Resveratrol-like Compounds Improve Mitochondrial Function and Insulin Resistance in a Lab Model of Alzheimer's Disease

Alzheimer's disease (AD) is the most common form of neurodegenerative dementia, affecting tens of millions of people worldwide, with a steadily increasing prevalence due to the aging population [1]. Histopathologically, AD is characterized by the presence of extracellular β-amyloid (Aβ) plaques and intraneuronal neurofibrillary tangles composed of hyperphosphorylated tau protein in the cerebral cortex and hippocampus, which are key markers of synaptic and neuronal damage [1,2]. However, the pathomechanism of AD is extremely complex. In recent years, it has become increasingly evident that, beyond the amyloid cascade, several other factors—including metabolic and vascular abnormalities, chronic inflammation, and oxidative damage—play central roles in the pathogenesis of the disease [1,3,4,5]. Consequently, AD is now considered a multifactorial neurodegenerative disorder in which age-related metabolic disturbances are particularly important [2,4,5].

Epidemiological and experimental evidence supports that components of the metabolic syndrome—especially type 2 diabetes mellitus (T2DM) and its associated insulin resistance—significantly increase the risk of developing AD [6,7]. A recent meta-analysis found that individuals with T2DM have an approximately 50% higher risk of AD compared to age-matched healthy controls [5]. Insulin also plays a key role in the brain; central insulin signaling supports neuronal survival, synaptic plasticity, and cognition [6,8,9]. In AD, impairment of insulin signaling was clearly demonstrated with reduced expression of insulin receptors and glucose transporters, enhanced inhibitory phosphorylation of IRS, and decreased activation of downstream pathways such as PI3K/Akt signaling [5,8,10]. Cerebral insulin resistance results in disinhibition of glycogen synthase kinase-3β (GSK-3β) that promotes tau hyperphosphorylation and amyloidogenic processing of the amyloid precursor protein (APP); thus, disrupted insulin signaling is linked to the classical AD pathologies [8,9,10]. Furthermore, chronic hyperglycemia and hyperinsulinemia associated with type 2 diabetes (T2DM) elicit oxidative stress, hinder Aβ degradation, and trigger microglia-mediated inflammation that further exacerbates neurodegeneration [3,4,8,11]. Furthermore, insulin-sensitizing therapies and intranasal insulin have been reported to induce cognitive improvements in multiple experimental AD models and early-phase clinical trials [3].

Oxidative stress and disrupted redox homeostasis are fundamental in AD. Signs of oxidative damage appear early in AD, and their markers (lipid peroxidation end products, oxidized proteins, and nucleic acids) rise in the brain and cerebrospinal fluid; meanwhile, antioxidant defenses are often depleted [3,11]. Oxidative stress and insulin resistance reinforce each other; hyperglycemia and mitochondrial dysfunction augment reactive oxygen species (ROS) generation and impair insulin signaling, while the insulin resistance is accompanied by elevated activity of pro-oxidant kinases (GSK-3, JNK), sustaining a vicious cycle [3,8]. Persistent microglial and astrocytic activation creates a pro-inflammatory milieu that promotes Aβ accumulation and neuronal damage [2,12].

Dysfunction of protein homeostasis, particularly the impaired autophagy, is another major contributor to neurodegeneration. Mitophagy, the selective degradation of mitochondria, is critical for neuronal survival, as accumulation of dysfunctional mitochondria leads to elevated ROS levels and energy deficits [13]. Aging and AD are characterized by impaired autophagy and mitophagy. Autophagosome accumulation and lysosomal dysfunction are evident in AD brains, compromising the clearance of Aβ and hyperphosphorylated tau [7,13]. Autophagy-related gene deficiencies (e.g., beclin-1) promote Aβ accumulation and neurodegeneration in experimental models, while enhancing autophagic flux (e.g., via rapamycin or resveratrol) improves cognitive performance and reduces AD-like pathology [2,7,13]. Thus, autophagy represents a promising therapeutic target in AD by promoting the clearance of toxic aggregates and damaged organelles.

