Phenolic Profiling and Biological Potential of Ficus curtipes Corner Leaves and Stem Bark: 5-Lipoxygenase Inhibition and Interference with NO Levels in LPS-Stimulated RAW 264.7 Macrophages

The economic value of fig trees has been globally acknowledged due to their utilization in the food industry, being also frequently used in traditional medicine. While ubiquitously distributed in Southeast Asia, Ficus curtipes Corner remains uninvestigated concerning its biological properties and chemical profile. HPLC-DAD-ESI/MSn characterization of methanol extracts obtained from the stem bark and leaves allowed the identification and quantitation of 21 phenolic compounds for the first time; the stem bark was predominantly rich in flavan-3-ols and apigenin derivatives, while solely apigenin-di-glycosides have been identified and quantitated on the leaf extract. Both extracts inhibited 5-lipoxygenase (5-LOX) activity in a concentration-dependent manner, the one obtained from the stem bark being significantly more active (IC50 = 10.75 μg/mL). The effect of both extracts on lipopolysaccharide (LPS)-stimulated RAW 264.7 macrophages was evaluated, and while the stem bark extract did not lead to a noticeable interference on nitric oxide (NO) levels, the extract obtained from the leaves notably decreased NO and L-citrulline levels at concentrations ranging from 250 to 500 μg/mL. Herein, F. curtipes is valorized due to its modulatory effects on inflammatory mediators and also as a source of bioactive phenols, which may fuel further studies on the development of nutraceuticals.


Introduction
Over the past few years, research on multi-target molecules, particularly those obtained from natural sources, has been gaining ground on the treatment and/or prevention of several multi-factorial diseases [1]. In fact, approximately 50% of the approved drugs in the past 30 years were derived either directly or indirectly from natural products, predominantly from medicinal plants, playing a unique role as sources of both simple and complex bioactives [2]. Systematic studies based on the (Department of Botany, Faculty of Science, Kasetsart University). A voucher specimen (PCERU_FC0005) was deposited at the Phyto-Chemodiversity and Ecology Research Unit, Faculty of Science, Kasetsart University.

Extraction
Leaves and stem barks were air-dried and ground to fine powder with a mean particle size ≤910 µm. A total of 51.1 and 71.8 g of leaves and stem bark powdered material, respectively, were macerated in 1 L of methanol for 7 days. The resulting extracts were filtered through Whatman ® grade 1 filtration paper (Sigma-Aldrich, St. Louis, MO, USA) using a Büchner funnel, and concentrated to dryness under reduced pressure in a Rotavapor ® R-210 (Büchi, Mumbai, India), yielding 4.8 and 4.0 g of extracts for leaves and stem bark, respectively. The yields of extraction with methanol from the leaves and stem bark were 9.4% and 5.6%, respectively.

HPLC-DAD-ESI (Ion Trap)/MS n Qualitative Analysis
Phenolic profiles were obtained under the same chromatographic conditions reported by Ferreres et al. [12], with minor modifications on the gradient. Elution was performed with water-formic acid (1%) (A) and acetonitrile (B) as mobile phase, starting with 5% B and using a gradient to obtain 30% B at 20 min and 50% B at 30 min. The flow rate was 800 µL/min and the injection volume 20 µL. Chromatograms were recorded at 280 and 340 nm, and the full scan mass covered the range from m/z 100 up to 1500.

UPLC-ESI-QTOF-MS Qualitative Analysis
Determination of the exact mass was carried out using an Agilent 1290 Infinity LC system coupled to the 6550 Accurate-Mass QTOF (Agilent Technologies, Waldbronn, Germany) with an electrospray interface (Jet Stream Technology). Samples (2 µL) were injected onto a reverse phase Kinetex column (1.7 µm, C18, 100 Å, 50 × 2.1 mm; Phenomenex, Macclesfield, UK) with a SecurityGuard ULTRA Cartridges of the same material operating at 30 • C and a flow rate of 0.5 mL min −1 . The mobile phase used was a mixture of acidified water (0.1% formic acid) (A) and acidified acetonitrile (0.1% formic acid) (B). Compounds were separated using the following gradient conditions: 0 min 5% B, obtaining 30% B at 12 min and 50% B at 15 min. The optimal conditions for the electrospray interface were the following: gas temperature 280 • C, drying gas 11 L min −1 , nebulizer pressure 45 psi, sheath gas temperature 400 • C, sheath gas flow 12 L min −1 . The MS system was operated in negative ion mode with the mass range set at m/z 100-1500 in full scan resolution mode. Further conditions as described in Garcia et al. [13].

