Synthesis and properties of the kojic acid dimer and its potential for the treatment of Alzheimer's disease

Abstract

The kojic acid dimer (KAD) is a metabolite derived from developing cottonseed when contaminated with aflatoxin. The KAD has been shown to exhibit bright greenish-yellow fluorescence, but little else is known about its biological activity. In this study, using kojic acid as a raw material, we developed a four-step synthetic route that achieved the gram-scale preparation of the KAD in approximately 25% total yield. The structure of the KAD was verified by single-crystal X-ray diffraction. The KAD showed good safety in a variety of cells and had a good protective effect in SH-SY5Y cells. At concentrations lower than 50 μM, the KAD was superior to vitamin C in ABTS⁺ free radical scavenging assay; the KAD resisted the production of reactive oxygen species induced by H₂O₂ as confirmed by fluorescence microscopy observation and flow cytometry analysis. Notably, the KAD could enhance the superoxide dismutase activity, which might be the mechanism of its antioxidant activity. The KAD also moderately inhibited the deposition of amyloid-β (Aβ) and selectively chelated Cu²⁺, Zn²⁺, Fe²⁺, Fe³⁺, and Al³⁺, which are related to the progress of Alzheimer's disease. Based on its good effects in terms of oxidative stress, neuroprotection, inhibition of Aβ deposition, and metal accumulation, the KAD shows potential for the multi-target treatment of Alzheimer's disease.

We developed a synthetic route that achieved the gram scale preparation of KAD, and the structure was verified by single crystal X-ray diffraction analysis, and the potential for the multi-target treatment of AD was identified for the first time.

1. Introduction

Alzheimer's disease (AD) has become a global problem in human health. The incidence of AD is increasing with human longevity, and the number of people with AD worldwide is expected to reach 131 million by 2050.¹ As a slowly progressing and ultimately fatal degenerative brain disorder,^2,3 AD can place intense financial and emotional pressure on patients and their families.⁴ Unfortunately, current drug-based therapies only moderately improve learning and memory, and there is no effective treatment to reduce or halt the progression of AD.^5–7 Consequently, the development of new therapies for AD has become a major challenge in modern medicine. Although the pathogenesis of AD is not completely clear, several hallmarks are associated with the pathological development of AD, including the abnormal accumulation of amyloid-β (Aβ) and tau protein, reactive oxygen species (ROS), and metal ions along with neuroinflammation.^8–11

Aβ is generated by the proteolysis of amyloid precursor protein (APP). The protein is first cleaved at its N-terminus by β-secretase (also called β-site APP cleavage enzyme1 (BACE1)),¹² and then the carboxy-terminal fragment (CTFβ) is cleaved by γ-secretase to form Aβ-peptide. In the amyloid hypothesis, Aβ-peptide in the form of an alpha-helical structure is successively transformed into a β-sheet structure, monomers, and oligomers, and finally form the fiber during the development of AD. The accumulation of Aβ aggregates stimulates a string of neurotoxic pathways such as apoptosis, inflammation, synaptic and neuronal damage, thus accelerating the process of AD.¹³ As the major pathological marker of AD, the inhibition of Aβ deposition is regarded as a potential therapeutic strategy for AD.¹⁴

Due to its high demand for oxygen and lack of antioxidant defense systems, the brain is very vulnerable to oxidative stress (OS). Multiple lines of evidence support overproduction of ROS leading to OS, which plays a critical role in the pathogenesis and pathophysiology of AD.^15,16 OS in the early stage of AD may be necessary for the deployment of other pathological mechanisms of AD, such as Aβ- and tau-induced neurotoxicity.¹⁷ Therefore, the removal of ROS or prevention of ROS formation may help halt AD progression. In addition to OS, the presence of excessive biological metals (e.g., iron, zinc, and copper) in the brain promotes the aggregation of Aβ and contributes to the generation of ROS, triggering the further progression of AD.^18,19 In this light, antioxidants and metal chelating agents have been evaluated as potential treatments for AD.^20,21

For thousands of years, people have been learning by observing nature, and many modern drugs are derived from secondary natural compounds produced by microorganisms. Kojic acid (KA), which has a typical γ-pyranone skeleton, is a natural fungal metabolite produced by a series of microorganisms. Khan et al.²² reported that KA reduced the expression of Aβ protein, decreased inflammation and OS, and exerted a neuroprotective effect to reverse pathological damage in HT-22 mouse hippocampal cells in vitro and in an AD mouse model. These findings suggest that KA might be a candidate for the prevention and treatment of AD. As a scaffold, γ-pyranone and the natural antioxidants it contains can scavenge ROS and reduce oxidative damage to reverse the pathological progress of AD.^23,24 For instance, Cheng et al.²⁵ and Hu et al.²⁶ designed and synthesized various γ-pyranone-based functional agents with antioxidant effects and the ability to chelate metal ions and inhibit Aβ aggregation, demonstrating their potential for AD treatment.

