Anti-inflammatory Effect of Curcuma longa and Allium hookeri Co-treatment via NF-κB and COX-2 Pathways

Although inflammation is a host defense mechanism, chronic inflammation mediates several diseases, including cancer, allergy, asthma, and autoimmune diseases, and reportedly, it is associated with a 60% mortality rate. There are several reports on the anti-inflammatory effects of Curcuma longa and Allium hookeri. However, although they can be used as culinary materials and have biological effects, they are not effective anti-inflammatory agents. In this study, we evaluated the synergic effect of C. longa and A. hookeri in order to confirm the possibility of a new anti-inflammatory agent. Based on cell viability and cytokine analyses, the appropriate ratio of C. longa and A. hookeri was confirmed using an air pouch animal model. Then, the anti-inflammatory effect of C. longa and A. hookeri co-treatment was evaluated by measuring the immune cell count and cytokines in the exudate and by comparing the morphological changes and cytokines in inflamed skin samples. Additionally, we evaluated the NF-κB/COX-2 pathway and iNOS levels. The active constituents detected in C. longa were demethoxycurcumin and bisdemethoxycurcumin, and that detected in A. hookeri was methylsulfonylmethane. An in vitro assessment determined the appropriate drug ratio as 3:7. In a carrageenan-induced inflammatory model, co-treatment effectively suppressed inflammatory cytokines, including IFN-γ, IL-1β, IL-6, IL-13, and IL-17, and recovered inflammation-related morphological changes in the skin. The anti-inflammatory effect of the co-treatment was mediated through the NF-κB/COX-2 pathway and iNOS inhibition. We concluded that co-treatment with C. longa and A. hookeri synergistically inhibited inflammation via the NF-κB/COX-2/iNOS pathway.

The ratio of C. longa vs. A. hookeri suppressing LPS-induced RAW 264.7 cell proliferation was found to be 3:7. In order to determine the effective ratio of C. longa and A. hookeri as well as curcumin alone (10 μg/mL) against LPS treatment-induced RAW 264.7 cell proliferation, serial concentrations of C. longa were used as shown in Fig. 2A. Depending on the increment of the relative concentration of A. hookeri, the cell viability was similar to that of the control group (100%). In order to evaluate the synergic effect of C. longa and A. hookeri, curcumin, the active constituent in C. longa, was analyzed. However, as the percentage of whole curcumins (includes curcumin metabolites) in C. longa was 0.34% (w/w) and 0.85 μg/mL whole curcumins existed in 250 μg/mL 30% EtOH-extracted C. longa (Supplementary File), and taking into consideration C. longa and A. hookeri combination treatment, the contribution effect of curcumin alone was confirmed ( Fig. 2 and Supplementary File). At first, in order to investigate the applied dose of curcumin alone, cell viability was evaluated after treatment with curcumin alone (Supplementary File), and although whole curcumin existed at an amount of 0.85 μg/mL in C. longa alone, 10 μg/mL curcumin was used for the treatment, for confirmation of its anti-inflammatory effect. However, the effect differed significantly when the levels of pro-inflammatory cytokines were compared (Fig. 2B). The most effective ratios against each cytokine were variable: 7:3 against IFN-γ, 10:0 against IL-6, 3:7 against IL-13, and 3:7 Cell proliferation suppression and pro-inflammatory cytokine regulation depending on the ratio of C. longa to A. hookeri. (A) Including curcumin alone treatment, the ratios of C. longa to A. hookeri were 10:0, 7:3, 5:5, 3:7, and 0:10, and controlled LPS-induced cell proliferation comparable to non-LPS-treated cells. The ratios (7:3, 5:5, 3.7, and 0:10) suppressed cell proliferation and the minimum ratio of cellular proliferation inhibition was 7:3. (B) In most groups, except the curcumin alone treatment group and the 5:5 ratio, IFN-γ levels, increased by LPS treatment, were suppressed by C. longa and A. hookeri co-treatment. Although the 7:3 ratio was most effective in suppressing IFN-γ expression, the ratios 7:3, 3:7, and A. hookeri alone downregulated IFN-γ levels. Unlike A. hookeri, C. longa dose-dependently inhibits IL-6 levels. The C. longa alone treatment most significantly inhibited IL-6 expression, but the 7:3, 5:5, or 3:7 ratios effectively suppressed IL-6 expression. IL-13 levels decreased only by the 3:7 C. longa to A. hookeri. In all groups, TNF-α levels were suppressed, with the lowest level by the 3:7 ratio. All values are presented as mean ± standard deviation. CON vs. * p < 0.05; CON vs. ** p < 0.001 LPS vs. $ p < 0.05; LPS vs. $$ p <0.001; C. longa: A. hookeri = 10:0 vs. # p < 0.05; C. longa: A. hookeri = 10:0 vs. ## p < 0.001; C. longa: A. hookeri = 7:3 vs. & p < 0.05; C . longa: A. hookeri = 5:5 vs. @ p < 0.05; C. longa: A. hookeri = 3:7 vs. % p < 0.05.

