Urolithin-A (3,8-dihydroxy-6H-benzo[c]chromen-6-one; UroA) produced by Tocris Bioscience (Bristol, United Kingdom) Witarix MCT (triglyceride of saturated vegetable oil-derived caprylic and capric fatty acids, oil phase, OF) from IOI Oleo GmbH (Hamburg, Germany) were employed. PEG 400 (polyethylene glycol 400), glycerol (GL), and polysorbate 80 (polyoxyethylene sorbitan monooleate, tween 80, TW80) were bought from Merck Life Science S.r.l. (Milan, Italy). Solvents were of HPLC grade, and all other chemicals were of analytical grade.

METHOD 1: The components of the dispersed phase (typically, to obtain 10 g of S-EM: OF 0.2 g, TW80 0.2 g, and ethanol 0.6 g) were weighed in a vial and subjected to magnetic stirring at 750 r.p.m. for 2–3 min (IKA RCT basic, IKA®-Werke GmbH and Co. KG, Staufen, Germany) to homogenize the mixture. The aqueous dispersing phase, typically composed of water 4.5 g and GL 4.5 g, was added dropwise to the dispersed phase (1:9, v/v) and magnetically stirred at 1,275 rpm for 30 min.

METHOD 2: The components of the dispersed phase (typically, to obtain 10 g of S-EM: OF 0.2 g, TW80 0.2 g, and ethanol 0.6 g) were weighed in a vial and subjected to magnetic stirring at 750 r.p.m. for 2–3 min to homogenize the mixture. The aqueous dispersing phase, typically composed of water 4.5 g and GL 4.5 g, was added dropwise to the dispersed phase (1:9, v/v) and magnetically stirred at 1,275 rpm for 30 min. Afterwards the emulsion was subjected to homogenization (ULTRA-TURRAX T 10 BASIC, IKA WERKE) at 25,000 rpm for 2 min.

In both cases the formulations were then stored in the dark at room temperature. For the formulative study A total of 38 different S-EMs were prepared in triplicate. Each formulation was prepared alternately with method 1 or method 2. Thus, 19 different compositions and two methods were tested. 15 compositions were based on OF, TW80, and EtOH, and 4 with the addition of glycerol to the aqueous phase.

One UroA-loaded S-EM was prepared adding the drug directly to the dispersed phase and mixing magnetically until complete dissolution, followed by the protocol described for method 2. The final UroA concentration was 0.2 mg/mL. The formulation was then stored in the dark at 4 °C.

Morphological analysis of S-EM was performed using TEM. The sample was negatively stained, depositing a sample drop on a grid covered with a collodion film. The excess drop was removed after 5 min from the grid with filter paper to keep a light veil of sample on the supporting substrate. A drop of 1% phosphotungstic acid was placed on the grid for 5 min and then removed with filter paper to surround the nanosystems deposited on the grid and adhere to their surface. Then the grid was observed with a Talos L120 G2 Trasmission Electron Microscope (Thermo Fisher Scientific Eindhoven-NL).

The size distribution of S-EM was evaluated measuring the values of Z-Average, i.e., the mean hydrodynamic diameter, and the polydispersity indexes (PI) by PCS (Zetasizer Nano-S90, Malvern Instr., Malvern, England). The instrument is equipped with a 5 mW helium neon laser, producing a wavelength output of 633 nm. The analyses were conducted at 25 °C with a 90° detection angle and 120 s equilibration time. The samples were diluted 1:10 (v/v) with bi-distilled water. Size distribution was obtained by the “CONTIN” method. Each analysis was carried out in triplicate from which the mean was derived.

Zeta potential (Z-potential) was evaluated using Zetasizer Pro (Malvern Instr., Malvern, England). This technique involves the measurement of scattering of incident laser light by moving particles according to their electrophoretic mobility. The instrument is equipped with a 4 mW helium neon laser that produces a wavelength output of 632.28 nm. The analyses were conducted at 25 °C. The samples were diluted 1:10 (v/v) with bi-distilled water inside vials and then put (1 mL) inside cuvette DTS1070 (Malvern Instr., Malvern, England).

The formulation pH was measured using Crison Basic C20 pH meter (Crison Instruments,S.A., Alella, Barcellona, Spain) equipped with a pH 50 12 - Hach electrode.