The KA dimer (KAD), in which two KA molecules are coupled at the 5,5′-position (Fig. 1A), is also a natural metabolite. The KAD is derived from developing cottonseed when aflatoxin invades it. The KAD was first discovered by Marsh et al.²⁷ in 1955 due to its bright greenish-yellow fluorescence under long-wave ultraviolet light. In 1999, Zeringue et al.²⁸ confirmed that the fluorescent substance was the KAD using nuclear magnetic resonance and mass spectrometry. However, except for its bright greenish-yellow fluorescence, the bioactivity of the KAD has not been thoroughly investigated. In this study, we developed an efficient synthetic route for the gram-scale preparation of the KAD and confirmed its structure for the first time by X-ray diffraction. In addition, the inhibition of the deposition of Aβ and the antioxidant, neuroprotective, and metal chelation effects of the KAD were investigated in detail to evaluate its potential as a treatment for AD.

Fig. 1. Natural source and synthetic route of the KAD. (A) Natural source, appearance, and molecular structure of the KAD; (B) synthetic route for the preparation of the KAD.

2. Results and discussion

2.1. Synthesis of the KAD

Self-dimerization is not uncommon in nature. Nevertheless, a general method for practical dimer synthesis is still lacking. Inspired by the biomimetic synthesis of the KAD, we attempted to find a suitable oxidant to dimerize KA in one step. The attempt to screen oxidants did not advance smoothly. Although the green-yellow fluorescence was monitored, the reactions were messy in the presence of oxidants such as H₂O₂, K₂S₂O₈, and tert-butyl hydroperoxide (TBHP), and the yield of the KAD detected by liquid chromatography-mass spectrometry was tiny. Consequently, an alternative protocol based on a cross-coupling strategy was developed.

Using KA as a raw material, 2a was produced in 75% yield by bromination with N-bromosuccinimide (NBS) in the presence of NH₄OAc. Subsequently, we attempted to directly convert 2a into the KAD via the Suzuki–Miyaura reaction; however, the hydroxyl group in combination with the adjacent carbonyl group, a metal coordinating site, poisoned the palladium catalyst, preventing the formation of the desired product. To avoid poisoning the catalyst, the hydroxyl group was protected by BnBr, and the corresponding compound 2b was obtained in 60% yield. As expected, the coupling of 2b by the Suzuki–Miyaura reaction proceeded smoothly, giving 2c in 70% yield. Notably, the copper-catalyzed Ullmann reaction was ineffective for this coupling. Finally, the treatment of 2c with 6 M hydrochloric acid in ethanol afforded the target product KAD (80%).

We achieved the gram-scale synthesis of the KAD using this four-step approach. For example, starting with 7.00 g KA, we obtained 1.45 g of KAD in approximately 25% overall yield. The KAD product could be further purified by recrystallization from 6 M HCl, providing a pale-yellow crystallite (Fig. 1B) with a purity of 97.8% based on HPLC.

2.2. Structural determination of KAD

The ultraviolet-visible (UV-vis) spectrum of the KAD collected from 200–600 nm was measured (Fig. 2A). The peak at 260–320 nm (λ_Max = 288 nm) was assigned to the n → π* transition of C O. The peak in the range of 320–470 nm (λ_Max = 378 nm) was assigned to the π → π* transition produced by the conjugated double bond. A solution of KAD (in HEPES buffer at pH 7.4) gave off bright greenish-yellow fluorescence when irradiated with UV light at 365 nm, consistent with previous reports.²⁷ Based on single-crystal X-ray diffraction, the synthesized KAD was symmetrical, and the carbonyl group formed an intramolecular hydrogen bond with the adjacent hydroxyl group (Fig. 2B). Notably, the KAD had a planar structure (Fig. 2C), which might explain its strong fluorescence.

Fig. 2. Optical properties and crystal structure of the KAD. (A) UV-vis spectrum of a solution of the KAD (50 μM) in HEPES buffer (pH 7.4); (B) front view and (C) side view of the molecular structure of the KAD derived from the single-crystal X-ray diffraction pattern (CCDC number 2113649).

2.3. Determination of free radical scavenging ability

The radical scavenging ability of the KAD was evaluated by [2,2-diphenyl-1-picrylhydrazyl] (DPPH) and [2,2′-azino-bis-3-ethylbenzothiazoline-6-sulfonic acid] (ABTS⁺) assays. In solution, antioxidants lighten the color of DPPH/ABTS⁺ by donating an electron/hydrogen atom to the unpaired electron of DPPH/ABTS⁺, which turns the purple/blue green colored DPPH/ABTS⁺ colorless. The color change can be simply monitored using UV-vis spectrophotometry. The absorbances at 517 and 734 nm were measured to calculate the free radical scavenging rates for DPPH (I_DPPH, %) and ABTS⁺ (I_ABTS, %), respectively. VC and KA were used as positive controls.