C. longa and A. hookeri co-treatment restored immune-related white blood cells in the exudate and completely recovered carrageenan-induced altered morphology in the skin via pro-inflammatory cytokine regulation.
With the exception of neutrophils, C. longa and A. hookeri co-treatment downregulated white blood cells, eosinophils, lymphocytes, and monocytes in a dose-dependent manner (Fig. 3A). Although the population of neutrophils was not regulated in a dose-dependent manner, a downregulation was observed. To measure the anti-inflammatory effect of the co-treatment, skin morphology was analyzed (Fig. 3B). Carrageenan changed skin morphology by inducing inflammation, which resulted in a thin membrane, muscle condensation, downregulation of adipocytes, and deep wrinkles, in comparison to the normal group. However, MSM and co-treatment ameliorated the altered skin membrane morphology, with C. longa and A. hookeri co-treatment dose-dependently restoring the altered skin membrane. Carrageenan stimulated inflammation-related cytokines such as IFN-γ, TNF-α, IL-1β, IL-6, IL-10, IL-13, and IL-17. IFN-γ is a pro-inflammatory cytokine and its expression increases during inflammatory conditions (Fig. 3C,D). Inflammation-related cytokines in the exudate, increased by carrageenan treatment, were regulated by co-treatment (Fig. 3C). The levels of IFN-γ, IL-1β, IL-6, IL-13, and IL-17 were downregulated by co-treatment; however, those of TNF-α and IL-10 were not suppressed. Notably, the level of IFN-γ decreased to the control level following co-treatment. Similarly, carrageenan stimulated IFN-γ and IL-1β expression in the exudate. However, with high dose co-treatment (500 mg/kg), IFN-γ and IL-1β levels were reduced to levels similar to those in the control group. Co-treatment dose-dependently suppressed the levels of IL-6, IL-13, and IL-17, which were upregulated by carrageenan. IL-10 was not suppressed by the treatment. In the case of TNF-α, the results obtained for the exudate and skin differed. Although TNF-α level was not suppressed by co-treatment in the exudate, its expression was completely inhibited in the skin of the animals in the 500 mg/kg co-treatment group.

C. longa and A. hookeri co-treatment effectively controlled both carrageenan-induced inflammations by regulating the NF-κB-COX-2 pathway and iNOS level. Changes in NF-κB and
COX-2 were assessed using the immunofluorescent method (Fig. 4). The results were analyzed by the image analyzing program in the K1-Fluo confocal microscope (Fig. 4A). Carrageenan increased NF-κB levels in the nucleus (1.79 ± 0.182) compared to those in the control group (1.00 ± 0.125); however, MSM suppressed the carrageen-induced increase in NF-κB (1.26 ± 0.181) (Fig. 4B). The decrease in the level of NF-κB in the C. longa and A. hookeri co-treatment group (1.17 ± 0.045) was comparable with that observed in the MSM treatment group. In the carrageenan treatment group (2.12 ± 0.320), the cytoplasmic level of COX-2 significantly increased compared to that in the control group (1.00 ± 0.112). In the co-treatment group, the expression of COX-2 was minimal (1.14 ± 0.059). The level of NF-κB in the nucleus or cytoplasm was evaluated using western blotting (Fig. 4C,D), and the level of NF-κB in the nucleus significantly increased compared to that in the carrageenan treatment group both inthe MSM treatment group and in 500 mg/kg C. longa and A. hookeri co-treatment group, nuclear and cytoplasmic NF-κB levels decreased. In order to measure the anti-inflammatory effect and mechanism of co-treatment, serum iNOS levels were measured (Fig. 4E). Co-treatment dose-dependently suppressed the carrageenan-induced serum iNOS level. Particularly, in the 500 mg/kg co-treatment group, iNOS levels were lower than those observed in the MSM treatment group and similar to those in the control group.

