Cytotoxic and Anti-Inflammatory Activities of Dihydroisocoumarin and Xanthone Derivatives from Garcinia picrorhiza

Garcinia picrorhiza, a woody plant native to Sulawesi and Maluku Islands, Indonesia, has been traditionally used as a wound healing ointment. In our continuous search for bioactive compounds from this plant, 15 phenolic compounds were isolated from its stem bark, including a previously undescribed dihydroisocoumarin, 2′-hydroxyannulatomarin, and two undescribed furanoxanthones, gerontoxanthone C hydrate and 3′-hydroxycalothorexanthone. The structures of the new metabolites were elucidated on the basis of spectroscopic analysis, including 1D and 2D NMR and HRESIMS. Gerontoxanthone C hydrate possessed cytotoxicity against four cancer cells (KB, HeLa S3, MCF-7, and Hep G2) with IC50 values ranging from 5.6 to 7.5 µM. Investigation on the anti-inflammatory activities showed that 3′-hydroxycalothorexanthone inhibited NO production in RAW 264.7 and BV-2 cell lines with IC50 values of 16.4 and 13.8 µM, respectively, whereas only (−)-annulatomarin possessed inhibition activity on COX-2 enzyme over 10% at 20 µM. This work describes the presence of 3,4-dihydroisocoumarin structures with a phenyl ring substituent at C-3, which are reported the first time in genus Garcinia. These findings also suggest the potential of furanxanthone derivatives as cytotoxic and anti-inflammatory agents for further pharmacological studies.


Introduction
The genus Garcinia, one of the largest genera to the family Clusiaceae, is widespread in tropical rain forests with the center points are in Southeast Asia and Madagascar. This genus is pantropical and comprises a high level of species diversity with more than 250 species of evergreen, lactiferous, dioecious, and small shrubs to medium-sized trees [1,2]. Phytochemical analysis of genus Garcinia resulted in the isolation of structurally diverse secondary metabolites, especially phenolic compounds, such as xanthones, polyprenylated benzoylphloroglucinols, biphenyls, depsidones, and biflavonoids and many of which gained great attention because of their biological and pharmacological activities [3]. For example, a caged xanthone named gambogic acid (GA) from Garcinia hanburyi, has completed its phase IIa clinical trial in China for patients with lung, colon, and renal cancers. GA has been proved to induce apoptosis, reverse multidrug resistance of cancer cells, inhibit cell proliferation, and possess anti-angiogenic activities [4].
Garcinia picrorhiza Miq. is a woody plant that can be found in Maluku Islands, Indonesia. The roots and latex have been traditionally used as a wound healing ointment and natural energy drink [5][6][7]. Previously, our group reported eleven polyprenylated benzoylphloroglucinols (PPBPs), including eight derivatives (Picrorhizone A-H) bearing a cyclobutane-containing side chain, with cytotoxic and anti-inflammatory activities. In particular, picrorhizone F was active against four human cancer cells (KB, HeLa S3, MCF-7, and Hep G2) with IC 50 values in the range of 5.9-9.4 µM, while picrorhizone H exhibited the highest COX-1 inhibitory activity (35.2 ± 9.6% inhibition) at 20 µM [8]. Encouraged by structurally diverse bioactive compounds from Garcinia species [8][9][10][11][12], we decided to revisit G. picrorhiza in part of a comprehensive phytochemical and biological investigation. Herein, we report the isolation, structural elucidation, and cytotoxic and anti-inflammatory effects of the phenolic compounds from the stem bark of G. picrorhiza.
Compound 5 was isolated as a pale yellow amorphous solid and the HRESIMS data showed a [M + H] + peak at m/z 345.0971, indicative of a molecular formula C 18 H 16 O 7 . A comparison of the NMR data of 5 with those of calothorexanthone [31] showed that the difference was restricted to the trimethylfuran moiety in ring A. An oxymethylene proton at δ H 3.91 (d, J = 5.6 Hz, H-3 ) was observed in 5 instead of a methyl proton resonance as in the reference ( Table 1). The COSY cross-peak of H-2 /H-3 and HMBC correlations of H-3 to C-1 and C-2 showed that the methyl group at C-2 was oxidized to be a hydroxymethyl unit (Figure 2), which was supported by its HRESIMS analysis indicating one more oxygen atom than calothorexanthone. Unfortunately, the ECD spectrum of 5 could not be obtained due to the lack of sample for analysis and the positive optical rotation value ([α] 20 D = +22.9) was used to determine 2 S configuration in 5 [30].
Among the known metabolites, compounds 1, 2, 8, 11, and 15 were isolated for the first time from genus Garcinia. To the best of our knowledge, compounds 1-3 with the 3-phenyl-3,4-dihydroisocoumarin structure are the first occurrence in genus Garcinia, while earlier works reported its analogs containing an alkyl chain at C-3 instead of a phenyl ring, such as (-)-mellein from G. bancana [32] and angelicoin B from G. xanthochymus [33]. Compounds 1 and 2 were previously obtained from the plant extracts of Montrouziera sephaeroidea, Hypericum annulatum, Cratoxylum sumatranum ssp. neriifolium, and some Dioscorea species [15,[34][35][36]. This class of compounds is common metabolites in fungi, lichens, and bacteria, while their existence in higher plants is limited [37].