As shown in Fig. 3, the KAD showed a better free radical scavenging ability than KA in both the DPPH and ABTS⁺ assays. For example, the I_DPPH value of the KAD was 50.33 ± 0.77% at 100 μM, while that of KA was 17.79 ± 0.62%. Similar results were obtained for ABTS⁺ (I_ABTS⁺ = 79.37 ± 4.42% for the KAD and 46.76 ± 6.40% for KA) at 100 μM. The antioxidant efficiency of the KAD was approximately two times greater than that of KA because the KAD has two sets of phenolic hydroxyl groups and a more planar π-conjugated structure compared to KA. Remarkably, the KAD (I_ABTS⁺ = 55.98 ± 5.37%) also showed a better free radical scavenging ability than VC (I_ABTS⁺ = 39.69 ± 3.50%) at 50 μM in the ABTS⁺ assay. These results suggest that the KAD, like other natural polyphenols, may effectively scavenge free radicals at a low dose, indicating promise for application in the treatment of AD via the antioxidant pathway.

Fig. 3. Determination of the free radical scavenging ability of the KAD and KA by DPPH (A) and ABTS⁺ (B) assays (at concentrations of 400, 200, 100, 50, 25, 12.5, and 6.25 μM). The scavenging rates are expressed as the average of three independent experiments. The data for VC and KA are shown for comparison.

2.4. Cytotoxicity

Given the importance of safety, the viability of several cells treated with the KAD was evaluated by 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay. The KAD was nontoxic in BV-2, SH-SY5Y, HT-22, H9c2, and L02 cells at concentrations up to 500 μM (Table 1).

Cytotoxicity (IC₅₀, μM) of KAD in the BV-2, SH-SY5Y, HT-22, H9c2, and L02 cell lines.

Compd.	BV-2	SH-SY5Y	HT-22	H9c2	L02
KAD	>500	>500	>500	>500	>500
KAa	>500	>500	>500	>500	>500

^aKA was selected as a positive control. Data were obtained based on three independent experiments.

2.5. Inhibitory effect of the KAD on Aβ_1–42 aggregation

Two principal forms of Aβ exist in the brain tissue of patients with AD: Aβ_1–40 and Aβ_1–42. Compared with Aβ_1–40, Aβ_1–42 is a more hydrophobic and fibrillogenic form of Aβ. It aggregates much faster than Aβ_1–40, accounting for about 10% of total Aβ. It was found to be the main component in senile plaque deposits. In this study, the effect of the KAD on senile plaques was evaluated by detecting the inhibition of Aβ_1–42.

Thioflavin T (ThT) dye fluorescence is commonly used to quantify the formation and inhibition of Aβ in the presence of anti-amyloidogenic compounds such as polyphenols. In this study, we used this dyeing method to detect the inhibitory effect of the KAD on Aβ_1–42. We chose KA and resveratrol (a natural polyphenolic phytochemical used for preventing and treating AD) as controls.²⁹ When Aβ_1–42 was pretreated with the KAD, KA, or resveratrol, the fluorescence intensity was reduced. Further, the inhibition rate for Aβ_1–42 (I_Aβ1–42, %) was calculated. Our study confirms that I_Aβ1–42, % for KAD was 16.98 ± 2.25% at 50 μM, with an effect similar to that of KA (I_Aβ1–42, % = 14.55 ± 1.78% at 50 μM), but with a lower inhibition rate compared with resveratrol (I_Aβ1–42, % = 64.64 ± 0.96% at 50 μM; Aβ_1–42 IC₅₀ = 13.15 ± 1.53 μM), as shown in Table 2.

KAD inhibited Aβ aggregation.

Compd.	Inhibition of Aβ1–42 aggregation (%)	Aβ1–42 IC50 (μM)
KAD	16.98 ± 2.25	n.db
KAa	14.55 ± 1.78	n.db
Resveratrola	69.64 ± 0.96	13.15 ± 1.53

^aKA and resveratrol were selected as positive controls. Data were obtained based on three independent experiments.

^bn.d. = not determined.

To further study the interaction mode of the KAD for Aβ_1–42, a molecular docking study was performed using the Ligand Docking program in the Maestro 11.5 software of Schrödinger, and visualization was performed using PyMOL. The molecular docking mode of the KAD compound with Aβ_1–42 is shown in Fig. 4. The carbonyl group of the KAD interacted with Lys28 (2.5 Å) through hydrogen bonds, and the two alcohol hydroxyl groups of the KAD also formed hydrogen bonds with Ala21 (2.0 Å) and Glu22 (1.7 Å).

Fig. 4. Docking study of the KAD with Aβ_1–42 (PDB: 1IYT). (A) Full view of KAD (colored yellow) binding to Aβ_1–42. (B) The possible hydrogen bonds between the KAD and residues Lys28, Ala21 and Glu22 are indicated by yellow dotted lines.

Aβ_1–42 units can easily form β-folded “hairpin” structures, which aggregate into fibers stabilized by the salt bridge resulting from electrostatic interactions between amino acid residues Asp23 and Lys28.³⁰ The docking results showed that the KAD compound could form hydrogen bonds with Lys28 to block the formation of the salt bridge between Lys28 and Asp23, which could effectively inhibit the aggregation of Aβ. Similarly, the inhibitory effect of KA on Aβ aggregation could rationalize the molecular docking results (Fig. S17 and S18†).

