Humiria balsamifera extract inhibits nitric oxide and tumor necrosis factor production in LPS-stimulated macrophages

Humiria balsamifera is used in traditional medicine as anthelmintic, expectorant, to treat hepatitis, diarrhea, hemorrhoids; to cure chronic wounds; and to alleviate toothaches. This species occurs in Jurubatiba shoal, Rio de Janeiro state-Brazil, a rich region which offers a variety of promising bioactive product sources. The present study focuses on the chemical and pharmacological evaluation of H. balsamifera . The n -hexane, dichloromethane and ethyl acetate leaf fractions exhibited higher inhibitory potential on NO production. Friedelin (1), quercetin (2) and quercetin-3- α - O -arabinopyranoside (3) were isolated and characterized; the latter is described for the first time for H. balsamifera . Quercetin (2) showed the best inhibitory activity on NO production and moderate inhibition of TNF- α production. These results contribute to the knowledge of Humiria balsamifera as a source of anti-inflammatory compounds. Furthermore, the identification of the terpenes ß-amyrone, betulin, citronellol, eremophillene, dihydroactinolide and borneol, and the isolation of quercetin-3- α - O -arabinopyranoside are being reported for the first time for this species.

Introduction nflammation is part of the survival strategy of the organism, however, there may be an excessive or inappropriate inflammatory response causing injuries to the body (Dinarello 2010). Steroidal and non-steroidal anti-inflammatory drugs (NSAIDs) have been commonly used for treating inflammation. However, the use of these drugs are associated with several adverse effects (Harirforoosh et al. 2013;Ronchetti et al. 2018).
Humiria balsamifera is popularly used in the Amazon as a perfume and in the treatment of diseases such as hepatitis, diarrhea and hemorrhoids (Coelho-Ferreira 2009). Its use has also been described as anthelmintic, expectorant, healing action against toothache and chronic wounds (Lorenzi & Matos 2008). Considering some of these popular applications, such as for the treatment of hemorrhoids and toothache, it may be directly related to the modulation of the inflammatory process, so, the evaluation of this species in terms of its anti-inflammatory potential became relevant.
This species can be found in Jurubatiba shoal, RJ, Brazil. The shoal is a complex ecosystem in very delicate balance that have a typical flora, well adapted to characteristic conditions. The combination of several physical and chemical factors of this region, such as high temperature, salinity and high exposure to light (Cogliatti-Carvalho et al. 2001;Kelecom et al. 2002) makes it a promising source in the search for new bioactive products and evidencing the importance of its preservation.
Therefore, this study aims to investigate the chemical composition of this species, to evaluate the inhibitory activity on pro-inflammatory mediators produced by LPS-stimulated RAW 264.7 macrophages for the first time, and to isolate and identify compounds which can be related to this activity. This study intends to contribute to the knowledge related to the popular use of H. balsamifera associated with inflammation.

Preparation of the crude extracts
The plant material was properly separated into leaves and bark and then dried in an oven with air circulation for 72 h at 45 ºC. Dried leaves (224 g) and bark (66.6 g) were macerated at room temperature with ethanol for 7 days, and the resulting extractive alcoholic solutions were subsequently evaporated under vacuum to provide the extracts (39.80 g and 3.55 g, respectively). Fresh leaves (50 g) were also extracted with boiling distilled water (2 L) for 15 min. This material was then subjected to a second extraction step at low temperature (40 °C) for 30 min with magnetic stirring. The aqueous extract (AqEx, 4.2 g) was lyophilized for quantification (Santos et al. 2020).

General experimental conditions GC-MS
The fractions were analyzed by GC-MS in a gas chromatograph (Shimadzu 2010) with a GCMS-QP2010 interface with electronic ionization energy of 70 eV, split ratio 1/20; mass range of m/z 30-500 D, and a scanning time of 1 s. We used a RTx-5MS column (30 m × 0.25 mm × 0.25 µm film; Restek Corporation, USA). The following conditions were adopted: helium as the carrier gas (1.1 mL min -1 ); temperature setting from 60 to 280 °C (15 °C min -1 ); injector temperature (250 °C) and interface (300 °C). The samples were prepared by adding 200 μL MSTFA in 2 mg of fraction. Each sample was dissolved in 2 ml of chloroform after 15 minutes, and 1 μl was then injected with an autosampler. Compound identification was based on comparing the mass spectra and the NIST spectrometer data bank together with a literature data comparison.