HPLC-DAD Quantitative Analysis
The gradient and mobile phases were the same as above described. Dried methanol extracts obtained from the stem bark (150 mg/mL) and the leaves (70 mg/mL) of F. curtipes were dissolved in methanol and filtered through a 0.45 µM pore size membrane (Millipore, Bedford, USA) before injection (20 µL). Samples were analyzed in triplicate on an analytical HPLC unit (Gilson Medical Electronics, Villers le Bel, France), using a RP-Kinetex C18 (150 × 4.6 mm, 5 µm particle size, 100 Å pore size) column (Phenomenex, USA). Detection was achieved with an Agilent 1260 series diode array detector (Agilent Technologies, Waldbronn, Germany). Spectral data were collected in the range of 200-700 nm, chromatograms being recorded at 280 and 340 nm. Data were processed on Clarity software system, version 5.04.158 (DataApex, Ltd., Prague, Czech Republic).

5-LOX Inhibition
5-LOX inhibition was evaluated by following the linoleic acid oxidation at 234 nm, according to a previously reported procedure [14]. Equal volumes of extract solution (20 µL) and 5-LOX (EC 1.13.11.12) from Glycine max (100 U/mL) were pre-incubated with 200 µL of phosphate buffer (0.1 M, pH 9), at room temperature for 5 min, on a 96 well-plate. The reaction was started by the addition of 20 µL of linolenic acid (4.18 mM in ethanol) and followed for 3 min at 234 nm (Multiskan ASCENT, Massachusetts, MA, USA). Results correspond to the mean ± SEM of three independent assays, each performed in triplicate. Quercetin was used as positive control.

Cell Viability
Cell viability was evaluated by monitoring the cell mitochondrial activity, using the MTT reduction assay, as previously reported [15]. Briefly, cells were cultured in 96-well plates (25,000 cells/well) and allowed to attach for 24 h. Posteriorly, cells were incubated with increasing concentrations of the plant extracts (62.5-1000 µg/mL) for 24 h. After, the medium was removed and MTT (final concentration 0.5 mg/mL) was added to each well, cells being incubated for 90 min at 37 • C. Formazan crystals were dissolved in a mixture of DMSO and isopropanol (3:1) and spectrophotometrically quantified at 560 nm in a microplate reader (Multiskan ASCENT, Massachusetts, MA, USA). Cell viability results were expressed as percentage of the control (untreated cells), corresponding to the mean ± SEM of three independent experiments, performed in triplicate.

Determination of NO Levels in Culture Medium
Interference with NO levels in murine macrophage cell line RAW 264.7 was performed according to Ferreres et al. [15]. Cells were cultured in 96-well plates (35,000 cells/well), allowed to attach for 24 h and exposed to increasing concentrations of extract. After 2 h, 1 µg/mL (final concentration) of LPS was added to each well and plates were incubated for further 22 h at 37 • C, in a humidified atmosphere of 5% CO 2 . The nitrite formed by the conversion of NO in the culture medium was spectrophotometrically quantified at 540 nm in a microplate reader (Multiskan ASCENT, Massachusetts, MA, USA). NO levels were expressed as percentage of the NO in cells exposed to LPS (positive control) and correspond to the mean ± SEM of four independent experiments, performed in triplicate.

Determination of l-Citrulline Levels in Culture Medium
Modulation of iNOS was determined by following the formation of l-arginine oxidative deamination bioproducts, NO and l-citrulline. Cells were cultured in 48-well plates (90,000 cells/well) and allowed to attach for 24 h. The medium was removed, and plates were incubated with medium containing increasing concentrations of the plant extracts. On the following 2 h, 1 µg/mL LPS was added to each well, plates being incubated for 22 h at 37 • C, in a humidified atmosphere of 5% CO 2 . After this period, the medium was removed and the iNOS substrate, l-arginine (50 µM) prepared in HBSS, was added to each well. The reaction was allowed to occur for 2 h, l-citrulline and NO levels being determined on the cell supernatant. NO levels were determined as described in Section 2.6.3., and l-citrulline was quantified as previously reported in [16]. Briefly, 250 µL of the culture supernatant was added to a mixture containing 79 mM diacetyl monoxime, 47.8 mM antipyrine E and 7.5 M H 2 SO 4 and incubated at 96 • C for 25 min. After, the solution was cooled down to room temperature and the absorbance was read at 405 nm. l-citrulline levels were expressed as percentage of the levels in cells exposed to LPS (positive control) and correspond to the mean ± SEM of five independent experiments, performed in duplicate.

Determination of • NO Levels in Non-Cellular System
Determination of • NO levels in non-cellular system was evaluated through the Griess diazotization reaction as in Ferreres et al. [14]. Equal volumes of SNP (100 µL) and plant extract, prepared in KH 2 PO 4 buffer (pH 7.4), were incubated for 1 h, under light at room temperature. Afterwards, the same volume of Griess reagent was added to each well and the plate was incubated for 10 min in the dark. Absorbance was read at 560 nm (Multiskan ASCENT, Massachusetts, MA, USA). • NO scavenging activity was determined through the comparison of the optic density between the extracts and the control (buffer instead of sample) and correspond to the mean ± SEM of three independent experiments, performed in triplicate. Quercetin was used as positive control.