ThT assay and molecular docking confirmed the inhibitory effect of the KAD on Aβ_1–42. This study was novel in evaluating the effect of the KAD on Aβ_1–42, thus providing the possibility to assess the KAD as an anti-AD drug.

2.6. The KAD inhibits H₂O₂-induced injury

Studies have suggested that natural polyphenols can effectively resist OS; therefore, H₂O₂, a widely used model for inducing cellular OS, was used to assess the effect of the KAD on OS.³¹ SH-SY5Y, a common cell line used in neurotoxic studies, was chosen to assess the protective effect of the KAD against oxidative damage.³² First, the appropriate dose and duration of H₂O₂ that cause the death of 50% of cells were obtained in preliminary experiments by MTT assay. As shown in Fig. 5A, the cell viability was 47% when SH-SY5Y cells were treated with 200 μM H₂O₂ for 2 h. Therefore, 200 μM H₂O₂ for 2 h was adopted to induce OS injury. When cells were pre-treated with different doses of the KAD for 24 h prior to exposure to 200 μM H₂O₂, we found that 500 μM KAD could significantly restore SH-SY5Y cell viability. This indicated that the KAD could protect neuronal cells from oxidative damage, which is of great significance in preventing the progression of AD (Fig. 5B). Also, we found that the KAD had a protective effect against oxidative damage in H9c2 cells (Fig. S19†).

Fig. 5. The KAD inhibits H₂O₂-induced injury in SH-SY5Y cells. (A) H₂O₂ reduced the SH-SY5Y cell viability (2 h). (B) Effect of the KAD on the viability of SH-SY5Y cells treated by H₂O₂. The cell viability was determined by MTT assay. The values are presented as the mean ± SEM (n = 3). ^###P < 0.001 vs. the control group; P* < 0.05 vs. the H₂O₂-treated group.

2.7. Effect of the KAD on ROS

Microglial BV-2 cells are sensitive to oxidative stress, and they are also a classical cell model for assessing AD.³³ Therefore, we chose BV-2 cells to assess the effect of the KAD against AD by reducing oxidative stress damage using an inverted fluorescence microscope (Fig. 6). When exposed to H₂O₂, BV-2 cells exhibited elevated intracellular ROS levels. At dosages of 500, 250, and 125 M, KAD pre-treatment dramatically reduced ROS. Similarly, we observed that the KAD can effectively combat the burst of ROS generated by H₂O₂ in H9c2 cells, which are commonly used to evaluate oxidative stress.³⁴ These results clearly demonstrate that the KAD is effective against oxidative stress (Fig. S20†).

Fig. 6. Inhibitory effect of the KAD on intracellular ROS accumulation in BV-2 cells exposed to H₂O₂ (10×). Images of cells stained with 2′,7′-dichlorofluorescein (DCF) showing the ROS content in BV-2 cells: (A) control; (B) H₂O₂; (C) 500 μM KAD; (D) 500 μM KAD + H₂O₂; (E) 250 μM KAD + H₂O₂; (F) 125 μM KAD + H₂O₂.

To further quantify the effect of the KAD on ROS, we conducted flow cytometry analysis (Fig. 7A and B). The ROS concentration was significantly higher in the cells stimulated with H₂O₂ compared to the control. Pretreating the cells with the KAD at 500, 250, and 125 μM for 24 h before exposure to H₂O₂ remarkably reduced the ROS concentration in a dose-dependent manner. In parallel, similar results were obtained for the KAD in H9c2 cells (Fig. S21A and B†). These results indicate that the KAD can efficiently prevent the accumulation of intracellular ROS, and this inhibition may explain the antioxidant effect of the KAD.

Fig. 7. The KAD alleviated the intracellular ROS accumulation and increased the SOD activity induced by H₂O₂ in BV-2 cells. (A) Flow cytometry images for each group: (a) control; (b) H₂O₂; (c) 500 μM KAD; (d) 500 μM KAD + H₂O₂; (e) 250 μM KAD + H₂O₂; (f) 125 μM KAD + H₂O₂; (B) ROS quantification by DCFH-DA using flow cytometry (10 000 cells); (C) effect of the KAD on SOD activity. All data are presented as the mean ± SEM of three independent experiments. ^#P < 0.05 and ^###P < 0.001 vs. the control group; P* < 0.05 and **P < 0.01 vs. the H₂O₂-treated group.

2.8. Effect of the KAD on superoxide dismutase (SOD) levels

SOD is an important free radical scavenging enzyme that exerts a protective function against OS injury. Herein, the mechanism of the KAD on H₂O₂-induced ROS burst was assessed by measuring the SOD activity in BV-2 cells. As shown in Fig. 7C, the SOD activity was reduced in cells treated with H₂O₂ compared to the control. Compared to the H₂O₂-treated cells, the SOD activity increased in a dose-dependent manner when the cells were pretreated with the KAD. Likewise, the KAD was found to have an enhancing effect on SOD in H9c2 cells (Fig. S21C†). These confirm that the KAD can combat OS by enhancing the free radical scavenging capacity, which is probably the mechanism to act as an antioxidant.

