NOx-, IL-1β-, TNF-α-, and IL-6-Inhibiting Effects and Trypanocidal Activity of Banana (Musa acuminata) Bracts and Flowers: UPLC-HRESI-MS Detection of Phenylpropanoid Sucrose Esters

Banana inflorescences are a byproduct of banana cultivation consumed in various regions of Brazil as a non-conventional food. This byproduct represents an alternative food supply that can contribute to the resolution of nutritional problems and hunger. This product is also used in Asia as a traditional remedy for the treatment of various illnesses such as bronchitis and dysentery. However, there is a lack of chemical and pharmacological data to support its consumption as a functional food. Therefore, this work aimed to study the anti-inflammatory action of Musa acuminata blossom by quantifying the cytokine levels (NOx, IL-1β, TNF-α, and IL-6) in peritoneal neutrophils, and to study its antiparasitic activities using the intracellular forms of T. cruzi, L. amazonensis, and L. infantum. This work also aimed to establish the chemical profile of the inflorescence using UPLC-ESI-MS analysis. Flowers and the crude bract extracts were partitioned in dichloromethane and n-butanol to afford four fractions (FDCM, FNBU, BDCM, and BNBU). FDCM showed moderate trypanocidal activity and promising anti-inflammatory properties by inhibiting IL-1β, TNF-α, and IL-6. BDCM significantly inhibited the secretion of TNF-α, while BNBU was active against IL-6 and NOx. LCMS data of these fractions revealed an unprecedented presence of arylpropanoid sucroses alongside flavonoids, triterpenes, benzofurans, stilbenes, and iridoids. The obtained results revealed that banana inflorescences could be used as an anti-inflammatory food ingredient to control inflammatory diseases.


