisolone; however, taking into consideration an increased time of immobility in the FST, we are more inclined to believe that the data obtained could reflect a panic-like reaction. Vitamin D₃ treatment diminished the immobility time and stimulated active behavior, suggesting anti-depressive effect of cholecalciferol. Exploratory behavior in the EPM was restored to control levels after VD₃ action.

Because lesions in the hippocampus are known to be capable of disrupting contextual fear conditioning (Chen et al., 1996), and given the observed GC-induced tendency to reduce freezing time in contextual conditioning, we first assessed the presence of histopathological changes in the hippocampus. For this purpose, histological staining with toluidine blue was carried out, followed by morphometric and statistical evaluation of the CA1–CA3 regions.

Histological examination revealed no visible signs of brain damage (neither hydropic neuronal dystrophy nor neurodegeneration) in all three groups (Figure 5A). Analysis of the morphometric parameters of pyramidal neurons in fields CA1–CA3 confirmed the presence of the following changes (Table 2). The perikaryon area increased by 55.1% in CA1, by 67.6% in CA2, and by 29.0% in the CA3 region in the prednisolone group compared with the control group (Student’s t-test, p < 0.001). In addition, after prednisolone administration, we observed intact, but at the same time, enlarged neuronal nuclei (by 62.55, 83.45, and 49.27% in CA1, CA2, and CA3, respectively) compared with the control (Student’s t-test, p < 0.001). The nuclei contained increased levels of euchromatin, indicating the activation of biosynthetic processes in neurons. Prednisolone also caused a slight decreasing effect on cell density in the CA1–CA3 hippocampal fields (Figure 5A). Vitamin D₃ administration led to a significant decrease in the mean areas of both the perikaryon and the nucleus in CA1 compared with the prednisolone group (by 68.12 and 73.79%, respectively, Student’s t-test, p < 0.001), while only a slight tendency to a reduction of these parameters in CA2 and CA3 sections was found.

Histological sections of the cerebral cortex demonstrate the preservation of the general morphological structure of the prefrontal and sensorimotor cortex (Figures 5B, C). In all studied samples, pyramidal neurons and glial cells were detected in all layers of the cerebral cortex, except for the molecular layer. Structurally intact microvessels, mostly hemocapillaries, were also present. We did not reveal any signs of inflammation, leukocyte infiltration, necrosis or hemorrhage (Figures 5B, C). However, in the groups of prednisolone and VD₃, clarification of the cytoplasm was observed in large pyramidal neurons, especially in the ganglionic cell layer, probably indicating impaired functional activity of cells, inhibition of synthetic processes, and disintegration of organelles of the protein-synthesizing apparatus of neurons. We did not detect a statistically significant difference in the number of neurons in the prefrontal and sensorimotor cortex between the studied groups (Table 3). When comparing the mean area of neurons in the ganglionic layer of the cerebral cortex, there was a tendency to its increase in the prefrontal cortex after prednisolone administration compared with the control (by 13.4%, Student’s t-test, p = 0.04). In the sensorimotor cortex, a trend toward an increase in the mean area of neurons was found after VD₃ treatment compared with the control (by 18.1%, Student’s t-test, p = 0.03) and prednisolone (by 24.1%, Student’s t-test, p = 0.035) groups.

Next, we studied the effect of prednisolone and cholecalciferol on the posterior thalamic nucleus. Histological evaluation did not reveal signs of dystrophic changes, necrosis or inflammation in the thalamus of all experimental groups. We also did not detect neuronal necrosis, however, single hyperchromic gliocytes and neurons were found. In general, the histological structure of the thalamic region in all the studied groups was similar (Figure 5D). These observations were confirmed by morphometric analysis, and there was no statistically significant between-group difference, as indicated in Table 4.

