Cognitive dysfunction refers to specific changes in a human’s mental abilities, including memory impairment, intellectual disability, attention deficits, language disorders, and other impairments. The most severe cases of irreversible cognitive decline fall into the category of dementia [1]. Among various dementia disorders, Alzheimer’s disease (AD) manifests the most severe clinical presentation [2]. The dominant hypothesis attributes its development to progressive amyloid-β peptide aggregation in brain structures [3]. AD is accompanied by mitochondrial dysfunction, which can arise from the direct effects of amyloid-β on mitochondrial proteins and membranes. It is known that there is a decrease in the activity of the Krebs cycle, mitochondrial respiratory chain, and ATP production in the brains of patients with AD [4]. Neuropathological examinations reveal significant mitochondrial genome abnormalities in Alzheimer’s disease. Quantitative analysis shows that cortical mitochondrial DNA (mtDNA) in AD patients below 75 years contains 15 times more deletions than neurologically intact individuals of comparable age [5]. Deletions in mtDNA result in a deficiency of cytochrome c oxidase in the brains of patients with AD [6]. Patients have 63% more mutations than healthy individuals, and notably, these mutations tend to accumulate in the regulatory regions of mtDNA [7]. The number of oxidized bases, especially 8-oxoguanine, was also approximately ten times higher in patients with AD compared to control subjects. Similarly, mtDNA had ten times more oxidative damage than nuclear DNA [8].

Proteolytic processing of APP by β-secretase and γ-secretase generates amyloid-β peptides, with presenilin proteins (PSEN1/2) forming the active site of γ-secretase [9]. AD predominantly (>90%) occurs as sporadic late-onset disease, though rare familial cases result from pathogenic variants in APP or presenilin genes [10]. Although laboratory animals, particularly mice, do not naturally develop AD, transgenic models have been created in which human mutant APP and PSEN1/2 genes are expressed [11]. In the APP/PS1 mouse model, progressive amyloid-β aggregation triggers neuroinflammatory cascades and ultimately leads to cognitive dysfunction [12]. Research indicates that these genetically modified mice develop mitochondrial abnormalities comparable to those found in individuals with AD [13]. It has also been shown that a large number of mitochondrial DNA deletions accumulate in the brains of mice in an AD model expressing an inducible mitochondrial-targeted endonuclease [14]. However, the accumulation of mutations, particularly mtDNA deletions, is not always preceded by the accumulation of oxidative damage [15]. Currently, there is no data on changes in the amount of mtDNA damage in animal models of AD.

Many signaling pathways that regulate mitochondrial metabolism and the clearance of amyloid-β are considered promising targets for pharmacological agents. In particular, nuclear factor erythroid 2-related factor 2 (Nrf2) plays a key role in the adaptive response to oxidative stress and can inhibit ferroptosis, which is a relatively new avenue in Alzheimer’s disease therapy [16]. Nrf2 has a close interplay with other signaling pathways. Specifically, Nrf2 is capable of regulating the expression of the gene that encodes a key component of the mechanistic target of rapamycin complex 1 (mTORC1). mTORC1-dependent pathways for regulating autophagy are also considered interesting targets for Alzheimer’s disease therapy [17]. Furthermore, these pathways are involved in maintaining synaptic plasticity and regulating long-term potentiation, thereby directly influencing cognitive functions [18]. Some polyphenolic compounds, including resveratrol, are considered promising for potentially slowing down the pathogenesis of AD. Resveratrol possesses anti-inflammatory properties and can activate signaling pathways that potentially protect brain mitochondria from damage caused by amyloid-β accumulation [19,20]. However, it should be noted that many polyphenols, including resveratrol, have limited bioavailability as they are largely metabolized in the intestine. The oral absorption of resveratrol in humans is approximately 75% through trans-epithelial diffusion. However, intensive metabolism in the gut and liver results in the bioavailability of resveratrol being significantly less than 1% when taken orally [21]. Resveratrol’s pleiotropic actions necessitate investigation within the gut-brain axis framework, recognizing this intricate signaling system coordinates neural, hormonal, and immunological communication between the intestine and brain. This axis is modulated through interconnected mechanisms involving neuronal signaling, endocrine pathways, immune responses, and metabolic regulation. The effects on the gut microbial composition may have remote consequences, including on inflammatory processes in the brain, and consequently, on cognitive functions [22]. The goal of this study was to investigate the ability of resveratrol to impact cognitive functions in APP/PS1 mice, amyloidogenesis, markers of inflammation, including in the gut microbiome, the level of mtDNA damage in the brain, as well as the expression level of genes that may influence cognitive functions and protect mitochondria from damage.