Introduction
Several nutritionists and food chemists have been working intensively with non-conventional food plants (NCFPs). Fruits, seeds, roots, flowers, leaves, or entire plants can constitute these foods and, in most cases, they represent a new cultural food value for various communities [1]. ty when tested from 10 to 1000 µ g/mL (p < 0.05). The other two fractions, FNBU and BD ed no effect on neutrophil viability at the tested concentrations (p > 0.05). , and bract raction from n-butanol partition (BNBU) (D) on neutrophil viability. Control: peritoneal neutrophils solated from mice treated only with vehicle; 10-1000: peritoneal neutrophils isolated from mice reated with concentrations of each specific extract ranging from 10 to 1000 µ g/mL. Each group epresents the mean ± standard error of the mean; n = 3/group. *p < 0.05, **p < 0.01, and ***p < 0.001 ompared to the control group (ctrl).
tion Study on Cytokine Secretion (IL-1β, TNF-α, and IL- 6) n the basis of the results obtained from the cell viability assay, inhibition effects of the frac evaluated on pro-inflammatory mediators at non-cytotoxic concentrations. Only FDCM sho icant inhibition of IL-1β levels, reducing the levels of this inflammatory mediator by 51. (p < 0.001) of when tested at 100 µ g/mL (Figure 2A). Other fractions in turn weakly affe ion of this inflammatory mediator (p > 0.05) ( Figure 2B-D). , and bract fraction from n-butanol partition (BNBU) (D) on neutrophil viability. Control: peritoneal neutrophils isolated from mice treated only with vehicle; 10-1000: peritoneal neutrophils isolated from mice treated with concentrations of each specific extract ranging from 10 to 1000 µg/mL. Each group represents the mean ± standard error of the mean; n = 3/group. * p < 0.05, ** p < 0.01, and *** p < 0.001 compared to the control group (ctrl).
Inhibition Study on Cytokine Secretion (IL-1β, TNF-α, and IL- 6) On the basis of the results obtained from the cell viability assay, inhibition effects of the fractions were evaluated on pro-inflammatory mediators at non-cytotoxic concentrations. Only FDCM showed significant inhibition of IL-1β levels, reducing the levels of this inflammatory mediator by 51.9% ± 7.2% (p < 0.001) of when tested at 100 µg/mL (Figure 2A). Other fractions in turn weakly affected secretion of this inflammatory mediator (p > 0.05) ( Figure 2B-D).
While the level of TNF-α was decreased by 46.1% ± 8.0% (p < 0.05) when neutrophils were treated with FDCM at 100 µg/mL ( Figure 3A), BDCM at concentrations of 30 and 100 µg/mL inhibited the secretion of the same cytokine by 46.5% ± 8.6% and 50.7% ± 6.2%, respectively (p < 0.01) ( Figure 3C). On the other hand, FNBU and BNBU did not reduce this inflammatory mediator (p > 0.05) ( Figure 3B , and BNBU (D) on IL-1β secretion by LPSstimulated peritoneal murine neutrophils. Control: peritoneal neutrophils isolated from mice treated only with vehicle; LPS: peritoneal neutrophils isolated from mice stimulated with LPS and treated with vehicle; 10-100: peritoneal neutrophils isolated from mice stimulated with LPS and treated with concentrations of each specific extract ranging from 10 to 100 µ g/mL. Each group represents the mean ± standard error of the mean; n = 3/group. ### p < 0.001 compared to the Ctrl group. *** p < 0.001 compared to the Ctrl group.
While the level of TNF-α was decreased by 46.1% ± 8.0% (p < 0.05) when neutrophils were treated with FDCM at 100 µ g/mL ( Figure 3A), BDCM at concentrations of 30 and 100 µ g/mL inhibited the secretion of the same cytokine by 46.5% ± 8.6% and 50.7% ± 6.2%, respectively (p < 0.01) ( Figure 3C). , and BNBU (D) on IL-1β secretion by LPS-stimulated peritoneal murine neutrophils. Control: peritoneal neutrophils isolated from mice treated only with vehicle; LPS: peritoneal neutrophils isolated from mice stimulated with LPS and treated with vehicle; 10-100: peritoneal neutrophils isolated from mice stimulated with LPS and treated with concentrations of each specific extract ranging from 10 to 100 µg/mL. Each group represents the mean ± standard error of the mean; n = 3/group. ### p < 0.001 compared to the Ctrl group. *** p < 0.001 compared to the Ctrl group. On the other hand, FNBU and BNBU did not reduce this inflammatory mediator (p > 0.05) ( Figure  3B,D). , and BNBU (D) on TNF-α secretion by LPSstimulated peritoneal murine neutrophils. Control: peritoneal neutrophils isolated from mice treated only with vehicle; LPS: peritoneal neutrophils isolated from mice stimulated with LPS and treated with vehicle; 10-100: peritoneal neutrophils isolated from mice stimulated with LPS and treated with concentrations of each specific extract ranging from 10 to 100 µ g/mL. Each group represents the mean ± standard error of the mean; n = 3/group. ## p < 0.001 compared to the Ctrl group. * p < 0.01 and ** p < 0.001 compared to the Ctrl group.

Measurement of NO x Production by Mouse Neutrophils
FDCM, FNBU, and BDCM were unable to reduce the levels of NO x secretion regardless of the concentrations tested (p > 0.05) ( Figure 5A-C), whereas BNBU was able to reduce the production of this inflammatory mediator by 40.2% ± 7.6% and 46.5% ± 5.2%, respectively, (p < 0.01) when tested at 30 and 100 µg/mL ( Figure 5D).
The relationship between the inflammatory cytokines produced by parasites during infection and their virulence [12][13][14][15] prompted us to evaluate all the fractions against the intracellular forms of T. cruzi, L. amazonensis, and L. infantum amastigotes, as well as the viability of the THP-1 cells (human monocytic cell line) used as host. with vehicle; 10-100: peritoneal neutrophils isolated from mice stimulated with LPS and treated with concentrations of each specific extract ranging from 10 to 100 µ g/mL. Each group represents the mean ± standard error of the mean; n = 3/group. ### p < 0.001 compared to the Ctrl group. * p < 0.01 and *** p < 0.001 compared to the Ctrl group.  10-100: peritoneal neutrophils isolated from mice stimulated with LPS and treated with concentrations of each specific extract ranging from 10 to 100 µg/mL. Each group represents the mean ± standard error of the mean; n = 3/group. ### p < 0.001 compared to the Ctrl group. ** p < 0.01 compared to the Ctrl group.