Biological assay Inhibition of NO production by LPS-stimulated macrophages
The murine macrophage cell line RAW 264.7 was obtained from the American Type Culture Collection (ATCC), grown at 37 °C and 5% CO 2 in DMEM/F-12 supplemented with 10% FBS. Macrophages (10 5 cells/mL) were seeded in 96well plates and incubated for 24 h under 5% CO 2 37 °C. Cells were treated with samples (0.8, 4, 20 and 100 µg/mL) and stimulated with 1 µg/mL LPS. Culture supernatants were collected for NO and TNF-α assays after 24 h (TNF-α assay as described below). Nitrite concentration was determined as an indicator of NO production, according to the Griess test (Griess 1879). Absorbance was spectrophotometrically measured at 540 nm. Nitrite concentration was calculated by comparison with a sodium nitrite standard curve.

Cytoxicity
The cytotoxic effects of samples on RAW 264.7 stimulated with LPS were determined using the lactate dehydrogenase (LDH) release assay (for extacts and fractions) and using the MTT assay (for isolated compounds). In both cases, the cytotoxicity percentage was calculated in relation to the negative control (untreated and LPS-stimulated macrophages), and to the positive control (stimulated macrophages treated with 1% (v/v) Triton X-100).

LDH assay
Cytoplasmic enzyme lactate dehydrogenase (LDH) release was determined using 50 µL of culture supernatant collected at the end of the assay (Muzitano et al. 2006). The LDH amount was colorimetrically verified using a commercial kit (Doles ® ).

MTT assay
3-(4,5-dimethylthiazol-2yl-)-2,5-diphenyl tetrazolium bromide (MTT) (Mosmann 1983) was used to evaluate the viability of the cells exposed to the isolated substances. The adherent macrophages remaining were treated with 10 µL of MTT (5 mg/mL), and further incubated for 2 h at 37 °C and 5% CO 2 . The viable cells reduced the MTT into formazan crystals, which were dissolved in HCl (4 mM) in isopropanol. This reduction was spectrophotometrically estimated at 570 nm to determine cell viability.

TNF-α assay
TNF-α was measured by a L929 fibroblast bioassay (Shiau et al. 2001), based on the sensitivity of L929 cells to the cytotoxic effect of TNF-α. First, L929 cells were seeded at a density of 2.5×10 5 cells/mL in a 96-well plate and incubated for 24 h in DMEM/F-12 with 10% FBS and 20 µg/mL of gentamicin at 37 °C and 5% CO 2 . The resulting cell monolayers were treated with the macrophage culture supernatants in the presence of actinomycin D (2 μg/mL) after 24 h of incubation. L929 viability was determined by the MTT method (Mosmann 1983). The reading was performed at 570 nm on a spectrophotometer plate. The cell viability percentage was determined using the culture supernatant of nonstimulated macrophages (negative control TNF-α production) and the culture supernatant of stimulated macrophages (positive control TNF-α production) as controls. A standard curve with recombinant mouse TNF-α was used to measure the TNF-α concentration found in the samples.

Statistical analysis
All experiments were performed in triplicate and the results were expressed as mean±standard error (M±SEM). The results were representative of three independent experiments. Statistical analyses were performed by one-way ANOVA followed by a Tukey's post-test. The results were considered statistically significant for p < 0.05. The IC 50 was determined by non-linear regression based on the results of the concentration-response curve.

