Intranasal Administration of Isoeugenol Improves Memory Deficits in APP/PS1 Old Mice—A Role Beyond Nrf2 Activation?

ABSTRACT

Alzheimer's disease (AD), a neurodegenerative disorder and the most common cause of dementia, has no cure or effective treatment; thus, identification of disease‐modifying therapeutics is crucial. Nrf2 is a master controller of homeostatic functions whose activity is compromised in AD. Most pharmacological Nrf2 activators are electrophilic molecules that covalently modify cysteine residues in the thiol‐rich Keap1, and many are Michael acceptors, such as low‐molecular‐weight (LMW) skin allergens. However, LMW allergens‐induced Nrf2 activation in a pharmacological setting has only recently attracted attention, exemplified by the clinical success of Dimethyl Fumarate. Hence, we investigated, for the first time, the potential of the skin allergen Isoeugenol to activate Nrf2 and reverse selected AD hallmarks, both in vitro and in vivo, in AD‐specific models. In vitro studies were performed using microglia cells exposed to LPS and neuronal cells overexpressing human APP with Swedish mutation, to evaluate Isoeugenol's potential in decreasing neuroinflammation and activating the Nrf2 pathway, respectively. In vivo studies were conducted in 10‐month‐old AD double‐transgenic mice (APP/PS1), which were intranasally administered Isougenol. Isoeugenol's pharmacokinetic and pharmacodynamic profile, and its effect on mice cognition were evaluated. The results showed that Isoeugenol (1) activated Nrf2 in AD neuronal cells (likely involving AKT signaling); (2) exhibited antioxidant and Nrf2‐dependent anti‐inflammatory activity, which was abolished after Nrf2 silencing; (3) exhibited good pharmacokinetic and pharmacodynamic profiles; (4) reduced the levels of Aβ peptides in vitro and in vivo; (5) reduced triglyceride and LDL cholesterol levels in treated mice; and (6) improved the memory deficits in old mice. This is the first study reporting Isoeugenol's intranasal administration, and on specific AD mice models. Overall, the results reinforce Isoeugenol as a pleiotropic molecule, with great potential for AD treatment.

Inactivation of the Nrf2/Are pathway in AD leads to decreased antioxidant and anti‐inflammatory response and to the accumulation of beta‐amyloid (Aβ) plaques, resulting in cognitive failure. Treatment with Isoeugenol activates Nrf2, reversing these outcomes by promoting antioxidant response in neuronal cells (N2a) and by increasing the anti‐inflammatory response in microglia cells (BV‐2). In addition to the safe profile and pharmacokinetics, intranasal administration of Isoeugenol resulted in reduced levels of secreted Aβ and enhanced cognition and memory in APP/PS1 old mice.

Abbreviations

ALP

Alkaline phosphatase

APP

Amyloid‐beta precursor protein

ARE

Antioxidant response element

BACE1

Beta‐secretase 1

BBB

Blood–brain barrier

BDNF

Brain‐derived neurotrophic factor

DAM

disease‐associated microglia

DMF

Dimethyl fumarate

GGT

Gamma‐Glutamyl Transferase

GOT

glutamic oxaloacetic transaminase

GSH

Glutathione

GSK‐3β

Glycogen synthase kinase‐3 beta

HMOX1

Heme Oxygenase 1

HPRT1

Hypoxanthine‐guanine phosphoribosyl transferase

HQM

Hydroxy quinone methide

Iba1

Ionized calcium‐binding adaptor molecule 1

IL

Interleukin

iNOS

Inducible nitric oxide synthase

Keap1

Kelch‐like ECH‐Associating protein 1

LPS

Lipopolysaccharides

Maf

musculoaponeurotic fibrosarcoma protein

NFTs

Neurofibrillary tangles

NF‐κB

Nuclear factor‐κB

NMDA

N‐methyl‐D‐aspartate

NO

Nitric oxide

NQO1

NAD(P)H quinone oxidoreductase 1

Nrf2

nuclear factor erythroid 2–related factor 2

PGC‐1α

Peroxisome proliferator‐activated receptor gamma coactivator 1‐alpha

QM

quinone methide

ROS

Reactive oxygen species

SOD1

superoxide dismutase 1

TREM2

Triggering receptor expressed on myeloid cells 2

1. Introduction

Alzheimer's disease (AD) is one of the most common neurodegenerative diseases and the most frequent cause of dementia in the elderly. With slowly progressive neurodegeneration in the cortex and the hippocampus [1, 2], AD neuropathologic hallmarks include Aβ peptide accumulation and amyloid plaque formation, and hyperphosphorylation of tau protein with neurofibrillary tangle formation [3, 4]. Neuroinflammation with microglial cell activation and oxidative stress occurs due to neurodegeneration and synaptic dysfunction [2, 5, 6].

To date, many advances have been made to better understand the neuropathology of AD and find efficacious disease‐modifying treatments with the capacity to prevent or cure it. The amyloid cascade hypothesis, with exacerbated production of Aβ peptides, is considered the main hypothesis underlying AD occurrence and has been intensively studied to combat the existing therapeutic gap [7, 8, 9, 10]. In fact, there are only four therapeutic drugs for AD treatment, which only relieve symptoms associated with memory loss or behavior [3, 11]. Recently, the United States Federal Agency Food and Drug Administration (FDA) approved two monoclonal antibodies designed to clear Aβ from the brain and block the formation of amyloid plaques [12]. However, both were rejected by the European Medicines Agency (EMA) due to a lack of proven safety and clinical benefit.

It has been shown that the Kelch‐like ECH‐Associating protein 1 (Keap1)‐Nuclear factor erythroid 2‐related factor 2 (Nrf2)‐antioxidant response element (ARE) (Keap1‐Nrf2‐ARE) signaling pathway is compromised in the context of AD [13]. This is supported by age‐dependent decreased activity and expression of Nrf2 and by its sustained cytoplasmic localization, suggestive of a defective nuclear translocation, as previously demonstrated by the team [14] and others [15, 16, 17]. Nrf2 is a transcription factor and one of the most important regulators of cellular defense mechanisms against oxidative and electrophilic stress [18]. Nrf2 activity is tightly regulated by two main mechanisms that control its stability in the cytoplasm under physiological conditions: (1) Nrf2 binding to Keap1, which comprises several redox‐ and electrophilic‐sensitive residues of cysteine, and (2) Nrf2 phosphorylation in specific serine residues by Glycogen synthase kinase‐3 beta (GSK‐3β), both culminating in Nrf2 ubiquitination and proteasomal degradation.

Low‐molecular‐weight skin allergens are small molecules capable of modifying the sulfhydryl groups of Keap1's cysteine residues and dissociating the Nrf2‐Keap1 complex [19]. This dissociation allows Nrf2 migration into the nucleus where it binds to the small musculoaponeurotic fibrosarcoma protein (Maf) and ARE sequences, initiating the transcription of genes involved in detoxification reactions, redox and energy metabolism pathways, inflammation, and proteostasis, all of which altered in neurodegenerative diseases [18, 20, 21], thus providing Nrf2 as an attractive molecular target for AD therapeutic research [22, 23]. Of notice, Dimethyl fumarate (DMF), a skin allergen and a typical Michael acceptor molecule known to react with cysteine residues [19], is a proven Nrf2 activator [24], and an orally available disease‐modifying drug approved by the FDA and EMA as first‐line therapy for the management of relapsing forms of multiple sclerosis [25]. DMF was also demonstrated to prevent spatial memory impairment and hippocampal neurodegeneration in rats [26]. Based on these pieces of evidence, we hypothesized that other skin allergens, which also act as Michael acceptors and have not been explored in drug discovery programs, may have a positive therapeutic role in AD when administered by non‐topical routes. Isoeugenol (2‐methoxy‐4‐(prop‐1‐en‐1‐yl) phenol), is a phenylpropanoid compound present in some essential oils of plants like clove and classified as a moderate to strong skin allergen [27]. Isoeugenol is not a direct Michael acceptor; instead, it has a pre‐Michael acceptor (thus considered a pre‐hapten). It can be biologically activated (by enzymes) or physically activated (by oxidation), originating reactive derivatives [28, 29]. It can scavenge reactive oxygen radicals by donating phenolic hydrogen atoms, forming a phenoxyl radical [27]. The phenoxyl radical and the Iso derivatives, quinone methide (QM) and hydroxy quinone methide (HQM), both behaving like Michael acceptors, were shown to react with nucleophilic residues of proteins (e.g., thiol groups present in cysteine residues). We recently demonstrated that Isoeugenol reacts with glutathione cysteine residues [30] and, according to our in silico data, it possesses drug‐like properties and the ability to cross the blood–brain barrier (BBB) [23]. Therefore, in this work, we used AD and neuroinflammation cell models to study the antioxidant and anti‐inflammatory properties of Isoeugenol and evaluated the effect of Isoeugenol administration on the memory of AD transgenic mice. Isoeugenol was intranasally administered, which is a simple, fast, and noninvasive method for drug delivery into the brain, including larger molecules that otherwise would not cross the BBB. Moreover, it avoids gastrointestinal metabolism and hepatic first‐pass effect, allowing lower‐dose administration, thus reducing potential drug toxicity [31, 32, 33]. To the best of our knowledge, this is the first pharmacokinetic study on intranasal administration of Isoeugenol, and a new HPLC method was validated in this work.

2. Materials and Methods

2.1. Materials—Source of Chemicals and Reagents

Isoeugenol (98% cis and trans racemic mixture), dimethyl sulfoxide (DMSO), Rosewell Park Memorial Institute (RPMI)‐1640 medium, lipopolysaccharide (LPS) from Escherichia coli (serotype 026:B6), sodium bicarbonate, L‐glutamine and D‐(+)‐glucose, MEM (Minimum Essential Medium) nonessential amino acid solution, Bovine Serum Albumin (BSA), Bicinchoninic Acid (BCA), penicillin–streptomycin antibiotic, ethylenediamine tetraacetic acid (EDTA), glycerol, Nonidet‐P‐40, sodium dodecyl sulfate (SDS), protease inhibitor cocktail for ex vivo assays, mouse monoclonal anti‐β‐Tubulin antibody were obtained from Sigma‐Aldrich (St. Louis, MO, USA). Dulbecco's Modified Eagle's Medium (DMEM)‐31600 (powder, low glucose with pyruvate; Gibco), geneticin G‐418 (Gibco), fetal bovine serum (FBS; Gibco), sodium chloride, Amyloid beta 40 Human ELISA (Invitrogen), Amyloid beta 42 Human ELISA (Invitrogen) and mouse monoclonal anti‐HMOX1 antibody were obtained from ThermoFisher Scientific (Walthman, Massachusetts, USA). NZYOL reagent, NZY First‐Strand cDNA Synthesis and NZYSpeedy qPCR Green Master Mix (2×) kits were from NZYTech genes & enzymes (Lisbon, Portugal). Nrf2 Transcription Factor Assay colorimetric kit and monoclonal antibodies against pro‐IL‐1β and lamin B1 were obtained from Abcam (Cambridge, UK). Nuclear Extract kit was purchased from Active Motif (Waterloo, Belgium). Phosphatase (PhosSTOP) and protease (Complete Mini) cocktail inhibitors were provided by Roche Diagnostics (Mannheim, Germany) and dithiothreitol (DTT) from Roche (Danvers, Massachusetts, USA). Polyvinylidene difluoride (PVDF) membranes were purchased from Millipore Corporation (Bedford, MA, USA). Mouse anti‐iNOS monoclonal antibody was obtained from R&D Systems (Mineapolis, MN, USA) and mouse anti‐Nrf2 monoclonal antibody from Santa Cruz Biotechnology (Dallas, TX, USA). Mouse phospho‐GSK3β, GSK3β, phospho‐AKT and AKT monoclonal antibodies, and secondary anti‐mouse and anti‐rabbit ECL antibodies were obtained from Cell Signaling Technology (Danvers, Massachusetts, USA). Clarity Western ECL (Enhanced Chemiluminescent) Substrate agent was purchased from Bio‐Rad Laboratories (Hercules, CA, USA).

2.1.1. Methods

Cell culture, Aβ40 and Aβ42 peptide levels quantification, gene expression, cell extracts and western‐blotting, Nrf2 activation and nitric oxide production followed the methods previously published by the team [23].

2.2. Cell Culture

The neuronal mouse neuroblastoma cell lines (N2a) wild‐type (N2a‐wt; Neuro‐2a, ATCC CCL‐131) and human Swedish‐mutant APP695 (N2a‐APPswe; a gift from Dr. Ciro Isidoro, Università del Piemonte Orientale “A. Avogadro,” Novara, Italy) were cultured in DMEM‐31600 medium supplemented with 3.7 g/L sodium bicarbonate, 4.5 g/L D‐glucose, 1% (v/v) nonessential amino acids and 10% (v/v) FBS inactivated at 56°C for 30 min. N2a‐wt cells were supplemented with 1% (v/v) penicillin–streptomycin and the N2a‐APPswe cells with 0.4 mg/mL geneticin G‐418. The murine microglia cell line (BV‐2; Interlab Cell Line Collection, ICLC ATL03001) was cultured and maintained in RPMI‐1640 medium supplemented with 10% (v/v) FBS (inactivated at 56°C for 30 min), 1% (v/v) penicillin–streptomycin, 2 g/L sodium bicarbonate and 2 mM L‐glutamine. All cell lines were subcultured every 2–3 days and maintained at 37°C in a 5% CO₂ atmosphere.