Neuronal inflammation is one of the characteristic features of AD pathogenesis. Chronic neuroinflammation observed in the brains of patients with AD can be attributed to activated microglial cells and the release of numerous cytokines [29,30]. Over the past decade, a substantial amount of experimental and clinical data has been obtained, indicating the involvement of inflammation in the pathogenesis of AD. Pro-inflammatory cytokine analysis revealed significantly elevated levels of IL-1, IL-6, TNF-α, and IL-18 in AD model systems compared to controls [31–34]. We found that the expression of Il-6 and Tnf genes was significantly increased in the prefrontal cortex of APP/PS1 mice (Fig. 5A). In the same gene cluster, we identified Ptgs2 (Fig. 5A), encoding cyclooxygenase-2, a crucial marker of inflammation [35]. Furthermore, the expression level of Gfap, a key biomarker for astrocytic activity and central nervous system (CNS) integrity [36], was significantly increased in the brains of APP/PS1 mice (Fig. 5A). The expression of these pro-inflammatory factors is regulated by the NF-κB signaling pathway. However, we did not observe an increase in the level of p50, a subunit of NF-κB (Fig. 5C). Although previous studies have shown an increase in the phosphorylated form of p65 (another subunit of NF-κB) in the brains of APP/PS1 mice [37,38]. It is likely that assessing the level of p65 is a more indicative marker of inflammatory pathway activation through the NF-κB signaling.

Previously, it has been shown that resveratrol can partially prevent the development of age-related cognitive dysfunction by reducing microglial activation [39]. In our study, resveratrol decreased the expression levels of the inflammatory markers, confirming its anti-inflammatory role (Fig. 5C). Moreover, when transgenic mice were given resveratrol, the amount of the cell-free mtDNA in blood plasma was similar to that of wild-type control mice (Fig. 5B). It is known that cell-free circulating mtDNA can act as a mediator of inflammatory processes. MtDNA functions as a damage-associated molecular pattern (DAMP) that can activate innate immune DNA sensors. When released during cellular or tissue damage, mtDNA serves as a potent signaling molecule that triggers inflammatory responses [40]. Levels of certain inflammatory markers, such as TNFα and Il6, are correlated with the level of extracellular mtDNA [40]. Multiple clinical studies have established a significant correlation between elevated circulating mtDNA concentrations and adverse clinical outcomes in acute respiratory distress syndrome (ARDS) patients, including increased mortality rates and prolonged ventilator dependence, sepsis [41], and COVID-19 [42]. Therefore, we observed an elevated level of cell-free mtDNA in the plasma of APP/PS1 mice, while resveratrol resulted in a significant reduction of its level (Fig. 5B), further confirming its anti-inflammatory properties. The significant increase in Proteobacteria levels in the gut microbiome of APP/PS1 mice (Fig. 7) further confirms the fact that inflammatory processes are more intense in transgenic mice compared to C57BL/6 mice. Increased levels of Proteobacteria are usually associated with dysbiosis, which is often linked to an elevation of pro-inflammatory markers due to disruption of the gut barrier function [43]. However, the level of Proteobacteria in resveratrol-treated APP/PS1 mice was equivalent to that of C57BL/6 mice (Fig. 7). Previous studies have demonstrated that resveratrol can reduce Proteobacteria levels in mice with induced obesity [44]. This suggests that the reduction in Proteobacteria levels could be considered as a result of the anti-inflammatory action of resveratrol.