Antitrypanosomal and Antileishmanial Activities
FDCM also revealed low toxicity to THP-1 cells with a CC 50 value of 341.5 µg/mL, leading to a selectivity index of 9.14 ( Table 1). As the cytotoxicity of FDCM was nearly 300 µg/mL, a non-toxic concentration of 50 µg/mL was considered to preserve the viability of the macrophage, with the aim of observing significant antiparasitic activity.
FDCM inhibited 90.37% of T. cruzi growth, corresponding to an IC 50 value of 37.35 µg/mL. It also showed weak effects against L. amazonensis, and L. infantum amastigotes, with inhibition of 37.13% and 11.04%, respectively. The remaining fractions showed weak to no activity against T. cruzi and the studied Leishmania species.

Chemistry
The hyphenated techniques UPLC-ESIMS and UPLC-ESI-MS 2 were used to establish the chemical compositions of FDCM, FNBU, BDCM, and BNBU, as they showed promising biological activity against inflammation mediators. Their constituents were exclusively sensitive to the negative ionization mode, and the base peak ionization (BPI) was used as the acquisition parameter of the chromatograms ( Figure 6). An error equal to or less than ±5 ppm was considered for the determination of the molecular formula. FDCM showed in its LC-ESI-MS chromatogram the presence of 15 metabolites, while 23 components were detected in FNBU (Tables S1 and S2). Two other metabolites appeared in the chromatogram at tR 7.32 and 7.58 min with the same mass value, m/z 655.1874, corresponding to C29H36O17. Both were different from the precedents at 6.66 and 6.99 min by 42 Da, indicating their acetylated derivatives. The tandem mass data of m/z 655.1874 showed ions consistent with structures of coumaric acid (m/z 163.0352), monoglycosylated coumaric acid (m/z 349.0909), and triacetylated disaccharide (m/z 467.1384) ( Figure S1). The metabolites at tR 7.32 and 7.58 min gave an almost similar fragmentation pattern, permitting their isomeric relationship to be deduced. A literature search of C29H36O17 led to the structures mumeose I, mumeose L, mumeose Q, mumeose U, and mumeose T, previously obtained from the flower buds of Prunus mume [17,18]. The exact structures of the compounds could not be determined based on the mass spectrometric data. However, the m/z 349.0909 fragment ( Figure S1), corresponding to the monoacetylated glycosyl coumaric acid ion, suggested muneose L and U as potential structures of the metabolite detected at 7.32 min. The isomer (tR 7.58 min) revealed in its MS 2 spectrum ( Figure S1) a fragment at m/z 391.0948, leading to the structure of 4,6,2',6'-O-tetraacetyl-3-O-p-coumaroylsucrose. This compound was previously obtained from the fruits of Prunus jamasakura [19]. The first two metabolites of FDCM were detected at 6.66 and 6.99 min with the same mass value, m/z 613.1731, corresponding to the elemental composition C 27 H 34 O 16 . Due to their low quantity in the extract, no fragment was obtained in the tandem mass analysis. However, on their MS spectra, a fragmentation pattern corresponding to a loss of ketene (42 Da) was observed. The obtained m/z 571.1544 fragment is a typical in-source-generated product, of which the mechanism has previously been studied and reported [16]. A literature search of this elemental composition led to the structure of four isomeric acetylated sucroses, namely mumeose G, mumeose S, mumeose H, and tomenside A. Up to now, these arylpropanoid sucroses have never been previously reported in banana species.
Two other metabolites appeared in the chromatogram at tR 7.32 and 7.58 min with the same mass value, m/z 655.1874, corresponding to C 29 H 36 O 17 . Both were different from the precedents at 6.66 and 6.99 min by 42 Da, indicating their acetylated derivatives. The tandem mass data of m/z 655.1874 showed ions consistent with structures of coumaric acid (m/z 163.0352), monoglycosylated coumaric acid (m/z 349.0909), and triacetylated disaccharide (m/z 467.1384) ( Figure S1). The metabolites at tR 7.32 and 7.58 min gave an almost similar fragmentation pattern, permitting their isomeric relationship to be deduced. A literature search of C 29 H 36 O 17 led to the structures mumeose I, mumeose L, mumeose Q, mumeose U, and mumeose T, previously obtained from the flower buds of Prunus mume [17,18]. The exact structures of the compounds could not be determined based on the mass spectrometric data. However, the m/z 349.0909 fragment ( Figure S1), corresponding to the monoacetylated glycosyl coumaric acid ion, suggested muneose L and U as potential structures of the metabolite detected at 7.32 min. The isomer (tR 7.58 min) revealed in its MS 2 spectrum ( Figure S1) a fragment at m/z 391.0948, leading to the structure of 4,6,2',6'-O-tetraacetyl-3-O-p-coumaroylsucrose. This compound was previously obtained from the fruits of Prunus jamasakura [19].  [20]. A literature search led to the structure of three isomeric metabolites, among which 3-O-p-coumaroylsucrose was assigned as the structure. This metabolite was previously identified in dried fruits of Prunus domestica [21].
Another metabolite was found at 2.37 min with m/z 487.1465 [C 21 H 28 O 13 -H] − , suggesting an isomer of 3-O-p-coumaroylsucrose. Tandem mass of the m/z 487.1465 data showed fragment ions at m/z 341.0868 and 179.0580, which were formed after the loss of a deoxyhexose (146 Da) and a disaccharide (deoxyhexose + hexose, 308 Da), respectively. The m/z 179.0580 aglycone was characterized as caffeic acid, and the structure of this metabolite was assigned as cistanoside F. Its fragmentation behavior was similar to that previously reported [22].  [20]. On the basis of the abovementioned information, the structures of these metabolites were assigned as positional isomers of mumeose B, P, or R, previously isolated from the flower buds of Prunus mune [17].
Four other positional isomers were also found in this fraction at 6.18, 6. to afford an m/z of 307.0814. As observed in the MS spectrum of the abovementioned metabolites, the aglycone was coumaric acid, consistent with an m/z 163.0378 and its dehydrated m/z 145.0278 fragment ion [20]. Because the position of the acetyl groups could not be determined using MS data, these four metabolite structures were deduced to be related to tomenside B based on the aforementioned information. Tomenside B is a triacetylated phenylpropanoid sucrose previously obtained from Prunus tomentosa leaves [23].
The  (Figure S3). This information led to the structure of diffusosides A or B, two diastereomeric iridoids previously obtained from Hedyotis diffusa [27].
The This information, together with that presented in Figure S4, Two other metabolites were found at 8.75 and 13.23 min with m/z values of 221.1178 and 447.2509, respectively. The lack of fragmentation limited their structural assignment; however, a literature search indicated that these compound structures were related to those of an alkylated phenol and a stilbene, respectively.