2.3. Cell Treatment

N2a cells were plated at a density of 5.3 × 10⁴ cells/cm² and BV‐2 cells at a density of 2.5 × 10⁴ cells/cm² and stabilized overnight (ON) before Isoeugenol (CAS 97‐54‐1; Sigma‐Aldrich #I17206) treatment.

N2a cells were treated with Isoeugenol for 24 h for quantification of Aβ40 and Aβ42 peptide levels and determination of HMOX1 protein levels, whereas Nrf2 nuclear levels and activation were studied after 2, 4, and 6 h treatment; for analysis of Hmox1 gene expression, the incubation times were 1, 3, and 6 h. The phosphorylated (p) levels of AKT and GSK3β proteins were analyzed after 5, 15, and 30 min of Isoeugenol incubation.

BV‐2 cells were stimulated with Lipopolysaccharides (LPS, from Escherichia coli O26:B6, Sigma‐Aldrich #L2654; 50 ng/mL, in sterile PBS), added 30 min after Isoeugenol, in a total of 18 h for gene expression analysis (Nos2, ll‐1β) or in a total of 24 h for protein levels analysis (iNOS and pro‐IL‐1β) and for nitrites quantification.

2.4. Cell Metabolism

N2a‐wt and BV‐2 cells were exposed to different concentrations of Isoeugenol (in μM: 500, 250, 100, 50 e 5) diluted in DMSO (up to a concentration of 0.1% v/v), for 24 h. Cell metabolic capacity was evaluated by the Alamar Blue (resazurin) reduction colorimetric assay, as described elsewhere [34]. Resazurin (50 μM, in sterile PBS) was added to each well 3 h before the end of the experiment (i.e., 24 h) and the absorbance at 570 nm and 620 nm (reference wavelength) was read in Biotek Synergy HT plate reader (Biotek, Winooski, VT, USA). The increase in fluorescence or absorbance reflects the increase in the number of viable cells, since metabolically active cells can reduce resazurin to resorufin (pink compound that emits fluorescence). Isoeugenol nontoxic concentration used in vitro was set at 250 μM (Supporting Information).

2.5. KeratinoSens Reporter Gene Assay

KeratinoSens assay (acCELLerate GmbH, Hamburg, Germany, Cat# RE232) is an OECD‐validated method for assessing the sensitization potential of a compound [35]. Cells harbor the luciferase gene under the transcriptional control of ARE sequence, and Nrf2 activation is quantified as luciferase fold‐induction above control (DMSO). The assay was performed according to the manufacturer's instructions, with modifications, to evaluate the contribution of the AKT pathway to NRF2 activation induced by Isoeugenol.

KeratinoSens cells were maintained in DMEM (liquid, low glucose, GlutaMAX supplement, with sodium pyruvate; Gibco # 21885‐025) supplemented with 9.1% fetal bovine serum, 500 μg/mL geneticin (G418 sulfate, Gibco #10131‐027), in a 5% CO2 incubator at 37°C. Cells were subcultured every 3 days at 80%–90% confluence, using a 1:6 split ratio in T75 flasks, and maintained up to passage 22.

Briefly, 1.0 × 10⁴ cells/well were plated (in triplicate) in 96‐well white plates and maintained a 5% CO₂ incubator at 37°C. After 24 h, culture medium was replaced with DMEM supplemented with 1% FBS without geneticin. Cells were incubated with Isoeugenol (250 μM), with or without Wortmannin (500 nM, in DMSO), added 1 h before Isoeugenol. DMF (50 μM) was used as a positive control in DMSO. The final DMSO concentration in the cells was up to 0.1%. After 6 h of exposure, cell viability was assessed by Alamar Blue assay (50 μM resazurin solution, for 4 h at 37°C). After, 180 μL of the supernatant was transferred to a 96‐well plate for absorbance measurements. The cells were then washed once with sterile PBS and lysis buffer (ONE‐Glo Luciferase Assay System; Promega #E6110) was added to the cells and incubated for 20 min at RT protected from light. Luciferase activity was assessed using FLUOstar OMEGA Microplate Reader (BMG Labtech GmbH, Ortenberg, Germany). Nrf2 activation was considered if luciferase induction was 1‐fold higher than the vehicle control (DMSO) (Supporting Information).

2.6. Blood–Brain Barrier (BBB) Permeability—Parallel Artificial Membrane Permeability Assay (PAMPA)

The capacity of Isoeugenol to cross the BBB was evaluated using a parallel artificial membrane permeability assay (PAMPA) model, as described by our team [36]. The reference drugs sulfasalazine, trazodone, and propranolol (Sigma‐Aldrich, St. Louis, MO, USA) with established BBB permeability were used as controls, as follows: sulfasalazine (not permeable (BBB−)) as a negative control; trazodone and propranolol (permeable (BBB+)) as positive controls. The donor solutions of the reference drugs and Isoeugenol were diluted (in phosphate buffer saline, PBS, pH 7.4) to a final concentration of 500 μM and 150 μg/mL, respectively (DMSO final concentration = 5%). The 2% (w/v) brain lipid extract solution of porcine polar brain lipid (PBL) (Avanti Polar Lipids, Alabaster, AL, USA) was diluted in n‐dodecane (Sigma‐Aldrich, St. Louis, MO, USA). The donor solution was prepared by adding 300 μL of the reference drugs or Isoeugenol per well in a microtiter plate (MultiScreen‐IP, catalog no. MATRNPS50, Millipore Corporation, Bedford, MA, USA). Next, 6 μL of 2% PBL was added to the hydrophobic filter of each acceptor well (MultiScreen‐IP, catalog no. MAIPNTR10, Millipore Corporation, Bedford, MA, USA), which were then filled with 150 μL of PBS‐containing DMSO in the same % of the donor solution. The 96‐well microfilter acceptor plate was carefully placed onto the microtiter donor plate, and the assembly was incubated for 16 h at room temperature, under gentle stirring. After incubation, the absorbance of the acceptor solutions was read by ultraviolet spectrophotometry, using a SpectraMax Plus 384 Spectrophotometer at the corresponding wavelengths previously determined for each compound, and the necessary dilutions were made to maintain an absorption inferior to 1. Three experiments were executed, and the apparent permeability (Papp) of each reference drug and Isoeugenol was calculated by applying the equation:

$$\text{Papp}=-2.303\times \frac{\mathrm{Vdn}\mathrm{Vac}}{\mathrm{Vdn}+\mathrm{Vac}}\times \frac{1}{S\times t}\times \log \left(1-\frac{\text{flux}\%}{100}\right)$$

$$\text{Flux}\%=\frac{\text{ODac}}{\text{ODeq}}$$

where Vdn is the volume of the donor solution (300 μL), Vac is the volume of the acceptor solution (150 μL), S is the surface area of the filter that separates both compartments (0.26 cm²), t is the incubation time (57 600 s), ODac is the optical density of the acceptor solution and ODeq is the optical density of the equilibrium solution, obtained after the assembly of the two 96‐well plates (450 μL). The permeability of each compound is predicted based on the calculated Papp value, being the cut‐off value of 2.0 × 10⁻⁶ cm per second (cm s⁻¹) (Supporting Information).

2.7. β‐Site APP Cleaving Enzyme 1 (BACE1) Activity Assay

The β Secretase 1 (BACE1) Activity Detection Kit CS0010 (Sigma‐Aldrich Co. LLC., St. Louis, MO, USA) allows a screening of BACE1 inhibitor molecules based on fluorescence resonance energy transfer that is observed when there is cleavage of the substrate by the enzyme. The BACE1 substrate stock solution (500 μM) was prepared in DMSO, aliquoted, and stored at −20°C. Just before the beginning of the assay, aliquots of BACE1 substrate and of the enzyme solution were brought to room temperature and further diluted (10‐fold) with the fluorescent assay buffer to a final concentration of 50 μM and 0.3 unit/μL, respectively, and kept in ice. A solution with an inhibitor provided with the kit was prepared in DMSO (1 mM) and then diluted to a concentration of 750 nM. Isoeugenol was tested at 50, 100 and 250 μM. The components were added to a 96‐well plate (according to the manufacturer's instructions). Blanks (without the addition of the enzyme and substrate and with Isoeugenol) were also prepared to evaluate the fluorescence of the test compound per se in the aforementioned concentrations. The plate was incubated for 2 h at 37°C and the fluorescence read in a fluorimeter at 320 nm (excitation) and 405 nm (emission) with an automatic cut‐off of 325 nm. Final fluorescence values were calculated by subtracting the negative control (with substrate and no enzyme) and the respective blank fluorescence to that of the tested compound and expressed as % of the positive control (enzyme and substrate, without inhibitor and considered as 100% of BACE activity). Results are shown as the mean of the obtained percentage of inhibition (%) ± standard error of the mean (SEM) for the three independent experiments performed in duplicate. β‐Secretase Inhibitor IV (Calbiochem, Merck KGaA, Darmstadt, Germany) was used as a reference inhibitor at its IC₅₀ concentration (15 nM) to validate the procedure (Supporting Information).

2.8. Aβ₄₀ and Aβ₄₂ Peptide Levels Quantification

Quantification of Aβ₄₀ and Aβ₄₂ peptide levels in N2a cells supernatants, and in animals' brain and plasma was performed using the ELISA kit Human Amyloid beta 40 (Invitrogen #KHB3481) and the ELISA kit Human Amyloid beta 42 (Invitrogen #KHB3441), respectively, and following the manufacturer's instructions. The absorbance was read at 450 nm in a Biotek Synergy HT plate reader (Biotek, Winooski, VT, USA), and the peptide concentration was determined according to an Aβ standard curve. The results were expressed in picograms per milligram of protein (brain tissues) or picograms per milliliter (plasma and supernatant).

2.9. Gene Expression Analysis by Real‐Time RT‐PCR

Total RNA was extracted with NZYol reagent (NZYTech #MB18501) and RNA concentration was quantified using a NanoDrop spectrophotometer (Thermo Scientific, Wilmington, DE, USA). Samples were stored at −80°C in an RNA storage solution (Ambion #AM7001) until use. 2 μg of RNA were transcribed to cDNA with NZY First‐Strand cDNA Synthesis kit (NZYTech #MB12501) in a C1000 Thermal Cycler (Bio‐Rad). The resulting products were amplified in duplicate with NZYSpeedy qPCR Green Master Mix (2×) kit (NZYTech #MB22402) by real‐time RT‐PCR in a CFX Connect Real‐Time System (Bio‐Rad). After amplification, a threshold for each gene and Ct (Cycle threshold) were calculated and normalized using Hprt‐1 (Hypoxanthine‐guanine phosphoribosyl transferase) as a reference gene. The results were analyzed using Bio‐Rad CFX Maestro 1.1 system software (Hercules, CA, USA). Mouse forward (F) and reverse (R) primers were designed with Beacon Designer software version 7.7 (Premier Biosoft International, Palo Alto, CA, USA) and synthesized by Eurofins (Eurofins Scientific, Luxembourg, Luxembourg) or purchased from Sigma (Sigma‐Aldrich, St. Louis, MO, USA), as follows:

Il1β (F: TCTATACCTGTCCTGTGTAATG and R: GCTTGTGCTCTGCTTGTG3); Trem2 (F: CAGGAATACTGGTGTGTA and R: TGTGGAGAATGTTTAATGTC); Hmox1 (F: CCAGTTCTACCAGAGTAA and R: ACAGAAGTTAGAGACCAA); Bdnf (F: CCACTAAGATACATCATAGC and R: CAGAACAGAACAGAACAG); Hprt1 (F: GTTGAAGATATAATTGACACTG and R: GGCATATCCAACAACAAAC) (Eurofins).

Nos2 (F: GCTGTTAGAGACACTTCTGAG and R: CACTTTGGTAGGATTTGACTTTG); Nfe2l2 (F: CATTCCCGAATTACAGTGTC and R: GGAGATCGATGAGTAAAAATGG); H‐APP (F: ATTGCCAAGAAGTCTACCC and R: AGTTCTGGATGGTCACTG); Bace1 (F: GAAGTGATCATTGTACGTGTG and R: TCTTGTCGTAGTTGTACTCC); Ppargc1α (F: TCCTCTTCAAGATCCTGTTAC and R: CACATACAAGGGAGAATTGC); Prkcg (F: CCTTTCTGTGTTCTAGATTCC and R: GCATAAAACTACCAAGTGGG); Aif/Iba1 (F: TTCATCCTCTCTCTTCCATC and R: TCAGCTTTTGAAATCTCCTC) (Sigma‐Aldrich).

Since the results are presented as ratios of treated samples/untreated (control) cells, a two‐base logarithmic transformation was used to make observations symmetric and closer to a normal distribution. If x represents the gene fold change in one sample, then the two‐base logarithmic transformation (log₂(x)) is ln(x)/ln(2). Hence, fold changes of 2 and 0.5 correspond to mean log₂ values of 1 and −1, respectively.

2.10. Cell Extracts and Protein Quantification

The cells were washed in ice‐cold PBS (pH = 7.4) and centrifuged at 300 × g for 5 min at 4°C. The total cell extracts were obtained by lysis of the cell pellet in RIPA buffer (150 mM sodium chloride, 50 mM Tris–HCl (pH = 8.0), 2 mM EDTA, 0.5% (w/v) sodium deoxycholate, 0.1% (w/v) SDS and 1% (v/v) Nonidet P‐40), supplemented with 1 mM dithiothreitol and phosphatase (PhosSTOP, Roche #04906837001; 1:10) and protease (Complete mini, Roche #04906837001; 1:7) inhibitors for 30 min in ice. Then, samples were centrifuged at 12 000 g for 10 min at 4°C, and the supernatant fraction (total extract) was collected and stored at −80°C until use.