Results and Discussion
Extracts and fractions analysis by chromatographic techniques and corresponding activities No significant differences (P > 0.05) were observed between the leaf and bark ethanolic extracts, nor for the AqEx (decoction), concerning NO and TNF-α inhibition, as well as cytotoxicity in LPS-stimulated RAW 264.7 cells (IC 50 >100 µg/ mL). The HEX, DCL and EtOAc Ppt leaf fractions exhibited the highest NO inhibition, with IC 50 of 9.26±0.83, 15.43±0.63 and 19.34±0.70 µg/mL, respectively. The DCL fraction and EtOAc Ppt. displayed the greatest promising TNF-α inhibition, with IC 50 equal to 44.92±0.84 and 52.52 ± 0.96 µg/ mL, respectively. Among the most active fractions, a cytotoxic effect was also verified (78.79±1.03 µg/ ml) for the DCL fraction, indicating that the activity described above is probably non-specific (Tab. 1).
Rodriguésia 72: e00292020. 2021 These results are being described for the first time to extracts and fractions from this species, adding an important pharmacological interest.
The following crude extracts EBC, EBF and AqEx were evaluated by HPLC-DAD and showed similar chromatographic profiles (Fig. 1), with compounds in T R less than 15.0 min (a -T R 10.2 min and b -T R 14.5 min) with UV suggestive of phenolic compounds such as bergenin (approx. 215 and 270 nm) (Qin et al. 2010), already described for this specie (Silva et al. 2004). There were also observed compounds with T R between 20.0 and 27.0 min (highlighting c -T R 24.9 min) showing to be compatible to flavonoid derivatives (approx. 205, 254 and 354 nm) (De Rijke et al. 2006). These results can justify the fact that the described extracts showed similar behavior when evaluated in NO and TNF-α inhibition, and in cytotoxicity test in LPS-stimulated RAW 264.7 cells. The ethyl acetate supernatant (EtOAc Sup.) when analyzed by HPLC-DAD also showed bergenin (T R 14.5 min) and the same flavonoid derivatives (T R 20.0 to 27.0 min) in a considerable higher amount compared to the extracts. While in the ethyl acetate precipitate (EtOAc Ppt.) it was observed two major compounds with T R 14.5 (a) and 24.9 min (b) (Fig. 1), compatible to bergenin and a flavonoid derivative, respectively.
In the dichloromethane fraction, other major compounds were identified in addition to friedelin and hexadecanoic acid ( Figure S2, available on supplementary material <https://doi.org/10.6084/ m9.figshare.16569393.v1>; Tab. 3). Among these compounds, the presence of some terpenes is highlighted, such as citronellol, eremophillene, dihydroactinolide and borneol, the last as the majority compound of this fraction (38.57% peak purity). This factor may be relevant to justify the observed activity since this compound has already been described for its anti-inflammatory potential (Zhong et al. 2014;Zou et al. 2017). In addition, the ability of borneol to increase cell uptake of other substances is also reported, which would lead to increased apoptosis in tumor cells (Su et al. 2013). This could contribute to understand the cytotoxic profile presented by this fraction against Raw 264.7 macrophages, which could be associated with an increase in cellular uptake of the other compounds present in this fraction. To sum up, the EtOAc Sup. and the HEX fractions showed to be the most actives, with less citotoxicity, which are in accordance with the literature, that has been described the antiinflammatory potential of some of the derivatives found in those fractions, such as the hexadecanoic acid (Aparna et al. 2012), phenolic (Gao et al. 2015Maleki et al. 2019) and terpenes derivatives, including betulin (Reyes et al. 2006;Liu et al. 2019), friedelin (Jin et al. 2018) and amyrone (Almeida et al. 2015). Some of the identified compounds have already been described for species from the same family Humiriaceae, such as bergenin in Sacoglottis gabonensis (Ogan 1971), and hexadecanoic acid, friedelin, betulin and bergenin identified in Endopleura uchi (Marx et al. 2002;Abreu et al. 2013), being bergenin associated to the anti-inflammatory activitiy of the plant (Nunomura et al. 2009). It is important to highlight that it is the first time that tetratetracontane, ß-amyrone, betulin, citronellol, eremophillene, dihydroactinolide and borneol are being reported for this species and also for the genus Humiria.  (Shiojima et al. 1992). 1 H-NMR (500 MHz, CDCl 3 ) showed signals between 0.72 and 2.49 ppm characteristic of the triterpene friedelin, which had already been earlier isolated from H. balsamifera (Silva et al. 2004 1H, m, H-18). HSQC (500 MHz, CDCl 3 ) indicated a characteristic correlation of a friedelane ring, a H-23 methyl group with C-23 in the most protected region of the spectra (δc 6.9) due to effect caused by the carbonyl group (C-3, δc 213.1). HMBC (500 MHz, CDCl 3 ) showed seven quaternary carbons δ C 213.1 (C-3), 42.0 (C-5), 37.6 (C-9), 39.7 (C-13), 38.3 (C-14), 29.9 (C-17), 28.2 (C-20). Characteristic correlations to the friedelan ring as H-23/C-3 and H-4/C-3 were also observed in this spectrum. The correlation between the H-4 and H-23 was verified in COSY spectrum (500 MHz, CDCl 3 ). These data are in agreement with literature (Queiroga et al. 2000;Almeida et al. 2011).