2.9. Metal chelating properties of the KAD

Due to the involvement of certain metal ions in neurodegenerative processes,^35,36 the metal chelation ability of the KAD was evaluated for 10 metal ions: Cu²⁺, Zn²⁺, Fe²⁺, Fe³⁺, and Al³⁺, which are recognized as being potentially involved in AD, and Na⁺, K⁺, Mg²⁺, Ca²⁺, and Mn²⁺, which play important roles in maintaining human physiological functions.

HEPES buffer (pH 7.4), a non-ionic amphoteric buffer that does not interfere with chemical reactions, was selected as the solvent to evaluate the metal chelation ability of the KAD by UV-vis spectroscopy. With the addition of one equivalent of metal ions, obvious optical responses of Cu²⁺, Zn²⁺, Fe²⁺, Fe³⁺, and Al³⁺ were observed (Fig. 8A), while Na⁺, K⁺, Mg²⁺, Ca²⁺, and Mn²⁺ did not produce optical responses (Fig. 8B). Thus, the KAD had no effect on the ions needed to maintain normal physiological functions and can be used to selectively chelate the ions related to AD (Cu²⁺, Zn²⁺, Fe²⁺, Fe³⁺, and Al³⁺).^37,38

Fig. 8. (A) UV-vis spectra of the KAD (100 μM) upon the addition of CuSO₄, ZnCl₂, FeCl₂, FeCl₃, and AlCl₃ (100 μM). (B) UV-vis spectra of the KAD (100 μM) upon the addition of NaCl, KCl, MgSO₄, CaCl₂, and Mn (CH₃COO)₂ (100 μM). (C) UV spectra of the KAD (100 μM) in the presence of various concentrations of Cu²⁺ (12.5, 25, 50, 75, 100, 125, 150, 175, and 200 μM) along with the stoichiometric ratio of KAD to Cu²⁺ determined at 475 nm. (D) UV spectra of the KAD (100 μM) in the presence of various concentrations of Fe²⁺ (12.5, 25, 50, 75, 100, 125, 150, 175, and 200 μM) and the stoichiometric ratio of KAD to Fe²⁺ determined at 475 nm. All spectra were measured in solutions of HEPES buffer (pH 7.4).

The abnormal deposition of copper and iron is strongly correlated with the development of AD.³⁹ Accordingly, we conducted an in-depth study on the coordination on Cu²⁺ and Fe²⁺ by the KAD. Upon the introduction of Cu²⁺ at concentrations of 12.5, 25, 50, 75, 100, 125, 150, 175, and 200 μM in HEPES buffer (pH 7.4), a new UV-vis absorption band appeared at 475 nm, and the peak at 370 nm decreased in intensity. We plotted the titration curve and used it to evaluate the stoichiometry of the interaction between the KAD and Cu²⁺. The absorbance at 475 nm increased as the ratio of Cu²⁺ to KAD increased from 0 to 0.8 and then reached a plateau after the ratio reached 1, indicating a 1 : 1 stoichiometry of KAD to Cu²⁺ (Fig. 8C).

Similarly, a KAD solution was mixed with different concentrations of Fe²⁺ (final concentrations of 12.5, 25, 50, 75, 100, 125, 150, 175, and 200 μM) in HEPES buffer (pH 7.4) to evaluate the chelation of Fe²⁺ by UV-vis spectrophotometry in the wavelength range of 350–650 nm. As the Fe²⁺ concentration increased, the absorption band of the KAD first decreased and then increased. The titration curve constructed using the absorbance at 475 nm showed the intersection of two straight lines at a mole fraction of 1.00, indicating a 1 : 1 stoichiometry for the KAD–Fe²⁺ complex (Fig. 8D).

3. Experimental

3.1. Chemicals and materials

Melting points were obtained using an MP100 melting point apparatus. The IR spectra were measured using a Nicolet is50 FT-IR spectrometer with a KBr disk. NMR spectra were recorded on a Bruker IMPACT-II 400 spectrometer operating at 400 MHz for ¹H NMR and 100 MHz for ¹³C NMR. High-resolution mass spectra (HRMS) were obtained using a Bruker FTICRMS 7T solariX spectrometer. Electrospray ionization mass spectra were recorded on a Shimadzu LC-20AP. HPLC was conducted using an Agilent Technologies 1290 Infinity II instrument. TLC was conducted using Energy Chemical GF254 silica gel plates. The absorbance values were recorded on a microplate reader (Epoch, BioTek, USA). The absorption spectra were recorded on a UV-vis spectrophotometer (UV-2450, Shimadzu, Japan). ThT assays were performed using a multi-mode microplate reader (SpectraMax Paradigm, Molecular Devices, USA). ROS were detected using an inverted fluorescence microscope (DMi8, Leica, Germany) and quantified by flow cytometry (FACSVerse, BD, USA).