The syringeability of S-EM was manually tested 24 h after preparation using 1-mL plastic syringe (plunger diameter 2 mm) equipped with a 26-G needle.

SAXS experiments were performed at the Austro-SAXS beamline of Elettra Synchrotron (Trieste, Italy) [27]. Samples were put in a quartz capillary with a diameter of 1.5 mm, thermostated within a KPR (Peltier heating/cooling) sample holder (Anton Paar, Graz, Austria). Experiments were performed at 25 and 37 °C using exposure times of 10 s, with 18 repeated frames acquired for each sample to minimize potential radiation damage. The obtained two-dimensional (2D) data were corrected for background and detector efficiency and then radially averaged to derive the scattered intensity I(Q) as a function of the modulus of the scattering vector, Q = 4π sinθ/λ (where 2θ is the scattering angle and λ the X-ray wavelength, equal to 0.154 nm). The used sample-to-detector distance provided a Q-range between 0.1 and 5 nm^-1. The detector used for the measurement was a 2D Pilatus3 1M Detector System with a pixel size of 172 × 172 μm².

The empty S-EMs, the UroA-loaded S-EM, and their components (GL, OF, TW80, and UroA) were analyzed by FT-IR spectroscopy. The analyses were performed by a BRUKER VERTEX 70 spectrophotometer (4000-400 cm⁻¹) (Milano, Italy) using the diffuse reflectance technique on a solid KBr sample. To avoid altering the structure of the formulations under examination and maintain their hydration, the samples were prepared by directly depositing 10 µL in 100 mg of anhydrous KBr and mixing with a mortar.

The amount of UroA effectively associated with the nanodroplets was evaluated by an ultrafiltration procedure. Briefly, 500 µL of S-EM were loaded into Microcon centrifugal filter devices equipped with YM-10 membranes (NMWCO 3 kDa, Sigma-Aldrich, St. Louis, MO, USA) and centrifuged in a Spectrafuge™ 24D Digital Microcentrifuge (Infitek Inc., Spokane, WA, USA) at 4000 rpm for 20 min.

The S-EM nanodroplet-retained fraction was then 10-fold diluted with ethanol, magnetically stirred for 30 min (IKA RCT basic, IKA®-Werke GmbH and Co. KG, Staufen, Germany), filtered through 0.22 μm nylon membranes (Whatman, Germany), and subsequently subjected to High-Performance Liquid Chromatography analysis (HPLC), according to the below reported method. In parallel, a reference aliquot (100 µL) of the whole S-EM was processed under the same conditions to determine the total drug content.

The entrapment efficiency (EE) was calculated according to the relationship reported in Equation 1: EE = D / T D × 100 (1) where D corresponds to the amount of UroA retained within the nanodroplets, and T_D is the total UroA content in the formulation.

The HPLC analyses were performed using Perkin Elmer Series 200 HPLC systems (PerkinElmer, Waltham, MA, USA), equipped with a micropump, an auto sampler, and an UV detector. A stainless-steel C-18 reverse-phase column (15 × 0.46 cm) packed with 5 µm particles (Hypersil BDSC18 Thermo Fisher Scientific S.p.A., Milan, Italy) was eluted at a flow rate of 1 mL/min with a mobile phase composed of 30% acetonitrile, 70% water, and 0.1% TFA; the wavelength for UroA was 305 nm. The mobile phase flow rate was maintained at 1 mL/min, and at this flow rate, the retention time of UroA was approximately 5.5 min. The quantitative analysis was performed by constructing a calibration curve based on a solution of UroA (4 mg/mL) in acetonitrile with 10% dimethyl sulfoxide (DMSO), and subsequently diluted in acetonitrile in various ratios. The straight line obtained with the equation y = 4939.4 x + 3.2856 showed a correlation coefficient R = 0.9991.

