A New Class of Hybrid Anti-inflammatory Agents of Silibinin A Modified Using Gamma Irradiation

Abstract

Silibinin is the major active constituent of the medicinal plant milk thistle seeds and possesses hepatoprotective functions. In this study, silibinin A was irradiated with gamma rays to produce 2 novel flavonolignans, silibinosins A (2) and B (3). The structures of these compounds were determined using spectroscopy and spectrometry. The anti-inflammatory effects of the flavonolignan derivatives were assessed using lipopolysaccharide (LPS)-stimulated RAW264.7 and DH82 macrophages. Silibinosin A (2) effectively suppressed the LPS-induced overproduction of pro-inflammatory mediators and cytokines in murine RAW264.7 cells. Western blot analysis revealed that compound 2 decreased the LPS-induced expression of inducible nitric oxide synthase, cyclooxygenase, and phosphorylated nuclear factor-κB and inhibitor-κBα compared to the original silibinin. Furthermore, the inhibitory effects on nitric oxide and prostaglandin E₂ production were observed in LPS-stimulated DH82 canine macrophages. Our results suggest that the newly generated flavonolignans can be novel anti-inflammatory agents for use as therapeutics or ingredients in functional foods.

Introduction

Inflammation is a major defense mechanism of the body against external stimuli; however, sustained inflammatory responses can lead to tissue and DNA damage, causing various diseases, including cancer.¹⁾ Macrophages in humans and animals produce large amounts of inflammatory mediators, such as nitric oxide (NO), synthesized by inducible NO synthase (iNOS), and prostaglandin E₂ (PGE₂), which are oxidized by cyclooxygenase-2 (COX-2).^2,3) External stimuli induce excessive production of inflammatory mediators, leading to the production of cytokines, such as interleukin-1β (IL-1β), IL-6, and tumor necrosis factor-α (TNF-α), which cause inflammatory responses.⁴⁾ When pro-inflammatory factors, such as lipopolysaccharide (LPS), are administered to macrophages, iNOS and COX-2 expression is induced, leading to the production of pro-inflammatory mediators that increase the secretion of pro-inflammatory cytokines.⁵⁾ Nuclear factor-kappa B (NF-κB) is a transcription factor involved in inflammatory responses. In normal cells, NF-κB is inactive and bound to inhibitor kappa B (IκB). However, inflammatory factors phosphorylate IκB, leading to its degradation and the release of NF-κB. Activated NF-κB translocates to the nucleus and promotes the expression of proteins that induce inflammatory responses, such as iNOS, COX-2, and TNF-α.⁶⁾ Therefore, the NF-κB signaling pathway is a key pathway that regulates inflammatory responses by regulating the transcription of pro-inflammatory mediators. In recent years, research into the discovery of functional materials for companion animals has been continuously conducted in line with the growth of the companion animal industry.⁷⁾ Therefore, studies evaluating anti-inflammatory activity in DH82 canine macrophage cells are rapidly emerging regarding methods for screening functional materials for companion animals.^8,9)

Milk thistle (Silybum marianum L.) seeds have been utilized in traditional medicine to protect the liver against toxins. They contain a mixture of flavonolignans, including silibinin, isosilibinin, silychristin, and silidianin.¹⁰⁾ These flavonolignans are produced via the biosynthetic pathways of taxifolin and coniferyl alcohol.¹¹⁾ The major active flavonolignan is silibinin, which is a mixture of 2 diastereomers, silbinins A (2R,3R,7″R,8″R) and B (2R,3R,7″S,8″S), in approximately equimolar ratios.¹²⁾ Silibinin possesses many biological activities, including hepatoprotective, anticancer, anti-inflammatory, antioxidant, and antibacterial properties.¹³⁾ Several recent studies have indicated that the base-catalyzed oxidation of silibinin A produces oxidized products, such as dehydrosilibinin A, silybinic acid, and dehydrosilybinic acid, which exhibit potent antioxidant capacities.^14,15)

Gamma irradiation is a food preservation technology that has been extensively studied and shown to enhance the physicochemical properties of food products. It can also be employed to address microbial contamination.¹⁶⁾ Irradiation is known to induce specific molecular changes associated with the formation of reactive oxygen species (ROS) and free radicals.¹⁷⁾ Recent research has suggested that gamma irradiation is a promising green chemistry technique for developing structurally modified naturally occurring products and new drugs with enhanced biological efficacies and yields.^9,18) In a previous study, we irradiated rosmarinic acid with gamma rays and confirmed that the new derivatives, rosmarinosins A, B, and C, were formed and inhibited preadipocyte differentiation in 3T3-L1 cells.¹⁹⁾ Additionally, we reported that irradiation of the natural flavonoid, baicalin, produced hydroxymethylated products with potent anti-inflammatory effects.²⁰⁾ To date, research on the chemical and biological effects of gamma irradiation on major polyphenols in foodstuffs is very limited. However, as polyphenols are closely related to human health, it is essential to investigate their effects under gamma irradiation. The primary objective of this study was to assess the effects of gamma irradiation at varying doses on the molecular modification and anti-inflammatory properties of silibinin A (1) in RAW264.7 and DH82 macrophages, which are related to immunity in humans and companion animals.

Results and Discussion

Several naturally occurring polyphenols are known to exhibit potent biological properties, and various polyphenol derivatives can undergo molecular transformations through physicochemical methods under oxidative conditions.²¹⁾ Gamma irradiation is an effective green chemical modification for the molecular transformation and functional enhancement of phytophenols, such as curcumin, phoridizin, and others.^{19,20,22–24)} Silibinin A, a flavonolignan with a unique hybrid structure based on dihydroflavonol, was selected to examine structural and physiological activity changes under gamma irradiation.

A sample solution containing silibinin A in methanol was irradiated with doses of 30, 50, and 70 kGy. The transformation pattern was examined directly using reversed-phase HPLC (Fig. 1). The reactant mixture irradiated at 30 kGy showed 2 predominant peaks at t_R 12.5 (3) and 15.1 (2) min in the HPLC chromatogram, with the exception of silibinin A (1, t_R 14.9 min) (Fig. 1B). The precursor, silibinin, was not detected in the irradiated reaction at a dose of 70 kGy, and the new peak 2 was generated in relatively greater amounts than peak 3 (Fig. 1D). These samples also exhibited the most anti-inflammatory activities, as determined by the NO production assay in LPS-stimulated RAW264.7 macrophages (Supplementary Table S1). Therefore, the 70 kGy-irradiated mixture was selected for further purification. Repeated column chromatography isolation and purification of this sample yielded 2 new degradation products, 2 and 3 (Supplementary Fig. S1). The structures of the novel flavonolignans were determined by interpreting their spectroscopic data (Supplementary Fig. 2A).