Cell nuclear and cytoplasmic extracts were obtained with the Nuclear Extract kit (Active Motif #40010), following the manufacturer's instructions. The extracts were stored at −80°C until use.

Brain (~150 mg) lysates were prepared with 1 mL of RIPA buffer supplemented with 1% protease inhibitor cocktail, sonicated to aid tissue fragmentation and placed on ice for 15 min for subsequent centrifugation at 14 000 rpm for 30 min at 4°C. The collected supernatants were stored at −80°C until their use.

Protein concentration was determined using the bicinchoninic acid method and extrapolated from a Bovine Serum Albumin (BSA) protein standard curve. Lysates were denatured in 4× concentrated loading buffer (0.25 M Tris (pH = 6.8), 4% (w/v) SDS, 200 mM DTT, 20% (v/v) glycerol and bromophenol blue) for 5 min at 95°C.

2.11. Protein Levels and Western Blotting Analysis

2.11.1. Cell Lysates

30 μg of denatured protein was electrophoretically separated on 10% (v/v) sodium dodecyl sulfate‐polyacrylamide gels (SDS‐PAGE) at 130 V. After, proteins were electrotransferred onto PVDF membranes using a wet transfer system (Bio‐Rad, USA) at 350 mA for 210 min at 4°C. Then, membranes were blocked with 5% (w/v) of non‐fat dry milk or bovine serum albumin (BSA; for analysis of phosphorylated proteins), diluted in Tris‐buffered saline [(TBS): 150 mM NaCl, 25 mM Tris–HCl (pH = 7.6) with 0.1% Tween‐20 (TBS‐T)], for 1 h at room temperature (RT). Then, membranes were incubated with the primary antibodies against: HMOX1 (Thermo Fisher Scientific #MA1‐112; 1:500), iNOS (R&D Systems #MAB9502; 1:500), pro‐IL‐1β (Abcam #ab9722; 1:1000), Nrf2 (Santa Cruz Biotechnology #sc‐722 (C‐20); 1:1000), diluted in skin‐milk (1% in TBS‐T); pAKT (Cell Signaling Technologies #9271; 1:1000), AKT (Cell Signaling Technologies #9272; 1:1000), pGSK3β (Ser 9) (Cell Signaling Technologies #9336; 1:1000), and GSK3β (Cell Signaling Technologies #12456; 1:1000) diluted in BSA (5% in TBS‐T), ON at 4°C. Membranes were washed for 30 min (3×, 10 min) with TBS‐T and incubated with anti‐mouse (Cell Signaling Technology #7076; 1:1000) or anti‐rabbit (Cell Signaling Technology #7074; 1:1000) horseradish peroxidase‐conjugated secondary antibodies for 1 h at RT. After washing with TBS‐T (3×, 10 min), membranes were incubated with antibodies against the loading control proteins Lamin B1 (Abcam #ab16048; 1:1000) for nuclear extracts or β‐Tubulin (Sigma‐Aldrich #T7816; 1:20000) for total or cytoplasmatic extracts during 1 h at RT. Immune complexes were detected with Clarity Western ECL (Enhanced Chemiluminescent) Substrate (Bio‐Rad #1705061) and the blots were visualized by chemiluminescence using ImageQuant TM LAS 500 (GE, Healthcare, Chicago, Illinois, USA) image system. The blots images were analyzed using TotalLab TL120 (Nonlinear Dynamics Ltd., Newcastle upon Tyne, UK).

2.12. Brain Lysates

35 μg of total protein were used. Membraned were incubated with the primary antibodies against APP (Sigma‐Aldrich #A8717; 1:4000), BACE1 (Biolegend #840101; 2 μg/mL), iNOS, pro‐IL‐1β, HMOX1, phospho‐p44/42 (Cell Signaling Technology #9101; 1:1000), p44/42 MAPK (Cell Signaling Technology #9102; 1:1000) and actin (Millipore Sigma #MAB1501; 1:20000), diluted in 5% BSA. Blots were visualized by chemiluminescence with ChemiDocTM Touch Imaging System (Bio‐Rad Laboratories Inc.) and analyzed using the ImageJ software (Java) (Supporting Information).

2.13. Nrf2 Activation

5 μg of protein from nuclear extracts were used to assess Nrf2 activation using the Nrf2 Transcription Factor Assay Kit (Abcam #ab207223), according to the manufacturer's instructions. The absorbance was read at 450 nm and 665 nm (reference wavelength) in a spectrophotometer, and the activation of Nrf2 was calculated relatively to the positive control provided with the kit.

2.14. Nitric Oxide Production

Nitric oxide (NO) production was measured in the supernatants of treated BV‐2 cells using the Griess assay [37]. Absorbance was read at 550 nm in a Biotek Synergy HT (Biotek, Winooski, VT, USA) plate reader, and the nitrite concentration of each sample was extrapolated from a standard curve of sodium nitrite.

2.15. Nrf2 Silencing

Two Nrf2 Small interfering RNAs (siRNA) were used: Silencer Select Predesigned siRNA Nfe2l2 from Invitrogen (#4390771‐s70521—F: GCAUGUUACGUGAUGAGGAtt and R: UCCUCAUCACGUAACAUGCtg; #4390771‐s70523—F: CAUUUUUACUCAUCGAUCUtt and R: AGAUCGAUGAGUAAAAAUGgt). Silencer Select GAPDH siRNA (Invitrogen #4404024) and scrambled siRNA (siSilencer Select Negative Control No. 1 siRNA, Invitrogen #4390843) were also used as positive and negative controls, respectively.

siRNA were resuspended in RNAse‐free water to a concentration of 100 μM (except for siGAPDH, whose stock concentration was 200 μM) and diluted to a working concentration of 10 μM. All solutions were stored at −20°C.

The silencing process followed the manufacturer's instructions with optimization. Cells were seeded at a density of 5.55 × 10⁴ cells/cm² in a 6‐well plate for 24 h. After, the culture medium was replaced with 1 mL of Opti‐MEM Reduced Serum Medium (ThermoFisher Scientific; CAT # 31985062) 1 h before transfection. The siRNAs and the Lipofectamine RNAiMAX Transfection Reagent (ThermoFisher Scientific; CAT #13778150) were prepared in different tubes. All procedures involving siRNA handling were performed in the dark. Each siRNA was diluted in Optimem to a final concentration of 75 nM (half of the volume of both siNfe2l2 was used to achieve this final concentration). Lipofectamine was also diluted in Optimem (1:17). After, the siRNAs were mixed with Lipofectamine (1:1) and left to rest for 20 min. Next, the complexes were added dropwise to the cells, which were left to incubate for 6 h. The medium was further replaced with culture medium for 12 h. Cells were left untreated for an additional 6 h (18 h after transfection) and Nfe2l2 mRNA levels were determined (N2a‐APPswe), or cells were treated with Isoeugenol (N2a‐APPswe) or Isoeugenol+LPS (BV‐2) for 24 h (36 h after transfection), for Aβ₄₀ peptide quantification and NO production, respectively.

2.16. Animals

All procedures involving animals were under the European Community guidelines for the use of animals in a laboratory (Directive 2010/63/EU) and were approved by the Direção Geral de Alimentação e Veterinária (DGAV; Ref: 0421/000/000/2021) and performed by users licensed by the Federation for European Laboratory Animal Science Association (FELASA).

APP/PS1 double‐transgenic (B6C3‐Tg (APPswe/PSEN1dE9)85Dbo/Mmjax) mice were obtained from Mutant Resource and Research Center (MMRC)—Jackson Laboratory (Bar Harbor, ME, USA). The in‐house colony at the Center for Neuroscience and Cell Biology (CNC), University of Coimbra, was maintained in a hemizygote state by crossing transgenic male mice to B6C3F1/J female mice provided by Charles River Laboratories (France). The APP/PS1 mice and the age‐matched Wild‐Type (WT) littermates were kept under standard conditions in individually ventilated cages, under standard conditions of temperature (22°C ± 3°C) and humidity (50%–60%) on a 12 h light/dark cycle and with free access to water and food (Breeding diet A03, Barcelona). The experiments were performed in the light phase of the circadian cycle.

2.17. Isoeugenol Administration to Mice—Dose Calculation

To calculate the dose to be used in laboratory animals, we applied the inverse of the reasoning used to calculate the Maximum Recommended Starting Dose in humans. Considering the in vitro biological effect level of 250 μM with which we obtained significant results in several parameters translating anti‐inflammatory activity in BV2 cells, the equivalent dose of isoeugenol in a 70 kg man with 2.5 L of plasma would be 1.466 mg/kg. Applying the standard safety factor of 10 and assuming that there is a 1:1 species ratio when normalized for body surface area, the dose to apply in APP/PS1 mice, equivalent to the biologically active dose in humans, would be 180 mg/kg using conversion tables between various species based on body weight [38]. Nevertheless, for safety reasons, we have used a final dose of 100 mg/kg in 10‐month‐old males and 50 mg/kg in the smaller 5‐month‐old females.

2.18. Pharmacokinetic Studies

Pharmacokinetic studies were performed in WT mice with the same background (B6C3F1) as transgenic APP/PS1 mice. Mice (25–30 g) were housed in local animal facilities under controlled conditions as previously mentioned, for at least 7 days before experiments, with ad libitum access to standard rodent diet and tap water. Mice (n = 24) were randomly divided into six groups (n = 4), representing six time points (5, 15, 30, 60, 120, and 180 min). The animals were anesthetized with a mixture of the anesthetic and analgesic ketamine (100 mg/kg) and the muscle relaxant xylazine (10 mg/kg) by intraperitoneal route [39], to ensure their immobility during Isoeugenol (100 mg/kg, dissolved in sterile PBS) intranasal administration, following procedures previously established by the team [40, 41, 42]. 15 μL of Isoeugenol were instilled into the left nostril with a polyurethane tube (24G × 19 mm) attached to a 1 mL syringe, while the animal was positioned in lateral decubitus. Mice were sacrificed by cervical dislocation and decapitation, and blood was immediately collected into heparinized tubes and centrifuged (1250 × g at 4°C for 10 min) to obtain plasma samples. Brain and lungs were excised, washed with sodium chloride 0.9% solution, dried and weighed. The samples were stored at −80°C until analysis. Tissues were further homogenized with NaCl 0.9% (brain: 3 mL/g of tissue; lung: 4 mL/g of tissue) using a tissue homogenizer with a Teflon pestle from Thomas Scientific (Swedesboro, NJ, United States), centrifuged (1803 g at 4°C for 15 min). Supernatant was analyzed by high‐performance liquid chromatography (HPLC) and validated according to international guidelines (Guideline M10 on Bioanalytical Method Validation) [43] (Supporting Information, Chromatographic conditions applied for Isoeugenol quantification and Table S2).

Mean experimental concentration versus time profiles were plotted in plasma, brain, and lung, and submitted to non‐compartmental pharmacokinetic analysis using WinNonlin software, version 5.2 (Pharsight Co, Mountain View, CA, USA). The following pharmacokinetic parameters were obtained in plasma and tissues: maximum concentration (C _max); time required to reach C _max (t _max); area under the concentration‐time curve from time zero to the time of last measurable concentration (AUC _t ), and from time zero to infinity (AUC_inf); extrapolated area under the concentration‐time curve (AUC_extrap); elimination rate constant (K_el); and apparent elimination half‐life (t _1/2β). Brain/plasma and lung/plasma ratios were also determined to compare tissue and systemic exposures.

2.19. Sample Extraction Procedure for Isoeugenol Extraction From Mouse Biological Matrices (Plasma, Brain, and Lung)

In this extraction method, 100 μL of each biological matrix were spiked with 10 μL of Isoeugenol spiking solution and 10 μL of internal standard (perampanel; 60 μg/mL) spiking solution. Afterwards, 200 μL of acetonitrile was added to precipitate plasma proteins. The mixture was vortexed for 1 min and centrifuged (5 min, 13 400 rpm). The supernatant was transferred to a tube and 200 μL were placed into a vial. In the end, 20 μL were injected into the high‐performance liquid chromatography system with diode array detection (HPLC‐DAD) (Supporting Information, Chromatographic conditions applied for Isoeugenol quantification and Table S2).

Samples from in vivo studies were processed with the same method as calibration standards and quality control samples.

2.20. Chromatographic Conditions Applied for Isoeugenol Quantification

Chromatographic analysis was achieved in a Shimadzu HPLC apparatus (Shimadzu Corporation, Kyoto, Japan) composed of a solvent distribution unit (LC‐20A), a degasser (DGU‐20A5), a column oven (CTO10ASVP), an autosampler (SIL‐20AHT), and a diode array detector (SPD‐M20A). The HPLC system and data acquisition were controlled by the LCsolution software (Shimadzu Corporation, Kyoto, Japan). The chromatographic column was LiChroCART Purospher Star‐C18 (55 × 4 mm; 3 μm pore size) from Merck Millipore (Darmstadt, Germany). Chromatographic separation was performed by isocratic elution, with a mobile phase composed of Milli‐Q water, acetonitrile, and methanol (60:35:5, v/v/v) pumped at 1 mL/min. Oven temperature was set at 25°C and run time was 8 min. Injection volume was 20 μL and Isoeugenol and internal standard detections were conducted at 258 nm.