Cytotoxic Activity against Human Cancer Cell Lines
The cytotoxic activity of the isolated compounds was evaluated using the MTT viability assay [38] with doxorubicin as the positive control and the results are presented in Table 2. The active compounds having IC 50 values lower than 30 µM against KB and HeLa S3 cell lines were further tested against the other three cancer cells (MCF-7, Hep G2, and HT-29). None of the dihydroisocoumarins 1-3 possessed potent cytotoxicity with IC 50 values over 30 µM. Among the group of xanthones, compound 4 was active against four cancer cells (KB, HeLa S3, MCF-7, and Hep G2) with IC 50 values ranging from 5.6 to 7.5 µM. Our earlier studies [11,39] reported that furanoxanthones 7 and 8 also significantly inhibited the growth of KB, HeLa S3, MCF-7, and Hep G2 cancer cells with IC 50 values less than 10 µM, while decreased cytotoxicity was observed for compound 6 with only one hydroxy group in the ring B of the xanthone skeleton. The hydration of the prenyl unit in 8 did not affect cytotoxic properties against the four cancer cells, as shown in schomburgone F [39]. These results suggested that trimethylfuran ring and ortho hydroxy unit might be required to enhance the cytotoxicity compared to the other isolated xanthones.

Nitric Oxide Inhibitory Activity
Compounds 1-8 were screened for their inhibitory effects on nitric oxide production in LPS-IFN-γ activated RAW 264.7 macrophages and BV-2 microglial cells at a final concentration of 50 µM. The preliminary results (Figure 3) indicated that dihydroisocoumarins 1-3 were weak to inactive on NO inhibitory activity, whereas 4, 5, and 7 suppressed NO production over 70% in both cell lines. A deprenylated furanoxanthone structure with ortho hydroxy group in ring B as in 7 might be important to enhance the inhibitory effect when compared to 6 and 8. None of the tested compounds showed obvious cytotoxicity towards RAW 264.7 cells (cell viability >90%), while the survival of BV-2 cells was reduced with cell viability in the range of 41.8-81.5% after compound treatment at 50 µM ( Figure 3). IC 50 values of the three active compounds are listed in Table 3. Compound 5 possessed the strongest inhibitory activity against NO production in RAW 264.7 and BV-2 cells with IC 50 values of 16.4 and 13.8 µM, respectively. Compounds 4 and 7 moderately inhibited NO production in RAW 264.7 cells, whereas their anti-inflammatory effect in BV-2 cells might be masked cytotoxicity (IC 50 values of their NO inhibition were only 1.4 and 1.1-fold lower than the cytotoxic activity).

Cyclooxygenase (COX) Inhibitory Activity
The anti-inflammatory effects of the isolated compounds were also screened on COX-1 and COX-2 enzymes inhibition at a final concentration of 20 µM. The results are given in Table 4. Only compound 2 exhibited activity against COX-2 with 10.4% inhibition. All furanoxanthones (4-8), except 5 with oxygenated trimethylfuran moiety, showed COX-1 inhibition higher than 10% with the new compound 4 possessing the strongest activity (32.4 ± 7.9% inhibition). Removal of the hydroxy group at C-2 in 9 reduced the COX-1 inhibition when compared with 10 and 11, while introducing a methoxy group at C-3 in 14 increased the inhibitory effect compared to 12 and 13.

General Experimental Procedures
Optical rotations were measured on a JASCO P-1010 polarimeter (JASCO, Easton, MD, USA). The experimental ECD data were recorded on a JASCO J-815 circular dichroism spectrometer (JASCO, Easton, MD, USA). The IR data were obtained with a Nicolet 6700 FT-IR spectrometer using an ATR technique (Thermo Fisher Scientific, Waltham, MA, USA). The NMR spectra were acquired on a Bruker 400 AVANCE spectrometer in acetone-d 6 (Merck, Darmstadt, Germany). The HRMS spectra were recorded using a Bruker MICROTOF model mass spectrometer (Bruker, Billerica, MA, USA) and Dionex Ultimate 3000 HPLC system hyphenated with a QExactive Hybrid Quadrupole Orbitrap MS (Thermo Fisher Scientific). Silica gel 70-230 mesh (Merck) and Sephadex LH-20 (GE Healthcare Bio-Sciences AB, Uppsala, Sweden) were used for column chromatography. Radial chromatography (Chromatotron model 7924 T, Harrison Research, Palo Alto, CA, USA) was carried out with silica gel 60 GF254 containing gypsum (Merck).
The cancer cells were cultured in MEM containing 10% fetal bovine serum in the presence of 100 U/mL penicillin and 100 µg/mL streptomycin sulphate (Gibco, Rockville, MD, USA), seeded in a 96-well plate (3000 cells/well), and pre-incubated at 37 • C for 24 h in a 5% CO 2 humidified atmosphere. The tested compounds with serial concentrations (0.3-100 µM) were added and incubated for a further 72 h in the same condition. At the end of treatment, 20 µL of MTT solution (5 mg/mL in PBS, Sigma, St. Louis, MI, USA) was added to each well and further incubated for 3 h. After centrifugation at 1400 rpm for 5 min at 4 • C, the supernatant was decanted and DMSO (100 µL/well) was added to allow formazan product solubilization, which was subsequently measured by a microplate reader (Tecan Trading AG, Switzerland) at wavelength 550 nm. Control cells were treated with 0.1% DMSO. Doxorubicin (tested concentrations: 0.01-3.0 µM, Sigma) was used as the positive control. The results are expressed as the mean values of three replicates for three independent experiments. IC 50 values were determined by graphical analysis using SigmaPlot 10 (Systat Software Inc., San Jose, CA, USA), obtained by plotting the cell viability percentage versus sample concentration.