Fig. 1. HPLC-DAD Chromatograms of the Irradiated Mixture of Silibinin A Using Gamma Rays

(A) 0, (B) 30, (C) 50, and (D) 70 kGy. 1: silibinin A; 2: silibinosin A; 3: silibinosin B.

Compound 2 was isolated as a yellow amorphous powder, ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ –28.0 (c 1.0, MeOH). The high-resolution electrospray ionization MS (HRESIMS) exhibited a sodium adduct ionic peak at m/z 535.1572 [M + Na]⁺, corresponding to the molecular formula C₂₇H₂₈O₁₀ (calculated for C₂₇H₂₈O₁₀Na, 535.1574). The ¹H-NMR spectrum (600 MHz) of compound 2 in acetone-d₆ + D₂O suggested the resonances for ABX-type aromatic signals at δ_H 7.00 (1H, d, J = 1.2 Hz, H-2), 6.95 (1H, d, J = 7.8 Hz, H-5), and 6.88 (1H, dd, J = 7.8, 1.2 Hz, H-6); meta-coupled aromatic signals at δ_H 5.92 (1H, d, J = 1.8 Hz, H-3′) and 5.91 (1H, d, J = 1.8 Hz, H-5′); an oxygenated methine signal at δ_H 5.41 (1H, dd, J = 12.0, 3.0 Hz, H-β); and methylene proton signals at δ_H 3.10 (1H, dd, J = 17.4, 12.0 Hz, H-α) and 2.77 (1H, dd, J = 17.4, 3.0 Hz, H-α), indicating the presence of a β-hydroxydihydrochalcone skeleton²⁵⁾ (Table 1). The combination of the ¹³C-NMR and heteronuclear single quantum coherence (HSQC) spectra of compound 2 supported the existence of signals for the β-hydroxydihydrochalcone moiety at δ_C 196.3 (C-7), 78.5 (C-β), and 42.5 (C-α). This partial structure was supported by the appearance of ¹³C-NMR resonance, which was further elucidated by key correlations of H-α/H-β in the ¹H–¹H correlated spectroscopy (COSY) spectrum, as well as the heteronuclear multiple bond correlations (HMBCs) of H-α/C-1, -1′, -7, and -β and H-β/C-1, -7, and -α (Fig. 2B).

Table 1. ¹H (600 MHz) and ¹³C (150 MHz) NMR Shifts of Compounds 2 and 3^a)

Position	2		Position	3
	δH (J in Hz)b)	δC, typec)		δH (J in Hz)b)	δC, typec)
1	—	131.8, C	—	—	—
2	7.00 (d, 1.2)	111.6, CH	2	5.10 (d, 10.2)	79.5, CH
3	—	147.0, C	3	4.07 (d, 10.2)	76.3, CH
4	—	143.8, C	4	—	75.1, C
5	6.95 (d, 7.8)	117.1, CH	5	—	158.5, C
6	6.88 (dd, 7.8, 1.2)	120.6, CH	6	5.77 (d, 1.8)	95.3, CH
7	—	196.3, C	7	—	159.2, C
α	3.10 (dd, 17.4, 12.0), 2.77 (dd, 17.4, 3.0)	42.5, CH2	8	5.90 (d, 1.8)	97.3, CH
β	5.41 (dd, 12.0, 3.0)	78.5, CH	9	—	156.4, C
1′	—	102.0, C	10	—	102.7, C
2′	—	163.2, C	11	4.03 (d, 11.4), 3.97 (d, 11.4)	65.1, CH2
3′	5.92 (d, 1.8)	95.2, CH	1′	—	132.9, C
4′	—	163.8, C	2′	6.98 (d, 1.2)	117.0, CH
5′	5.91 (d, 1.8)	96.0, CH	3′	—	148.3, C
6′	—	167.0, C	4′	—	141.1, C
7′	3.52 (q, 7.2)	56.8, CH2	5′	6.96 (d, 7.8)	121.7, CH
8′	1.07 (t, 7.2)	17.6, CH3	6′	6.94 (dd, 7.8, 1.2)	117.2, CH
1″	—	128.0, C	1″	—	128.8, C
2″	7.02 (d, 1.2)	115.1, CH	2″	7.03 (d, 1.2)	111.8, CH
3″	—	147.9, C	3″	—	148.4, C
4″	—	143.9, C	4″	—	147.6, C
5″	6.84 (d, 7.8)	119.8, CH	5″	6.85 (d, 7.8)	115.7, CH
6″	6.98 (dd, 7.8, 1.2)	115.2, CH	6″	6.90 (dd, 7.8, 1.2)	121.6, CH
7″	4.90 (d, 7.8)	76.3, CH	7″	4.90 (d, 7.8)	76.8, CH
8″	4.12 (ddd, 7.8, 4.8, 1.8)	48.4, CH	8″	4.10 (m)	79.0, CH
9″	3.65 (dd, 12.0, 1.8), 3.46 (dd, 12.0, 4.8)	60.7, CH2	9″	3.66 (dd, 12.0, 1.8), 3.45 (dd, 12.0, 4.8)	61.3, CH2
3″-OCH3	3.80 (s)	55.5, CH3	3″-OCH3	3.80 (s)	56.1, CH3

^a) Measured in acetone-d₆ + D₂O and assignments of chemical shifts are based on the analysis of 1D- and 2D-NMR spectra. The overlapped signals were assigned from HSQC, HMBC, and ¹H–¹H COSY spectra without designating multiplicity. ^b) Data (δ) measured at 600 MHz. ^c) Data (δ) measured at 150 MHz.

Fig. 2. Structures (A) and 2D-NMR Correlations (B) of the Novel Hybrid Products 2 and 3 Using Gamma Ray

Blue arrow: HMBC correlation; red line: ¹H-¹H COSY correlation; and double arrow: NOESY correlation.