2.21. Biochemical and Histological Analyses

The safety of Isoeugenol was evaluated in an additional group of 10‐month‐old WT female mice (from the same littermate as the males further used). Awake animals were daily administered intranasally with 100 mg/kg of Isoeugenol (n = 3), or PBS (n = 4) for 1 month. After treatment and after 6 h fast, the animals were anesthetized with 100 mg/kg of ketamine (anesthetic and analgesic) and 10 mg/kg xylazine (muscle relaxant) [39], to ensure their immobility during cardiac puncture for total blood collection, organs perfusion with paraformaldehyde (PAF) and lungs insufflation. Blood was collected into serum tubes with clot accelerator, and granule serum separator, allowed to clot for 30 min, and centrifuged at 4000 × g for 20 min at RT. Metabolic parameters, namely glucose, triglycerides, cholesterol, high‐density lipoprotein (HDL), and low‐density lipoprotein (LDL) cholesterol levels were evaluated by spectrophotometric determination of chemical reactions in an Autoanalyzer [(ARCHITECT ci‐1000 Chemistry Analyzer (Abbott Laboratories, Chicago, IL, USA))] at the Clinical Analysis Laboratory at the University of Coimbra. In addition, markers of liver (glutamic‐oxaloacetic transaminase (GOT), gamma‐glutamyl transferase (GGT), and alkaline phosphatase (ALP)), and kidney (albumin, creatinine, and urea) function were also measured using the same technique to discard hepatic and renal toxicities. The results were expressed as the amount of the parameter in milligrams per deciliter or international units per liter detected in the serum of the animals (Supporting Information).

After blood collection, histopathological analysis of excised lungs, kidney, brain, heart, and liver tissues was conducted. The animals were perfused by cardiac route with PBS, and 4% (w/v) PAF solution in PBS. The organs were removed and fixed for 24 h in PFA solution at 4°C and then stored in PBS (pH 7.4) containing 0.1% (w/v) sodium azide, at 4°C until processing. After cardiac perfusion, a small incision was made on the ventral surface of the trachea, where the sheath of a 20 G angiocatheter was slid to insufflate the lungs with 1 mL of PBS, and 1 mL of the PAF. The lungs were collected and stored in PAF at room temperature until histological analysis. Then, a series of gradient concentrations of ethanol were used to dehydrate the specimens, followed by xylene clearing and impregnation in paraffin. After, tissue sections (4 μm) of the lung, kidney, brain, heart and liver were embedded in paraffin and stained with hematoxylin and eosin (H&E) [44]. Images were obtained using a Leica DM1000 LED microscope (Leica Microsystems, Wetzlar, Germany) and blind‐analyzed by an expert pathologist.

2.22. 5‐Month‐Old (Mo) Females—In Vivo Procedures and Sample Collection

Five‐month‐old females were intranasally administered with Isoeugenol (50 mg/kg, in sterile PBS) daily for 1 month. The animals were subdivided into the following groups: 5 WT mice administered with PBS (WT VEH, as disease control); 5 APP/PS1 mice administered with PBS (APP/PS1 VEH, as vehicle control); 5 APP/PS1 mice administered with Isoeugenol (APP/PS1 Iso). Intranasal administration was performed without anesthesia, using a micropipette, according to the protocol by Hanson et al. [33]. Before administration, the animals had a period of conditioning to handle. The administered volume was 6 μL per nostril, totaling 12 μL. The animals were sacrificed under anesthesia with 4% of isoflurane in an induction chamber [5], followed by decapitation for total blood collection, and the brain was quickly dissected and frozen in iced nitrogen. The blood collected in Vacuette K3EDTA tubes (Greiner Bio‐One, Kremsmünster, Austria) was centrifuged at 1000 g for 15 min at 4°C. The aliquoted samples were stored at −80°C until their use. Biological samples (brain and blood) were collected for neurochemical analysis of AD biomarkers and evaluation of the metabolic parameters (triglycerides and glycemia), respectively (Supporting Information).

2.23. 10‐Month‐Old (Mo) Males—In Vivo Procedures and Sample Collection

Ten months male APP/PS1 double‐transgenic mice and age‐matched WT littermates were randomly divided into 4 groups (n = 5 per group): WT Vehicle (VEH) and APP/PS1 Vehicle (VHE), submitted to PBS via intranasal administration for 1 month; WT Isoeugenol (Iso) and APP/PS1 Isoeugenol (Iso), submitted intranasally to 100 mg/kg/day of Isoeugenol (emulsified in PBS) for 1 month. The PBS or Isoeugenol administration followed the methodology for intranasal administration of CNS therapeutics to awake mice described by Hanson et al. [33]. Body weight was monitored during treatment and on the day before the end of treatment the intraperitoneal glucose tolerance test (ipGTT) was performed after a 6 h fast using 1.8 mg glucose per kg body weight and the evaluation of glycemia was performed at 0, 15, 30, 60, and 120 min using a glucose meter and test strips (Contour, Bayer, Leverkusen, Germany). Response to glucose was expressed by area under the curve (AUC). Serum triglycerides were measured on the same day before glucose administration (Accutrend Plus Meter mg/dl, Roche, Basel, Switzerland). In the next day, 2 h after the last dose of Isoeugenol, animals were sacrificed by decapitation for total blood collection in Vacuette K3EDTA tubes (Greiner Bio‐One, Kremsmünster, Austria) and organs removal after being anesthetized with 4% of isoflurane in an induction chamber [39]. Plasma was obtained by centrifugation of collected blood at 1000 × g for 15 min at 4°C and stored at −80°C until use. After brain removal, the cerebral cortex and the hippocampus were dissected, and the samples were immediately frozen, and stored at −80°C. Lungs, liver, and kidney were immediately frozen and stored at −80°C until use (Supporting Information).

2.24. Behavioral Tests

All the behavioral tests were performed during the light phase of the circadian cycle (between 9.00 am–5.00 pm), under red/low light. The intranasal administration of vehicle (VHE; PBS) or Isoeugenol was performed around 6.00 pm, to avoid repercussions of the acute effect of the drug on behavioral tasks and on animal sleeping cycle. For all the tests, mice were acclimated into the test room for a 1 h period. Between each trial, the apparatus was cleaned with a 10% ethanol solution to avoid odor cues. All experimental data were analyzed using the ANY‐maze video tracking system, Stoelting, US.

“Hot plate” and “Open field” were performed to ensure animal equivalence between experimental groups regarding pain sensitivity and locomotor activity, respectively, and “Elevated plus maze” was done to evaluate animals' anxiety levels (Supporting Information).

The modified “Y‐maze” and “Novel object recognition” tasks, which rely on the innate preference of rodents to explore novelty, were performed as a measurement of short‐term spatial and social recognition memory, respectively. The “Fear condition” test was performed to evaluate the hippocampal‐dependent long‐term contextual/associative memory.

2.24.1. Hot Plate Test

The hot‐plate test (LE7406 Hot‐Plate, Panlab, Barcelona, Spain) evaluates thermal pain reflexes due to footpad contact with a heated surface and it was used to ensure animal equivalence between experimental groups in terms of pain sensitivity, since fear condition test implies the use of footstock. During the experiments, the animal was placed on a thick aluminium plate (10 mm) at 55°C and confined in a removable clear acrylic cylinder where the latency time to the first hind paw and/or jumping responses were measured (Supporting Information).

2.24.2. Open Field

Mice were placed in a 42 × 42 cm arena with 42 cm high walls and they were allowed to explore for 10 min. The time spent on the central and peripheral zones of the chamber, the number of entries in each one of them, and the values of mean speed and total distance traveled were analyzed (Supporting Information).

2.24.3. Elevated Plus Maze

The apparatus was composed of two open arms and two enclosed arms (30 × 5) with 15 cm high walls arranged so that the arms of the same type are opposite to each other with a central square of 5 cm. The apparatus was elevated to a height of 50 cm above floor level. The trial involved placing the individual animal on the central platform of the maze facing an open arm. The number of entries and the time spent in closed and open arms were recorded for a 5 min test. Entry into an arm was defined as the animal placing all four paws onto the arm (Supporting Information).

2.24.3.1. Y‐Maze

The Y‐maze apparatus, made from impermeable black formica, consisted of three arms (30 cm long, 20 cm height, and 5 cm large) placed at 120° intervals. Arms were equipped with different internal visual cues placed on the side and end walls of each arm. During the training phase, one arm was blocked by a removable door. In this phase, the mice were positioned in the start arm, facing the center of the maze and allowed to explore only two arms (start and other) for 8 min. The test was performed 2 h later (with the door blocking the novel arm removed), and the animals were placed again in the start arm and allowed to explore the three arms during 8 min. The number of entries and the time spent in the novel arm were measured.

2.24.3.2. Novel Object Recognition

Mice were habituated for a 10‐min period in the empty chamber during the open field test. Then, in the training phase, performed 1 h after the habituation phase, two identical 200 mL brown glass bottles were placed 15 cm away from the walls of the arena. The mice were placed at the center of the apparatus and allowed to explore these two identical objects for 10 min, and we recorded the time spent sniffing/whisking or looking (≤ 1 cm) at the objects. The test was performed 1 h and 30 min later, where one of the objects was replaced by a 50 mL volumetric flask filled with dyed sterile water with the same diameter creating a novelty for the animal to explore during 5 min; again, the time spent exploring both the novel and the familiar objects was recorded. The recognition index was determined by the percentage of time spent investigating the novel object over the time spent exploring both objects.

2.24.3.3. Fear Conditioning

Fear conditioning is a form of associative learning that occurs when a previously neutral stimulus (e.g., tone) elicits fear responses after it has been paired with an aversive stimulus (e.g., electric paw shock). On day one (training day), the mice were submitted individually to an A context, and after 2 min of habituation, to two presentations of an auditory conditioned stimulus (95 dB for 30 s), paired with a paw shock unconditioned stimulus (0.5 mA for 2 s), applied at the end of the auditory stimulus, with a 60s interval between presentations. One minute after the last pair of auditory‐shock stimuli, the animals were placed back in the housing cage. On day two (contextual memory) mice were returned to context A in the conditioning chamber and “freezing” behavior was measured over a period of 5 min. On the third day (associative memory) animals were placed in a different context B (conditioning chamber with perspex floor and walls with different patterns from those of context A and change of lighting) and “freezing” behavior was measured over a period of 3 min, after which the auditory stimulus was introduced for a further 3 min, during which the “freezing” behavior was also evaluated.

2.25. Estimated Animal Sample Size

Considering the mean and standard deviation values of Aβ peptides levels observed in 6 months APP/PS1 untreated and Iso‐administered animals, an effect size of 2.49 was obtained by using a one‐tail t‐test and assuming a normal distribution. For a power set to 0.95 and a level of significance of 0.05, the estimated sample size obtained was 5 individuals per group (using G*Power version 3.1.9.7. software).

2.26. Statistical Analysis

Results are presented as mean ± SEM of the number of experiments mentioned in the Figures and were analyzed with ANOVA or unpaired t‐test for comparisons between two groups.

The in vitro results, analyzed with one‐way ANOVA followed by Dunnet's multiple comparison posttest, were compared to untreated cells (Ctr). Comparisons between Ctr and the vehicle (DMSO) were performed, and no differences were detected.

The in vivo results, analyzed with two‐way ANOVA followed by a Tukey's multiple comparison posttest, were compared relative to healthy animals of the control group (WT) or transgenic control group (APP/PS1). GraphPad Prism 8.0.2 (GraphPad Software, San Diego, CA, USA) was used to perform the statistical analysis of the results. A value of p < 0.05 was considered significant.

3. Results

3.1. Isoeugenol Reduces Aβ₄₀ Peptide Levels in AD Neuronal Cells

To validate and support the use of Amyloid‐beta precursor protein (APP)‐overexpressing cells with Swedish mutation (N2A‐APPswe) as an AD model, we measured the levels of secreted Aβ₄₀ and Aβ₄₂, the main components of senile plaques in AD. N2a‐wt and N2a‐APPswe cells were left untreated or treated with Isoeugenol (Iso) at the highest, nontoxic concentration tested (250 μM; Figure S1A) for 24 h, and peptide levels were quantified by ELISA (Figure 1).

FIGURE 1.

Effect of Isoeugenol on Aβ peptides levels in neuronal cells. (A) Aβ₄₀ and (B) Aβ₄₂ levels in N2a‐wt and N2a‐APPswe cells exposed to Isoeugenol (Iso, 250 μM). Data corresponds to the mean ± SEM of four independent experiments. Statistics: One‐way ANOVA with Tukey's multiple comparisons test. p < 0.05 was considered significant. **p < 0.01 and ****p < 0.0001, compared to N2a‐wt; ##p < 0.01, compared to N2a‐APPswe.

As expected, N2a‐APPswe cells secreted higher levels of the peptides Aβ₄₀ (~8 fold increase; Figure 1A) and Aβ₄₂ (~2.4 fold increase; Figure 1B) than N2a‐wt cells. Moreover, Iso treatment significantly decreased Aβ₄₀ levels in N2a‐APPswe, compared to untreated cells (Figure 1A), although no effect was observed on Aβ₄₂ peptide levels (Figure 1B).

3.2. Isoeugenol Activates the Nrf2 Pathway in an AD Cell Model—Involvement of AKT/GSK3β Signaling

As an electrophilic molecule, with a pre‐Michael acceptor domain, Iso reacts with cysteine residues [28], potentially leading to Nrf2‐Keap1 detachment and further Nrf2 activation. However, it has been suggested that the activation of serine/threonine‐specific protein kinases such as AKT and consequent inactivation of GSK‐3β (through Ser9 phosphorylation by AKT) might also upregulate Nrf2 [20, 22]. Hence, AKT and GSK3β phosphorylation levels were determined in N2a‐wt and N2a‐APPswe cells after Iso exposure for 5, 15, and 30 min (Figure 2A–C).

FIGURE 2.