Statistical Analysis
All biological data were analyzed using Graph Pad Prism 6 software (San Diego, USA). Grubbs' test was performed to detect significant outliers, differences at p < 0.05 being considered significant. On cellular assays, D'Agostino-Pearson normality test was performed and one-way analysis of variance (ANOVA) with Dunnett as post hoc test was used to compare significant differences amongst samples and controls (no extract treatment), p values lower than 0.05 being considered statistically significant. Compounds 1, 3, and 4 exhibited the same UV spectrum (300 sh, 326 nm) and the same deprotonated molecular ion (at m/z 353.0878), with molecular formula (C 16 H 18 O 9 ) corresponding to caffeoylquinic acids. MS fragmentation of these compounds revealed a base peak at m/z 191; nevertheless, while in compounds 3 and 4 the ion at m/z 179 was weak or undetectable, in compound 1 fragmentation this ion had a high relative abundance (50%). Thus, according to Clifford et al. [17], compound 3 can be labeled as a chlorogenic acid isomer, while 1 and 4, with the same chromatographic behavior of authentic standards, can unequivocally be identified as 3-O-caffeoylquinic acid and 5-O-caffeoylquinic acid, respectively.  (Table 1), which, for mono-C-glycosylflavones, characterize the nature of the aglycone as apigenin [18]. As so, these compounds are apigenin-6/8-C-hexosides. The higher relative abundance of the ion at m/z 341 in compound 18 fragmentation comparing to that found for compound 16 (30% vs. 3%), the higher RP mobility of compound 16 (Table 1), and the occurrence of the [(M − H)-18] − ion in 18 and its absence in 16 (data not shown in Table 1), indicate that the C-glycosylation in compound 16 occurs at position 8, while in 18 it is at position 6. Thus, considering the hexose to be glucose, compound 16 can be apigenin-8-C-glucoside (vitexin) and compound 18 apigenin-6-C-glucoside (isovitexin). Comparison with commercial standards allowed to unequivocally identify compounds 16 and 18 as vitexin and isovitexin, respectively.  (Table 1), which, for mono-C-glycosylflavones, characterize the nature of the aglycone as apigenin [18]. As so, these compounds are apigenin-6/8-C-hexosides. The higher relative abundance of the ion at m/z 341 in compound 18 fragmentation comparing to that found for compound 16 (30% vs. 3%), the higher RP mobility of compound 16 (Table 1), and the occurrence of the [(M − H)-18] − ion in 18 and its absence in 16 (data not shown in Table 1), indicate that the C-glycosylation in compound 16 occurs at position 8, while in 18 it is at position 6. Thus, considering the hexose to be glucose, compound 16 can be apigenin-8-C-glucoside (vitexin) and compound 18 apigenin-6-C-glucoside (isovitexin). Comparison with commercial standards allowed to unequivocally identify compounds 16 and 18 as vitexin and isovitexin, respectively.   (Table 1), indicating the presence of another hexose linked to a phenolic hydroxyl [19], are also detected. Consequently, compound 10 can correspond to an apigenin-O-hexoside-C-hexoside. While O-glycosylation generally occurs at position 7, C-glycosylation can appear either at position 6 or 8; thus, compound 10 can be labeled as apigenin-7-O-hexoside-6/8-C-hexoside corresponding, in the case of glucose being the hexose, to apigenin-7-O-glucoside-6/8-C-glucoside (isovitexin-7-O-glucoside or vitexin-7-O-glucoside). Unlike the previously described compounds, 10 was detected in both methanol extracts ( Figure 1).
Showing the same ions at m/z 383/353, as in 8, and having the deprotonated molecular ion at m/z~563.1400, compounds 11 and 14 correspond to apigenin-di-C-glycosides with both hexose and pentose on the sugar moiety. For both compounds, the relative abundance of the ion [(M − H)-60] − is higher than the one found for compound 8, due to the pentose; the higher abundance of this ion in compound 11 (15%), than in 14 (7%), indicates that the pentose is found at position 6 in compound 11, as this is the preferential fragmentation position. On the other hand, the elution order of asymmetric di-C-glycosylflavones in RP indicates that the 6-C-pentosyl-8-C-hexosyl derivatives elute before the isomeric 6-C-hexosyl-8-C-pentosyl [18]. Thus, compound 11 can be labeled as apigenin-6-C-pentoside-8-C-hexoside and 14 as apigenin-6-C-hexoside-8-C-pentoside.
Not only 10, but also compounds 14 and 11 were detected in the two methanol extracts, while the remaining compounds were solely identified on the extract obtained from the stem bark of F. curtipes ( Figure 1B).
With [M − H] − at m/z 289.0714 (C 15 H 14 O 6 ), compounds 2, 6, and 7 have the same MS fragmentation pattern, characteristic of the most common natural monomers, catechin and epicatechin ( Table 2). Comparison with authentic standards allowed the unequivocal identification of 2 and 7 as catechin and epicatechin, respectively.
Compounds 15, 20, and 21 (C 24 H 20 O 9 ), displaying the same MS fragmentation, are substituted monomers coincident with cinchonain I type (a phenylpropanoid linked through a carbon-carbon linkage to the C-8 A-ring of epicatechin). Having the same deprotonated molecular ion (C 30 H 26 O 12 ) and the same MS fragmentation, 5 and 17 correspond to catechin/epicatechin dimers, being known as procyanidin B (Table 2). Compounds 12 and 13 (C 39 H 32 O 15 ) share the same MS fragmentation pattern, corresponding to cinchonain II type compounds, structurally related with procyanidin B but having a cinchonain I type similar substitution (a phenylpropanoid moiety on the C-8 position of the flavan superior unit). Furthermore, a catechin/epicatechin trimer (C 45 H 38 O 18 ), procyanidin C type, was also detected (9). Compound (19)