It is known that resveratrol can activate the Nrf2/ARE signaling pathway [45]. We found that the transcript and protein Nrf2 level were decreased in 10-month-old APP/PS1 mice compared to wild-type mice of the same age (Figs. 5A and 6A). Previous studies have also demonstrated that the Nrf2/ARE pathway is attenuated in the brains of APP/PS1 transgenic mice during amyloid-β plaque deposition [46]. Downregulation of Nrf2/ARE signaling has also been observed in the brains of patients with AD [47]. This decrease may be associated with inflammatory processes. It is known that certain inflammatory cytokines, such as TNF-α, have a bidirectional effect on Nrf2/ARE signaling. Excessive production of TNF-α reduces the expression of antioxidant genes by inhibiting Nrf2/ARE signaling, while at lower concentrations, this pro-inflammatory cytokine is associated with Nrf2 activation [48].

The Nrf2/ARE signaling pathway is closely associated with maintaining the integrity of mtDNA. Nrf2 regulates the expression of many antioxidant proteins targeted to mitochondria [49], as well as genes involved in mitochondrial turnover [50], mtDNA repair [51], and enzymes for detoxification of drugs and xenobiotics [52], which are often targeted to mitochondria and mtDNA [53,54]. Indeed, we found that the ventral midbrain, cerebellum, and prefrontal cortex of APP/PS1 mice had significantly more mtDNA damage compared to C57BL/6 mice (Fig. 3). Similar findings were obtained in post-mortem brain analyses of patients with AD. It has been shown that the number of oxidized bases, particularly 8-oxoguanine, was approximately ten times higher in patients with AD compared to control patients. Specifically, mtDNA had ten times more oxidative damage than nuclear DNA [55]. In the cortex of patients up to 75 years old, the amount of mtDNA deletions was fifteen times higher than in healthy individuals of the same age. Interestingly, in patients over 75 years old, the level of deletions was slightly lower than in the control group [56]. Subsequently, it was revealed that deletions in mtDNA caused a deficiency of cytochrome c oxidase in the hippocampus of patients with AD [57]. It is worth noting that no increase in mtDNA damage was observed in the hippocampus (Fig. 3). This suggests that alterations in mtDNA integrity are not the primary cause of cognitive function impairments in the AD model.

In transgenic mice treated with resveratrol, the level of mtDNA damage in the prefrontal cortex was reduced (Fig. 3). It is known that resveratrol induces conformational changes in SIRT1, which affect its activity and specificity towards acetylated substrates [58]. SIRT1, in turn, deacetylates Nrf2, leading to its activation [59]. Several studies have demonstrated that the regulation of Nrf2 and SIRT1 is bidirectional. Under conditions of depressive-like behavior induced by systemic inflammation with lipopolysaccharides, melatonin reduced the level of inflammation and oxidative stress in mitochondria. The addition of siRNA targeting both Nrf2 and SIRT1 neutralized the positive effects of melatonin [60]. Earlier studies have shown that Nrf2 positively influences the deacetylase activity of SIRT1 towards proteins such as fibronectin (FN) and transforming growth factor-β1 (TGF-β1) [61]. It has not been previously explored whether Nrf2 can increase the expression of the Sirt1 gene. Computational analysis using FIMO (p value threshold < 1 × 10⁻⁵) identified a high-probability antioxidant response element (ARE) at position −9407/−9393 bp relative to the Sirt1 transcription start site, with a statistically significant match (p < 0.00001) to the canonical ARE consensus sequence. The fact that the expression of Nfe2l2 and Sirt1 genes (Fig. 5A) was found in the same cluster suggests the existence of a feedback loop in which Sirt1 expression depends on Nrf2/ARE signaling.