In recent years, several review articles have addressed the neuroprotective effects of natural compounds, particularly polyphenols, in neurodegenerative diseases. A recent review discusses the structural features and aggregation of α-synuclein, as well as the ability of various natural polyphenols—such as epigallocatechin gallate (EGCG), baicalein, or nordihydroguaiaretic acid—to interact with the monomeric and oligomeric forms of the protein, thereby inhibiting fibrillar aggregation and formation of toxic intermediates [14]. Although synucleinopathies (e.g., Parkinson's disease) are the main focus, these mechanisms are also relevant to AD, as protein aggregation, oxidative stress, and inflammatory processes represent common pathomechanisms across different neurodegenerative disorders. In parallel, another comprehensive review highlights that naturally derived bioactive compounds (flavonoids, alkaloids, terpenoids, etc.) can interfere with the pathogenesis of AD at multiple levels. According to the reviewed literature, these molecules can reduce β-amyloid overproduction, attenuate tau hyperphosphorylation, inhibit neuroinflammation and apoptosis, and improve autophagy and mitochondrial function. In addition, through their antioxidant activity, they can alleviate oxidative stress and the related endoplasmic reticulum (ER) stress, which are key factors in the progression of neurodegeneration [15]. Among these natural compounds, resveratrol (trans-3,5,4′-trihydroxystilbene) deserves particular attention as one of the most promising molecules due to its antioxidant, anti-inflammatory, anti-apoptotic, and cognitive-enhancing properties [16,17,18]. Preclinical studies have demonstrated that resveratrol improves cognitive functions, reduces amyloid and tau pathology, and mitigates inflammation and oxidative stress in AD models [19]. In our recent research, we have demonstrated the protective effect of resveratrol in an in vivo insulin-resistant neurodegeneration model developed in our laboratory [8]. It significantly reversed the insulin resistance induced by streptozotocin (STZ) in SH-SY5Y neuroblastoma cells and improved mitochondrial quality control by improving both mitogenesis and autophagy [20]. Despite its beneficial effects, the clinical use of resveratrol is hindered by its poor bioavailability due to its rapid metabolism and limited central nervous system penetration [1,21]. To overcome these limitations, several structural derivatives of resveratrol have been developed. Modifications such as hydroxyl group methylation or the addition of extra hydroxyl groups significantly affect both its pharmacokinetics and pharmacodynamics [22,23]. However, the activity of resveratrol analogs can also differ considerably. According to our previous results, monomethylated resveratrol retains cytoprotective effects, potentially through autophagy induction, despite its reduced antioxidant activity. In contrast, trimethylated resveratrol, which lacks free hydroxyl groups, exhibits no antioxidant effects but even acts as a pro-oxidant. Conversely, the hydroxylated analog oxyresveratrol exhibits stronger antioxidant properties without autophagy induction [23]. Yet, the effect of structural modifications on the neuroprotective activity of resveratrol in the insulin resistance-associated neurodegenerative processes remains unclear, as few studies have evaluated the efficacy of resveratrol analogs in AD-relevant in vitro models [24,25].

In our study, we aimed to compare the neuroprotective effects of three resveratrol analogs—oxyresveratrol, monomethylated resveratrol, and trimethylated resveratrol— to that of the parent compound in an in vitro model of AD based on STZ-induced insulin resistance in retinoic acid-differentiated SH-SY5Y cells [8]. Their ability to counteract key pathological processes of AD—oxidative stress, autophagy dysfunction, and insulin signaling impairment—was studied. Our results are expected to contribute to a better understanding of the structure-activity relationships of resveratrol derivatives and pave the way to the development of multitarget neuroprotective strategies against AD.

The neurotoxic effect of streptozotocin (STZ) is well documented. It impairs insulin signaling in the nervous system and induces neurodegenerative alterations resembling AD [26]. In our results, STZ induced a significant increase in ROS generation, mitochondrial depolarization, insulin resistance, and accumulation of protein aggregates in line with the previous literature [26,27]. The found alterations are consistent with the AD pathomechanisms and support the applicability of the model for studying the neuroprotective effect of resveratrol analogs.

Resveratrol possesses numerous neuroprotective properties, which have been demonstrated in various neurodegeneration models (for reviews, see [28,29,30,31]). In our previous work, we showed a strong protective effect in the same model. It had effectively reversed insulin resistance induced by STZ treatment via augmenting insulin signaling. It also significantly reduced the levels of ROS, upregulated the antioxidant machineries, and improved the mitochondrial quality control via enhanced autophagy and mitogenesis [20]. In the present study, similarly to resveratrol, its three analogs also showed neuroprotective activity by improving insulin sensitivity and reversing STZ-induced AD-like pathologies, although with different efficacy and potency and likely via different mechanisms. All the examined compounds effectively upregulated the compromised insulin signaling in the STZ model of neurodegeneration. They improved the inhibitory phosphorylation of GSK-3 that is accompanied by a significant reduction in protein aggregates characteristic of neurodegeneration in AD. Phosphorylation of tau protein by GSK-3β and the protective effect of insulin are well established [32,33], while involvement of GSK-3α in amyloid plaque formation is also suggested [34,35]; thus, its inhibition is clearly connected to the anti-AD effect.