HPLC-DAD Quantitative Analysis
HPLC-DAD quantitative analysis of the methanol extracts obtained from the stem bark and leaves was attainted using external calibration curves, allowing the quantitation of 17 flavonoids (2, 5-18, 20, 21), three caffeoylquinic acid derivatives (1, 3, 4) and aviculin (19). Linear relationships with coefficient of determination (R 2 ) above 0.998 were obtained for the external calibration curves built with the standard compounds (Table 3). Table 3. Linear regression equation analysis, limit of detection (LOD) a and limit of quantification (LOQ) b , for external standards.

Standard Regression Equation Linearity
Range (µg/mL) The methanol extract obtained from the stem bark of the plant is predominantly characterized by the presence of cinchonain type I (15, 20, and 21) and type II (12 and 13) derivatives, corresponding to ca. 34% and 19% of the total quantifiable phenolic content, respectively ( Table 4). As seen in Table 4, the cinchonain type I derivative 21 is the main component (1478.00 ± 18.67 mg/kg dry extract), the stem bark extract being also characterized by significant amounts of aviculin (19; 1024.17 ± 81.73 mg/kg dry extract). Despite the qualitative and quantitative dissimilarities between the extracts obtained from the two vegetal materials, the total amount of apigenin derivatives was similar (stem bark (SB): 618.67 ± 36.26 mg/kg dry extract; leaves (LV): 611.65 ± 29.81 mg/kg dry extract) ( Table 4). While vicenin-2 (8) was the apigenin derivative identified in higher amounts in the stem bark extract, the main apigenin derivative, and simultaneously the major phenolic constituent present in the extract obtained from the leaves, was the apigenin-di-C-glycoside 11 (Table 4).

5-LOX Inhibition
As observed in Figure 2, both extracts were able to inhibit the activity of 5-LOX at the range of concentrations tested. The extract obtained from the stem bark proved to be more active, causing 91.72% ± 4.51% of inhibition at 250 µg/mL, while exposure to the leaf extract resulted in 59.57% ± 16.80% inhibition at the same concentration ( Figure 2). Concentration-dependent inhibition allowed to determine the IC 50 value for both extracts, 10.75 and 200.80 µg/mL being recorded upon treatment with the stem bark and leaf extract, respectively ( Figure 2). While notably more potent than the leaf extract (Figure 2), the methanol extract obtained from the stem bark was 5-fold less effective than quercetin (positive control) with an IC 50 value of 2.45 ± 0.65 µg/mL.

Effect on RAW 264.7 Macrophages
In order to select working concentrations to evaluate the effects on NO levels, potential interference with the mitochondrial viability of RAW 264.7 macrophages was assessed through the MTT assay. While the extract obtained from the stem bark did not lead to cytotoxicity at concentrations ranging from 62.5 to 1000 μg/mL, a significant (p < 0.0001) reduction in the cell viability was observed upon treatment with the leaf extract at 1000 μg/mL (Figure 3).

Interference with NO Levels
Considering the potential cytotoxicity previously assessed, 1000 and 500 μg/mL were selected as the highest concentrations to evaluate, respectively, the effects of stem bark and leaf extracts on NO levels in LPS-challenged RAW 264.7 macrophages, measured as nitrite formation. After 22 h of concomitant exposure to LPS and extracts obtained from F. curtipes, no interference on the NO levels was observed with the stem bark extract (Figure 4). In contrast, the leaf extract was able to cause a significant concentration-dependent reduction on NO levels, at concentrations ranging from 31.25 to 500 μg/mL (Figure 4). It is worth also mentioning that at the highest tested concentration (500 μg/mL) the extract obtained from the leaves caused a significant (p < 0.0001) NO reduction (39.32% ± 8.69%), compared to the respective control (no extract treatment).