The in vitro release of UroA from S-EM was determined by dialysis method. Briefly, 1 mL of S-EM UroA (0.2 mg/g) and a UroA suspension in DMSO/saline 1:10 (v/v) at the same concentration (SUSP UroA) were injected into the dialysis tube (Pur-A-Lyzer™ Maxi Dialysis Tube 0.1–3 mL, MWCO, 6–8 kDa, Merck, Milan, Italy), previously filled with 1 mL of bidistilled water. The dialysis tube was then placed into 200 mL of receiving phase constituted of 0.9% NaCl and 1% TW80 maintained under constant magnetic stirring at 37 ± 1 °C. At predetermined time intervals, 1 mL of the receiving phase was then sampled and subjected to HPLC analysis to determine the amount of active ingredient released over time. Each sample was replaced with an equal volume of receiving phase. UroA released amounts at each predetermined time were determined five times in independent experiments, calculating the mean values ± s.d. The cumulative percentage of drug released with respect to the total amount of drug loaded in the formulation was plotted against the time to obtain the release profile. Statistical comparison was evaluated by a Paired samples t-Student test considering 5 different experiments (n = 5 per group) comparing two different samples, making the paired test appropriate.

To gain insight into the release mechanism, the fitting of the experimental release data was analyzed using four different mathematical release models: zero-order kinetics (cumulative % drug released vs. time), first-order kinetics (log cumulative % drug remaining vs. time), Higuchi and Peppas (log cumulative % drug released vs. log time) models. The suitability of the model fits was verified using the DDsolver add-in for Excel 2016 (Version 2312 Build 16.0.17126.20126), using the correlation coefficient (R²) as indicator of the best fitting, for each of the considered models.

A stability study was carried out over a period of 3 months, evaluating once a month size distribution and Z-potential of S-EM UroA stored at 4 °C and protected from light (n = 4). In addition, the chemical stability of UroA was evaluated by HPLC analysis in S-EM samples stored in the same conditions (n = 4).

The cytotoxicity study was performed on primary dermal human fibroblasts obtained from 3-mm skin biopsies taken from healthy donors [28], grown in Dulbecco’s modified Eagle’s medium (DMEM, Lonza, Milan, Italy) low glucose (1.0 g/L) supplemented with 10% foetal bovine serum (FBS, EuroClone, Milan, Italy), 1% L-glutamine (Lonza, Milan, Italy), and 1% penicillin-streptomycin (Pen-Strep, Lonza, Milan, Italy). The cells were incubated at 37 °C in 95% air and 5% CO₂ until confluence.

To determine the viability of human fibroblasts, the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) cell viability assay, was used. Namely, S-EM, S-EM UroA 0.2 mg/L, SUSP UroA, and DMSO/saline (1:10 v/v) were evaluated in two independent experiments performed in triplicate. Fibroblasts were plated in 96-well plates at a density of 2 × 10⁴ cells/well in 200 μL of medium. Twenty-four h after seeding, the cells were treated with increasing concentrations from 1 to 200 μM of the above formulations for 24 h. The medium containing the treatments was then discarded, while 110 µL of 0.5 mg/mL MTT solution was added to each well. The plate was then incubated for 4 h at 37 °C in a dark incubator. The MTT solution was then removed, taking care not to remove the formazan crystals that had formed, which were dissolved by adding 100 µL of DMSO to the well. The plate was then placed at 37 °C for 15 min, in the dark. After 5 min of shaking, the absorbance of the solution was measured using a spectrophotometer at 590 nm, using 670 nm as the reference wavelength. This was then converted into a percentage of viability compared to the control cells.

Curve fitting and statistical analysis were performed using GraphPad Prism software, version 8 (GraphPad Software Inc., San Diego, CA). Data were fitted by nonlinear regression using a four-parameter logistic (4PL) model with a variable slope, constraining the bottom to values greater than 0.0 to avoid negative viability estimated on log-transformed x-axes. The statistical significance of concentration-dependent effects was assessed by one-way ANOVA, followed by Dunnett’s post hoc test for multiple comparisons and statistical significance was indicated by a two-tailed p-value 0.05. All data are presented as mean ± standard deviation (s.d.).

The anti-inflammatory activity of SUSP UroA and S-EM UroA was compared using the carrageenan-induced paw edema model in male C57BL/6 mice (30–35 g body weight). Animals were housed under standard laboratory conditions (12:12 h light–dark cycle, 22 ± 1 °C, 50–60% relative humidity) with ad libitum access to food and water.

Mice were randomly assigned to groups of six and treated (intraperitoneally; i.p.) 2 h prior to carrageenan injection with one of the following: DMSO/saline (1:10 v/v), SUSP UroA 2 mg/kg, SUSP UroA 4 mg/kg, S-EM, or S-EM loaded with UroA at 2 or 4 mg/kg. Edema was induced via intraplantar injection (30 μL) of λ-carrageenan (15 mg/mL; Sigma Chemical Co.) into the right hind paw, under anesthesia (1.5% mixture of isoflurane and air). The contralateral paw received 30 μL of saline and served as the negative control, showing no measurable edema.