Discussion
The cytotoxic effect of the studied extracts on the isolated mouse neutrophils showed that FDCM and BNBU reduced these cells' viability at concentrations equal to or greater than 300 µM. Alongside acetylated arylpropanoid sucroses, FDCM was also composed of fatty acids and other phenolics, while FNBU was formed of flavonoids, triterpenes, cyclohexanetetrol, and a low quantity of acetylated arylpropanoid sucroses ( Table 2). The most concentrated sucrose, at 8.27 min (m/z 697.1997), and its positional isomers were presumably responsible for the cytotoxicity observed against neutrophil cells. However, their weak concentration in FNBU might support why this fraction lacked cytotoxic activity. Interestingly, compounds related to 3-phenylpropanoid-triacetyl sucrose esters, such as tomensides A-D and numeose C, demonstrated cytotoxicity against four human cancer cell lines in a previous study, although no information was provided about their selectivity towards normal cells [23]. The LCMS data of BDCM showed the presence of arylpropanoids, glycolipids, arylbenzofurans, fatty acids, and stilbenes; among them, an m/z 447.2509 stilbene derivative was the major component. No cytotoxicity was observed for this fraction. However, neutrophil cells responded slightly to BNBU, which was rich in glycolipids, stilbenes, flavonoids, arylbenzofurans, other phenolics, and iridoids, among which 6,8-di-C-glycosylated luteolin and an O-acyl glycoside were the main components. A previous study revealed that 6,8-di-C-β-d-glucopyranosyl-luteolin (lucenin-2) is weakly or not cytotoxic against five cancer cell lines [28]. Therefore, iridoids might be responsible for the cytotoxic effect on neutrophils, since some of these metabolites have been described as antiproliferative agents [29]. No toxicity study was found in the literature on banana inflorescences; however, previous bioassays have shown that its incorporation in rat diets might modulate serum cholesterol and glucose [8].
The anti-inflammatory effect of these fractions was evaluated at concentrations ranging from 10 to 100 µM. A different inhibitory profile was observed when these fractions were tested on the anti-inflammatory mediators IL-1β, TNF-α, NO x , and IL-6. FDCM, rich in phenylpropanoid sucroses (m/z 613.1731, 655.1882, and 697.1997), fatty acids, and other phenolic compounds, inhibited the mediators IL-1β, TNF-α, and IL-6; its anti-inflammatory activity was presumably related to the presence of these phenolic glycosides. This conclusion is supported by former studies reporting similar metabolites with the same pharmacological property [30]. These arylpropanoid sucroses have also been described as inhibitors of aldose reductase, which is involved in various inflammatory disorders [17,18]. In fact, inhibition of aldose reductase might reduce reactive oxygen species and, therefore, prevent the inflammatory signals induced by cytokines and other factors [17,18]. Despite the presence of these metabolites in FNBU, no inhibition effect was observed on IL-1β, TNF-α, IL-6, and NO x levels. FNBU showed traces of metabolites at m/z 613.1731, 655.1882, and 697.1997 in its LCMS data, alongside other arylpropanoid sucroses (3-O-p-coumaroylsucrose, cistanoside F, and acetyl cistanoside F derivative), flavonoid derivatives, and pentacyclic triterpenes. These classes of metabolites are recognized to possess anti-inflammatory properties [31,32]. Therefore, the lack of anti-inflammatory activity of FNBU might have been due to the low concentrations of these components, which were not sufficient to produce the expected effect.