Finally, we found a number of morphological changes in the cerebellar cortex caused by the prednisolone action, with a slight ameliorative effect of VD₃. In general, for all experimental groups the cytoarchitecture of the cerebellar cortex was well preserved; the molecular, ganglionic, and granular layers were clearly visible (Figure 5E). In the molecular layer, single nuclei of gliocytes and processes of neurons were detected, which together formed the outer layer of the cerebellar cortex. Separate piriform neurons (Purkinje cells) in the ganglionic layer were located in a one-layered manner. The granular layer was formed by numerous small granular neurons. Thus, we did not find any neurodystrophic changes in the molecular and granular layer after prednisolone administration, while in the ganglionic layer there were strained neurons with hyperchromic staining of the cytoplasm and nuclei. Although we did not observe intergroup differences in the number of neurons and density of Purkinje cells in the cerebellar cortex (Table 5), the size of strained neurons was 43.2% and 33.2% smaller in the prednisolone group (406.1 ± 14.7 μm², Student’s t-test, p = 0.009) and after VD₃ treatment (477.7 ± 27.8 μm², Student’s t-test, p = 0.012), respectively, compared with the control group (714.9 ± 48.3 μm²). These observations indicate that neurodegenerative changes may occur under the influence of prednisolone in the cerebellar cortex.

Thus, we did not establish any valuable differences in the cytoarchitectonics of the cerebral cortex and the posterior nucleus of the thalamus between the studied groups. However, GC-induced changes were found in the hippocampus and cerebellum, namely, a decrease in cell density in the hippocampal zones, as well as a decrease in the size of neurons in the cerebellar cortex. These changes in general may indicate the onset of neurodegenerative changes under the influence of prednisolone. Vitamin D₃ partially normalized the cytoarchitecture of the hippocampus and cerebellum.

It is most likely that despite the absence of striking prednisolone-induced histopathological changes in the brain structures they, nevertheless, can be accompanied to some extent by impaired neuronal function. The harmful effects of GC on the brain and the impact of vitamin D₃ therapy may primarily concern some important aspects of synaptic function, such as depolarization-induced exocytosis from isolated neurons, synaptic vesicle fusion with plasma membrane of nerve terminals and changes in long-term potentiation (LTP).

First, we characterized the process of depolarization-induced exocytosis from isolated neurons and monitored vesicle-to-membrane fusion. As shown in Figures 6A, B, following high K⁺-depolarization, the nerve endings isolated from the brain of prednisolone-administered rats demonstrated a 57.5%-decrease in exocytosis amplitude compared with the control (Tukey’s test, p = 0.00023).

One of the reasons for the decrease in the exocytosis amplitude might be changes in the mechanisms involved in providing the fusion of synaptic vesicles with the plasma membranes of synaptic terminals. This process strictly depends on the presence of soluble cytosolic synaptic proteins and involves successive stages of interaction of synaptic vesicles with the target membrane: docking, priming, and fusion. We isolated synaptic vesicles and synaptic plasma membranes and evaluated their Ca²⁺-triggered fusion in a cell-free system. As is seen in Figure 6C, synaptic vesicles from prednisolone-administered rats show a significantly lower rate of Ca²⁺-dependent synaptic plasma membrane fusion (by 29.53%) compared to the control (Tukey’s test, p = 0.01).

Glucocorticoids have been reported to impair cholesterol metabolism in the brain, and its level is one of the important factors influencing the rate of synaptic vesicle fusion with each other and with the synaptic plasma membrane (Gumenyuk et al., 2013). With this in mind, we studied the content of total cholesterol in synaptic membranes isolated from the cerebral hemispheres and found an increase in its level in prednisolone-supplemented animals (0.81 ± 0.07 μmol/mg protein) compared to the control level (0.65 ± 0.05 μmol/mg protein) (Figure 6D). Elevated synaptic plasma membrane cholesterol may indicate abnormal cholesterol metabolism associated with the administration of prednisolone and is comparable to that observed in alimentary vitamin D deficiency (Kasatkina et al., 2018).

After supplementation with VD₃, we observed an increase in mean amplitude of 103.07% compared with the prednisolone group (Tukey’s test, p = 0.00002, Figure 6B), which almost reached the level of control values. Recovery of depolarization-induced release of synaptic vesicles after VD₃ treatment (Figure 6A) indicates the reversibility of prednisolone-induced changes in stimulated secretion. Contrary to our expectations, we failed to find statistically significant changes in cholesterol levels after VD₃ supplementation compared with the prednisolone group (p > 0.5, Figure 6D).