Paw volume was measured using a plethysmometer (Ugo Basile, Comerio, Italy) before carrageenan injection (baseline) and at 3-, 6-, and 24-h post-injection. Edema was expressed as the increase in paw volume at each time point relative to the baseline measurement.

Edema volumes were expressed as mean ± standard error of the mean (SEM) for each group (n = 6) and analyzed using one-way analysis of variance (ANOVA), followed by Tukey’s test for multiple comparisons.

All in vivo experiments were performed in accordance with the European Communities Council Directive of September 2010 (2010/63/EU), as verified by the Animal Welfare Committee (OPBA) of the University of Ferrara. The experimental procedures have been approved by the Italian Ministry of Health. Adequate measures were taken to minimize the number of animals used and animal pain and discomfort.

A preformulation study was conducted with the goal of obtaining a homogeneous disperse system constituted of biocompatible excipients suitable for UroA solubilization.

As reported in Table 1, for the disperse phase of S-EM the following components were employed: a biocompatible oil phase constituted of a blend of capric and caprylic acid triglycerides (40:60, v/v) (OF), the surfactant TW80 as an emulsifier, and EtOH as a cosurfactant and a cosolvent, to stabilize the system and improve drug solubilization [29, 30].

As a first approach, S-EMs were produced using Method (1), maintaining constant the volume of the aqueous dispersing phase, while varying the percentages of the dispersed phase components, to evaluate the effect of the composition on the macroscopic appearance of S-EMs (homogeneity and absence of phase separation).

As reported in Table 1, S-EM 1, S-EM 2, S-EM 5, S-EM 6, S-EM 9, and S-EM 13 appeared macroscopically homogeneous, while phase separation phenomena due to creaming of the oil phase were observed in the case of the other S-EMs.

The size distribution of S-EMs was evaluated to study the effect of composition on both droplet mean diameter, expressed as Z-Average (Z-AV), and PI values (Table 1). Z-AV of S-EMs was inversely proportional to the surfactant content. Specifically, TW80 content ≥ 7%, w/w led to Z-AV ≤ 200 nm, with the smallest droplets in the case of TW80 8%, w/w (Z-AV 13.81 nm). Conversely, TW80 ≤ 3%, w/w resulted in droplet Z-AV ≥ 400 nm.

Both EtOH and OF amounts negatively affected droplet size distribution, respectively increasing droplet diameter and PI. Specifically, EtOH ≥ 5% resulted in droplets with Z-AV > 400 nm, while OF content ≥ 4% w/w resulted in PI values ≥ 0.35, indicating an inhomogeneous size distribution. The results reported in Table 1 suggested that (i) an OF content ≥ 3% w/w hampered the production of macroscopically stable systems; (ii) an OF content 1–3% allowed to produce homogeneous S-EMs only in the presence of TW80 ≥ 7% w/w; (iii) an OF content ≤ 1%, and TW80 4% w/w led to stable and homogeneous S-EMs; (iv) TW80 ≤ 4% prevented to obtain homogeneous S-EMs.

Notably, to ensure safety for intraperitoneal administration in mice, the concentration of TW80 should not exceed 2%, as established by prior biocompatibility studies [31]. Since at this TW80 concentration, the S-EM exhibited phase separation, to address this instability without exceeding the TW80 limit, a homogenization step (Method 2) was introduced as an alternative solution, using the same S-EM compositions employed by Method (1). Nonetheless, the homogenization step did not improve the aspect of S-EMs, in some cases worsening the homogeneity, possibly due to aggregation phenomena (Table 1).

Particularly, in the case of S-EMs containing 2% TW80, despite a reduction of Z-AV, creaming phenomena were still observed (S-EM 4, S-EM 7).

Therefore, as a second attempt to stabilize the ultrastructure of S-EMs, the addition of glycerol (GL) to the composition was evaluated. Notably, considering the UroA high lipophilicity (logP 2.3), to facilitate its solubilization, the highest OF and EtOH concentrations were maintained. Namely, based on the size distribution values reported in Table 2, the composition of S-EM 7 (2% OF, 2% TW80, 6% EtOH), characterized by a mean diameter ≤ 500 nm and a PI ≤ 0.2, was taken as reference.