The remaining ¹H-NMR signals of compound 2 illustrated the presence of extra ABX-type aromatic systems at δ_H 7.02 (1H, d, J = 1.2 Hz, H-2″), 6.98 (1H, dd, J = 7.8, 1.2 Hz, H-5″), and 6.84 (1H, d, J = 7.8 Hz, H-5″) and a methoxy group at δ_H 3.80 (3H, s, 3″-OCH₃). In addition to these proton signals, the ¹H-¹H COSY spectrum revealed aliphatic AMXY-type signals at δ_H 4.90 (1H, d, J = 7.8 Hz, H-7″), 4.12 (1H, ddd, J = 7.8, 4.8, 1.8 Hz, H-8″), 3.65 (1H, dd, J = 12.0, 1.8 Hz, H-9″), and 3.46 (1H, dd, J = 12.0, 4.8 Hz, H-9″), indicating the presence of a dihydroconiferyl alcohol moiety.²⁶⁾ The deshielded oxymethine doublet at δ_H 4.90 (H-7″) and doublet of double doublet at δ_H 4.12 (H-8″) implied the linkage of dihydrochalcone and phenylpropanoid units by a 1,4-dioxane bridge.²⁷⁾ Additionally, ethoxyl groups were observed at δ_H 3.52 (2H, q, J = 7.2 Hz, H-7′) and 1.07 (3H, t, J = 7.2 Hz, H-8′). The ¹³C-NMR and HSQC spectra of compound 2 revealed 27 carbons, including 15 dihydrochalcone skeleton carbons and 9 phenylpropanoid skeleton carbons, along with an ethoxy group and a methoxy carbon. This suggests that compound 2 was a hybrid of phenylpropanoids and dihydrochalcones formed through a 1,4-dioxane bridge. The ethoxy and methoxy groups were located at the C-6′ and C-3″ positions, respectively, and supported by the key HMBC spectrum, which showed H-7′ to C-6′ and 3″-OCH₃ to C-3″ correlations, respectively. The connection points of the dihydroconiferyl alcohol and ethylated β-hydroxydihydrochalcone moieties were unambiguously elucidated by the key HMBCs of H-7″ to C-3 and H-8″ to C-4 (Fig. 2B).

The large coupling constant (J_7″,8″ = 7.8 Hz) between H-7″ and H-8″, along with the nuclear Overhauser effect spectroscopy (NOESY) correlations between H-2″/H-8″ and H-7″/H-9″ clearly indicate the trans-configuration of the chiral center on the dioxane ring²⁸⁾ (Fig. 2B). The absolute configuration of compound 2 was determined through circular dichroism (CD) spectroscopy. A comparison of the CD spectra of the 2 possible stereoisomers (βS,7″R,8″R and βR,7″S,8″S) with the experimental data for compound 2 indicated that the most probable configuration of compound 2 was βS,7″R,8″R-configuration (Fig. 3A). Therefore, the absolute structure of the novel chalcone lignan 2 was assigned as silibinosin A, which is a new modified product of the original compound 1 (Fig. 2A).

Fig. 3. CD Spectra of Compounds 2 (A) and 3 (B)

Compound 3 was obtained as a white amorphous powder with a molecular formula of C₂₆H₂₆O₁₁ determined based on a combination of 1-dimensional (1D) NMR spectroscopic data and positive-ion mode HRESIMS at m/z 537.1362 [M + Na]⁺ (calculated for C₂₆H₂₆O₁₁Na, 537.1367). The ¹H-NMR spectrum (600 MHz) of compound 3 in acetone-d₆ + D₂O showed the resonance of 2 ABX-type aromatic protons at δ_H 7.03 (1H, d, J = 1.2 Hz, H-2″), 6.98 (1H, d, J = 1.2 Hz, H-2′), 6.96 (1H, d, J = 7.8 Hz, H-5′), 6.94 (1H, dd, J = 7.8, 1.2 Hz, H-6′), 6.90 (1H, dd, J = 7.8, 1.2 Hz, H-6″), and 6.85 (1H, d, J = 7.8 Hz, H-5″), indicating the presence of two 1,3,4-trisubstituted aromatic rings (Table 1). The spectrum also revealed meta-coupled aromatic signals at δ_H 5.90 (1H, d, J = 1.8 Hz, H-8) and 5.77 (1H, d, J = 1.8 Hz, H-6); 2 oxygenated methine signals at δ_H 5.10 (1H, d, J = 10.2 Hz, H-2) and 4.07 (1H, d, J = 10.2 Hz, H-3); a methoxyl group at δ_H 3.80 (3H, s, 3″-OCH₃); and additional AMXY-type signals at δ_H 4.90 (1H, d, J = 7.8 Hz, H-7″), 4.10 (1H, m, H-8″), 3.66 (1H, dd, J = 12.0, 1.8 Hz, H-9″), and 3.45 (1H, dd, J = 12.0, 4.8 Hz, H-9″). Furthermore, the 1D NMR spectrum of compound 3 displayed resonances corresponding to hydroxymethyl groups at δ_H 4.03 (1H, d, J = 11.4 Hz, H-11), 3.97 (1H, d, J = 11.4 Hz, H-11), and δ_C 65.1 (C-11) and quaternary carbon at δ_C 75.1 (C-4). Consistent with these ¹H-NMR observations, the ¹³C-NMR and HSQC spectra of compound 3 closely resembled those of the original compound, silibinin A (1),^29,30) except for the presence of a hydroxymethyl and hydroxyl group instead of a ketone group at the C-4 position in compound 3. The linkage points of the hydroxymethyl residue at the C-4 position in compound 3 were clearly determined from the key HMBCs, which displayed H-3/C-2, -4, and -11 and H-11/C-3, -4, and -10 relationships (Fig. 2B).

The relative configuration of the flavan 3,4-diols skeleton of compound 3 was established by analyzing the NOESY correlations observed between H-2 and H-11, as well as H-3 and H-2′, along with the large coupling constant (J_2,3 = 10.2 Hz), which confirmed the trans–trans relationship^30,31) (Fig. 2B). Furthermore, an energy-minimized model generated using Chem3D Ultra 10.0, based on the presumed configuration, aligned well with the observed NOESY correlations. The large coupling constant (J_7″,8″ = 7.8 Hz) in the dioxane ring also indicated a trans-configuration between the C-7″ and C-8″ positions of compound 3.²⁸⁾ The absolute configuration of compound 3 was determined by comparing the experimental and calculated CD data. The calculated CD spectrum of the 2R,3S,4R,7″R,8″R-stereoisomer of compound 3 matched well with the experimental curve. However, its enantiomer displayed mirror-like cotton effects (Fig. 3B). Thus, the absolute structure of the unusual flavonolignan 3 was proposed as silibinosin B, as shown in Fig. 2A.