In the next series of experiments, we studied changes in long-term potentiation in animals in response to chronic prednisolone action and after vitamin D₃ administration. For the LTP study, we used a common tetanic stimulation protocol, which is classic for studying plasticity in the hippocampal CA1 area (100 stimuli, at a frequency of 100 Hz). In this experiment, we used sections of the septal part of the hippocampus involved in spatial memory processes. Analysis of the LTP parameters (Figure 7A) showed a decrease in fEPSP by 16.6% after GC administration compared with the control (p = 0.0025; Figure 7B) that reflects an alteration of the LTP. Vitamin D₃ treatment restored the prednisolone-induced LTP (prednisolone vs. vitamin D₃, p = 0.033) to nearly control levels (VD₃ vs. control, p = 0.56).

Thus, in relation to functional disorders, we have established a GC-induced decrease in the amplitude of high K⁺-stimulated exocytosis, the rate of Ca²⁺-dependent fusion of isolated synaptic vesicles with synaptic plasma membranes that may be partially associated with an increase in cholesterol levels in the latter. A reduced level of LTP was found in the synapses of the hippocampus. Taken together, all these impairments may underlie the observed changes in behavior after glucocorticoid loading. In turn, vitamin D₃ therapy tended to restore, at least partially, the above changes, regardless insignificant effect on cholesterol levels.

Excess formation of reactive oxygen/nitrogen species (ROS/RNS) is considered to be one of the universal mechanisms leading to cell injury and CNS dysfunction (Oosthuizen et al., 2005). It is known that nitric monooxide (NO) can interact with superoxide-anion radicals to form an unstable and highly reactive peroxynitrite (ONOO-). Tyrosine residues of protein molecules are sensitive targets for the action of peroxynitrite. Thus, the content of 3-nitrotyrosine can serve as a specific marker of peroxynitrite activity and the development of oxidative-nitrosative stress (Bartesaghi and Radi, 2018). We have shown that prednisolone elevated protein nitration at tyrosine residues in the brain tissue by 28% (Tukey’s test, p = 0.0052) compared with the control (Figures 8A, C). This reflects NO-mediated protein damage that may lead to neuronal dysfunction. VD₃ administration decreased the level of nitrated proteins by 61.7% compared with the prednisolone group (Tukey’s test, p = 0.0005) and by 21% compared with the control (Tukey’s test, p = 0.013). This is clearly consistent with previous data on the antioxidant activity of cholecalciferol (Martín Giménez et al., 2021).

Next, we assessed the degree of oxidative modification of proteins based on the content of carbonyl groups (>C=O), which are formed within protein molecules upon ROS-mediated oxidative modifications of aminoacid residues, especially Pro, Arg, Lys, Trh, as well as a result of the interaction of macromolecules with reactive carbonyl products or reducing sugars (Hauck et al., 2019). Immunoblot analysis of carbonylated proteins did not reveal, in contrast to the level of nitrated proteins, a statistically significant difference between the three groups, as shown in Figures 8B, D. In our opinion, this can be explained by the augmented neutralization of nitric monooxide by superoxide anion radicals with the formation of peroxynitrite, which, in turn, can cause an increase in the content of nitrated proteins in the brain under the GC influence (Figure 8C).