Effect on RAW 264.7 Macrophages
In order to select working concentrations to evaluate the effects on NO levels, potential interference with the mitochondrial viability of RAW 264.7 macrophages was assessed through the MTT assay. While the extract obtained from the stem bark did not lead to cytotoxicity at concentrations ranging from 62.5 to 1000 µg/mL, a significant (p < 0.0001) reduction in the cell viability was observed upon treatment with the leaf extract at 1000 µg/mL (Figure 3).

Effect on RAW 264.7 Macrophages
In order to select working concentrations to evaluate the effects on NO levels, potential interference with the mitochondrial viability of RAW 264.7 macrophages was assessed through the MTT assay. While the extract obtained from the stem bark did not lead to cytotoxicity at concentrations ranging from 62.5 to 1000 μg/mL, a significant (p < 0.0001) reduction in the cell viability was observed upon treatment with the leaf extract at 1000 μg/mL (Figure 3).

Interference with NO Levels
Considering the potential cytotoxicity previously assessed, 1000 and 500 μg/mL were selected as the highest concentrations to evaluate, respectively, the effects of stem bark and leaf extracts on NO levels in LPS-challenged RAW 264.7 macrophages, measured as nitrite formation. After 22 h of concomitant exposure to LPS and extracts obtained from F. curtipes, no interference on the NO levels was observed with the stem bark extract (Figure 4). In contrast, the leaf extract was able to cause a significant concentration-dependent reduction on NO levels, at concentrations ranging from 31.25 to 500 μg/mL (Figure 4). It is worth also mentioning that at the highest tested concentration (500 μg/mL) the extract obtained from the leaves caused a significant (p < 0.0001) NO reduction (39.32% ± 8.69%), compared to the respective control (no extract treatment).

Interference with NO Levels
Considering the potential cytotoxicity previously assessed, 1000 and 500 µg/mL were selected as the highest concentrations to evaluate, respectively, the effects of stem bark and leaf extracts on NO levels in LPS-challenged RAW 264.7 macrophages, measured as nitrite formation. After 22 h of concomitant exposure to LPS and extracts obtained from F. curtipes, no interference on the NO levels was observed with the stem bark extract (Figure 4). In contrast, the leaf extract was able to cause a significant concentration-dependent reduction on NO levels, at concentrations ranging from 31.25 to 500 µg/mL (Figure 4). It is worth also mentioning that at the highest tested concentration (500 µg/mL) the extract obtained from the leaves caused a significant (p < 0.0001) NO reduction (39.32% ± 8.69%), compared to the respective control (no extract treatment).

Interference with L-Citrulline Levels
To assess the leaf extract interference with iNOS expression and/or its direct activity, L-citrulline levels were determined on LPS-stimulated RAW 264.7. As observed in Figure 5, there was a significant reduction in L-citrulline levels at concentrations higher than 125 μg/mL, a strong reduction (p < 0.0001) of ca. 58.90% ± 10.12% being observed at 500 μg/mL.

Nitric Oxide Radical Levels in Non-Cellular System
Considering the in vitro results obtained with RAW 264.7 cells, we found relevant to assess the ability of the methanol extract obtained from the leaves to scavenge the nitric oxide radical ( • NO), produced through the photolytic decomposition of sodium nitroprusside. The leaf extract exhibited significant scavenging activity in a concentration-dependent manner (Figure 6), displaying an IC50 value of 304.8 μg/mL, ca. six times higher than the positive control (quercetin, IC50 = 54.42 μg/mL). At the range of concentrations used on the cellular assay (500-31.25 μg/mL) the scavenging activity varied between 12.92% ± 1.78% and 65.34% ± 5.92%.

Interference with l-Citrulline Levels
To assess the leaf extract interference with iNOS expression and/or its direct activity, l-citrulline levels were determined on LPS-stimulated RAW 264.7. As observed in Figure 5, there was a significant reduction in l-citrulline levels at concentrations higher than 125 µg/mL, a strong reduction (p < 0.0001) of ca. 58.90% ± 10.12% being observed at 500 µg/mL.

Interference with L-Citrulline Levels
To assess the leaf extract interference with iNOS expression and/or its direct activity, L-citrulline levels were determined on LPS-stimulated RAW 264.7. As observed in Figure 5, there was a significant reduction in L-citrulline levels at concentrations higher than 125 μg/mL, a strong reduction (p < 0.0001) of ca. 58.90% ± 10.12% being observed at 500 μg/mL.