The ROS produced during gamma irradiation in alcoholic conditions have been reported to be hydroxyl (HO^•), superoxide anion (O₂^•–), hydrogen peroxide (H₂O₂), hydroxyalkyl (^•CH₂OH), and methoxyl (^•OCH₃) radicals.^32,33) Although gamma irradiation is effective for microbial inactivation, it has only recently been applied to the development of functional materials.^18,19 Recently, gamma-irradiated rutin and mangiferin were readily converted into hydroxymethylation products with a hydroxymethyl functionality substituted instead of a ketone moiety,^23,34) and our results were similar to those of previous studies (Chart 1). In addition, natural representative flavonols, including quercetin, morin, kaempferol, and galangin, were irradiated with gamma rays to produce ethylated depside benzoic acid when the C-ring was broken.^35,36) However, our current study is the first to report the major conversion of a flavonoid-based derivative to a dihydrochalcone through ring cleavage at the C-ring (Fig. 2A). Furthermore, β-hydroxydihydrochalcones represent a rare subclass of chalcones, with only a few reported in natural products, yet they exhibit promising pharmacological activities.^25,37) Based on the observed structural transformations and previous studies on gamma ray-induced radical reactions, we propose a detailed reaction mechanism for the formation of compound 2. Under gamma irradiation in methanolic conditions, silibinin A (1) undergoes hydroxymethyl radical (^•CH₂OH)-mediated substitution at the C-2 position, generating a reactive intermediate.^20,23) This intermediate then undergoes C-ring cleavage, leading to the formation of a dihydrochalcone structure. During the ring cleavage process, formaldehyde (HCHO) is generated from methanol radiolysis under gamma ray exposure, following the established radical recombination mechanism.^32,38) We propose that this methanol-derived formaldehyde reacts with the dihydrochalcone-like intermediate, leading to the formation of a β-hydroxyl group (C-β position) at the newly formed dihydrochalcone structure 2. In the final step, the hydroxyl radical (HO^•) generated by gamma irradiation facilitates a dehydroxylation reaction that removes hydroxyl groups (α-OH and 8′-OH), allowing structural rearrangement into the final product.^32,39) These sequential transformations result in the formation of silibinosin A (2), which contains a β-hydroxydihydrochalcone moiety (Chart 1). This additional insight further supports the role of gamma ray-induced methanol-derived radicals in facilitating novel structural modifications that contribute to the unique transformation of silibinin A into novel chalcone-lignan hybrid 2.

Chart 1. Proposed Mechanism of the Formation of the Novel Flavonolignans Using Gamma Ray

The formation of compound 3 follows a distinct pathway, primarily involving hydroxymethylation at the C-4 ketone position of silibinin A. The gamma ray-induced hydroxymethyl radical preferentially reacts at this site to form a hydroxymethylated product.³⁴⁾ Given the potential for diastereomeric variations, the observed product distribution suggests a stereoselective preference driven by steric hindrance and electronic effects. Additionally, hydrogen bonding interactions may contribute to stabilizing a specific diastereomer. These factors collectively influence the final structural outcome, leading to the selective formation of compound 3 with a hydroxymethylated flavonolignan skeleton (Chart 1).

RAW 264.7 macrophages were treated with various concentrations of silibinosin A (2), modified from silibinin by gamma irradiation, for 24 h. Cytotoxicity was assessed using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Silibinosin A showed no cytotoxic activity in the cells compared to the untreated control cells. The highest concentration of silibinosin A (2) that caused no more than a 5% loss in cell viability was 20 µM (Fig. 4A). Based on this result, a silibinosin A concentration below 20 µM was chosen for further studies. The novel flavonolignan silibinosin B (3), based on flavan 3,4-diols, exhibited lower inhibitory activity against NO production in RAW264.7 cells at a concentration of 20 µM compared to compounds 1 and 2 (Supplementary Fig. S18).

Fig. 4. Effects of Silibinosin A on Anti-inflammatory Activities in LPS-Stimulated RAW264.7 Cells

Cell were pretreated with silibinin A (1) and silibinosin A (2) at concentrations of 5, 10, and 20 µM for 1 h, and then stimulated with LPS (0.1 µg/mL) for 24 h. (A) Cell viability detection by MTT assay in RAW264.7 cells. (B) The nitrite content of culture media was analyzed by a Griess reagent assay. (C) The PGE₂ production content of culture media was analyzed by ELISA. (D) The protein levels of iNOS and COX-2 in cell lysate were analyzed by Western blotting analysis. GAPDH was used as a loading control. Western blotting data are shown as a representative plot from 3 independent experiments. (E, F) The relative band intensity of each protein is expressed as a percentage. The amount of (G) TNF-α, (H) IL-1β, and (I) IL-6 in the culture medium was measured by ELISA. The results are expressed as mean ± standard deviation (S.D.) (n = 3). *p < 0.05; **p < 0.01 vs. LPS group by unpaired Student’s t-test.

The potential anti-inflammatory capacities of silibinosin A (2) in RAW264.7 macrophages were examined after 24 h of treatment with a mixture of compounds (5, 10, and 20 µM) and LPS (0.1 µg/mL). Enzyme-linked immunosorbent assay (ELISA) kits were used to determine the concentrations of NO (µM) and PGE₂ (pg/mL) in cell supernatants. The novel flavonolignan, silibinosin A (2), based on dihydrochalcone, considerably inhibited NO and PGE₂ production in LPS-stimulated RAW264.7 cells in a dose-dependent manner compared to the mother compound, silibinin (Figs. 4B, 4C).

We also investigated changes in the iNOS and COX-2 protein expression following silibinosin A treatment. iNOS is a synthetic protein involved in NO generation, whereas COX-2 is an oxidative protein involved in PGE₂ generation. To determine whether silibinosin A inhibits pro-inflammatory repertoires at the protein levels, RAW264.7 cells were induced with LPS for 24 h, and the expression of iNOS and COX-2 proteins was examined in the presence of compound 2. Western blot analysis was used to assess the protein levels. As shown in Fig. 4D, RAW264.7 cells were highly activated by LPS. However, the protein expression levels of iNOS and COX-2 were significantly inhibited at silibinosin A (2) concentrations of 5, 10, and 20 µM, compared to silibinin A (Figs. 4E, 4F). These results closely correlated with the inhibitory effects on NO and PGE₂ production.