The total superoxide dismutase and reactive oxygen species assay kits were obtained from Beyotime (Shanghai, China). Recombinant human HFIP-pretreated Aβ_1–42 peptide and thioflavin T were obtained from Sigma-Aldrich (Germany). 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT), Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F12) and phosphate-buffered saline (PBS) were purchased from Meilunbio (China). Dulbecco's Modified Eagle Medium (DMEM) and Roswell Park Memorial Institute 1640 Medium (RPMI) were purchased from HyClone (USA). Fetal bovine serum (FBS) was obtained from Thermo Scientific (Gibco, Germany).

All reagents and solvents were purchased from commercial sources and used without further purification, unless indicated otherwise. Where necessary, reactions were performed under a nitrogen atmosphere.

3.2. Synthesis of the KAD

2-Bromo-3-hydroxy-6-(hydroxymethyl)-4H-pyran-4-ketone (2a)

KA (7.00 g, 50 mmol), NBS (13.35 g, 75 mmol), NH₄OAc (3.85 g, 50 mmol), and tetrahydrofuran (70 mL) were introduced into a 200 mL flask. The reaction was stirred at 70 °C for 2 h. Upon completion, the solvent was removed under vacuum, and the residue was recrystallized from ethyl acetate to give 2a as a yellowish solid (8.13 g, 75%): ¹H NMR (400 MHz, DMSO-d₆) δ 9.94 (s, 1H), 6.38 (s, 1H), 5.80 (s, 1H), 4.32 (s, 2H). ¹³C NMR (100 MHz, DMSO-d₆) δ 172.72 (s), 169.35 (s), 144.62 (s), 129.64 (s), 110.01 (s), 59.74 (s); HRMS (ESI) m/z: calculated for C₆H₅BrNaO₄ [M + H]⁺ 220.9450, found 220.9444.

3-Benzyloxy-2-bromo-6-hydroxymethyl-4H-pyran-4-ketone (2b)

2a (8.13 g, 37.5 mmol) was dissolved in methanol (45 mL) and H₂O (45 mL) followed by the addition of NaOH (2.25 g, 56.30 mmol). To this solution, benzyl bromide (6.74 g, 39.40 mmol) was added dropwise, and the reaction was stirred at 80 °C for 2 h, when all of the starting material was consumed (as monitored by TLC). After the removal of solvent under reduced pressure, the residue was diluted with 30 mL H₂O and extracted with ethyl acetate (3 × 50 mL). The organic phases were combined and dried over anhydrous Na₂SO₄. The crude mixture was purified by column chromatography on silica gel to give the product 2b as a yellow oily substance (7.00 g, 60%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.41–7.30 (m, 5H), 6.37 (d, J = 0.9 Hz, 1H), 5.76 (t, J = 6.2 Hz, 1H), 5.08 (s, 2H), 4.27 (dd, J = 6.1, 0.9 Hz, 2H). ¹³C NMR (100 MHz, DMSO-d₆) δ 179.87 (s), 173.84 (s), 169.58 (s), 144.73 (s), 140.36 (s), 136.74 (s), 129.08 (s), 128.89 (s), 128.82 (s), 128.79 (s), 112.63 (s), 73.60 (s), 59.53 (s); HRMS (ESI) m/z: calculated for C₁₃H₁₀BrO₄ [M + H]⁺ 310.9919, found 310.9912.

3,3′-Bis(benzyloxy)-6,6′-bis(hydroxymethyl)-4H,4′H-[2,2′-bipyran]-4,4′-diketone (2c)

Potassium acetate (4.40 g, 45.01 mmol), bis(pinacolato)diboron (6.86 g, 27.00 mmol), and a 1,1′-bis(diphenylphosphino) ferrocene–palladium dichloride dichloromethane adduct (1.23 g, 1.69 mmol) were added to 2b (7.00 g, 22.51 mmol) in N,N-dimethylformamide (70 mL). After stirring at 70 °C for 6 h under a nitrogen atmosphere, the reaction was quenched with water and extracted with ethyl acetate (3 × 50 mL). The organic layers were dried over anhydrous Na₂SO₄ and concentrated in vacuo. After purification by column chromatography on silica gel (petroleum ether/ethyl acetate, 2 : 1), 2c was obtained as a yellowish solid (3.64 g, 70%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.33–7.27(m, 6H), 7.22 (dd, J = 6.6, 3.0 Hz, 4H), 6.47 (s, 2H), 5.80 (t, J = 6.1 Hz, 2H), 5.11 (s, 4H), 4.20 (d, J = 6.0 Hz, 4H). ¹³C NMR (100 MHz, DMSO-d₆) δ 175.19 (s), 169.11 (s), 146.63 (s), 145.26 (s), 136.77 (s), 128.82 (s), 128.73 (s), 128.65 (s), 112.77 (s), 73.52 (s), 59.64 (s); HRMS (ESI) m/z: calculated for C₂₆H₂₃O₈ [M + H]⁺ 463.1393, found 463.1386.