BDCM contained a stilbene derivative which, among other metabolites, inhibited only TNF-α. This fraction displayed a chemical profile different from those of FDCM and FNBU, and the lack of sucroses might be why this fraction showed a different inhibition profile. On the other hand, its effect on TNF-α level could have been associated with the presence of a stilbene, of which the analogues, such as resveratrol, are known to be inhibitors of TNF-α [33]. In addition, the presence in BDCM of coumaric acid methyl ester (7.36 min, m/z 177.0550) related to the aglycone of the arylpropanoid sucroses could also have contributed to the inhibition of TNF-α levels. The similarity of BNBU and FDCM relied on an unidentified phenolic, although BNBU was able to inhibit the increase of NO x and IL-6 levels. Considering the chemical profiles and the inhibition effects of FNBU and BDCM, iridoids and the phenolic derivative might have been responsible for the anti-inflammatory activity of BNBU. Iridoids structurally related to those found in BNBU, such as morroniside and geniposide, have been described as anti-inflammatory agents, and morroniside in particular is a NO x inhibitor [34]. The presence of a diffusoside derivative and 7-O-ethylmorroniside might support the observed anti-inflammatory activity of this fraction. Lucenin-2, a 6,8-di-C-glycosylated flavone, could also have contributed to this activity based on previous results describing its anti-inflammatory properties [35].
Since inflammatory cytokines are produced during parasite infections and these cytokines are also manifestly related to their virulence [12][13][14][15], this study also aimed to investigate whether fractions from banana blossom could exert antiparasitic effects against intracellular T. cruzi, L. amazonensis, and L. infantum amastigotes.
As human monocyte THP-1 cells were used as the macrophage, their viability when treated with the only active fraction (FDCM) was evaluated. This fraction was weakly cytotoxic to the THP-1 cell line. As observed with the neutrophil cells, FDCM, composed essentially of phenylpropanoid sucroses, required a high concentration to affect cell viability.
Only FDCM showed antitrypanosomal activity against the intracellular form of T. cruzi with a good selectivity index. It has been reported in the literature that human macrophages infected with T. cruzi display an increased level of MMP-9, which has a strong relationship with the production of inflammatory cytokines such as IL-1β, TNF-α, and IL-6 [14]. Therefore, the trypanocidal activity observed for FDCM might have had a relationship with its anti-inflammatory effect by inhibiting these three cytokines. In contrast to FDCM, which concomitantly inhibited three cytokines, BDCM solely inhibited the cytokine TNF-α and showed no effect against T. cruzi. This observation led to the conclusion that FDCM displayed antitrypanosomal activity, because it could reduce the levels of these three cytokines without inhibiting the level of NO x . It has been reported that nitrogen-derived species (NO x ) have a crucial role for the immune system by protecting cells against intracellular T. cruzi infection [36]. Therefore, the selective effect on the cytokines but not NO x is important for antitrypanosomal activity. Nitrogen oxide species chemically specifically modify cysteine-containing proteins in T. cruzi, and can potentially interact with the metalloproteins that mediate crucial metabolic processes [36]. This might support why BNBU did not show any trypanocidal activity. None of the fractions were active against the studied Leishmania species, indicating that the inhibition of IL-1β, TNF-α, and IL-6 cytokines and NO x species might affect the growth of Leishmania.
As this non-conventional food showed various biological benefits, it can be classified as a functional food, although more studies including toxicology and balanced diet studies need to be performed.