Intensification of ROS and peroxynitrite formation, in addition to protein oxidative modification, can also lead to DNA damage, inducing the development of genotoxic stress. In turn, the appearance of single- and double-strand breaks in DNA activates one of the key enzymes of the cellular DNA repair system, PARP-1, which provides covalent modification of histone and non-histone proteins in the process of their poly(ADP)-ribosylation (Wang et al., 2019). As PARP-1 overactivation was previously reported to be detrimental in multiple types of experimental brain injury, including the impact of oxidative stress (Salech et al., 2020), the brain levels of protein poly(ADP)-ribosylation were studied. We found a significant increase (by 45%, Figures 8E, G) in the level of PAR after prednisolone administration (Tukey’s test, p = 0.00055). As shown in Figures 8F, H, the PARP-1 protein level did not change compared to the control (Tukey’s test, p = 0.537) that, however, do not exclude an increase in the overall activity of this enzyme. Interestingly, when comparing rats supplemented with prednisolone and prednisolone plus VD₃, there were no intergroup differences in PAR (Tukey’s test, p = 0.44), suggesting that VD₃ does not provide effective inhibition of this protein modification. However, VD₃ supplementation diminished PARP-1 content by 25.3% (Tukey’s test, p = 0.015) compared with the prednisolone group, thus affecting protein levels rather than enzyme activity.

Collectively our data suggest that prednisolone caused the development of oxidative-nitrosative stress in brain tissue that may contribute to glucocorticoid-induced neurotoxicity. The ability of VD₃ to prevent protein nitration and to reduce PARP-1 level was established.

Glucocorticoids act through their particular receptor, and in rodents its expression is ubiquitous throughout the brain (Kadmiel and Cidlowski, 2013). It was important to determine whether GR is involved in prednisolone-induced neuropathological changes. Unexpectedly, we observed a significant increase (by 80%, Figures 9A, B) in the GR level after prednisolone administration compared with the control group (Tukey’s test, p = 0.01). This may either reflect or contribute to impaired GR-signaling and partially explain the neurotoxic effects of long-term prednisolone administration.

Since NF-κB is reported to be a redox-sensitive transcription factor (Morgan and Liu, 2011), and in addition, GR may interact with NF-κB p65 subunit through direct protein-protein interactions (Ray and Prefontaine, 1994; Nelson et al., 2003), thus influencing its transcriptional activity, we assessed the components of the NF-κB system. Using RT-PCR (Figure 9C), we established a significant 4.23-fold increase in Nf-κb p65 mRNA compared with the control rats (Tukey’s test, p = 0.0009). The same elevating effect was observed for the NF-κB p65 phosphorylated at Ser311 protein level, which is considered to be transcriptionally competent (Figures 9A, C). The level of phosphoNF-κB p65 was increased by 78.3% after prednisolone compared with the control group (Tukey’s test, p = 0.008).

An important role in the regulation of NF-κB p65 activity belongs to the NF-κB inhibitor, which binds to NF-κB dimers and keep them inactive in the cytoplasm (Solt and May, 2008). We studied IκB expression in rat brain and found conflicting results concerning effects of GC on IκB mRNA (Figure 9D) and IκB protein (Figures 9A, D) levels. With a strong decrease in the gene expression of IκB under the GC action by 88.2 ± 5.5% (Tukey’s test, p = 0.001), the level of IκB protein was 1.6-fold elevated compared with the control (Tukey’s test, p = 0.015). This finding suggests GC-induced dysregulation of transcriptional and translational processes in nerve cells. Vitamin D₃ treatment restored the GR to control values (Figures 9A, B). In addition, VD₃ almost completely normalized Nf-κB mRNA expression and significantly reduced the protein content of phosphoNF-κB p65 (by 66.3% compared vs. prednisolone, and by 40% compared vs. control), suggesting VD₃ ability to counteract prednisolone-induced NF-κB activation in brain tissue. At the same time, we also observed contradictory effect of VD₃ supplementation on IκB. It increased IκB mRNA (by 9.4-fold compared vs. prednisolone), while decreasing IκB protein level (by 65.6% compared vs. prednisolone, and by 45% compared vs. control). This uncoordinated action of vitamin D₃ on IκB expression at transcriptional and translational levels renders the entire pattern of NF-κB activation unclear, indicating the need for further study.

To conclude, at the molecular level prednisolone caused an increase in the expression of the p65 subunit of NF-κB and an increase in its specific phosphorylation at Ser311 against the background of an increase in the IκB protein content. VD₃ diminished the intensity of oxidative-nitrosative stress, possibly by suppressing the activation of NF-κB signaling in the brain.