Effect of Isoeugenol on AKT, GSK3β and Nrf2 activation in AD neuronal cells. (A, B) WB analysis of AKT and (C) GSK3β phosphorylated (p) protein levels, in N2a‐APPswe cells exposed to Iso (250 μM) for 5, 15, or 30 min. The basal levels of pAKT (A) in untreated N2a‐wt and N2a‐APPswe neuronal cell lines were determined at the longest time‐point evaluated (30 min). (D) WB analysis of cytoplasmatic (cyt) and nuclear (nuc) Nrf2 levels, in N2a‐APPswe cells exposed to Iso for 2, 4, and 6 h. (E) Nrf2 activation determined in N2a‐APPswe cells after 2, 4, and 6 h of Iso exposure, using a commercial kit. (F, G) Hmox1 mRNA levels determined by real‐time RT‐PCR, in N2a‐APPswe cells exposed to Iso for 1, 3, and 6 h. The basal levels of Hmox1 gene (F) in untreated N2a‐wt and N2a‐APPswe neuronal cell lines were determined at the longest time‐point evaluated (6 h). (H) HMOX1 protein levels determined by WB, in N2a‐APPswe cells exposed to Iso for 24 h. Representative blot images are presented (bands correspond to the proteins run in the same gel. See Supporting Information for uncropped images). Values are the mean ± SEM of three to four independent experiments and expressed relatively to untreated cells (control, Ctr). Statistics: Unpaired t‐test (t), one‐way ANOVA with Dunnett's multiple comparisons test (*) and two‐way ANOVA with Sidak's multiple comparisons test (§). p < 0.05 was considered significant. t and *p < 0.05, compared to N2a‐wt or Ctr cells; tt and **p < 0.01, §§§p < 0.001, ****p < 0.0001 compared to Ctr cells (in H, Ctr = 0 [Log₂1]).

As depicted in Figure 2A, AKT phosphorylated levels are significantly reduced in N2a‐APPswe cells (compared to N2a‐wt), suggesting defective AKT activation. Iso‐induced AKT activation (Figure 2B) paralleled GSK3β inactivation (Figure 2C) in N2a‐APPswe cells exposed to Iso for 30 min. We further analyzed Nrf2 cytoplasmatic (Nrf2 cyt) and nuclear (Nrf2 nuc) protein levels, as well as Nrf2 activation (by ELISA), in these cells exposed to Iso for 2, 4, and 6 h (Figure 2D,E, respectively). The results showed that Nrf2 nuclear levels were significantly increased in N2a‐APPswe cells exposed to Iso for 6 h (Figure 2D) compared to untreated cells (Ctr), suggesting its nuclear translocation and activation, as corroborated by the results obtained with the ELISA kit (Figure 2E). We also analyzed ARE‐Nrf2‐dependent Heme Oxygenase 1 (Hmox1) gene expression and protein levels (HMOX1), in N2a‐APPswe cells exposed to Iso for 1, 3, and 6 h (Figure 2F,G) or 24 h (Figure 2H). Hmox1 gene expression was reduced in N2a‐APPswe cells (compared to N2a‐wt; Figure 2F), which was significantly reverted by Iso after 3 h and 6 h of cell exposure (Figure 2G), and reflected the significant increase in HMOX1 protein levels observed after 24 h in N2a‐APPswe cells (compared to the control, Ctr; Figure 2H). Moreover, the expression of other Nrf2‐dependent genes (i.e., NAD(P)H quinone oxidoreductase 1 (Nqo1) and superoxide dismutase 1 (Sod1)) was also induced by this molecule (Figure S2). Altogether, these results strongly suggest that the Nrf2‐dependent antioxidant response is activated by Iso, in our AD cellular model.

3.3. Isoeugenol Reduces Pro‐Inflammatory Parameters in a Neuroinflammation Cell Model

To study the effect of Iso on neuroinflammation, LPS‐stimulated microglia were treated with Iso and the production of pro‐inflammatory mediators was evaluated (Figure 3).

FIGURE 3.

Effect of Isoeugenol on inflammatory parameters in microglia cells exposed to LPS. BV‐2 cells were exposed to Iso (250 μM), in the presence or absence of LPS (50 ng/mL, 30 min after Iso incubation). (A) Nos2 and (D) Il‐1β mRNA levels determined by real‐time RT‐PCR, in BV‐2 cells exposed to Iso for 18 h. (B) iNOS and (E) pro‐IL‐1β protein levels determined by WB, in BV‐2 cells exposed to Iso for 24 h. (C) Nitric Oxide production inferred by determining nitrite levels through the Griess assay, in BV‐2 cells exposed to Iso for 24 h. Representative blot images are presented (bands correspond to the proteins run in the same gel. See Supporting Information for uncropped images). Values are the mean ± SEM of four independent experiments and expressed relatively to control (Ctr) untreated cells. Statistics: One‐way ANOVA with Tukey's multiple comparisons test. p < 0.05 was considered significant. **p < 0.01, ***p < 0.001 and ****p < 0.0001, compared to Ctr cells (in A and D, Ctr = 0 [Log₂1]); ##p < 0.01, ###p < 0.001 and ####p < 0.0001, compared to LPS.

As expected, the Toll‐like receptor agonist LPS (50 ng/mL) induced a significant increase in Nos2 (Figure 3A) and Il‐1β (Figure 3D) gene expression (after 18 h), as well as in the levels of their encoded proteins, inducible nitric oxide synthase (iNOS) (Figure 3B), and pro‐interleukin (IL)‐1β (Figure 3E) (after 24 h), which was reduced by the highest, nontoxic concentration tested of Iso (250 μM; Figure S1B). Moreover, LPS‐triggered nitrite production was significantly decreased in microglia cells exposed to Iso (Figure 3C). These results suggest that Iso has a potent anti‐inflammatory role, capable of counteracting neuroinflammation, a chronic pathological event in AD.

3.4. Isoeugenol‐Induced Decrease in NO Production Is Mediated by Nrf2

To determine if Iso‐induced reduction of Aβ₄₀ and NO levels was mediated by Nrf2, we performed an independent set of experiments (N = 3) in which Nrf2 was silenced (siNrf2) in N2a‐APPswe and BV‐2 cells, respectively. The cells were also transfected with scrambled siRNA as a negative control (siNeg) and siGAPDH to demonstrate the efficiency and specificity of the silencing.

Nrf2 silencing in N2a‐APPswe cells was confirmed by determining the reduction in Nfe2l2 mRNA levels, 18 h after transfection (siNrf2 vs. siNeg; Figure 4A). In addition, we inferred the stability and specificity of the silencing after 36 h by determining GAPDH protein expression levels. Accordingly, GAPDH protein was reduced in cells transfected with siGAPDH (51% in N2a‐APPswe and 44% in BV‐2 cells, Figure S3A,B) compared to siNeg, with no alterations detected in siNrf2 cultures.

FIGURE 4.

Effect of Nrf2 silencing on Iso‐induced reduction of Aβ40 and NO production. (A) Nfe2l2 mRNA levels in N2a‐APPswe cells transfected with siNrf2, siGAPDH or siNeg for 18 h. Aβ₄₀ levels in not transfected (B) or siRNA‐transfected (C) cells for 36 h, untreated or exposed to Iso. (D) Nitrite levels determined by the Griess assay in BV‐2 cells not transfected or transfected with siNrf2, siGAPDH or siNeg for 36 h, untreated or exposed to Iso and/or LPS. Data corresponds to the mean ± SEM of three independent experiments. Statistics: One‐way ANOVA with Dunnett's or Tukey's (D) multiple comparisons test. p < 0.05 was considered significant (*p < 0.05, **p < 0.01 and ****p < 0.0001). In (D): ****p < 0.0001, compared to Ctr and ####p < 0.0001, compared to LPS.

In this set of experiments, the levels of Aβ₄₀ peptide observed in N2a‐APPswe cells exposed to Iso were significantly decreased (vs N2a‐APPswe), and comparable to those observed in these cells transfected with siNeg + Iso (N2a‐APP‐swe + Iso vs. siNeg + Iso; Figure 4B) and siGAPDH + Iso (siNeg + Iso vs. siGAPDH+Iso; Figure 4C), implying that the silencing did not affect the Iso effect. Strikingly, Nrf2‐silenced cells also exhibited equivalent Aβ₄₀ levels (siNrf2 vs. siNeg + Iso and siGAPDH + Iso; Figure 4C), which were further decreased in the presence of Iso (siNrf2 + Iso; Figure 4C), in a synergistic‐like effect.

In BV‐2 cells, the Iso‐induced NO decrease in LPS‐activated microglia was reversed in cells transfected with siNrf2, while it was not affected in cells transfected with siGAPDH or siNeg and exposed to Iso + LPS (Figure 4D), aligning with the pattern of iNOS expression in these cultures (Figure S3C). These results demonstrate that Iso inhibition of NO production in BV‐2 cells is mediated through Nrf2 activation.

3.5. Intranasal Administration of Isoeugenol Is Safe for Mice

Considering the previous results and our in chemico observations indicating that Iso could cross the BBB (Figure S4) and inhibited Beta‐secretase 1 (BACE1) activity (Table S1), the enzyme responsible for the cleavage of APP and production of Aβ peptides, the potential therapeutic effect of Iso in AD was investigated in APP/PS1 transgenic mice. These animals display a predisposition to produce insoluble forms of Aβ, with astrocytosis and severe gliosis in the proximity of the developing plaques at 6 months [45], exhibiting abundant plaques in the hippocampus and cortex at 9 months of age [46].

To evaluate Iso safety, we performed pharmacokinetic studies and organ histological analysis of WT animals (with the same genetic background as APP/PS1 mice) intranasally administered with Iso (100 mg/kg, single dose) (Figure 5).

FIGURE 5.

Pharmacokinetic studies and histopathological analysis. Concentration‐time profile in biological matrices of plasma (A), brain (B), and lung (C) of WT mice after 5, 15, 30, 60, 120, and 180 min of Iso intranasal administration (100 mg/kg). Symbols correspond to mean values ± SEM (n = 4). Histopathology of the heart (D), hippocampus (E), lung (F), liver (G), and kidney (H) of 11‐month‐old WT female mice (from the same littermate as APP/PS1 mice) intranasally administered with Vehicle (PBS) or Iso (100 mg/kg) for 1 month. Representative images of each organ sections stained with hematoxylin and eosin at 40×, 100×, or 200× magnification.

The concentration‐time profiles of Iso in mice plasma, brain, and lung are presented in Figure 5A–C, and pharmacokinetic parameters are shown in Table 1.

TABLE 1.

Pharmacokinetic parameters of Iso in mouse plasma, brain and lung after intranasal administration (100 mg/kg).

Pharmacokinetic parameters a	Plasma	Brain	Lung
t max (min)	5.00	5.00	5.00
C max (μg/mL)	26.41	4.34 b	348.16 b
AUC t (μg min/mL)	599.05	140.87 c	4030.85 c
AUCinf (μg min/mL)	604.28	142.15 c	4091.75 c
AUCextrap (%)	0.86	0.90	1.49
kel (min−1)	0.038	0.039	0.022
t 1/2β (min)	18.12	17.65	31.27
AUC t ratios
AUCbrain/plasma	0.23	—	—
AUClung/plasma	6.73	—	—

Abbreviations: AUC_extrap, extrapolated area under the concentration‐time curve; AUC_inf, area under the concentration‐time curve from time zero to infinity; AUC _t , area under the concentration‐time curve from time zero to the time of last measurable concentration; C _max, maximum concentration; k_el, elimination rate constant; t _1/2β, apparent elimination half‐life; t _max, time required to reach the maximum concentration.

^a Parameters estimated using mean concentration‐time profiles obtained from 4 mice per time point (n = 4).

^b Values expressed in μg/g.

^c Values expressed in μg.min/g.

Regarding AUC_extrap values (< 20%) in plasma, brain and lung, it was confirmed that the number of analyzed samples was adequate for a reliable calculation of K_el, AUC _t and AUC_inf values (Table 1). Accordingly, Iso was absorbed quickly (t _max = 5 min in plasma) and removed from the blood due to either a fast distribution into peripheral tissues (t _max = 5 min in brain and lungs) or elimination (t _1/2β = 18.12) (Table 1). These results agree with those observed by others, after intravenous and oral administration of Iso [47, 48].

Although Iso levels in the brain and lung were not determined by Hong et al. [48], the authors suggested that the molecule might display a fast distribution to tissues due to its low alpha half‐life distribution time (7–10 min) and steady‐state apparent volume of distribution (11–25.2 L/kg) after intravenous dosing, which exceeds total body water volume (0.6 L/kg) and indicates extensive distribution to extravascular tissues. Iso fast distribution might also be related to its low molecular weight (164.2 g/mol) and lipophilicity (octanol–water partition coefficient (logP) = 3.04), allowing it to quickly spread across cell membranes. On the other hand, its fast elimination may be due to Iso‐facilitated conversion into phase II metabolites [47], hence promoting its excretion as glucuronide or sulfate derivatives.

Considering the doses of Iso administered by Hong et al. [48] to mice by intravenous (35 mg/kg) and oral routes (35, 70, and 140 mg/kg), AUC values in plasma were dose‐normalized to compare with AUC_inf and AUC _t values of the present study (100 mg/kg) after intranasal administration. The dose‐normalized AUC_inf value after intranasal administration (6.04 μg min/mL/kg) was similar to that observed after intravenous administration (6.76 μg min/mL/kg). In contrast, dose‐normalized AUC _t values (1.91, 1.68, 1.45 μg min/mL/kg) after oral administration were lower than those of the present study (5.99 μg min/mL/kg), suggesting that the systemic exposure of Iso after intranasal and intravenous administration routes is identical, and both are higher than the oral route.