Monomethyl resveratrol was the most similar to the parent compound, showing similar antioxidant and autophagy augmenting effects. Both compounds induced a concentration-dependent mitochondrial depolarization that, in turn, may contribute to the induction of autophagy and mitochondrial quality control. The effect of resveratrol on mitochondrial function is extensively discussed in the literature. It has a weak inhibitory effect on ATP synthase [36,37] as well as other mitochondrial electron transport chain complexes [38] was demonstrated. These findings are in line with our previous [20,23,39] and present results. Mild inhibition of the electron transport chain can impair mitochondrial membrane potential and reduce the ATP–AMP ratio. This, in turn, may enhance AMPK activity, which is another proposed mechanism behind the effects of resveratrol [40,41,42,43]. In our previous work, we also demonstrated the importance of AMPK activity in the effect of resveratrol using the same model that may contribute to both improved mitochondrial quality control via upregulating autophagy and reversal of insulin resistance [20]. Activation of AMPK by monomethyl resveratrol was also reported [44] which further confirms the similar mechanism of the two compounds. Induction of autophagy to a similar extent by both resveratrol and its monomethyl analog is found in our present and previous works [23]. Autophagy is involved in multiple cellular processes that can participate in the neuroprotective effect of the examined compounds. It contributes to the elimination of damaged organelles, e.g., mitochondria, and enhances the clearance of pathological protein aggregates (for review, see [45]). A bidirectional connection between insulin resistance and defective autophagy is also proposed [46]; thus, we can speculate the paramount importance of autophagy induction in the insulin-sensitizing and neuroprotective effects of resveratrol and monomethyl resveratrol.

The two compounds also shared some similarity in their antioxidant effect in the early phase of STZ exposure that declined after continued treatment, and in the case of monomethyl resveratrol, even turned to prooxidant activity. As a polyphenolic compound, resveratrol was demonstrated to be able to directly scavenge free radicals [47,48,49], and its effect on endogenous antioxidant machineries, such as superoxide dismutase, glutathione peroxidase, and catalase, was also reported (for review [50]). Activation of Nrf2 signaling was identified as a crucial upstream step in the strengthened antioxidant defense induced by resveratrol [51,52]. Similar free radical scavenger activity [53] and facilitation of the Nrf2 pathway by monomethyl resveratrol [54] were also shown; however, its prooxidant and Nrf2 inhibitory action were also reported [55]. In line with the literature data, our present results suggest that direct scavenger activity of the compounds is more prominent compared to the upregulation of antioxidant enzyme systems because of the short-term effect. Based on the present results, we can infer that although ROS neutralizing activity contributes to the insulin-sensitizing and neuroprotective effect of resveratrol and monomethyl resveratrol, they are more considerably derived from autophagy induction.

Trimethyl resveratrol lacks free phenolic hydroxyl groups, which is consistent with its abolished antioxidant activity and reduced neuroprotective effect found in the present experiments. Direct free radical scavenger activity of stilbene derivatives relies on their phenolic hydroxyl groups, with 4′ one being the most important in the case of resveratrol [56], which can explain the lack of antioxidant activity. This is in line with the report of Kim et al., who demonstrated the failure of trimethyl resveratrol to protect HT22 neuronal cells from glutamate neurotoxicity and associated oxidative stress, contrary to resveratrol [57]. On the other hand, the trimethylated analog is still capable for modulation of mitochondrial membrane potential and inducing autophagy [23], improving AMPK activity [58] and insulin sensitivity [59]. Our results, together with the literature data, suggest that trimethyl resveratrol is a strong activator of autophagy without significant antioxidant activity that can improve insulin resistance and accumulation of protein aggregates; however, its neuroprotective effect lags behind its parent compound.