Pro-inflammatory cytokines have small molecular weights and are mostly generated through transcriptional and translational regulation via induction by immunogens or other promoters.⁴⁰⁾ The effects of silibinosin A (2) on LPS-stimulated pro-inflammatory cytokines, including TNF-α, IL-1β, and IL-6 production, in RAW 264.7 cells were analyzed using ELISA assay kits. Treatment of RAW 264.7 cells with LPS alone resulted in a significant increase in cytokine expression relative to the untreated control group. Compared to its parent compound, silibinin A, the newly generated silibinosin A (2) significantly inhibited the production of TNF-α (Fig. 4G), IL-1β (Fig. 4H), and IL-6 (Fig. 4I) in LPS-treated macrophages in a dose-dependent manner.

NF-κB is a major transcription factor that regulates the production of various cytokines. Upon activation by external stimuli, NF-κB rapidly translocates to the nucleus of cells, where it stimulates the expression of inflammatory mediators.⁴¹⁾ Therefore, we investigated whether the novel flavonolignan, silibinosin A (2), could inhibit LPS-induced NF-κB activation and translocation to the nucleus of macrophages or the degradation of phosphorylated IκBα in RAW 264.7 cells using Western blot analysis (Fig. 5A). The expression levels of NF-κB protein in the nuclei of the cells increased after LPS treatment compared to that in the untreated control group. However, when the LPS-induced macrophages were treated with silibinosin A (2), the amount of NF-κB protein translocating to the nucleus was markedly reduced compared to when the cells were treated with the parent compound, silibinin (Fig. 5B). Additionally, the phosphorylated forms of IκBα were lowly expressed in the untreated groups. However, upon exposure to LPS alone, IκBα phosphorylation increased, and silibinosin A (2) decreased the LPS-induced IκBα phosphorylation in a concentration-dependent manner (Fig. 5C).

Fig. 5. Effects of Silibinosin A on the Activation of NF-κB in LPS-Stimulated RAW264.7 Cells

Cells were pretreated with silibinin A (1) and silibinosin A (2) at concentrations of 5, 10, and 20 µM for 1 h, and then stimulated with LPS (0.1 µg/mL) for 24 h. (A) The protein levels of p-IκBα, IκBα, p-NF-κB, and NF-κB in cell lysate were analyzed by Western blotting analysis. GAPDH was used as a loading control. Western blotting data are shown as a representative plot from 3 independent experiments. (B, C) The relative band intensity of each protein is expressed as a percentage. The results are expressed as mean ± S.D. (n = 3). *p < 0.05; **p < 0.01 vs. LPS-treated group by unpaired Student’s t-test.

Kang et al. and Youn et al. reported that the flavonolignan silibinin exhibits potent anti-inflammatory properties through the NF-κB activation pathway in RAW 264.7 macrophages. However, silibinin was active at a concentration of at least 25 µg/mL (ca. 51.8 µM) or higher.^42,43) Our results confirmed that the new dihydrochalcone–lignan hybrid, silibinosin A (2), modified from silibinin A (1) by gamma irradiation, exhibited enhanced anti-inflammatory activities at a relatively low concentration of 20 µM in LPS-stimulated RAW264.7 cells, compared to parent silibinin A. These results suggest that ring cleavage at the C-ring in the flavonolignan may have influenced its anti-inflammatory activities (Chart 1).

Milk thistle (Silybum marianum L.), which is rich in flavonolignans such as silibinin, isosilibinin, and silidianin, is also used as a companion animal drug because of its many beneficial properties.⁴³⁾ In this study, a novel chalcone–lignan hybrid, compound 2, was isolated by irradiating a flavonolignan with gamma rays. Compound 2 exhibited significantly enhanced anti-inflammatory effects in LPS-stimulated RAW 264.7 mouse macrophages. Additionally, silibinosin A (2) was evaluated as an anti-inflammatory agent in companion animals using DH82 canine macrophages (Fig. 6).

Fig. 6. Effects of Silibinosin A on Anti-inflammatory Activities in LPS-Stimulated DH82 Cells

Cells were pretreated with silibinin A (1) and silibinosin A (2) at concentrations of 5, 10, and 20 µM for 1 h, and then stimulated with LPS (0.1 µg/mL) for 24 h. (A) Cell viability detection by MTT assay in DH82 cells. (B) The nitrite content of culture media was analyzed by a Griess reagent assay. (C) The PGE₂ production content of culture media was analyzed by ELISA. The results are expressed as mean ± S.D. (n = 3). *p < 0.05; **p < 0.01 vs. LPS group by unpaired Student’s t-test.

An MTT assay was performed to determine the cytotoxicity of silibinin (1) and silibinosin A (2) on DH82 cells. The cells were cultured in the presence of a range of concentrations of compounds (5, 10, and 20 µM) or LPS (0.1 µg/mL) for 24 h at 37°C. Neither the concentrations of the compounds nor LPS treatment affected the viability of DH82 macrophages (Fig. 6A).

The inhibitory effects of compounds 1 and 2 on NO production in LPS-stimulated DH82 cells were investigated. As shown in Fig. 6B, NO production increased drastically after LPS treatment of canine DH82 macrophages. The unique hybrid based on the β-hydroxydihydrochalcone, silibinosin A (2), remarkably inhibited LPS-treated NO production in a dose-dependent manner compared to silibinin. ELISA was performed to determine whether silibinosin A (2) could affect the secretion of PGE₂ in LPS-induced DH82 macrophages. As depicted in Fig. 6C, the levels of PGE₂ were increased sharply after LPS treatment compared to those in the control group. In contrast, treatment with silibinosin A (2) at concentrations of 10 and 20 µM significantly inhibited the production of PGE₂ compared to the LPS-treated group. Therefore, the enhancement of the anti-inflammatory properties of irradiated silibinin A in canine macrophages might be due to irradiation processing, resulting in its transformation from a flavonolignan to a new class of hybrids.

Alcoholic radiolysis produces a wide range of free radicals that can cause chemical reactions in natural products and withdrawal drugs. Ionizing radiation is a non-thermal process that can be used to inactivate pathogenic bacteria in food and related industries. It has also been recognized as a discovery and development tool for incrementally modified drugs.¹⁸⁾ Our group recently reported that the hydroxymethyl (^•CH₂OH) radicals produced by methanolic radiolysis hydroxymethylated minaprine and baicalin, and these compounds showed the greatest anti-inflammatory activity in LPS-stimulated RAW 264.7 and DH82 macrophages.^9,20) In this study, we found that the irradiation of silibinin resulted in the modification of a new class of flavonolignans, silibinosins A (2) and B (3), through hydroxymethylation, ethylation, and ring-cleavage reactions (Chart 1). These compounds showed potential anti-inflammatory effects in RAW264.7 and DH82 macrophage cells. These findings provide useful evidence for the interaction between energetically highly reactive gamma irradiation and principal dietary bioactive ingredients in foods.