Discussion
Inflammation is a homeostatic response in organisms against various conditions. However, when this response is excessive, it could result in conditions such as atherosclerosis, rheumatoid arthritis, and asthma [17][18][19] . Hence, there are several ongoing trials aimed at decreasing the occurrence of excessive inflammation. Several cytokines are related to the inflammatory process. Pro-inflammatory cytokines include IL-1β, IL-6, IL-13, and TNF-α 4 . IFN-γ is an important cytokine that stimulates cell-mediated immune reactions 20 and regulates inflammatory diseases as a pro-inflammatory cytokine. However, reports contrary to the induction of inflammation have been reported, and the yin and yang theory of IFN-γ has been published 21 . For example, during IFN-γ-mediated inflammation control, the downregulatory effect of IFN-γ was induced by an increase in programmed cell death 1 (PD-1) and PD-1 ligand 1 (PD-L1) in the rheumatoid synovium; 22 however, an upregulatory effect was induced by the antagonizing IL-17 function 23,24 . Additionally, TNF-α is released by macrophage and neutrophil activation induced by IFN-γ 25 . Thus, IFN-γ demonstrates a "double-edged sword" effect; it induces inflammatory diseases such as rheumatoid arthritis and inflammatory bowel disease 26,27 related to TNF receptor 1, and it is involved in homeostasis, for example, survival and regeneration of cells or tissues through TNF receptor 2 28 . IL-6 activation has a pivotal role in inflammation and inflammation-induced cell damage. Hence, IL-6 modulation has been used as a target for anti-inflammatory drug development, including drugs against rheumatoid arthritis and juvenile idiopathic arthritis 29 . IL-13 is a key cytokine in allergic and inflammatory diseases that induces an inflammatory cascade 30 . IL-13 is a drug target in several inflammatory diseases, including asthma, atopic dermatitis, and inflammatory bowel disease 31 .
C. longa has been used as a culinary material worldwide and curcumin is one of its well-known active components. It is commonly used as a spice, food additive, or dietary pigment. Curcumin has been known to have several pharmacological effects, including anti-inflammatory, antioxidant, and anti-cancer activities [34][35][36] . The molecular mechanisms of its effects are diverse, involving various signaling pathways (such as NF-κB and STAT3 signaling) 37 . Thus, curcumin and other C. longa components are promising inflammation-targeting (2020) 10:5718 | https://doi.org/10.1038/s41598-020-62749-7 www.nature.com/scientificreports www.nature.com/scientificreports/ phytochemicals. Sandur et al. 38 reported that curcumin, demethoxycurcumin, and bisdemethoxycurcumin are the active substances in C. longa that suppressed TNF-induced NF-κB activation. Their effects were found to be due to the methoxy groups on the phenyl ring.