Some differences were observed between the three biological matrices analyzed (Table 1). Importantly, the drug was quantified in the brain, where the therapeutic target is located, almost immediately after being administered (Figure 5B). In lung, C _max and AUC _t values were higher than in other matrices, as well as lung‐plasma ratios, suggesting direct passage of Iso to the lungs, when administered intranasally. Although t _1/2β values were short for all matrices, t _1/2β was higher in lung (31.27 min vs. 18.12 e 17.65 min; Table 1), suggesting that Iso elimination from the lung is slower compared to the plasma and the brain. For this reason, histopathological analysis of the heart, hippocampus, lung, liver and kidney of 11‐month‐old female WT mice (from the same littermate as APP/PS1 mice), administered with vehicle (VHE, PBS) or Iso (100 mg/kg) for 1 month, were assessed in hematoxylin‐ and eosin‐stained sections of each organ by microscopy (Figure 5D–H). No histopathological changes were observed in the heart (Figure 5D) of Iso‐treated mice relative to the VEH group. We specifically assessed the interventricular septum as the anatomical structure controlling the transmission of the action potential between the atria and ventricles to exclude any pro‐arrhythmic potential. Increased cell activation in the hippocampus was observed after Iso administration (filled black arrows; Figure 5E). Some infiltration of lymphocytes (filled black arrow) and bronchiolar hyperplasia (empty black arrow) were detected in the lung slices of the Iso‐treated mice (Figure 5F), although the latter was observed in some animals of the VEH group. Histopathological analysis of liver slides showed hepatic steatosis in the control animals (VEH), revealing a pattern of histopathological alterations in the livers of these animals, which was also maintained in Iso‐treated mice (filled black arrows; Figure 5G). In addition, cellular eosinophilia and tubular epithelial hypertrophy were observed in the kidneys of Iso‐treated animals (filled black arrows; Figure 5H). Yet, no alterations were detected in these animals regarding serum levels of liver parameters (i.e., serum glutamic oxaloacetic transaminase (GOT), gamma‐glutamyl transferase (GGT), and alkaline phosphatase (ALP); Figure S5F–H) and kidney parameters (i.e., albumin, creatinine, and urea levels; Figure S5I–K), sustaining the safety of Iso administration.

3.6. Intranasal Administration of Iso Improves Memory Deficits in 11‐Month‐Old APP/PS1 Mice

Both male and female APP/PS1 mice display similar, sex‐independent learning and memory deficits at 9 months of age, although, in males, the memory deficits are slightly more severe [49]. In fact, we conducted a preliminary study with 6‐month‐old females intranasally administered with Iso (50 mg/kg) for 1 month and no relevant behavioral alterations were observed (Figure S6). APP/PS1 untreated animals showed increased anxious behavior compared to WT animals (inferred by the time spent to make the first entry in the central zone of the open field arena; Figure S6D), which was expected because anxiety is a prodromic symptom of AD. Also, diseased animals showed enhanced “freezing” behavior in an environment associated with an aversive stimulus, which was not observed in untreated mice. However, this was only detected on the training day of the Fear Conditioning test (Figure S6G), with no alterations registered on day two (evaluation of contextual memory) or three (evaluation of associative memory). We hypothesized that the % of total freezing observed in non‐treated APP/PS1 mice on the first day might be related to the animal's distrustfulness (compared to WT animals). As expected, these animals displayed increased Aβ₄₀ and Aβ₄₂ levels in the brain and both were decreased in Iso‐treated mice, compared to untreated animals (Figure S7A,B). Moreover, these animals also showed an increase in APP protein and a decrease in p44/42 MAPK phosphorylated levels (which was reverted by Iso) (Figure S8A,F, respectively), compared to WT mice, with no major phenotypical‐related differences detected on the other proteins analyzed (iNOS, pro‐lL‐1β, BACE1 and HMOX1).

Thus, we evaluated the effect of Iso treatment on the memory of WT and APP/PS1 male mice at 10 months of age. To ensure animal welfare, the body weight of the animals was measured every week during the intranasal administration period (1 month) and the metabolic parameters (glucose levels, fasting glucose and triglycerides; Figure S9) were determined. Apart from a decrease in body weight of Iso‐treated animals (< 15%, hence not compromising animal well‐being; Figure S9B), no other relevant differences were observed.

These animals were equivalent regarding pain sensitivity (assessed by the Hot Plate Test), locomotion and natural exploratory behavior (assessed by the Open Field), and displayed similar anxiety levels (assessed by the Elevated plus‐maze test) (Figure S10). The results demonstrated that VEH‐treated APP/PS1 mice spent less time and apparently entered less in the novel arm compared to VEH‐exposed WT mice (Figure 6A,B, respectively) and, although not statistically significant, Iso reversed this trend to levels comparable to those of WT mice (APP/PS1 Iso vs. WT VEH; Figure 6A,B). The distance traveled was not affected in the different experimental conditions (Figure 6C). Moreover, the recognition index of VEH‐treated APP/PS1 mice was significantly lower than that of VEH‐exposed WT mice, which was reversed in mice treated with Iso (Figure 6D). A qualitative decrease in the percentage of investigations in the novel object was also observed in VEH‐treated APP/PS1 (vs. VEH‐exposed WT mice), which were reverted by Iso administration to similar levels as those of WT animals (Figure 6E). Overall, these results indicate that the impaired spatial and nonemotional memory in APP/PS1 mice is improved by Iso treatment. APP/PS1 mice showed reduced contextual memory since the % of freezing when placed in the aversive environment on day 2 was significantly reduced, compared to the WT VHE (Figure 6F). In contrast, treated mice displayed a clear but not significant increase in the % of freezing, suggesting that Iso might enhance APP/PS1 mice contextual memory. Regarding animals' associative memory, no differences were detected between groups (Figure 6G).

FIGURE 6.

Effect of Isoeugenol on WT and APP/PS1 memory. (A–C) Y‐maze test, performed to evaluate spatial memory, (D, E) Novel Object Recognition test, performed to evaluate social recognition memory, and (F, G) Fear Conditioning test performed to evaluate hippocampal‐dependent long‐term contextual/associative memory of 11‐month‐old male mice intranasally administrated with VEH or Iso (100 mg/kg) for 1 month. The results represent the mean ± SEM of four to seven animals. Statistics: Two‐way ANOVA with Tukey's multiple comparisons test (* and #). p < 0.05 was considered significant. (*) p < 0.05 and (**) p < 0.001, compared to WT VEH; (#) p < 0.05, compared to APP/PS1 VEH.

3.7. Isoeugenol Modulates the Hippocampal Transcriptional Signature of APP/PS1 Mice and Reduces Aβ Peptide Levels in the Plasma

The hippocampus is the initial region for Aβ accumulation and neuronal damage [50] and a key brain region involved in learning and memory. Thus, we wondered if the memory improvement in Iso‐treated APP/PS1 mice could be related to gene alterations in the hippocampus of these animals. We evaluated the transcription of genes related to the (1) amyloidogenic pathway (i.e., hAPP, Bace1), (2) inflammatory pathway (i.e., Il1β), (3) microglia function (i.e., ionized calcium‐binding adaptor molecule 1, Iba1 and Triggering receptor expressed on myeloid cells 2, Trem2), (4) antioxidant cell defense (i.e., Hmox1), (5) energy metabolism (i.e., peroxisome proliferator‐activated receptor gamma coactivator 1‐alpha, Ppargc1α), and (6) neuroprotection (i.e., brain‐derived neurotrophic factor, Bdnf) (Figure 7).

FIGURE 7.

Effect of Isoeugenol on gene expression in the hippocampus of WT and AD mice. mRNA levels of the genes (A) hAPP, (B) Bace1, (C) Il‐1β, (D) Iba1, (E) Trem2, (F) Hmox1, (G) Ppargc1α, and (H) Bdnf determined in the hippocampus of 11‐month‐old WT and APP/PS1 male mice, intranasally administered with Iso (100 mg/kg) or PBS (VEH), for 1 month. The results represent the mean ± SEM of four animals. Statistics: Two‐way ANOVA with Tukey's multiple comparisons test. p < 0.05 was considered significant. (*) p < 0.05 and **p < 0.01, compared to WT VEH; #p < 0.05, ##p < 0.01, and ####p < 0.0001, compared to APP/PS1 VEH.

As anticipated for this AD mouse model, the h‐APP gene expression was significantly augmented in APP/PS1 transgenic animals (compared with WT) (Figure 7A), which was not affected by Iso treatment. These results corroborate the increase in APP protein levels in the brain of 6‐month‐old mice, which was also not altered by Iso (Figure S8). No significant differences were detected in the expression of Bace1 and Il1β among groups, although a clear tendency toward a reduction in the mRNA levels of these genes was observed in WT and APP/PS1‐treated animals compared to untreated mice (Figure 7B,C). Moreover, Iso increased the expression of Iba1 in APP/PS1 (Figure 7D), decreased Trem2 mRNA levels in WT mice (Figure 7E) and increased the expression of the antioxidant gene, Hmox1, in both WT and APP/PS1‐treated mice (Figure 7F). Interestingly, phenotype‐related alterations in the expression of Ppargc1α and Bdnf genes were detected (Figure 7G,H, respectively). Transgenic mice (APP/PS1 VEH) showed a significant reduction in the mRNA levels of Ppargc1α as well as a nonsignificant decrease (p = 0.0791) in Bdnf gene, compared to WT VEH, which was reversed by Iso treatment (APP/PS1 + Iso vs. APP/PS1). These results suggest that Iso might modulate microglia function (e.g., Iba1 and Trem2), antioxidant mechanisms (e.g., Hmox1), energy metabolism (e.g., Ppargc1α) and neuroprotection (e.g., Bdnf) at the transcriptional level.

Since brain‐derived Aβ can be effluxed to plasma [51] and further cleared by peripheral tissues and organs, resulting in improved memory in AD [52] we analyzed Aβ₄₀ and Aβ₄₂ levels in the cortex, hippocampus, and plasma of WT and APP/PS1 mice (Figure 8).

FIGURE 8.

Effect of Isoeugenol on Aβ peptides levels in the cortex, hippocampus, and plasma of 11‐month‐old AD mice. (A) Aβ40 levels in the cortex, (B) hippocampus, and (C) plasma, and (D) Aβ42 levels detected in the cortex, (E) hippocampus, and (F) plasma of WT and APP/PS1 11‐month‐old male mice, administered with Iso (100 mg/kg) for 1 month. (G) Aβ42/Aβ40 ratio in the plasma. The results represent the mean ± SEM of five animals. Statistics: Two‐way ANOVA with Tukey's multiple comparisons test. p < 0.05 was considered significant. ***p < 0.001 and ****p < 0.0001, compared to WT VEH; ##p < 0.01, compared to APP/PS1 VEH.

As expected, Aβ₄₀ (Figure 8A–C) and Aβ₄₂ (Figure 7D–F) peptide levels were significantly higher in the cortex, hippocampus, and plasma of APP/PS1 mice compared to WT mice. Iso treatment decreased both Aβ₄₀ and Aβ₄₂ levels in the plasma of APP/PS1 mice (Figure 8C,F, respectively), with no alterations observed in the cortex and hippocampus. We further calculated the plasma Aβ42/Aβ40 ratio (Figure 8G), which is increasingly recognized as a clinically relevant biomarker for Alzheimer's disease. APP/PS1 mice exhibited a significantly reduced Aβ42/Aβ40 ratio compared with WT mice, and Iso treatment produced a partial increase in the ratio, although this did not completely restore WT values. Given the lack of statistical significance, these findings further support the notion that the cognitive benefits of Iso are more likely attributable to modulation of neuroinflammatory, antioxidant, and neurotrophic pathways rather than reversal of established amyloid pathology.

4. Discussion

The research data gathered from diverse neurodegenerative disease models (e.g., AD) in the last decade strongly support that Nrf2 activation in the brain might contribute to the attenuation of disease progression. Excellent results have been obtained with a panel of small molecules that activate Nrf2 in preclinical neurodegenerative disease models. These Nrf2 activators are electrophilic molecules that react with Keap1 cysteine residues, hindering Keap1 from presenting Nrf2 for ubiquitination and degradation. However, molecules able to cross the BBB with good pharmacokinetic and pharmacodynamic profiles need to be identified [53]. Nrf2 activation is a key event triggered by all electrophilic low molecular weight skin allergens. Yet, electrophilic drugs can nonspecifically react with thiol groups, depleting glutathione (GSH, the predominant thiol in normal cells) and increasing toxicity in normal and unstressed cells. Iso is not a direct Michael acceptor, instead it is considered a pro‐electrophilic molecule. Pro‐electrophilic drugs are activated by oxidation in redox‐stressed cells where GSH is already depleted, thus being advantageous and expected to be better clinically tolerated over conventional electrophilic compounds (including DMF). Iso's anti‐inflammatory and antioxidant properties, attributed to the presence of a phenolic group in its structure [54], have been previously reported in several cellular and animal models. It was shown to inhibit the expression of iNOS and decrease NO levels in macrophages exposed to LPS [55], reduce the levels of NO and reactive oxygen species (ROS) in the brain of rats exposed to acrylamide [56], decrease lipid peroxidation in the liver and brain of normal rats [57], and to have neurocognitive benefits in animal models of amnesia [58]. Also, according to a molecular simulation model, Iso protects DNA against oxidative damage caused by hydroxyl radicals [59].