Oxyresveratrol was found to be a strong antioxidant that can be explained by the presence of an additional phenolic hydroxyl group on the stilbene core [60,61]. Its similar but partially synergistic direct antioxidant effect compared to that of resveratrol was reported [62], indicating some difference in their scavenger mechanism. Enhanced antioxidant defense via activation of Nrf2-mediated transcription was also described in the case of oxyresveratrol [63,64], further confirming its comparable activity to its parent compound. On the other hand, contrary to resveratrol and the other analogs, it had no effect on the mitochondrial membrane potential and was not able to induce autophagy. These results are in line with our previous findings on mouse embryonic fibroblast [23]. Although the induction of autophagy by oxyresveratrol was reported by some papers [25,65], it might be condition-dependent. The considerable effect of the compound on insulin signaling and protein aggregation suggests an alternative neuroprotective mechanism that warrants further studies. Involvement of AMPK activation and alleviation of glutamate toxicity by oxyresveratrol were reported in different neuroprotective models [25,66,67] that might also have contributed to its effect in the present model.

Altogether, our results indicate significant differences in the mechanisms underlying the neuroprotective effects of resveratrol derivatives, summarized in Table 1. Although all the examined compounds reversed insulin resistance induced by STZ treatment, comparing the effects of resveratrol, its methylated and hydroxylated analogs revealed important structure–activity relationship data. The neuroprotective effect of resveratrol is strongly connected to the presence of free hydroxyl groups, which are critical to its antioxidant capacity [68,69]. Its molecular structure, however, is also appropriate for the induction of autophagy [17] and mitochondrial quality control [36,70] likely by influencing the mitochondrial respiratory chain [36,37] and activating AMPK [40,41,42,43]. This dual mechanism—antioxidant defense combined with autophagy induction—explains the complex protective action of the parent compound [23]. The introduction of a single methyl group can reduce the antioxidant capacity of resveratrol by decreasing the number of free hydroxyl groups available for free radical scavenging [53,71]. This structural modification may contribute to its altered redox profile, showing antioxidant activity in the short term [72] but even prooxidant effects at later stages [55]. However, the presence of a single methyl substitution does not modify its autophagy-inducing activity, which explains its substantial similarity to resveratrol [23]. In trimethyl resveratrol, the complete methylation of phenolic hydroxyl groups abolishes its direct antioxidant activity, due to the lack of functional groups required for the free radical scavenging effect [56]. This structural feature accounts for its weaker neuroprotective effect and enhanced ROS generation. On the other hand, it has a more potent effect on the mitochondria and more effectively activates autophagy. The lack of antioxidant capacity, however, still limits its protective potential compared to the parent compound [23]. In the case of oxyresveratrol, the addition of an extra hydroxyl group to the stilbene core enhances the radical scavenging capacity and provides a strong antioxidant activity [60,61,73]. However, this structural modification results in a distinct mechanism of action, as it cannot induce autophagy, and its neuroprotective effect is likely mediated primarily through antioxidant pathways rather than mitochondrial quality control [63].

This study has certain limitations. Protein aggregation was assessed using Thioflavin S staining, which detects β-sheet-rich amyloid-like structures in a protein-independent manner. Therefore, the observed signal was interpreted as a marker of global protein aggregation burden, without assigning it to a specific protein such as tau, Aβ, α-synuclein, or hIAPP. In addition, treatments were administered in parallel with STZ exposure (co-treatment) rather than in a post-treatment setting. While this approach is commonly applied in in vitro studies, it does not fully reproduce therapeutic conditions, where interventions are given after pathological alterations have already developed. Future studies will be required to address these aspects for a more comprehensive evaluation of the neuroprotective potential of resveratrol analogs.

Our findings clearly demonstrate the neuroprotective effect of the examined resveratrol analogs in an in vitro AD model based on STZ-induced damage and insulin resistance. All the compounds prevented the accumulation of protein aggregates and restored the response to insulin, which is consistent with ameliorating AD-like pathology. Resveratrol and monomethyl resveratrol had rather similar effects, while trimethyl resveratrol, which lacks direct antioxidant activity, was less potent in protecting neuron-like cells. These results suggest that the preservation of some phenolic hydroxyl groups on the stilbene core is critical for achieving optimal neuroprotective activity. Oxyresveratrol, on the other hand, showed only antioxidant activity among the studied mechanisms that may contribute to its considerable neuroprotective effect.

Our results indicate the importance of structure–activity relationship experiments in exploring the neuroprotective effect of stilbenes, as small structural differences had a significant impact on their biological activity.