Resveratrol is also associated with other signaling pathways that may be involved in protecting mitochondria and mtDNA. SIRT1-mediated deacetylation of PGC-1α serves as a critical regulatory mechanism governing metabolic homeostasis and mitochondrial biogenesis, facilitating the restoration of mitochondrial function under stress conditions [62]. Resveratrol-induced activation of SIRT1 can initiate antioxidant pathways through the SIRT1/FOXO1 axis [63]. Notably, all these genes (Nfe2l2, Sirt1, Pparagc1a, Foxo1) are clustered together, and their expression is downregulated in APP/PS1 mice, which is partially reversed by resveratrol in the brain (Fig. 5A). The Mtor gene is also located in this cluster, and its product is a core component of the mTORC1 complex. The mTORC1 complex is responsible for ribosome assembly and protein synthesis, which are necessary for the formation of new neuronal connections that underlie long-term memory formation [64], which may explain the partial improvement in memory observed in transgenic mice treated with resveratrol (Figs. 1 and 2). Another indicator supporting this hypothesis is the increased level of phosphorylated AKT (Fig. 6B). The PI3K/AKT signaling pathway stimulates mTORC1 activity by inhibiting tuberous sclerosis complex 2 (TSC2), a negative regulator of mTORC1 [65]. The total level of AKT remained unchanged, but the level of phosphorylated AKT was increased (Fig. 6B), confirming the idea that resveratrol activates the PI3K/AKT/mTORC1 axis. Previous data on the role of resveratrol in activating the PI3K/AKT/mTORC1 axis have been somewhat contradictory. In some studies, it has been shown that resveratrol inhibits the phosphorylation of AKT, but these results were mostly obtained in tumor tissues [66–68]. On the other hand, resveratrol can activate the PI3K/AKT/mTORC1 axis in models of cardiac ischemic diseases [69] and Parkinson’s disease [70]. Additionally, there is cross-talk between the PI3K/AKT and Nrf2/ARE signaling pathways through the inhibition of GSK3β. Previous studies have shown downregulation of these signaling pathways in the AD pathogenesis [71].

However, the activation of the PI3K/AKT/mTORC1 pathway may have negative consequences in the context of potential AD therapy. The mTORC1 complex inhibits the process of autophagy by suppressing autophagic membrane formation [72]. Nevertheless, autophagy is crucial for amyloid-β clearance. In recent years, it has been suggested that inhibition of mTORC1 could serve as a therapeutic approach for AD by promoting autophagy and removing amyloid-β [73]. Indeed, we observed partial improvement in cognitive function of APP/PS1 mice treated with resveratrol (Figs. 1 and 2). At the same time, resveratrol only contributed to a 15% reduction in the accumulation of amyloid-β in the brains of mice (Fig. 4B). Additionally, APP/PS1 mice treated with resveratrol showed reduced expression of Sqstm1 and Pink1 compared to untreated mice of the same genotype (Fig. 5A). The mentioned proteins are key participants in mitophagy, which is a specific type of autophagy. Mitophagy is also crucial for the removal of dysfunctional mitochondria, which are affected in AD [74].

The principal limitation of our study resides in the insufficient examination of the resveratrol-associated mechanisms pertaining to amyloid-β. Our investigation predominantly concentrated on the mTORC1-dependent regulatory pathway of autophagy, while alternative pathways involved in the clearance of amyloid-β were not explored within the scope of this experiment.

Therefore, due to the potential mTORC1-dependent inhibition of autophagy, resveratrol only slightly contributed to the reduction of amyloid-β levels. However, the observed resveratrol-induced amelioration of cognitive deficits in the AD model may be attributed to the activation of the PI3K/AKT/mTORC1 and Nrf2/ARE signaling pathways, which directly participate in the formation of long-term memory and the mitigation of mitochondrial dysfunction (Fig. 8). Thus, we believe that the use of resveratrol may modulate cognitive properties not only in the context of Alzheimer-type dementias but potentially also in other memory disorders.