In this study, we observed that silibinin A (1) could be easily transformed into 2 novel hybrids, silibinosins A (2) and B (3), based on the β-hydroxydihydrochalcone and flavan 3,4-diols, respectively. The structures of the newly generated compounds were elucidated using a combination of NMR and MS spectroscopic data, and their absolute configurations were determined. The results demonstrated that the chalcone–lignan hybrid, silibinosin A (2), was the most effective anti-inflammatory agent among the isolated flavonolignans. Silibinosin A can inhibit the production of the inflammatory mediators NO and PGE₂ in LPS-stimulated RAW264.7 macrophages and downregulate the expression of their synthetic enzymes (iNOS and COX-2, respectively). It can also suppress the production of pro-inflammatory cytokines such as TNF-α, IL-1β, and IL-6 more effectively than the original silibinin A (1). These effects corresponded to the inhibition of IκBα phosphorylation through the downregulation of the NF-κB signaling pathway. Furthermore, silibinosin A (2) inhibited the production of NO and PGE₂ in LPS-stimulated DH82 canine macrophages more effectively than silibinin. Further systematic investigation of major food ingredients is essential to improve the safety and biological activity of not only human foods but also functional foods for companion animals.

Experimental

Chemicals and Instruments

Silibinin A (99% purity), acetonitrile (MeCN), methanol (MeOH), acetone-d₆, deuterium oxide (D₂O), LPSs, and Griess reagent were purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.). The UV spectrum was obtained using MeOH as the solvent with a T-60 spectrophotometer (PG Instruments, Leicestershire, U.K.). CD spectra and optical rotations were recorded on JASCO J-1500 and P-2000 spectrometers (JASCO, Tokyo, Japan). NMR spectroscopy was performed on an Avance NEO-600 instrument (Bruker, Karlsruhe, Germany) operating at 600 MHz (¹H) and 150 MHz (¹³C) and processed using MestReNova software (version 9.0). Chemical shifts were referenced to the carbon and residual proton signals of the deuterated solvent, acetone-d₆ (δ_H 2.04; δ_C 29.8), as the internal standard. The HRESIMS were measured using a Vanquish UPLS System (Thermo Fisher Scientific, Waltham, MA, U.S.A.). Column chromatography was conducted using Toyopearl HW-40 (coarse grade; Tosoh Co., Tokyo, Japan), Sephadex LH-20 (particle size 25–100 µm; GE Healthcare Biosciences AB, Uppsala, Sweden), and YMC gel ODS AQ 120-50S (particle size 50 µm; YMC Co., Kyoto, Japan) gel columns. A microplate reader (Infinite F200, Tecan Austria GmBH, Grodig, Austria) was utilized to measure the absorbance. Semi-preparative HPLC was performed using a Agilent HPLC 1200 system (Agilent Technologies, Palo Alto, CA, U.S.A.), equipped with a photodiode array detector (1200 Infinity series, Agilent Technologies) and a series of YMC-Pack ODS A-302 columns (4.6 mm i.d. × 150 mm, particle size 5 µm; YMC Co.). The solvent system that was used to purify the compounds comprised a gradient mode with an initial 0.1% of HCOOH in H₂O, which was changed to MeCN over 27 min (temperature: 40°C; flow rate: 1.0 mL/min; UV detection: 280 nm). All procedures were performed using solvents purchased from commercial sources without further purification.

Sample Preparation Procedure

Ionizing radiation was applied using a previously reported method at 25°C with a cobalt-60 experimental irradiator (point source AECL, IR-79, MDS Nordion International Co., Ltd., Ottawa, ON, Canada) located at the Advanced Radiation Technology Institute, Korea Atomic Energy Research Institute, Jeongup, Korea.¹⁹⁾ The source strength was approximately 320 kCi at a dose rate of 10 kGy/h at the sample position. Dosimetry was conducted using 5 mm diameter alanine dosimeters (Bruker Instruments, Rheinstetten, Germany) calibrated against an International Standard set by the International Atomic Energy Agency (Vienna, Austria). Pure silibinin A (200 mg) in MeOH (200 mL) was placed in conical test tubes and irradiated with gamma rays at doses of 30, 50, and 70 kGy. The solutions irradiated at different doses were immediately concentrated using a rotary vacuum evaporator to remove methanol and then lyophilized.

Isolation of Modified Compounds

The irradiated mixtures (195.0 mg) at a dose of 70 kGy were directly subjected to column chromatography over a Toyopearl HW-40 column (coarse grade; 2.5 cm i.d. × 30 cm) and eluted in a gradient manner with H₂O-MeOH (0:100 to 100:0, then aqueous acetone) to produce 3 subfractions ISA01 to ISA03. Subfraction ISA01 (37.7 mg) was further purified by passage over a YMC GEL ODS AQ column (1 cm i.d. × 40 cm) with aqueous MeOH using reversed-phase HPLC to yield pure compound 3 (3.6 mg, t_R 12.5 min). Subfraction ISA03 (130.1 mg), which contained the major modified products, was subjected to column chromatography over a Sephadex LH-20 (1.0 cm i.d. × 37 cm) with EtOH to yield pure compound 2 (53.7 mg, t_R 15.1 min). The yields of compounds 2 and 3 are 27.5 and 1.8%, respectively.

Silibinosin A (2) comprised the following properties: yellow amorphous powder, ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ –28.0 (c 1.0, MeOH); UV λ_max (MeOH) (logε): 210 (3.56), 232 (sh), 284 (1.12) nm; CD (MeOH) Δε: 220 (−1.0), 291 (–4.1), 328 (+1.3) nm; ¹H- and ¹³C-NMR, see Table 1; ESIMS m/z 535 [M + Na]⁺, HRESIMS m/z 535.1572 [M + Na]⁺ (calculated for C₂₇H₂₈O₁₀Na, 535.1574) (Supplementary Figs. S2–S9).

Silibinosin B (3) comprised the following properties: white amorphous powder, ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ +19.5 (c 1.0, MeOH); UV λ_max (MeOH) (logε): 238 (2.02), 279 (2.50) nm; CD (MeOH) Δε: 222 (−3.2), 236 (−7.2) 290 (−1.0) nm; ¹H- and ¹³C-NMR, see Table 1; ESIMS m/z 537 [M + Na]⁺, HRESIMS m/z 537.1362 [M + Na]⁺ (calculated for C₂₆H₂₆O₁₁Na, 537.1367) (Supplementary Figs. S10–S17).