3,3′-Dihydroxy-6,6′-bis(hydroxymethyl)-4H,4′H-[2,2′-bipyran]-4,4′-diketone (kojic acid dimer, KAD)

The compound 2c (3.00 g, 6.50 mmol) was dissolved in ethanol (20 mL), and 20 mL of 6 M HCl (aq.) was added. The reaction mixture was stirred at 90 °C overnight. After completion, the reaction was cooled to room temperature, and the precipitate was filtered and purified by column chromatography to afford the KAD as a yellowish brown solid (1.45 g, 80%): m.p. >300 °C; IR (KBr) max 3186, 3074, 2919, 2816, 1630, 1577, 1433, 1238, 1154, 1084, 970; ¹H NMR (400 MHz, DMSO-d₆) δ 6.45 (s, 2H), 4.34 (s, 4H); ¹³C NMR (100 MHz, DMSO-d₆) δ 174.44 (s), 169.05 (s), 145.07 (s), 137.78 (s), 109.68 (s), 60.04 (s); HRMS (ESI) m/z: calculated for C₁₂H₁₀O₈ [M + H]⁺ 283.0454, found 283.0448.

3.3. Structure determination and optical characteristics of the KAD

The structure of the KAD was characterized by NMR, and its absolute configuration was further validated by single-crystal X-ray diffraction analysis. The UV-vis spectrum was measured to evaluate the optical characteristics of the KAD.

3.4. Determination of DPPH and ABTS⁺ free radical scavenging ability

A series of solutions of varying KAD concentrations were prepared by sequential dilution with dimethyl sulfoxide (DMSO). Vitamin C (VC) and KA were used as positive control drugs. In the experiment, three duplicate holes were created, and three parallel measurements were performed.

DPPH assay was conducted using the 96-well plate method of Jeong et al.⁴⁰ with slight modification. DPPH (180 μL, 100 μM) in ethanol solution was added to 20 μL of the sample. The absorbance at 517 nm was determined using a microplate reader after shaking the mixture in the dark for 30 min at room temperature.

ABTS⁺ assay was conducted using the 96-well plate method of Pinto⁴¹ with some modifications. A solution of ABTS in deionized water (7 mM) was reacted with 4.96 mM potassium persulfate solution and stored in the dark at room temperature for 16 h to obtain the ABTS⁺ working solution. The absorbance was read at 734 nm and adjusted to 0.70 ± 0.02. Next, 180 μL of the ABTS⁺ working solution was added to 20 μL of the sample, and the mixture was shaken for 6 min at room temperature in the dark. The absorbance at 734 nm was determined using a microplate reader.

3.5. Cytotoxicity evaluation

3.5.1. Cell culture

SH-SY5Y cells were cultured in DMEM/F12. BV-2 cells, HT-22 cells, and H9c2 cells were cultured in DMEM. L02 cells were cultured in RPMI 1640 Medium. These cells (Department of Pharmacology, School of Pharmacy, Fujian Medical University) were cultured with 10% FBS, penicillin (100 U mL⁻¹), and streptomycin (100 μg mL⁻¹). They were incubated in an incubator containing a 5% CO₂ atmosphere at 37 °C. All cells were subcultured to 90% confluence with 0.25% trypsin (w/v) every 1–3 days.

3.5.2. Cell viability assay

Cell viability was evaluated by MTT assay. BV-2, SH-SY5Y, HT-22, H9c2, and L02 cells were seeded at a density of 1 × 10⁴ cells per well. After culturing for 24 h, the cells were treated with different concentrations of compounds diluted in the culture medium for 24 h. Subsequently, 20 μL MTT (5 mg mL⁻¹ in PBS) was added to each well for 4 h at 37 °C. After the formation of formazan dye crystals, the supernatant was discarded, and the dye crystals were dissolved in 150 μL of DMSO under shaking for 15 min. A microplate reader was used to measure the absorbance at 490 nm.

3.6. Inhibition of Aβ-aggregation

To investigate the inhibitory effect of the KAD on Aβ aggregation, a thioflavin T based fluorometric assay was performed. Recombinant human HFIP-pre-treated Aβ_1–42 peptide was dissolved in 1% NH₄OH at a concentration of 1000 μM, sonicated for 1 h in a cold-water bath, and immediately stored at −20 °C. The stock solution was further diluted in PBS (10 mM, pH 7.4) to 50 μM when used. 10 μL Aβ_1–42 (25 μM, final concentration) and 10 μL of the compounds (50 μM, final concentration) or PBS to be measured were incubated for 24 h at 37 °C in the dark. Then, 20 μL sample was extracted into a black-walled 96-well plate, and 180 μL Gly-NaOH buffer containing 5 μM ThT solution (50 mM, pH 8.5) was added. After shaking at low speed for 5 min, the fluorescence intensity was determined using a multi-mode microplate reader (excitation = 450 nm; emission = 490 nm). KA and resveratrol were used as positive control drugs.

The percentage of inhibition of aggregation was calculated using the following formula: I_Aβ1–42, % = 1 − (F_s/F₀) × 100%, in which F_s and F₀ were the fluorescence intensities gained for absorbance in the presence and absence of inhibitors which had subtracted the absorbance of the background of ThT.