A. hookeri has been cultivated for consumption in East Asia. MSM is an active constituent of A. hookeri, and known to have anti-inflammatory, anti-arthritic, anti-allergic, and anti-asthmatic effects 16  Inhibitory effect of co-treatment against blood cell proliferation in the exudate. Co-treatment with C. longa and A. hookeri dose-dependently inhibits white blood cells and monocyte proliferation, controls eosinophils levels similar to MSM treatment, completely decreases carrageenan treatment-upregulated lymphocyte levels, dose-dependently controls monocyte levels, and only demonstrates a tendency for neutrophil suppression with no statistical comparison. (B) Recovery effect of co-treatment on the air pouch membrane's morphological changes. (C) Co-treatment suppresses the expression of inflammatory-related cytokines such as IFN-γ, IL-1β, IL-6, IL-13, and IL-17, but not TNF-α. Co-treatment with ≥250 mg/kg C. longa and A. hookeri exerts a stronger inhibitory effect on IFN-γ levels than MSM treatment. Co-treatment dose-dependently suppresses IL-1β, IL-6, and IL-13 levels. However, no significant decrease in TNF-α and IL-10 expression was observed with co-treatment. In the case of IL-17, there is a dose-dependent decrease pattern, which is hard to obtain with cotreatment. (D) Significant inhibitory effects against inflammatory-related cytokines in the dermis such as IFNγ, TNF-α, IL-1β, IL-6, IL-13, and IL-17. Co-treatment dose-dependently suppresses the expression of cytokines such as IFN-γ, IL-1β, IL-10, and IL-17, and 500 mg/kg treatment completely inhibits the expression of IFN-γ, IL-10, and IL-17. In a dose-dependent manner, the levels of IL-1β, IL-6, and IL-13 are effectively downregulated. Co-treatment effectively inhibits the expression of TNF-α, with complete suppression observed with 250 mg/ kg and 500 mg/kg co-treatment. a, control; b, 1% carrageenan treatment; c, 25 mg/kg MSM treatment; d, 50 mg/kg C. longa and A. hookeri co-treatment; e, 250 mg/kg C. longa and A. hookeri co-treatment; f, 500 mg/ kg C. longa and A. hookeri co-treatment. Magnification, 200×. Scare bar, 100 μm. All values are presented as mean ± standard deviation. CON vs. ** p < 0.001; Carrageenan vs. $ p < 0.05; MSM vs. # p < 0.05.
reported that MSM inhibits nitric oxide prostaglandin 2 production via suppressing iNOS and COX-2 expression. Moreover, MSM strongly inhibits IL-6 and TNF-α production through the transcription factor NF-κ B 16 .
In the present study, we identified curcumin and its two derivatives in C. longa extract. Curcumin is a well-known bioactive ingredient, reported to have a synergistic effect with various bioactive components. Morgana et al. reported that curcumin and piperine increased TGF-β levels, significantly improved collagen repair, and decreased cellularity and activation of NF-ĸB in periodontal tissues 41 . Nishtha et al. found that curcumin and quercetin synergistically inhibit cancer cell proliferation and modulate Wnt signaling pathways in cancer cells 42 . Chen et al. conducted a study similar to ours. They elucidated the synergistic anti-inflammatory effects of a curcumin, tetramethylpyrazine, and resveratrol mixture 43 . Based on previous reports, curcumin and other C. longa extract substances exert synergistic effects in combination with A. hookeri extract. For future studies, the interaction between the main bioactive substances of C. longa and the single substance present in A. hookeri, should be investigated.
We previously reported that the anti-inflammatory efficacy of A. hookeri is via the regulation of pro-inflammatory cytokines such as IL-1β, IL-6, IL-13, and TNF-α 14 . In our previous study, we demonstrated that 300 mg/kg A. hookeri treatment suppressed carrageenan-induced inflammation, with notable downregulation of skin pro-inflammatory cytokine levels. However, in this study, we observed a synergistic effect of C. longa and A. hookeri co-treatment. Although the 50 mg/kg C. longa and A. hookeri co-treatment (3:7) consisted of only 35 mg/ kg A. hookeri extract, the levels of several inflammation-related cytokines, including IFN-γ, IL-1β, IL-6, IL-13, and IL-17, were suppressed (Fig. 3D). Specifically, 250 mg/kg co-treatment (175 mg/kg A. hookeri) significantly controlled TNF-α expression, and 500 mg/kg co-treatment (350 mg/kg A. hookeri) effectively suppressed IFN-γ and TNF-α levels. In the case of the exudate, 250 and 500 mg/kg co-treatment modulated IFN-γ levels to those observed in the control group, 500 mg/kg co-treatment effectively controlled IL-1β and IL-6 levels compared to those in the positive control (MSM), and IL-13 and IL-17 expression in the 500 mg/kg co-treatment were comparable to those observed with MSM treatment (Fig. 3C).