DFT/TDDFT Calculations

To generate conformers for the subsequent DFT/TDDFT calculations, Merck molecular force field (MMFF) calculations were carried out using the Spartan ‘14 software package (Wavefunction Inc., Irvine, CA, U.S.A.). From compounds 2 and 3, a total of 21, 68, and 12 low-energy conformers were generated, respectively. All DFT/TDDFT computations were performed using the Gaussian 09 software.⁴⁴⁾ Initially, the low-energy conformers derived from the MMFF calculations underwent geometry optimization using the DFT method at the B3LYP/6-31G(d) level, incorporating the polarizable continuum model (PCM) in methanol. Following this, each optimized conformer was subjected to frequency calculations at the same theoretical level (B3LYP/6-31G(d) with PCM in methanol) to estimate the thermal free energy (ΔG) and verify the absence of imaginary frequencies.⁴⁵⁾ The relative abundance of each conformer was determined based on thermal energy calculations using the Boltzmann distribution. Ultimately, 2 dominant conformers were selected from compound 2 (comprising approximately 90% of the total) and 2 from compound 3 (accounting for around 95%) for electronic CD (ECD) calculations. ECD spectra for all conformers were simulated using the TDDFT method at the B3LYP/6-31G(d) level with PCM in methanol. The resulting weighted-average spectra were subsequently compared with the experimentally recorded ECD spectra in methanol.

Cell Culture

The RAW 264.7 cell line was obtained from the Korean Cell Line Bank (Seoul National University, Seoul, Korea), and the DH82 cell line was obtained from the American Type Culture Collection (ATCC, Rockville, MD, U.S.A.). Macrophages were cultured under sterile conditions at 37°C in a humidified atmosphere containing 5% CO₂. The macrophage cell culture medium consisted of Dulbecco’s Modified Eagle’s Medium supplemented with 1% penicillin/streptomycin and 10% fetal bovine serum (GIBCO, Carlsbad, CA, U.S.A.).

Cell Viability Assay

The cell viability in RAW 264.7 and DH82 cells was evaluated using the MTT method.⁴⁶⁾ Cells were seeded at a density of 5 × 10⁴ cells/well into 96-well plates and incubated for 24 h at 37°C. The cells were treated with isolated compounds (5, 10, and 20 µM) in dissolved free medium and incubated for 24 h at 37°C. A solution of MTT (0.5mg/mL) was added to each well, and the plates were incubated for 4 h at 37°C to allow the reaction to take place before removal of the culture medium. Then, the produced formazan blue was dissolved in dimethyl sulfoxide. Cell viability was determined using a spectrophotometer, and the absorbance was measured at 570 nm. The control group was considered to be 100%.

Pro-inflammatory Mediator Content Assay

RAW264.7 and DH82 cells were plated in a 96-well plate at a density of 5 × 10⁴ cells/well and incubated for 24 h at 37°C. The cells were pre-treated with various concentrations of isolated compounds for 2 h before incubating with LPS (0.1 µg/mL) for 24 h at 37°C. NO production was determined by the reaction of the macrophage culture supernatant with Griess reagent.⁴⁷⁾ The culture supernatant (100 µL) was mixed with Griess reagent (100 µL) at room temperature and shaken gently for 20 min. Finally, the absorbance of the reactants was measured at 548 nm using a microplate reader. PGE₂ and cytokine levels were determined by ELISA using commercial reagent kits (BD Biosciences, San Jose, CA, U.S.A.) according to the manufacturer’s instructions.⁴⁸⁾

Western Blot Analysis

Sample-treated RAW264.7 macrophage cells were harvested and lysed using radioimmunoprecipitation assay buffer (Rockland Immunochemicals, Inc., Limerick, PA, U.S.A.). Cell debris was removed by centrifugation, and protein concentration was determined using a bicinchoninic acid protein assay kit (Thermo Fisher Scientific, Inc., Waltham, MA, U.S.A.) according to the manufacturer’s instructions. Cell lysates containing an equal amount of protein (30 µg) were prepared and separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) along with 10 µg of cytosolic fractions. Moreover, 15 µg of both mitochondria and debris fractions were separated using 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% nonfat milk in Tris-buffered saline containing Tween-20 for 1 h at room temperature. The membranes were probed overnight at 4°C with primary antibodies at a dilution of 1:1000 for anti-COX-2 (Cat. #4842), anti-iNOS (Cat. #2977), anti-NF-κB (Cat. #8242), anti-p-NF-κB (Cat. #3033), anti-IκB (Cat. #4812), anti-p-IκB (Cat. #2859), and anti-GAPDH (Cat. #2118) (all primary antibodies; Cell Signaling Technology, Danvers, MA, U.S.A.). The blots were then incubated with horseradish peroxidase-conjugated anti-rabbit IgG (secondary antibody; Cat. #7074 at 1:3000; both from Cell Signaling Technology) for 2h at room temperature. The membranes were incubated with a chemiluminescence reagent for protein band detection.

Statistical Analysis

All data were evaluated using the Student’s 2-tailed t-test, and results were considered statistically significant at *p < 0.05 and **p < 0.01. All experiments were independently performed at least 3 times.

Acknowledgments

This research was supported by a KAERI Institutional Program (Project No. 523310-25) Grant funded by the Nuclear R&D Program of the Ministry of Science and ICT and the “regional innovation mega project” program through the Korea Innovation Foundation funded by the Ministry of Science and ICT (Project Number: 2023-DD-UP0031).