Plant Identification
The inflorescences of Musa acuminata were collected in Itacorubi/Florianópolis in March 2017. A voucher was deposited under the number RB 02574A in the Jardim Botanico (Botanical Garden) of Rio de Janeiro Herbarium (RB).

Mouse Neutrophil Isolation and Primary Culture
Mouse neutrophils were collected from mouse peritoneal leakage and maintained in Dulbecco's Modified Eagle Medium (DMEM) (Gibco, Grand Island, NY, USA) with 10% fetal bovine serum, 100 U/mL of penicillin, and 100 mg/mL of streptomycin incubated at 37 • C in a humidified CO 2 incubator. The peritoneal neutrophils were obtained after injection of oyster glycogen into the peritoneal mouse cavity, as described by Silva and co-workers with some modifications [37]. A total 3 mL of oyster glycogen at 1% (w/v) dissolved in sterile phosphate-buffered saline (PBS) was injected into the peritoneal mouse cavity, and after 4 h the animals were euthanized by overdose of xylazine and ketamine administered intravenously (i.v.). After euthanasia, 3 mL of sterile PBS was injected into the peritoneal cavity and the cavity was massaged for 10 s to suspend the neutrophils. An incision was made using sterile surgical material and the peritoneal leakage was collected in 50 mL sterile tubes and stored immediately in an ice bath. Furthermore, a pool of peritoneally collected neutrophils was made in order to obtain 1 × 10 6 neutrophils/well. A reduced number of animals were used with respect to the 3Rs (Replacement, Reduction and Refinement) principle [38]. The procedures were approved by the Committee for Ethics in Animal Research from UFSC (Protocol 8665141117) and were in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals.

Lipopolysaccharide Stimulation of Isolated Neutrophils
The neutrophils were preincubated after plate distribution with or without different concentrations (10,30, and 100 µg/mL) of the studied fractions for 1 h, and then the medium was exchanged with fresh DMEM mixed with lipopolysaccharide (LPS) at a final concentration of 5 µg/mL and incubated for 16 h at 37 • C in a CO 2 incubator (5%).