Based on the above results, we conclude that C. longa and A. hookeri co-treatment synergistically inhibit inflammation by regulating the NF-κB/COX-2/iNOS pathway.

Materials and Methods
Plant material preparation. Dried C. longa roots were obtained from the Chonnam Medical Herb-Agricultural Cooperation (Hwasun, Chonnam, South Korea). The extraction was performed 10 times with 30% ethanol at 80 °C for 1 h, and the extract was concentrated at 40 mmHg at 50 °C. The extract was pre-freezed in a deep freezer for 2 days as 2-L aliquots. Subsequently, after the extraction process, the extract was freeze-dried at 0.06 mbar at −70 °C for 48 h and stored at −50 °C. A. hookeri was supplied by the College of Pharmacy, Mokpo National University. The voucher specimen MNUCSS-SC-01 was recorded. Briefly, the root was separated for the study. Air-dried, powdered A. hookeri roots (1000 g) were extracted twice with 70% ethanol (4 L) at room temperature for 3 days. After filtration, ethanol evaporated, and the sample was freeze-dried and stored at 50 °C 14 .
The extraction was performed 10 times with either 70% ethanol (A. hookeri) or 30% ethanol (C. longa) at 80 °C for 1 h. The extract was concentrated at 40 mmHg at 50 °C. Pre-freezing of the high concentrated extract was performed in a deep freezer for 2 days by the 2-L aliquot respectively. Subsequently, after the extraction process, freeze-drying was carried out at 0.06 mbar at −70 °C for 48 h. The freeze-dried extract was stored at −20 °C with grinding and packing.

Instrumentation and chromatographic conditions for C. longa and A. hookeri. All analysis for C.
longa was performed using an Alliance 2695 HPLC system (Waters, Milford, MA, USA) equipped with a photodiode array detector 44 . An Agilent Zorbax extended C18 analytical column (5 µm, 150 mm × 5 mm) was used with a mobile phase consisting of a mixture of solvent A (acetonitrile) and B (water containing 0.2% phosphoric acid, pH 3.5). A gradient elution (20% A in B solvent ~ 100% A solvent) at a flow rate of 0.8 mL/min (Table 1) was employed. The detection wavelength was 450 nm. The solvent was filtered through a 0.22-µm filter and degassed. The sample injection volume was 10 µL. Agilent 7890 gas chromatography (GC) and Agilent 5975 quadrupole mass spectrometry (MS) system (Agilent Technologies, Palo Alto, CA, USA) were used to analyze molecular mass fragments of A. hookeri root 14 . The mass fragments were ionized in an Agilent DB-1 capillary column (30 m l. × 0.32 mm i.d., 0.25-µm film thickness) under electron ionization (EI) conditions. GC oven was thermally programmed as described in Table 2. All the scanned mass spectra were compared with the data system library (NIST 2017).

Evaluation of an appropriate ratio of C. longa and A. hookeri.
To confirm the appropriate ratio of C. longa and A. hookeri, we conducted two measurements. We assessed their suppression effect against lipopolysaccharide (LPS)-induced proliferation of RAW 264.7 cells (Korean Cell Line Bank, Seoul, Korea), a murine macrophage cell line, and measured the pro-inflammatory cytokine levels at different ratios.
To analyze the pro-inflammatory cytokine levels after each treatment, enzyme-linked immunoassays (ELISA) were conducted. To analyze the levels of IFN-γ, TNF-α, and IL-6 OptEIA, mouse ELISA kits were purchased from BD Biosciences. IL-1β, IL-10, IL-13, and IL-17 levels were assessed using mouse ELISA kits purchased from Thermo Fisher Scientific (MA, USA). All assays were conducted in accordance with the manufacturer's guidelines.
Carrageenan-induced air pouch model. The animal study was conducted twice using the same method.