Author Contributions

Gyeong Han Jeong: conceptualization, methodology, investigation, resources, writing—original draft, writing—review and editing. Hanui Lee: conceptualization, methodology, formal analysis, investigation, writing—original. Byung Yeoup Chung: supervision, resources, writing—review and editing. Hyoung-Woo Bai: project administration, validation, visualization, writing—review and editing.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1)  Nathan C., Ding A., Cell, 140, 871–882 (2010).
• 2)  Murakami A., Ohigashi H., Int. J. Cancer, 121, 2357–2363 (2007).
• 3)  German A. J., Ryan V. H., German A. C., Wood I. S., Trayhurn P., Net. J., 185, 4–9 (2010).
• 4)  Dinarello C. A., Chest, 118, 503–508 (2000).
• 5)  Huang Z., Kraus V. B., Nat. Rev. Rheumatol., 12, 123–129 (2016).
• 6)  Shih R. H., Wang C. Y., Yang C. M., Front. Mol. Neurosci., 8, 77 (2015).
• 7)  de Godoy M. R., Kerr K. R., Fahey G. C. Jr., Nutrients, 5, 3099–3117 (2013).
• 8)  Park W. B., Kim S., Kyung S. M., Lee E. S., Lee Y. J., Yoo H. S., Net. Immunol. Immunop., 260, 110607 (2023).
• 9)  Jeong G. H., Lee H., Woo S. Y., Lee H. K., Chung B. Y., Bai H. W., Sci. Rep., 13, 10887 (2023).
• 10)  Bijak M., Molecules, 22, 1942 (2017).
• 11)  Lv Y., Xu S., Lyu Y., Zhou S., Du G., Chen J., Zhou J., Green Chem., 21, 1660–1667 (2019).
• 12)  Lee D. Y. W., Liu Y., J. Nat. Prod., 66, 1171–1174 (2003).
• 13)  Gazák R., Walterova D., Kren V., Curr. Med. Chem., 14, 315–338 (2007).
• 14)  Gažák R., Trouillas P., Biedermann D., Fuksová K., Marhol P., Kuzma M., Křen V., Tetrahedron Lett., 54, 315–317 (2013).
• 15)  Gažák R., Svobodová A., Psotová J., Sedmera P., Přikrylová V., Walterová D., Křen V., Bioorg. Med. Chem., 12, 5677–5687 (2004).
• 16)  Farkas J., Int. J. Food Microbiol., 44, 189–204 (1998).
• 17)  Shah N. S., Khan J. A., Nawaz S., Khan H. M., J. Hazard. Mater., 278, 40–48 (2014).
• 18)  Song H. Y., Kim K. I., Han J. M., Park W. Y., Seo H. S., Lim S., Byun E. B., Radiat. Phys. Chem., 194, 110013 (2022).
• 19)  Jeong G. H., Cho J. H., Jo C., Lee S., Lee S. S., Bai H. W., Chung B. Y., Kim T. H., Food Chem., 258, 181–188 (2018).
• 20)  Jeong G. H., Lee H., Lee H. K., Choi H. J., Chung B. Y., Bai H. W., Bioorg. Med. Chem. Lett., 96, 129491 (2023).
• 21)  Wang L., Yang X., Yu X., Yao Y., Ren G., J. Agric. Food Chem., 61, 12274–12282 (2013).
• 22)  Jeong G. H., Kim T. H., Chem. Biodivers., 17, e2000305 (2020).
• 23)  Jeong G. H., Kim T. H., Chem. Pharm. Bull., 65, 678–682 (2017).
• 24)  Kim T. H., Kim J. K., Ito H., Jo C., Bioorg. Med. Chem. Lett., 21, 1512–1514 (2011).
• 25)  Chepkirui C., Bourgard C., Gilissen P. J., Ndakala A., Derese S., Gütlin Y., Erdélyi M., Yenesew A., Fitoterapia, 158, 105166 (2022).
• 26)  Savidge R. A., Förster H., Phytochemistry, 57, 1095–1103 (2001).
• 27)  Khan S. A., Ahmad B., Alam T., Pak. J. Pharm. Sci., 19, 290–294 (2006).
• 28)  Kim T. H., Ito H., Hayashi K., Hasegawa T., Machiguchi T., Yoshida T., Chem. Pharm. Bull., 53, 641–644 (2005).
• 29)  Monti D., Gaž́ák R., Marhol P., Biedermann D., Purchartová K., Fedrigo M., Riva S., Křen V., J. Nat. Prod., 73, 613–619 (2010).
• 30)  Lee T. H., Hong S. M., Yoon D. H., Kim S. Y., Kim C. S., Lee K. R., J. Nat. Prod., 85, 917–926 (2022).
• 31)  Li M., Zhang Y., Han T., Guan L., Fan D., Wu Q., Liu J., Xu Y., Fan Y., Phytochemistry, 205, 113487 (2023).
• 32)  Getoff N., Ritter A., Schwörer F., Bayer P., Radiat. Phys. Chem., 41, 797–801 (1993).
• 33)  Simic M., Neta P., Hayon E., J. Phys. Chem., 73, 3794–3800 (1969).
• 34)  Jo C., Yoon K. Y., Jang E. J., Kim T. H., Biosci. Biotechnol. Biochem., 80, 2022–2024 (2016).
• 35)  Kozlowski D., Marsal P., Steel M., Mokrini R., Duroux J. L., Lazzaroni R., Trouillas P., Radiat. Res., 168, 243–252 (2007).
• 36)  Marfak A., Trouillas P., Allais D. P., Calliste C. A., Cook-Moreau J., Duroux J. L., Biochim. Biophys. Acta, Gen. Subj., 1670, 28–39 (2004).
• 37)  Muiva L. M., Yenesew A., Derese S., Heydenreich M., Peter M. G., Akala H. M., Eyase F., Waters N. C., Mutai C., Keriko J. M., Walsh D., Phytochem. Lett., 2, 99–102 (2009).
• 38)  Chen J., Zhu G., Wu J. Acta Chimi. Sin., 81, 1609 (2023).
• 39)  Getoff N., Ritter A., Schworer F., J. Chem. Phys., 4, 503–510 (1983).
• 40)  Johnson R. W., J. Anim. Sci., 75, 1244–1255 (1997).
• 41)  Devary Y., Rosette C., DiDonato J. A., Karin M., Science, 261, 1442–1445 (1993).
• 42)  Kang J. S., Jeon Y. J., Park S. K., Yang K. H., Kim H. M., Biochem. Pharmacol., 67, 175–181 (2004).
• 43)  Youn C. K., Park S. J., Lee M. Y., Cha M. J., Kim O. H., You H. J., Chang I. Y., Yoon S. P., Jeon Y. J., Biomol. Ther. (Seoul), 21, 258–263 (2013).
• 44)  Stephens P. J., Harada N., Chirality, 22, 229–233 (2010).
• 45)  Nugroho A. E., Morita H., J. Nat. Med., 68, 1–10 (2014).
• 46)  Morgan D. M., Polyamine Protocols, 79, 184 (1998).
• 47)  Dirsch V. M., Stuppner H., Vollmar A. M., Planta Med., 64, 423–426 (1998).
• 48)  duPont N. C., Wang K., Wadhwa P. D., Culhane J. F., Nelson E. L., J. Reprod. Immunol., 66, 175–191 (2005).