3.7. The KAD inhibits H₂O₂-induced injury

SH-SY5Y cells were pre-treated with different concentrations of the KAD followed by exposure to H₂O₂. The cell viability was then determined by MTT assay. The SH-SY5Y cells were collected, re-suspended, and seeded in 96-well plates at a final density of 1 × 10⁴ cells per well. After incubation for 24 h, cells were exposed to different concentrations of H₂O₂ (0, 50, 100, 200, 300, 400, 500, 600, and 700 μM) for 2 h. Subsequently, 20 μL of MTT (5 mg mL⁻¹) was added into each well. After incubation for 4 h, the cell supernatants were removed, and DMSO (150 μL per well) was added to dissolve the formazan crystals by shaking at low speed for 15 min. The absorbance at 490 nm was then measured using a microplate reader. The relative cell viability was expressed as the percentage relative to the control well (not treated with drugs).

3.8. Determination of ROS

3.8.1. Fluorescence microscopy observation

2′,7′-Dichlorodihydrofluorescein diacetate (DCFH-DA), a ROS-sensitive probe, was used to detect intracellular ROS accumulation in cells. BV-2 cells (2 × 10⁵ cells per well) or H9c2 cells (1 × 10⁵ cells per well) were pre-treated with different concentrations of the KAD for 24 h followed by treatment with 1000 μM H₂O₂ for 12 h or 400 μM H₂O₂ for 4 h, respectively. The cell supernatants were then removed and continually incubated with 10 μM DCFH-DA in serum-free medium at 37 °C for 30 min. After the DCFH-DA was removed, the cells were washed three times with PBS and maintained in serum-free medium. Fluorescence analysis was observed using an inverted fluorescence microscope (excitation = 488 nm; emission = 525 nm).

3.8.2. Flow cytometry

The intracellular accumulation of ROS was evaluated according to the manufacturer's instruction. Briefly, the sample preparation procedure was the same as that of the fluorescence microscopy observation. Finally, after the addition of 500 μL PBS, the cell solutions were blown evenly and passed through 300-mesh sieves. The intracellular ROS concentration was immediately detected by flow cytometry (excitation = 488 nm; emission = 525 nm).

3.9. Determination of SOD activity

The SOD activity was determined using commercially available kits in accordance with the manufacturer's instructions.

3.10. Metal chelating properties

The metal chelating ability of the KAD was studied using the method of Peng⁴² with some modifications. The KAD was dissolved in HEPES buffer (pH 7.4) at a concentration of 100 μM followed by the addition of solutions of CuSO₄, ZnCl₂, FeCl₂, FeCl₃·7H₂O, AlCl₃, Mn(CH₃COO)₂, MgSO₄, NaCl, KCl, and CaCl₂. The UV-vis absorption spectra at 200–800 nm were measured using a UV-vis spectrometer.

3.11. Molecular modeling with Aβ_1–42

The crystal structures of Aβ_1–42 (PDB: 1IYT) were retrieved from the RCSB Protein Data Bank (https://www.rcsb.org/pdb/) with accession codes. Protein was prepared with the Protein Preparation Wizard, while the KAD and KA were prepared with LigPrep. We referenced the key amino acid Lys28,⁴³ including all residues located within 20 Å to generate the receptor grid, and the GlideScore was obtained from the semi-flexible docking method in Glide in standard precision (SP) mode. All the above operations were performed in the Maestro 11.5 software package. Subsequently, the docking results were visualized using PyMOL software.

3.12. Statistical analysis

All results are expressed as the mean ± standard error of the mean. One-way analysis of variance was used for multiple comparisons. Dunnett's test was applied if there was a difference among the treated groups. P < 0.05 was considered significant. All analyses were performed with GraphPad Prism 8.0 or Origin 9.0 software.

Conclusions

In conclusion, OS is one of the early events leading to AD and an important factor in other pathological mechanisms of AD. Redox-active metal ions such as Cu²⁺, Zn²⁺, Al³⁺, and Fe²⁺ contribute to the production of ROS, which further promote OS and contribute to AD pathogenesis. Therefore, therapeutic agents with antioxidant activity and the ability to inhibit the abnormal deposition of metal ions in the brain have been considered for AD treatment. Polyphenols have shown promise in the treatment of AD based on their antioxidant properties and ability to chelate metals and inhibit protein aggregation.^44,45 Thus, natural polyphenols show promise as drugs for the multi-target treatment of AD.

The KAD is a secondary metabolite of plants with multiple phenolic hydroxyl groups. Our results confirm that the KAD exhibits a free radical scavenging ability, antioxidant properties, inhibition of Aβ deposition, neuroprotective effects, and metal chelating properties. Consequently, the KAD can combat OS by reducing ROS production. The KAD thus has anti-AD potential and could be further structurally modified to create naturally inspired, multifunctional drugs for AD treatment.

Author contributions

Xueyan Liu conducted the flow cytometry and molecular docking and wrote the manuscript; Chuanyu Yu conducted the in vitro activity determination; Biling Su synthesized the compounds; Daijun Zha designed and guided the entire project and edited the manuscript. All the authors have read and approved the final manuscript.

Conflicts of interest

The authors declare no conflicts of interest.