NO Production Inhibition Assay
The iNOS inhibitory activity of the isolated compounds and respective cytotoxicity was determined using a previously described protocol [40]. Briefly, BV-2 mouse microglial (25,000 cells/well) and RAW 264.7 mouse macrophage (100,000 cells/well) cell lines were seeded into 96-well plates in high glucose Dulbecco's Modified Eagle's Medium (DMEM, Gibco) containing 4 mM L-glutamine, 10% FBS, 2% HEPES, 1% penicillin/streptomycin, and 0.5% amphotericin B. After incubation for 24 h at 37 • C in a 5% CO 2 humidified atmosphere, the cells were stimulated with 0.05% LPS (Sigma) and 0.025% interferon-γ (Roche Diagnostics, Mannheim, Germany) and treated with different concentrations (5-100 µM) of the tested compounds. The level of NO production in cell culture supernatants was determined using Griess reagent (Sigma) by photocolorimetric analysis after 16 h of incubation, using a Hidex Sense Microplate Reader (Hidex, Turku, Finland). Fifty microliters of XTT reagent (Sigma, Kit II) was then added into the same plates for cell viability measurement after 60 min incubation. L-NMMA (Alexis, Grünberg, Germany) 100 µM was used as the positive control, while 1% DMSO and digitonin (Roth, Karlsruhe, Germany) 400 µM were used as the vehicle and negative control, respectively. The assays were conducted in triplicates and repeated in two independent experiments. IC 50 values were determined by graphical analysis using GraphPad Prism 7 (GraphPad Software, La Jolla, CA, USA).

COX Enzymes Inhibition Assay
Purified PGHS-1 from ram seminal vesicles for COX-1 and human recombinant PGHS-2 for COX-2 (Cayman Chemical Co., Ann Arbor, MI, USA) were used for the COX inhibition assays. Experiments were performed in a 96-well plate format, as previously described [40,41]. Briefly, the incubation mixture contained 180 µL of 0.1 M TRIS/HClbuffer pH 8.0 (Roth, Karlsruhe, Germany), 50 µM Na 2 EDTA (only for COX-2, Titriplex III, Merck, Darmstadt, Germany), 18 mM L-epinepherine bitartrate (Fluka, Buchs, Switzerland), 5 µM porcine hematin (MP Biomedicals LLC, Solon, OH, USA), and COX-1 or COX-2 enzymes (0.2 U/well). Ten microliters of tested compounds (dissolved in DMSO with a final concentration of 20 µM) was added and the mixture was pre-incubated for 5 min at room temperature. Thereafter, 10 µL of 5 µM arachidonic acid (Ann Arbor, Michigan USA) was added to the mixture and incubated for 20 min at 37 • C. The reaction was subsequently stopped by adding 10 µL of 10% (v/v) formic acid. The concentration of PGE 2 generated in the reaction was quantitatively measured by a competitive PGE 2 ELISA kit (Enzo Life Sciences Inc., Farmingdale, NY, USA) according to the manufacturer's protocol and previous procedure [40]. The absorbance at 405 nm was measured using a Hidex Sense Microplate Reader (Hidex, Turku, Finland). Indomethacin (1.25 µM, dissolved in ethanol; MP Biomedicals, Solon, OH, USA) and celecoxib (8.8 µM, dissolved in DMSO; Sigma) were used as a positive control for COX-1 and COX-2, respectively. The assays were performed in duplicates and repeated in two independent experiments.

Conclusions
Phytochemical investigation of the stem bark of G. picrorhiza led to the isolation of 15 dihydroisocoumarins and xanthones, including three new analogs. The presence of dihydroisocoumarin structures with a phenyl moiety at C-3 is expected to enrich information about the chemical diversity of the genus Garcinia. This work also indicates that furanoxanthone structures are promising candidates and could be used as a template for discovering potential anticancer and anti-inflammatory agents.