Nitric Oxide Radical Levels in Non-Cellular System
Considering the in vitro results obtained with RAW 264.7 cells, we found relevant to assess the ability of the methanol extract obtained from the leaves to scavenge the nitric oxide radical ( • NO), produced through the photolytic decomposition of sodium nitroprusside. The leaf extract exhibited significant scavenging activity in a concentration-dependent manner (Figure 6), displaying an IC50 value of 304.8 μg/mL, ca. six times higher than the positive control (quercetin, IC50 = 54.42 μg/mL). At the range of concentrations used on the cellular assay (500-31.25 μg/mL) the scavenging activity varied between 12.92% ± 1.78% and 65.34% ± 5.92%.

Nitric Oxide Radical Levels in Non-Cellular System
Considering the in vitro results obtained with RAW 264.7 cells, we found relevant to assess the ability of the methanol extract obtained from the leaves to scavenge the nitric oxide radical ( NO), produced through the photolytic decomposition of sodium nitroprusside. The leaf extract exhibited significant scavenging activity in a concentration-dependent manner (Figure 6), displaying an IC 50 value of 304.8 µg/mL, ca. six times higher than the positive control (quercetin, IC 50 = 54.42 µg/mL).

Phenolic Profile Characterization
In contrast with widely investigated Ficus species, such as F. carica, F. racemosa, or F. religiosa, there are no available records on the chemical profile of F. curtipes. The stem bark extract of this species is characterized by the occurrence of caffeoylquinic acids (1, 3, and 4), apigenin derivatives (8, 10, 11, 14, 16, and 18), and predominantly flavan-3-ols (2, 5-7, 9, 12, 13, 15, 17, 20, and 21), while only apigenin derivatives (10, 11, and 14) were detected on the extract obtained from the leaves ( Figure 1). As seen in Table 4, the main structural group of components occurring in the stem bark are phenylpropanoid-substituted monomeric (15, 20, and 21) and dimeric (12 and 13) flavan-3-ols, high amounts of procyanidins (5, 9, and 17) being also quantitated. Analogously, previous studies demonstrate that the occurrence of monomeric and condensed flavan-3-ols is common in Ficus spp., namely in external tissues; Konai et al. [21] reported that an aqueous extract obtained from Ficus sycomorus bark was also characterized by the occurrence of condensed tannins, particularly procyanidin type with 2,3-cis stereochemistry; the majority of the compounds identified in a methanol extract obtained from the stem bark of Ficus microcarpa correspond to catechin and epicatechin compounds, procyanidin B1 and B3 being also detected [22].
Curiously, while flavan-3-ols and apigenin derivatives are commonly reported in the genus, there was no evidence on the occurrence of the lignan glycoside aviculin (19) until the current study, herein identified as one of the main constituents of the methanol extract obtained from the stem bark of F. curtipes (Table 4).
Quantitative profiles of the two extracts were found to be distinct, total quantifiable phenolic content of the extract obtained from the stem bark being approximately nine times higher than the leaves (5374.15 ± 436.61 vs. 611.65 ± 29.81 mg/kg dry extract) ( Table 4). It is relevant to emphasize that, despite its lower phenolic content, the extract obtained from the leaves presented a higher concentration of the apigenin derivatives 10, 11, and 14, the apigenin-6-C-hexoside-8-C-pentoside (14) content being ca. eight times higher on this extract. On the other hand, considering the total

Phenolic Profile Characterization
In contrast with widely investigated Ficus species, such as F. carica, F. racemosa, or F. religiosa, there are no available records on the chemical profile of F. curtipes. The stem bark extract of this species is characterized by the occurrence of caffeoylquinic acids (1, 3, and 4), apigenin derivatives (8, 10, 11, 14, 16, and 18), and predominantly flavan-3-ols (2, 5-7, 9, 12, 13, 15, 17, 20, and 21), while only apigenin derivatives (10, 11, and 14) were detected on the extract obtained from the leaves ( Figure 1). As seen in Table 4, the main structural group of components occurring in the stem bark are phenylpropanoid-substituted monomeric (15, 20, and 21) and dimeric (12 and 13) flavan-3-ols, high amounts of procyanidins (5, 9, and 17) being also quantitated. Analogously, previous studies demonstrate that the occurrence of monomeric and condensed flavan-3-ols is common in Ficus spp., namely in external tissues; Konai et al. [21] reported that an aqueous extract obtained from Ficus sycomorus bark was also characterized by the occurrence of condensed tannins, particularly procyanidin type with 2,3-cis stereochemistry; the majority of the compounds identified in a methanol extract obtained from the stem bark of Ficus microcarpa correspond to catechin and epicatechin compounds, procyanidin B1 and B3 being also detected [22].
Curiously, while flavan-3-ols and apigenin derivatives are commonly reported in the genus, there was no evidence on the occurrence of the lignan glycoside aviculin (19) until the current study, herein identified as one of the main constituents of the methanol extract obtained from the stem bark of F. curtipes (Table 4).
Quantitative profiles of the two extracts were found to be distinct, total quantifiable phenolic content of the extract obtained from the stem bark being approximately nine times higher than the leaves (5374.15 ± 436.61 vs. 611.65 ± 29.81 mg/kg dry extract) ( Table 4). It is relevant to emphasize that, despite its lower phenolic content, the extract obtained from the leaves presented a higher concentration of the apigenin derivatives 10, 11, and 14, the apigenin-6-C-hexoside-8-C-pentoside (14) content being ca. eight times higher on this extract. On the other hand, considering the total quantifiable phenolic content, the great abundance of cinchonain type compounds in the stem bark extract points this vegetal material of F. curtipes as a great source flavan-3-ols.