GL containing S-EMs were prepared by Methods (1) and (2), previously mixing GL with water, according to the percentages reported in Table 2.

As reported in Table 2, while the addition of GL caused a significant increase in the droplet diameters, ranging between 540 nm and 860 nm for S-EMs obtained by Method (1), the homogenization step (Method 2) enabled to reduce droplet diameter and to improve size distribution. Particularly, S-EM 18, with GL in a 1:1 weight ratio with water, underwent a 1.7-fold Z-AV decrease, and achieved a homogeneous aspect.

To further reduce the droplet size, (i) the homogenization step was prolonged from 2 to 4 min; (ii) PEG 400 (2–4%, w/w) was used both as an alternative to GL, or in addition to GL. Although both strategies led to apparently homogeneous S-EMs, with Z-AV 400–440 nm and PI 0.1-0.19, the formulations underwent phase separation phenomena after 15 days of storage.

Based on homogeneity and size distribution (Z-AV < 500 nm, and PI < 0.2), S-EM 18 prepared by Method 2, hereafter named S-EM, was selected for the solubilization of UroA (S-EM UroA).

UroA was loaded in S-EM at a concentration of 0.2 mg/mL. The presence of UroA did not affect the homogeneous aspect of S-EM.

Three replicates were prepared and analyzed the following day in terms of appearance, size distribution, Z-potential and pH (Supplementary Table S1).

The results of the PCS analyses revealed a homogeneous size distribution, evidenced by the low PI values. The presence of UroA did not affect the droplet size, while it increased the absolute value of the zeta potential, keeping it negative. In each case, the pH value was weakly acidic, thus compatible with intraperitoneal administration, for which a pH between 5 and 9 is appropriate [32].

To gain insights into the morphology of the produced S-EM, visualization was carried out using TEM (Supplementary Figure S1). The images evidence a disperse phase consisting of spherical droplets with diameters in agreement with PCS data.

The assessment of UroA content in the whole S-EM and of its EE in the dispersed phase was evaluated the day after preparation, by ultrafiltration, disaggregation and HPLC. As expected, the simple S-EM preparation procedure avoided possible drug degradation, resulting in a total recovery of UroA (0.2 mg/mL, 100% with respect to the weight of the drug). Notably, an almost complete UroA association within the nanodroplets was achieved (EE = 98.3%, w/w).

Both S-EM and S-EM UroA were syringeable, thus suitable for parenteral administration through a 26-G needle.

To investigate the effect of UroA on the structural organization of S-EM, and to further validate PCS results, SAXS experiments were performed. The profiles observed for S-EM and UroA-loaded S-EM are shown in Figure 1 (panels A and B, respectively). For each sample, experiments were performed at 25 and 37 °C, to mimic conditions relevant to formulation storage and systemic administration.

SAXS profiles are very similar to those reported by Cruz et al [33] confirming the presence of nanostructures, in which TW80 is surrounding the oil core. The absence of any peak around 0.1 nm^-1 demonstrates the homogeneity of the internal structure of the nanodroplets.

To further characterize S-EM, a study was conducted via FT-IR spectroscopy analysis. Figure 2 shows the FT-IR spectra of the excipients (A-C), UroA (D), S-EM (E), and S-EM UroA (F). It is evident that the FT-IR spectrum profiles of S-EM (E,F) are strongly influenced by the high concentration of GL (45%, w/w) in the formulation.

In the S-EM FT-IR spectrum (E), a broad band centred at 3,350 cm⁻¹ is observed, characteristic of O–H stretching, along with a strong absorption at 2,935–2,881 cm⁻¹ due to the stretching of aliphatic C–H bonds. A medium-intensity band at 1,745 cm⁻¹, not present in GL spectrum, is typical of the C=O stretching of ester functional groups, characteristic of both OF and TW80. At 1,650 cm⁻¹ the bending vibration of water is still visible, due to the natural hygroscopicity of GL. The bands at 1,110 cm⁻¹ and 1,040 cm⁻¹, partially overlapping, are attributable to the C–O stretching of alcoholic, ether, and ester groups present. Notwithstanding the low amount of UroA did not substantially alter the spectral profile of S-EM UroA (F), a weak shoulder at 1,706 cm⁻¹ can be observed, corresponding to the intense C=O stretching band of the UroA lactone, whilst the other characteristic absorptions of the molecule are instead masked by the bands mainly attributable to GL.