Compound 3 (70%): HPLC/UV-DAD 25.49 min and λ máx 202, 255 and 355 nm in the UV spectra (Fig. 3). Spectroscopy data showed signals characteristic of quercetin-3-Oα-arabinopyranoside (Wollenweber et al. 1997;Ahmadu et al. 2007     NO and TNF-α inhibition and cytotoxicity of isolated compounds in LPS-stimulated RAW 264.7 cells Quercetin (2) presented the highest potential in NO (IC 50 4.58±0.95 µg/mL) and TNF-α (IC 50 33.25±0.88 µg/mL) inhibition. In addition, this compound was more active than the L-NMMA inhibitor and showed no cytotoxic effect up to a concentration of 20 µg/mL, in which excellent inhibitory activity was observed for NO production     (84.87±1.47%). Friedelin (1) and quercetin-3-O-α-arabinopyranosyl (3) demonstrated no significant NO and TNF-α inhibition (IC 50 > 100 µg/mL). Besides, we observed low cytotoxicity for compounds 1 and 3 (Tab. 4). Friedelin (1) has already been briefly described in the literature for its anti-inflammatory activity. Friedelin isolated from the stems of Heritiera littoralis Aiton. (Malvaceae) exhibited IC 50 > 100 μM (approximately 42.67 μg/mL) in the NO inhibition (Tewtrakul et al. 2010), a result which can be considered compatible with the observed in this study (at 100 μg/mL; 24.23 ± 2.41% inhibition). This substance was isolated from the leaves and twigs of Acer mandshuricum Maxim. (Aceraceae), being evaluated at 100 nM (approximately 0.04 μg/mL) and inhibiting 23.5% of TNF-α release (Ding et al. 2010). A slightly reduced TNF-α inhibition percentage (14.98 ± 0.20%) was verified in this study at 0.8 μg/mL and, in addition to that, it was also observed that even with the increase of the evaluated concentrations, no increase in the inhibitory effect was noted, suggesting that it may be the maximum effect caused by friedelin.
Quercetin-3-O-α-arabinopyranoside (3) is herein described for the first time in H. balsamifera extracts. The non-expressive effects of this compound on NO and TNF-α production compared with the correspondent aglicone quercetin agree with literature data. Quercetin-3-O-α-arabinopyranoside isolated from guava leaves (Kim et al. 2015) and Acer tegmentosum Maxim. (Aceraceae) (Lee et al. 2014) evaluated at 10-100 μM demonstrated no significant NO and TNF-α inhibition in RAW 264.7 cells. This compound similarly isolated from Psidium acutangulum DC. (Myrtaceae) did not inhibit NO when tested at 50 μg/mL (Houël et al. 2016). A higher inhibitory activity of pro-inflammatory mediators for quercetin (2) compared to quercetin-3-O-α-arabinopyranoside (3) was also observed in the present study. Therefore, it can be suggested that glycosylation leads to an increase in the polarity and molecular size, and hence may alter the interaction with the target molecule and its passage through the phospholipidic membrane, modifying the possible activity on intracellular sites.
Other activities previously described in the literature for these compounds such as analgesic, antipyretic, antimicrobial, and wound healing (Antonisamy et al. 2011;Maalik et al. 2014;Hatahet et al. 2016;Noundou et al. 2016;Sa et al. 2017;Özbilgin et al. 2018;Singh et al. 2018) may contribute to justify the use of this species in traditional medicine to treat hepatitis, diarrhea, hemorrhoids, in curing chronic wounds, and alleviating toothaches.
In summary, this study describes the NO and TNF-α inhibition by H. balsamifera, the isolation of quercetin-3-O-α-arabinopyranoside and identification of terpenes ß-amyrone, betulin, citronellol, eremophillene, dihydroactinolide and borneol for the first time in this species. Thus, promising results for a species of Jurubatiba shoal are demonstrated in this work, evidencing this region as being a relevant source in the search for new bioactive products. These aspects reinforce the importance of biodiversity preservation of Jurubatiba shoal. The results also strengthen the proposal that inhibition of NO and TNF-α production might be a useful screening strategy in searching for new anti-inflammatory compounds, especially by using medicinal plant extracts.