5-LOX Inhibition
The enzyme 5-LOX is involved in the first steps of the conversion of arachidonic acid into leukotrienes, which are primarily responsible for the initiation of the inflammatory process [23]. Besides their chemokinetic and chemotactic responses, leukotrienes also play a role in the recruitment, migration, adhesion, and degranulation of granulocyte cells, acting as potent modulators of the inflammatory process [28]. In fact, leukotrienes are intimately related with the pathophysiology of diseases, such as atherosclerosis, asthma, arthritis, and skin disorders with an inflammatory background, and became the focal point of many therapeutic approaches for the treatment of these conditions [28,29]. Hence, there is a rationale on the inhibition of 5-LOX as a proper therapeutic strategy for these leukotriene-related disorders management [29].
Our results demonstrate that the two extracts obtained from F. curtipes materials were able to inhibit 5-LOX in a concentration-dependent manner (Figure 2), the extract obtained from the stem bark being markedly more effective (IC 50 = 10.75 µg/mL). At a first glance, it is tempting to relate the stronger inhibitory effects observed upon exposure to the stem bark extract to the higher amounts of phenolic constituents identified in the extract (5374.15 ± 436.61 mg/kg dry extract) in comparison with the leaf extract (611.5 ± 29.81 mg/kg dry extract) ( Table 4). In fact, while phenolic compounds are generally described as prominent plant-derived inhibitors of 5-LOX product synthesis [28,30], individual constituents identified in stem bark extract can directly contribute to the observed inhibitory effects. Specifically concerning monomeric flavan-3-ols, it is worth referring to the effects upon 5-LOX dependent leukotriene C 4 (LTC 4 ) production in bone marrow-derived mast cells upon treatment with catechin (2) (20 µM), mediating 99.8% inhibition [31], being also reported to inhibit potato 5-LOX activity (IC 50 value of 16.5 µg/mL) [32]. Relevantly, catechin (2) happens to be one of the main active compounds found in flavocoxid, an FDA-approved medical food for the treatment of osteoarthritis [33]. Additionally, epicatechin (7) has been reported to inhibit both human 5-LOX and rabbit 15-LOX, with IC 50 values of 22 and 60 µM, correspondingly [34]. Such findings are in agreement with the reported decay of plasmatic pro-inflammatory cysteinyl leukotrienes observed after human intake of flavonoid-rich chocolate with substantial amounts of epicatechin (7) and related oligomers (procyanidins) [35]. Interestingly, Schewe and colleagues demonstrated that epicatechin (7) and its lower-molecular procyanidins inhibit both steps of 5-LOX catalysis: the deoxygenation of arachidonic acid to 5-hydroperoxy-6E,8Z,11Z,14Z-eicosatetraenoic acid and the following conversion to 5,6-leukotriene A 4 [34]. Additionally, the potential contribution of the caffeoylquinic acids, constituting ca. 5% of the stem bark phenolic content (Table 4), should be taken into account; as recently reported by Gawlik-dziki et al. [36], 5-O-caffeoylquinic acid (4) acts as a potent soybean 5-LOX inhibitor, with an IC 50 of 18.56 µg/mL. It is also worth mentioning the possible interference of the identified apigenin derivatives on the stem bark extract that were not detected in the leaves. While there are no available studies on the inhibitory capacity of vicenin-2 (8), both vitexin (16) and isovitexin (18) have already been reported as soybean 5-LOX [37] and 15-LOX inhibitors [38], with IC 50 values of 5.1 and 107.1 µM, respectively. Hence, it is conceivable to assume the observed inhibitory effects towards 5-LOX to be related, at least partially, with the identified phenolic compounds.