To evaluate the stability of S-EM UroA, PCS analyses and Z-potential evaluations were repeated 3 months after S-EM preparation. In addition, the ability of the formulation to control UroA degradation was assessed by quantifying the drug concentration over time. All formulations were stored at 4 °C and protected from light.

As reported in Supplementary Table S2, the presence of UroA did not affect the size stability of the system, as no significant changes in Z-AV and PI were recorded after 90 days. The slight variation in Z potential was compatible with the stability range for dispersed systems (≥± 25 mV). Notably, the decrease of UroA in 90 days was around 10% with respect to day one, suggesting the capability of S-EM to control UroA degradation.

The release of UroA was determined in vitro via dialysis method, comparing S-EM UroA with SUSP UroA. As receiving phase, 0.9% NaCl and 1% TW80 was employed, maintained under constant magnetic stirring at 37 ± 1 °C. The release profiles, obtained plotted as cumulative % of drug released vs. time, are reported in Figure 3.

The release kinetics of UroA exhibited a biphasic profile, characterized by a faster release within 2 and 7 h, for S-EM UroA and SUSP UroA respectively, followed by a gradually slower drug release phase. Notably, the UroA release kinetic in the linear tract (0–7 h) was 2.35-fold faster in the case of S-EM UroA with respect to SUSP UroA, while the percentage of UroA released at 96 h was 1.7-fold higher in the case of S-EM UroA than SUSP UroA. The statistical comparison evidenced a t-value −5.3, and a p-value < 0.0001, indicating a very high statistical significance.

Zero-order plot, first-order plot, Higuchi plot and Peppas plot were applied to gain insight on the mechanism of UroA release. Supplementary Table S3 shows the R² values obtained plotting the release profile according to the different mathematical models [34]. For both S-EM UroA and SUSP UroA the profile can be better described by Higuchi model, while the “n” values obtained from the Peppas plot revealed a non-Fickian diffusion mechanism. Thus, this mechanistic evaluation indicates diffusion as the primary driver. However, the Peppas model, being more sensitive to subtle changes in drug release kinetics, revealed a not purely Fickian diffusion, suggesting the involvement of additional factors that slightly modulate UroA release rate as a function of time.

Figure 4 shows the results of the MTT test performed on fibroblasts treated with S-EM, S-EM UroA, SUSP UroA, and DMSO/saline 1:10 (v/v) at the concentration range 1–200 μm. As expected, for all treatments the cytotoxicity increased in a concentration-dependent manner. Specifically, the most toxic concentrations were 100 and 200 µM for both S-EM (unloaded and loaded with UroA) and for the simple SUSP UroA. For concentrations up to 5 μM, S-EM UroA maintained 100% viability. Interestingly, comparing IC₅₀ values of SUSP UroA and S-EM UroA, we did not observe significant differences (IC₅₀ = 35.53 µM and 36.61 µM respectively), confirming that S-EM Uro A did not induce cell cytotoxicity in the concentration range 1–50 µM. This result was also confirmed by the high value of IC₅₀ for fibroblasts treated with unloaded S-EM (IC₅₀ = 71.14 µM).

As expected, intraplantar (i.pl.) injection of carrageenan induced a marked paw edema, observable at 3- and 6-h post-injection (Figures 5A,B), which moderately persisted at 24 h (Figure 5C). In contrast, the contralateral paw injected i.p. with saline showed no measurable edema.

Pretreatment with DMSO/saline (1:10 v/v), SUSP UroA 2 mg/kg, SUSP UroA 4 mg/kg or S-EM had no effect on carrageenan-induced paw edema at any time point (Figure 5). On the contrary, pretreatment with S-EM UroA at 2 mg/kg showed a trend toward edema reduction at 3 and 6 h (Figures 5A,B), with a statistically significant effect observed at 24 h post-injection (Figure 5C). Notably, S-EM UroA at 4 mg/kg significantly reduced paw edema at all measured time points, indicating a dose-dependent anti-inflammatory effect (Figure 5).