In this study, we evaluated the therapeutic potential of Iso in activating Nrf2 and reverting AD hallmarks. Iso was shown to activate Nrf2 in the KeratinoSens gene reporter cell line [28], but the antioxidant properties of Iso associated with Nrf2 activation in vivo were only demonstrated in a Nrf2‐mutant zebrafish‐based assay system [60]. Recently, Methyl Isoeugenol (CAS 93‐16‐3), an analog of Iso, was shown to activate Nrf2 and ameliorate cerebral oxidative stress in a rat model of cerebral ischemia–reperfusion injury [61].

In our study, we demonstrated that Iso induced Nrf2 activation in N2a‐APPswe cells by (1) increased nuclear levels of Nrf2 protein, (2) increased Nrf2 activation (determined by ELISA), and (3) increased Nrf2‐dependent antioxidant gene expression (Hmox1; Nqo1 and Sod1, Figure S2). Both Nrf2 activation and the increase in HMOX1 protein levels triggered by Iso in the neuronal cell line suggest its potential role in neuroprotection since HMOX1 is associated with the neuroprotective effect of Nrf2 activation, as reported in several cellular and animal models of neurodegenerative diseases [62, 63, 64]. Although Iso might contribute to Nrf2 activation due to Keap1 modification, we demonstrated that it also activates the AKT‐GSK3β signaling pathway. AKT activation (due to phosphorylation by PI3K) is known to further inactivate GSK3β, allowing Nrf2 translocation to the nucleus [22, 65, 66]. We showed that in N2a‐APPswe cells treated with Iso for 30 min, there was an increase in the phosphorylation levels of AKT, which were significantly lower in untreated cells, reflecting a dysfunctional pathway [67], and also in GSK3β. In addition, through the Keratinosens reporter gene assay, we confirmed that Iso induced Nrf2 activation after 6 h (as the positive control, DMF; Figure S11A). In this assay, NRF2 activators act mainly by covalently modifying cysteine residues of Keap1. However, Iso‐induced Nrf2 activation was partially inhibited by Wortmannin, an irreversible PI3K inhibitor (Figure S11B), supporting the involvement of AKT.

The anti‐inflammatory potential of Iso to counteract neuroinflammation was demonstrated in LPS‐stimulated microglia, with an observed decrease in iNOS and pro‐IL‐1β, at the transcriptional and posttranscriptional levels, as well as in NO production. Inhibition of iNOS expression and decreased NO levels by Iso have been previously observed in LPS‐exposed macrophages [55, 68] and shown to be related to the downregulation of the transcription factor nuclear factor‐κB (NF‐κB) [68]. In our study, we demonstrated that NO levels decrease in BV‐2 cells treated with Iso was dependent on Nrf2 since Nrf2 knockdown completely reverted this effect.

The potential benefits of Iso as a therapeutic molecule for AD were validated in 10‐month‐old APP/PS1 male mice, intranasally administered with Iso for 1 month. Iso displayed a good pharmacokinetic profile. It was absorbed quickly and removed from the blood due to either fast distribution into peripheral tissues or elimination, which agrees with the results obtained by other authors after intravenous and oral administration [47, 48]. Our results showed that Iso was present in the brain, the target for drug activity, though it was rapidly cleared. However, transient exposure may be sufficient to initiate signaling cascades that produce sustained downstream effects. Activation of the AKT/GSK3β/Nrf2 pathway (as observed in vitro in this study) can induce transcriptional programs and antioxidant responses (as observed in vivo) that persist beyond the period of detectable drug exposure, such that pharmacodynamic effects may outlast pharmacokinetic exposure. Histological analyses of organ cryosections of 11‐month‐old WT female mice (from the same littermate as APP/PS1 male mice, used in line with the 3R′ principle), administered with Iso or PBS (VEH), revealed no major alterations. However, lymphocyte infiltration and bronchiolar hyperplasia were observed in the lung, the latter also occurring in animals treated with PBS, suggesting that these events might be due to the administration route (intranasal) rather than the compound itself. In contrast, kidney alterations (cellular eosinophilia and tubular epithelial hypertrophy) were observed only in treated animals. Notably, some adverse renal and liver reactions have also been reported for DMF [69], a skin allergen that behaves like a Michael acceptor and a drug approved for the treatment of multiple sclerosis. However, no alterations in liver and kidney health parameters were detected in the serum of these animals, sustaining the safety of Iso administration. Yet, the relatively high dose employed in this proof‐of‐concept study and the short brain residence time highlight the need for future pharmacokinetic/pharmacodynamic studies, dose optimization, and repeated‐dose investigations to establish clinically relevant dosing strategies and sustained target engagement with potential mitigation of related histopathological alterations.

Obesity is associated with vascular dementia and cognitive impairment, increasing the risk of AD [70]. In our study, we observed that APP/PS1 mice weighed more than WT (both at 6‐ and 11‐month‐old). Interestingly, Iso administration led to weight reduction in older WT and APP/PS1 animals', which might be associated with metabolic alterations, as observed herein (e.g., reduction of LDL cholesterol in 11‐month‐old WT females and reduction of triglycerides in 11‐month‐old APP/PS1 males), and by others (e.g., reduction of glucose blood levels in diabetic neuropathic rats) [71]. Consistent with the activation of the AKT‐GSK3β pathway by Iso, as suggested above, it is conceivable that Iso influences additional PI3K‐AKT‐GSK3β‐dependent cellular processes, including neuronal survival and metabolism [72, 73], which should be explored in future studies. In line with our results, it was recently demonstrated that Iso blocked lipid accumulation and inhibited adipocyte differentiation in 3T3‐L1 cells [74], suggesting that it exerts a beneficial effect on adipogenesis, which plays a critical role in obesity. In our study, Iso did not reduce the adipose tissue weight of 6‐month‐old APP/PS1 female mice, but a trend toward a decrease was observed in 11‐month‐old WT and APP/PS1 males administered with a higher dose of Iso (100 mg/kg vs. 50 mg/kg at 6 months of age).

Eleven‐month‐old APP/PS1 animals showed memory deficits (compared to their WT counterpart), which were partially reversed by Iso treatment. The slight improvement of Iso on hippocampal‐dependent memory of APP/PS1 mice could be associated with the positive transcriptional regulation of Hmox1, Ppargc1α, and Bdnf that we observed in the hippocampus. Hmox1 gene expression was increased by Iso in both WT and APP/PS1‐treated mice (compared to their respective untreated group). Although lacking statistical significance, an evident increase (1.5‐fold; p = 0.068) in HMOX1 protein levels was also detected in 6‐month‐old APP/PS1 mice brain, treated with Iso. Overall, these observations align with our in vitro results, supporting the neuroprotective role of Nrf2‐HMOX1 activation via. In line with the potential neuroprotection promoted by Iso, the lower levels of Bdnf mRNA in the hippocampus of APP/PS1 mice (compared to WT) were increased after Iso treatment. Bdnf encodes the brain‐derived neurotrophic factor (BDNF), crucial for neuronal support. BDNF levels were shown to be reduced in the serum and brain of patients with AD and in a mouse model of tauopathy [75], and BDNF administration ameliorated the learning deficits in an AD rat model [76]. BDNF is also involved in ROS‐induced Nrf2 nuclear translocation, which requires PI3K/AKT and ERK1/2 activation, as observed in primary hippocampal neurons [77] and in the SH‐SY5Y neuronal cell line [78]. Consistent with this, we observed a significant decrease in ERK1/2 phosphorylation in the brain of APP/PS1 females at 6 months old, which was increased by Iso. Moreover, Ppargc1α gene expression, which was also reduced in APP/PS1 animals, was induced by Iso. This gene encodes the PGC‐1α protein, which is involved in lipid, glucose, and energy metabolism and is reduced in the human AD brain and AD mouse models, including the APP/PS1 used in this study [79]. In addition, PGC‐1α overexpression has been shown to diminish Aβ production, particularly by regulating the expression of the BACE1 enzyme, as reviewed in [79]. We did not find significant alterations in the expression of the Bace1 gene between groups. However, Bace1 mRNA levels were ~4‐fold reduced in Iso‐treated animals, in a similar pattern to that observed for BACE1 protein levels in the brain of 6‐month‐old mice. Nevertheless, we hypothesize that Iso‐induced Ppargc1α gene expression and BACE1 activity reduction (as determined in the cell‐free assay) might underlie the reduction of Aβ peptides in 6‐month‐old APP/PS1 females' brains. Yet, at 11 months old, Iso failed to reduce Aβ₄₀ and Aβ₄₂ levels in the cortex and hippocampus of APP/PS1 mice, which might be related to the overaccumulation of Aβ in the brain of aged mice that could not be reversed by Iso in such an advanced state. Consistent with this, Iso was able to reduce Aβ₄₀ in APP‐overproducing N2a‐APPswe cells, without affecting Aβ₄₂ levels, which is more prone to aggregate and more resistant to degradation [80]. Unexpectedly and unexplainably, Nrf2 knockdown in these cells also reduced Aβ₄₀ peptide levels, which were further decreased in the presence of Iso (with no effect on Aβ₄₂ levels; not shown). One possible mechanism is that Iso selectively reduced Aβ₄₀ production by modulating γ‐secretase activity through electrophilic and membrane‐associated effects. This is usually counteracted by NRF2‐dependent redox homeostasis, and removal of this compensatory buffering system would result in a more pronounced Aβ40 decrease. Nevertheless, the observed result should be further confirmed in other AD models.

In Sakimura and colleagues' work, 5xFAD AD mice administered with a pyruvate dehydrogenase kinase inhibitor at 7 months of age, for 3 months, also showed improved cognitive impairment (with limited neuron loss), without a reduction in Aβ deposits [81]. The authors suggested that the results were due to the observed amelioration in glucose metabolism, which could be a beneficial therapeutic strategy beyond Aβ‐reducing therapies, especially in patients with established Aβ pathology. Nevertheless, we observed a decrease in Aβ levels in the plasma of Iso‐treated APP/PS1 animals, suggesting that brain‐derived Aβ might not be effluxed to the plasma. Consistent with these results, Izco et al. showed that in APP/PS1 mice, an increase in Aβ₄₀ and Aβ₄₂ peptides in the brain paralleled a decrease in Aβ plasma levels, which could be related to an impaired Aβ clearance from the brain [82]. In fact, the reduction in Aβ peptides in 6‐month‐old female brain observed in our study was accompanied by a slightly increased level of both Aβ₄₀ and Aβ₄₂ levels in the plasma of Iso‐treated APP/PS1, compared to untreated mice (Aβ₄₀: APP/PS1—60.20 ± 2.162; APP/PS1 + Iso—72.29 ± 7.271. Aβ₄₂: APP/PS1—57.67 ± 6.944; APP/PS1 + Iso—83.88 ± 9.498). Though we cannot infer the role of Iso in Aβ clearance, our results highly suggest a potential effect of Iso in circumventing Aβ accumulation in the brain at early AD stages.

We did not find phenotypically related differences in Il1β gene expression between WT and APP/PS1 mice at 11 months old or in the pro‐IL‐1β protein levels determined in 6‐month‐old animals, in line with what was observed in a postmortem study, where no alterations in the levels of IL‐1β were found in the AD brain [83]. However, mRNA levels were markedly reduced in the hippocampus of both WT and APP/PS1 old animals treated with Iso, supporting its role in decreasing brain pro‐inflammatory mediators, as also suggested by the results obtained in microglia cells exposed to LPS. Moreover, Iso also modulated genes related to microglia function, namely Iba1 and Trem2. Iba1 is essential for microglial migration and phagocytosis, functions that have been demonstrated to be altered in AD [84]. Despite the lack of differences in Iba1 expression between WT and APP/PS1 untreated mice, this gene was significantly induced in transgenic animals treated with Iso. These results are in accordance with the work of Franco‐Bocanegra et al., who also did not observe significant differences between human AD samples and the control group but detected an increase in Iba1 in the brains of AD patients who underwent Aβ immunotherapy [83]. The authors suggested that this increase might reflect microglial motility changes associated with phagocytosis and showed that the Iba1 increase after Aβ immunotherapy was associated with a less pro‐inflammatory environment than that observed in AD. Additionally, we observed that Trem2 gene expression was significantly decreased in Iso‐treated WT animals, while a trend toward an increase (by 1.6‐fold) was observed in APP/PS1‐treated mice. Trem2 gene encodes the TREM2 immune receptor, which is mainly expressed in microglia and is implicated in complex signaling pathways related to inflammatory signaling, microglial metabolism and function, and lipid homeostasis [85], and TREM2 variants are related to AD pathogenesis [86]. Recently, Keren‐Shaul et al. [87] identified a novel and unique microglial phenotype associated with AD, using the 5XFAD mouse model. The authors showed that under stressful conditions, microglia shift to a disease‐associated microglia (DAM) profile, and a functional TREM2 receptor is required for this transition. DAM was found near Aβ plaques and is positively stained for intracellular Aβ particles (in both 5XFAD mice and human AD postmortem brains), hence considered neuroprotective. More recently, the overexpression of TREM2 in the brain of APP/PS1 transgenic mice was shown to rescue spatial cognitive deficits in diseased animals, decrease Aβ plaque burden, and ameliorate inflammation [88]. Therefore, Iso appeared to modulate Trem2 expression in a context‐dependent manner—rather than uniformly increasing or decreasing Trem2 expression, Iso may help restore an appropriate microglial response according to the pathological environment. Although the functional significance of these transcriptional changes remains to be determined, this pattern is consistent with the context‐dependent role of TREM2 in regulating microglial activation. One possible explanation for this context‐dependent effect of Iso is its differential electrophilic reactivity as a pre‐Michael acceptor, which may selectively modulate redox‐sensitive signaling pathways depending on the cellular environment, as extensively reviewed by the team [89]. However, this mechanistic link warrants further investigation. Nevertheless, the slight increase in Trem2, along with the significant increase in Iba1 mRNA levels observed in the hippocampus of APP/PS1 mice administered with Iso, suggests that this molecule might modulate microglial profile and function toward neuroprotection in AD. Although these transcriptional changes are consistent with modulation of microglial function, they do not establish a specific microglial phenotype. Future studies should investigate whether Iso promotes the acquisition of a DAM phenotype by evaluating additional DAM‐associated markers, together with functional assays of microglial phagocytosis, cytokine secretion and single‐cell transcriptomic profiling. Yet, and consistent with our results, it is plausible that the increased cell activation we observed in hippocampal histological slices of Iso‐treated APP/PS1 mice might be linked to its potential modulatory effect on microglia.