Effect on RAW 264.7 Macrophages and NO Scavenging Capacity
The innate immunity system responds to the presence of pathogens or tissue injury through immune cells recruitment, namely monocytes and macrophages [39]. These cells are commonly responsible for the development of inflammatory-derived symptoms, namely edema, heat, redness, and pain [39]. When triggered by microbial endotoxins and/or endogenous inflammatory cytokines, macrophages become activated and start to express genes involved in host defense, leading to the production of reactive nitrogen species and bioactive lipids derived from arachidonic acid, contributing to the inflammatory process [39].
Appropriate cell models to assess anti-inflammatory properties include LPS-challenged murine macrophage RAW 264.7 cells [14,16]. LPS-derived activation leads to the upregulation of pro-inflammatory cytokines, such as IL-1, IL-6, or TNF-α, as well as pro-inflammatory proteins, including iNOS [39]. This enzyme catalyzes the conversion of the substrates O 2 and l-arginine to l-citrulline and NO, being the main responsible for their overproduction [39]. Increased NO production is actively involved in the inflammatory process, and can be either beneficial, acting as a bactericidal, or harmful, causing DNA damage and protein oxidation [40]. This cellular mediator is a well-acknowledged inflammatory biomarker, inhibition of its production being a widely validated approach for the amelioration of the inflammatory response [39,40].
In order to ensure that the potential reduction of NO levels in RAW 264.7 macrophages was not related with cytotoxicity, interference with mitochondrial viability was firstly evaluated through the MTT reduction assay. As seen in Figure 3, solely the extract obtained from the leaves of F. curtipes caused a significant decrease on the cell viability of RAW 264.7 cells, a range of working concentrations between 31.25 and 500 µg/mL being consequently selected. The 24 h-exposure to the leaf extract led to a decrease on NO levels at all tested concentrations, a reduction to 39.32% ± 8.69% being recorded at 500 µg/mL (Figure 4). No effects were observed upon treatment with the extract obtained from the stem bark ( Figure 4).
Since cellular NO levels decrement results from the interference with the iNOS expression and/or activity, and/or the direct scavenging of this reactive species, additional in vitro assays were geared as an attempt to explain the observed effects upon treatment with the leaf extract. This extract was further evaluated for its ability to reduce l-citrulline levels on the cells' supernatant ( Figure 5) and to scavenge • NO in a non-cellular system ( Figure 6). Results evidence that, at 250 and 500 µg/mL, the extract acts through a bimodal mechanism, directly interfering with iNOS activity and also displaying a significant scavenging effect towards • NO, thus explaining the reduction in NO levels in LPS-activated RAW 264.7 macrophages (Figure 4). However, the effects observed at lower concentrations ( Figure 4) are likely to be predominantly related with a radical scavenging effect ( Figure 6).
It is plausible to consider that the di-C-glycosides 10, 11, and 14 contribute to the direct modulation of iNOS ( Figure 5), as apigenin derivatives are known to interfere with the enzyme [41,42]. SAR studies with a series of flavonoids allowed to ascribe the structural features resulting in iNOS downregulation in RAW 264.7 macrophages, such as the importance of the C-2,3 unsaturation, the A-ring di-hydroxy meta-substitution and the 4'-hydroxyl group, flavones being within the optimal structures, if not the best [43,44]. Additionally, the strong antiradical effects observed upon treatment with F. curtipes leaf extract may be partially mediated by the presence of the apigenin di-C-glycosides (10, 11, and 14), as flavones have long been identified as efficient • NO scavengers [45]. In fact, apigenin derivatives follow the optimal structural requirements, namely the presence of hydroxyl substituents in C5 and C4 , as well as the C-ring carbonyl function [46].
Overall, our data prove that the extract obtained from the leaves of F. curtipes significantly reduced NO levels in RAW 264.7 macrophages, acting through the combination of radical scavenging properties and iNOS modulation, which can thereby result from direct enzyme inhibition and/or enzyme expression inhibition. The results recorded here are reported for the first time in this species, providing initial proof on its effect upon a well-recognized inflammatory pathway and stimulating further molecular approaches to thoroughly comprehend the conceivable anti-inflammatory potential of this plant material.

Conclusions
Biological data generated here provides preliminary evidence on the conceivable modulatory effect of F. curtipes on inflammatory-related conditions. While the extract obtained from the stem bark proved to efficiently inhibit 5-LOX activity, the leaf extract was found to significantly interfere with NO levels in LPS-challenged RAW 264.7 macrophages, as a consequence of a significant anti-radical effect and modulation of iNOS expression and/or activity.
Furthermore, the species is also valorized as a source of a series of structurally diverse phenolic compounds. The phenolic profile of the species is disclosed for the first time, evidencing a chemical richness characterized by the occurrence of apigenin, as well as flavan-3-ols, some of which are well known for their anti-inflammatory ability and therapeutic utility.
Taking the above into account, the present work provides further insights not only on F. curtipes polyphenolic composition, but also on its biological properties, encouraging further molecular approaches in order to thoroughly comprehend the anti-inflammatory potential of this plant species.