Cell Viability Assay Using the Isolated Neutrophilis
The extracts were added to each well at different final concentrations (10, 30, 100, 300, and 1000 µM) and incubated for 16 h at 37 • C in a CO 2 incubator (5%). This procedure was performed after the neutrophils were plated in a 96 well plate with DMEM culture medium enriched with 10% fetal bovine serum and 1% (w/w) penicillin/streptomycin. The entire experiment was conducted in triplicate and repeated on two different days of experimentation. The viability of the neutrophils after treating with the blossom fractions was evaluated using the colorimetric (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) MTT assay. The supernatant was discarded after incubation for 16 h and MTT solution (5 mg/mL) was added to each well, followed by incubation for a further 3 h at 37 • C in a CO 2 incubator (5%). The medium was then discarded again and dimethylsufoxide (DMSO) was added to dissolve the formazan dye. The optical density was checked at 540 nm using an ELISA reader (Infinite M200, Tecan, Männedorf, Switzerland).

Cell Inflammation Assay on Isolated Neutropils
In order to evaluate the effect of the standards and the fractions on inflamed ex vivo mouse neutrophils, cells were designated to different groups (n = 4/group) consisting of the following: (a) blank control (Ctrl, uninflamed neutrophils), cells treated only with vehicle; (b) negative control (LPS, lipopolysaccharide-inflamed neutrophils), cells stimulated only with LPS (5 µg/mL); (c) positive controls (dexamethasone: Dexa, reference anti-inflammatory drug treatment), cells pre-treated with Dexa (10 µM) and after 0.5 h stimulated with LPS (5 µg/mL); and (d) experimental groups (studied extracts), cells pre-treated with the extracts at 10, 30, and 100 µg/mL and stimulated after 0.5 h with LPS (5 µg/mL). All experimental groups were incubated for 16 h at 37 • C in a CO 2 atmosphere (5%). The supernatant was collected for further inflammatory analysis and comparisons (NO x , IL-1β, TNF-α, and IL-6).

Measurement of NO x Production in Neutrophils
The production of NO metabolites by mouse neutrophils (n = 10 per experiment) was determined using Griess reagent. Measures of 100 µL of the Griess reagent were mixed with 50 µL of cell supernatant and incubated for 40 min at 37 • C. Absorbance at 540 nm was measured with interpolation from the nitrite standard curve (0-20 µM), and the results are expressed in µM.
3.2.6. Quantification of Pro-Inflammatory Cytokines Levels (IL-1β, TNF-α, and IL-6) in Neutrophils The interleukin-1β (IL-1β), tumoral necrosis factor alpha (TNF-α), and interleukin 6 (IL-6) levels in the neutrophil supernatants were quantified as follows. The supernatant was removed and submitted to determination of the concentrations of IL-1β, TNF-α, and IL-6 using a commercially available enzyme-linked immunosorbent assay kit (Peprotech, Rocky Hill, NJ, USA) according to the manufacturer's instructions. Cytokine level was estimated by interpolation from the standard curve and the results are expressed in pg/mL.

Cell Viability Assay (MTT)
THP-1 cells were grown and cultivated in 96 well microplates (4.0 × 10 4 cells/well), treated with the compounds serially diluted in concentrations ranging from 2 µg/mL to 500 µg/mL, and incubated for 72 h at 37 • C, 5% CO 2 . The plates were centrifuged (3700×g/7 min), the supernatant was removed, and the cells were resuspended in 50 µL of a solution of MTT (Amresco) at 3 mg/mL in saline buffer and incubated for 4 h at 37 • C, 5% CO 2 before being centrifuged (3700×g/7 min), and the formazan salt was solubilized in 100 µL DMSO. Optical density was determined at 540 nm in a Tecan ® Infinite M200 spectrophotometer. DMSO 1% (v/v) and DMSO 50% (v/v) were the negative and positive controls, respectively. The IC 50 values were calculated by non-linear regression using the GraphPad Prism program [40].

Extraction, Fractionation, and Sample Preparation
The banana inflorescences were separated into bracts (228 g) and flowers (60 g), which were extracted in methanol (500 and 100 mL respectively). Both extractions furnished crude extracts of 70 mg from the petals and 40 mg from flowers. Each crude extract was diluted in water and separated of m/z 433.1710, Table S1: Chemical constituents of the flower fractions, Table S2: Chemical constituents of the bract fractions.