The present work presents some limitations, especially regarding in vivo experiments. The main issue concerns the use of APP/PS1 mice lacking Tau pathology, an important AD hallmark. Second, though the Power analysis indicates five mice as a robust number of animals to be used, the size of the groups should be increased to better support the findings. In fact, in behavioral experiments, the analysis between the two groups APP/PS1 VEH vs. APPS/PS1 Iso animals (with one‐way ANOVA) revealed a significant improvement in the treated mice's memory. However, with four groups and only about five animals per group, two‐way ANOVA has limited power, particularly for detecting a genotype × treatment interaction. Loss of significance can therefore reflect limited statistical power rather than absence of a treatment effect. Third, the observed outcomes were attained through direct intranasal administration of Iso to APP/PS1 animals, lacking a formulation facilitating targeted delivery to the brain. Therefore, it is imperative to devise a formulation enabling the selective delivery of Iso to the brain, enhancing its effectiveness while mitigating potential peripheral toxicity. Fourth, we used female mice at 6 months old (in the preliminary study) and male mice at 11 months old, limiting direct comparisons due to sex‐related differences. Finally, the authors suggest that Iso administration should be investigated in younger animals for a longer period (e.g., administered at 5 months of age for 3 months) to validate the beneficial effect of Iso in reducing Aβ accumulation in the brain at an early AD stage, and potentiate the positive outcomes of Iso in older animals. In addition, dose–response studies or longer recovery analyses would help determine whether the histological alterations observed are adaptive, route‐related, or indicative of early toxicity. This study was subject to institutional constraints, notably limited access to the mouse colony and the denial of approval for Iso administration beyond 1 month, hindering the suggested investigation.

5. Conclusions

In summary, and as far as we know, this is the first report showing the neuroprotective role of Iso, both in vitro and in vivo AD models, underscoring its promising therapeutic utility for AD treatment (Graphical abstract). Of note, this is the first pharmacokinetic study on intranasal administration of Isoeugenol, with a new HPLC method being validated. Although we have demonstrated that Iso reduced (neuro)inflammation via the Nrf2 pathway, we were not able to link other positive outcomes with it. Nevertheless, our results revealed Iso as a pleiotropic molecule that exerts beneficial effects through different pathways. The cognitive improvement observed in 11‐month‐old APP/PS1 mice occurred despite the absence of significant reductions in cortical or hippocampal Aβ burden, suggesting that Iso exerts its beneficial effects through mechanisms other than amyloid clearance at advanced disease stages. This interpretation is supported by the increased expression of Hmox1, Bdnf, and Ppargc1α, which are involved in antioxidant defense, neuronal plasticity, mitochondrial function, and energy metabolism. Furthermore, Iso reduced inflammatory mediators in vitro and modulated microglial‐associated genes in vivo, supporting an anti‐inflammatory effect that may help preserve neuronal function. Together, these findings suggest that Iso enhances neuronal resilience by targeting oxidative stress, neuroinflammation, and metabolic dysfunction, thereby improving cognitive performance independently of reducing established amyloid pathology. This interpretation is consistent with the growing recognition that therapies targeting multiple pathological mechanisms may remain beneficial even after extensive Aβ deposition has occurred.

Author Contributions

Ana Silva: investigation, writing – manuscript preparation and reviewing and editing. Sónia Silva: investigation, writing – draft preparation. Jéssica Macedo: investigation, writing – draft preparation. Patrícia Moreira: investigation, writing – draft preparation. Diana Baptista: investigation. Joana Bicker: investigation, writing – draft preparation. Ana Fortuna: methodology, validation, reviewing and editing. Joana Liberal: investigation. Beatriz Rodrigues: investigation. Rosa Resende: investigation. Inês Vitorino: investigation. Armanda E. Santos: resources, reviewing and editing. Bruno Miguel Neves: reviewing and editing. Cláudia Pereira: funding acquisition, reviewing and editing. Maria Teresa Cruz: conceptualization, funding acquisition, supervision, project administration, resources, reviewing and editing. All authors reviewed and edited the manuscript and approved the final version.

Funding

This work was financed by COMPETE 2020—Operational Programme for Competitiveness and Internationalization and by the EU Recovery and Resilience Facility and Portuguese national funds via FCT—Fundação para a Ciência e a Tecnologia, under projects, LA/P/0058/2020 [DOI: 10.54499/LA/P/0058/2020], UID/PRR/4539/2025 [DOI: 10.54499/UID/PRR/04539/2025] and UID/04539/2025 and also supported by the COST action BenBedPhar: Translating NRF2 research from bench to bed and HORIZON‐RIA—HORIZON Research and Innovation Actions [Grant agreement ID: 101080329].

Ethics Statement

All procedures involving animals were under the European Community guidelines for the use of animals in a laboratory (Directive 2010/63/EU) and were approved by the Direção Geral de Alimentação e Veterinária (DGAV; Ref: 0421/000/000/2021) and performed by users licensed by the Federation for European Laboratory Animal Science Association (FELASA).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Determination of a nontoxic concentration of Isoeugenol in neuronal and microglia cells. Cellular metabolic capacity was evaluated by Alamar Blue assay (resazurin) in neuronal cells N2a‐wt (A) and in microglia cells BV‐2 (B), exposed to Isoeugenol at different concentrations (500 μM, 250 μM, 100 μM, 50 μM, and 5 μM), for 24 h. The bars of the graphs represent the value of mean ± SEM of two to seven independent experiments. p < 0.05 was considered significant. Statistics: One‐way ANOVA with Dunnett's multiple comparisons test. *p < 0.05, compared to control cells (DMSO).

Figure S2: Effect of Isoeugenol on Nrf2‐dependent antioxidant Nqo1 and Sod1 gene expression in AD neuronal cells. (A) Nqo1 (Forward primer: CCTCTATGCTATGAACTT; Reverse primer: 5′GTCCTTCCTTATATGCTA3′) and (B) Sod1 (Forward primer: CACTCTAAGAAACATGGTGG; Reverse primer: GATCACACGATCTTCAATGG) mRNA levels determined by real time RT‐PCR, in N2a‐APPswe cells exposed to Iso for 1, 3, and 6 h. Values are the mean ± SEM of four independent experiments and expressed relatively to control (Ctr) cells. Statistics: Unpaired t‐test (t) and one‐way ANOVA with Dunnett's multiple comparisons test (*). p < 0.05 was considered significant. (t) and (*) p < 0.05, and (**) p < 0.01, compared to Ctr cells (Log₂1 = 0).

Figure S3: Protein levels in siRNA‐transfected cells. siRNA transfection efficiency was assessed by determining the protein levels of GAPDH (R&D Systems #NB300‐221; 1:1000) in N2a‐APPswe cells (A) and BV‐2 cells (B) transfected with siNRF2, siGAPDH and scrambled siRNA (siNeg). (C) iNOS protein levels determined in siRNA‐transfected BV2 cells exposed to Iso for 24 h and LPS (50 ng/mL, 30 min after Iso incubation). Representative blot images are presented (see Western Blot files, for uncropped images).

Figure S4: Blood–brain barrier (BBB) permeability of Isoeugenol. Iso permeability through BBB, expressed as experimental Papp (Permeability coefficient through the artificial membrane) values obtained with PAMPA models with 2% (w/v) porcine polar brain lipid (PBL). Data represent the mean of Papp in centimeters per second of three independent experiments performed in duplicate. (BBB−), non‐BBB permeable; (BBB+), BBB‐permeable.

Figure S5: Metabolic and biochemical parameters of Iso treated 11 months female WT mice. Metabolic parameters of (A) glucose, (B) triglycerides (TG), (C) total cholesterol (CHOL), (D) high‐density lipoprotein cholesterol (HDL‐C), and (E) low‐density lipoprotein cholesterol (LDL‐C). Markers of liver function (F) glutamic‐oxaloacetic transaminase (GOT), (G) gamma‐glutamyl transferase (GGT), and (H) alkaline phosphatase (ALP). Markers of kidney function (I) albumin, (J) creatinine, and (K) urea were determined in 11 months female mice administered with Vehicle (PBS) or Iso (100 mg/kg) for 1 month. Data corresponds to the mean ± SEM of five animals. Statistics: Unpaired t‐test. p < 0.05 was considered significant. (*) p < 0.05, compared to VEH.

Figure S6: Effect of Isoeugenol on locomotor activity and cognitive function of 6 months AD mice. (A) Hot plate, (B–F) Open Field and (G–I) Fear Conditioning behavioral tests performed with WT and APP/PS1 6 months female mice, administered with Vehicle (PBS) or Iso (50 mg/kg; APP/PS1 only) for 1 month. Data correspond to the mean ± SEM of five animals. Statistics: Unpaired t‐test (t) and Two‐way ANOVA (* and #) with Tukey's multiple comparisons test. p < 0.05 was considered significant. (*) p < 0.05 and (ttt) p < 0.001, compared to WT; (#) p < 0.05, compared to APP/PS1.

Figure S7: Effect of Isoeugenol on brain Aβ peptide levels and on demographic and metabolic parameters of 6 months AD mice. (A) Aβ40 and (B) Aβ42 levels were detected in the brain of WT and APP/PS1 6 months female mice, administered with Vehicle (PBS) or Iso (50 mg/kg; APP/PS1 only) for 1 month. (C) Body weight, (D) Adipose tissue weight, (E) Fasting Glucose and (F) Triglycerides (TG) were also measured. Data correspond to the mean ± SEM of five to six animals. Statistics: Unpaired t‐test (t) and One‐way ANOVA with Tukey's multiple comparisons test (* and #). p < 0.05 was considered significant. (t) p < 0.05 and (****) p < 0.0001, compared to WT; (#) p < 0.05 and (###) p < 0.001, compared to APP/PS1.

Figure S8: Effect of Isoeugenol on 6 months mice protein levels. hAPP (A), BACE1 (B), iNOS (C), Pro‐IL‐1β (D), HMOX1 (E) and p44/42 (ERK) MAPK protein levels were determined in 6 months females mice brain, after Iso (50 mg/kg; APP/PS1 only) intranasal administration, for 1 month. Values are the mean ± SEM of five animals. Statistics: One‐way ANOVA with Dunnett's (A) or Tukey's (F) multiple comparisons test (*). p < 0.05 was considered significant. (*) and (#) p < 0.05, compared to WT and APP/PS1, respectively.

Figure S9: Effect of Isoeugenol on biochemical parameters of 11 months male mice. (A) Body weight, (B) percentage of body weight loss, (C) adipose tissue weight, (D) brain tissue weight, (E–G) Glucose levels and (H) Triglycerides (TG) were measured in 11 months male mice, administered with Vehicle (PBS) or Iso (100 mg/kg; APP/PS1 only) for 1 month. Data corresponds to the mean ± SEM of five animals. Statistics: One‐way ANOVA with Tukey's multiple comparisons test. p < 0.05 was considered significant. (*) p < 0.05, compared to WT and (#) p < 0.05, compared to APP/PS1. AUC, Area Under the Curve; GTT, Glucose Tolerance Test.

Figure S10: Effect of Isoeugenol on pain sensitivity, anxiety levels and locomotor activity of 11 months WT and APP/PS1 mice. (A) Pain sensitivity evaluated by Hot plate test; (B–D) anxiety levels evaluated by Elevated plus‐maze test; (E–H) locomotor activity evaluated by Open field test, performed with 11 months male mice intranasally administrated with VEH or Iso (100 mg/kg) for 1 month. The results represent the mean ± SEM of five animals.

Figure S11: Contribution of AKT to Nrf2 activation induced by Isoeugenol evaluated through the KeratinoSens reporter assay. Keratinosens cells were incubated with Isoeugenol (Iso 250 μM) for 6 h, without or with Wortmannin (+Wortm; 500 nM), a PI3K/AKT inhibitor (added 1 h before Isoeugenol). DMF, a known Nrf2 activator, was used as positive control. Nrf2 activation was assessed by measuring Luciferase activity and presented as fold‐induction of vehicle (DMSO). Metabolic activity of the cells was determined by Alamar Blue assay and presented as % of control (Ctr) cells. The results represent the mean ± SEM of three independent experiments in triplicate. Statistics: Unpaired t‐test and one‐way ANOVA with Dunnett's multiple comparisons test (*). p < 0.05 was considered significant. (*) p < 0.05, compared to Vehicle and (****) p < 0.0001, compared to Iso alone (‐Wortm; filled dark blue bar).

Table S1: BACE1 activity inhibition induced by Isoeugenol.

Table S2: Summary of validation parameters of the HPLC‐DAD method applied to quantify Isoeugenol in mouse matrices.

Data Availability Statement

The datasets supporting the conclusions of this article are included within the article and its additional files.