In each study, 36 male ICR mice (20-25 g) were purchased from Samtako Korea (Osan, Korea) and acclimatized for 8 days. To induce a subcutaneous air pouch from the ninth day, 3 mL of air was injected three times into the intra-scapular area for 6 days, and all mice were classified into two categories; the control mice (CON) were not treated with carrageenan, while the others received treatment. The carrageenan-treated category consisted of five groups including the normal saline oral administration group (Carrageenan), 25 mg/kg methylsulfonylmethane (MSM)-treated group (used as an anti-inflammatory drug), 50 mg/kg C. longa and A. hookeri-treated group, 250 mg/kg C. longa and A. hookeri-treated group, and 500 mg/kg C. longa and A. hookeri-treated group. In the animal study, the appropriate anti-inflammatory ratio of C. longa vs. A. hookeri was confirmed as 3:7. Two hours after the above treatments, 1 mL carrageenan solution (2%) was injected into the air pouch. After 24 h of the carrageenan injection, all mice were sacrificed using Zoletil (tiletamine HCl and zolazepam HCl; Virbac, Carros, France) via intraperitoneal injection. To collect the exudate, the pouches were flushed with 2 mL of phosphate-buffered saline. The number of total and differential cells in the pouch exudate were counted using the Hemavet Multispecies Hematology System (Drew Scientific Inc., Waterbury, CT, USA).
Histopathological analysis. Histopathological measurement was conducted according to our previous study 14 . After exudate and blood collection, the skin tissues were retrieved, fixed in 10% (v/v) formaldehyde solution, dehydrated in a graded ethanol series (99.9%, 90%, 80%, and 70%), and embedded in paraffin. Paraffin-embedded skin tissues were then sectioned (5 µm) and stained with hematoxylin and eosin. Immunohistochemical analysis. Immunohistochemical analysis was performed according to our previous study 14 . Deparaffinized tissue sections were treated with 3% hydrogen peroxide in methanol for 10 min, to remove endogenous peroxidase. Antigen retrieval was performed with the sodium citrate buffer (0.1 M), using the boiling method. The slides were incubated with normal serum to prevent nonspecific binding and then incubated overnight at 4 °C with the following primary antibodies (diluted 1:100 or 1:200): IFN-γ (Santa Cruz, sc-74104), TNF-α (MY BioSource, CA, USA, MBS175453), IL-1β (Santa Cruz, sc-1251), IL-6 (Santa Cruz,
Western blotting. Each skin sample was prepared with 500 μL lysis buffer, centrifuged at 9,000 x g for 20 min, and the supernatant was collected. For NF-κB measurement, the nuclear or cytoplasmic proteins were separated by NE-PER Nuclear and Cytoplasmic Extraction Reagents (ThermoFisher Scientific, 78833), but COX-2 level was evaluated using the whole protein. For each group, 20 μg protein was diluted in SDS sample buffer, boiled, and electrophoresed on 10% acrylamide SDS-PAGE gels. Samples were transferred to nitrocellulose membranes and blocked with 5% skim milk overnight such that each protein in the sample could bind with primary antibodies in the 5% skim milk. After washing three times with Tris-buffered saline containing 0.1% Tween-20, the membranes were treated with peroxidase-conjugated secondary antibodies and visualized using ECL reagents. NF-κB (ThermoFisher Scientific, PA5-16545), COX-2 (Abcam, ab15191), GAPDH (Thermo Fisher Scientific, MA5-15738), and Peroxidase-conjugated Affinipure rabbit anti-mouse IgG (Jackson Immunoresearch, 315-035-003) were used. iNOS analysis. The animals were anesthetized and the exudate in their air pouches were collected. Blood samples were obtained from the heart and allowed to coagulate for 2 h at room temperature. Next, the samples were centrifuged at 1000×g for 15 min, and the serum iNOS levels were measured using a mouse iNOS ELISA kit (Mybiosource, MBS723353) according to the manufacturer's guidelines. Statistics. Results are expressed as mean ± standard deviation (SD). Group differences were evaluated using one-way analysis of variance, followed by Dunnett's multiple comparison test. p < 0.05 or p < 0.001 was considered statistically significant.