6-Bromoindole Derivatives from the Icelandic Marine Sponge Geodia barretti: Isolation and Anti-Inflammatory Activity

An UPLC-qTOF-MS-based dereplication study led to the targeted isolation of seven bromoindole alkaloids from the sub-Arctic sponge Geodia barretti. This includes three new metabolites, namely geobarrettin A–C (1–3) and four known compounds, barettin (4), 8,9-dihydrobarettin (5), 6-bromoconicamin (6), and l-6-bromohypaphorine (7). The chemical structures of compounds 1–7 were elucidated by extensive analysis of the NMR and HRESIMS data. The absolute stereochemistry of geobarrettin A (1) was assigned by ECD analysis and Marfey’s method employing the new reagent l-Nα-(1-fluoro-2,4-dinitrophenyl)tryptophanamide (l-FDTA). The isolated compounds were screened for anti-inflammatory activity using human dendritic cells (DCs). Both 2 and 3 reduced DC secretion of IL-12p40, but 3 concomitantly increased IL-10 production. Maturing DCs treated with 2 or 3 before co-culturing with allogeneic CD4+ T cells decreased T cell secretion of IFN-γ, indicating a reduction in Th1 differentiation. Although barettin (4) reduced DC secretion of IL-12p40 and IL-10 (IC50 values 11.8 and 21.0 μM for IL-10 and IL-12p40, respectively), maturing DCs in the presence of 4 did not affect the ability of T cells to secrete IFN-γ or IL-17, but reduced their secretion of IL-10. These results indicate that 2 and 3 may be useful for the treatment of inflammation, mainly of the Th1 type.


Introduction
Many chronic illnesses, including cancer, neurological diseases, diabetes, and autoimmune diseases, exhibit dysregulation of pathways that have been linked to inflammation [1,2]. A vast number of unique marine natural products possessing in vitro and in vivo anti-inflammatory activity

Structural Elucidation
Compound 1 was obtained as a yellow solid. Its UV spectrum was characteristic for an oxindole skeleton with absorption λ max at 217 and 263 nm [31,32], while the IR spectrum showed the presence of OH (ν max 3352 cm −1 ), NH 2 (ν max 3285, 3214, and 1614 cm −1 ), and lactam (ν max 1679 cm −1 ) functionalities. The characteristic isotope pseudo-molecular ion peaks at m/z 451. 0728 , suggesting the presence of a trisubstituted benzene ring. The chemical shift of the quaternary aromatic carbon C-7a (δ C 143.9) indicated that C-7a was substituted by an NH group. The characteristic resonances of the amide carbonyl C-2 at δ C 179.3 and the oxygenated carbon (C-3, δ C 78.3), as well as the HMBC correlations observed from H-4 to C-3, C-6, and C-7a, and from H-7 to C-3a, assisted the construction of a 3-hydroxy-2-oxindole skeleton. The chemical shifts for carbon atoms were in good agreement with those of previously published 3-hydroxy-2-oxindole scaffolds [33,34]. The position of the bromine atom at C-6 was determined based on the HMBC correlations H-4/C-3, H-4/C-6, and H-4/C-7a; H-5/C-3a and H-5/C-7; as well as H-7/C-3a, H-7/C-5, and H-7/C-6. This was supported by the NOE correlations between H-4/H-8 observed in spectrum run in CD 3 OD and a weak cross peak between H-7/NH-1 in spectrum run in DMSO-d 6 , in combination with characteristic coupling constant values, namely J 4,5 (7.9 Hz) and J 5,7 (1.5 Hz). A similar weak cross peak between H-7/NH-1 has previously been observed for 6-bromoindole cyclic guanidine alkaloids [35]. The presence of two amide carbonyl signals C-11 (δ C 167.2) and C-14 (δ C 160.5) suggested the presence of a DKP moiety attached to the oxindole skeleton and the chemical shift of C-14 indicated that this carbon was conjugated to a double bond (∆ 8  , t, J = 6.0 Hz). The molecular fragments described above account for all atoms except three nitrogen atoms and one quaternary carbon atom (δ C 158.7, C-19) and for 10 of 11 degrees of unsaturation. The position of the remaining three nitrogen atoms and the carbon atom was assigned to a guanidine group connected to the linear spin system at C-17 based on the HMBC correlation between H 2 -17/C-19 ( Figure 2). This suggested that the aliphatic side chain linked to the DKP moiety was an Arg residue. Further examination of the NMR data revealed that spectroscopic data of compound 1 were very similar to those of compound 4 [10], with the main differences being the upfield shift of the H-7 resonance (δ H 7.03 (1H d, J = 1.5 Hz) in compound 1; δ H 7.67 (1H d, J = 1.6 Hz) in compound 4) and the replacement of the indole ∆ 2 double bond by a hydroxyl group at C-3 and a carbonyl group at C-2. The latter was confirmed by the HMBC correlation observed between H-8 and C-2 ( Figure 2 Further examination of the NMR data revealed that spectroscopic data of compound 1 were very similar to those of compound 4 [10], with the main differences being the upfield shift of the H-7 resonance (δH 7.03 (1H d, J = 1.5 Hz) in compound 1; δH 7.67 (1H d, J = 1.6 Hz) in compound 4) and the replacement of the indole Δ 2 double bond by a hydroxyl group at C-3 and a carbonyl group at C-2. The latter was confirmed by the HMBC correlation observed between H-8 and C-2 ( Figure 2    The configurations of two chiral centers at C-3 and C-12 in compound 1 were determined using a combination of techniques. The absolute configuration of C-3 was determined by comparison of the ECD spectra ( Figure 3) of the hydrogenolysis product of compound 1 with the compound (R)-3-propyldioxindole (8), which share the dioxindole core structure. (R)-8 was obtained by hydrogenolysis of R-dioxindole (8a, a gift from Dr. Annaliese K. Franz (University of California, Davis)) (Scheme 1) [37]. The CD spectra of (R)-8 (EtOH, 23 • C) revealed Cotton effects (λ 208 nm (∆ε +13.1), 232 (−10.7), 259 (+3.5)) similar to dioxindoles recently investigated [38]. Hydrogenolysis of geobarretin A (1) (Scheme 1) gave debromodihydrogeobarrettin A (1a) with a CD spectrum identical to that of (R)-8 ( Figure 3). Given the identical CD spectra of 1a and (R)-8, whose absolute configuration was established by X-ray crystallography, and the similarity of their Cotton effects with those of (R)-5-methyldioxindole [39] and the natural product (+)-trikentramide I [38], the C-3 configuration of compound 1 was assigned as R, unambiguously. The configurations of two chiral centers at C-3 and C-12 in compound 1 were determined using a combination of techniques. The absolute configuration of C-3 was determined by comparison of the ECD spectra (Figure 3) of the hydrogenolysis product of compound 1 with the compound (R)-3propyldioxindole (8), which share the dioxindole core structure. (R)-8 was obtained by hydrogenolysis of R-dioxindole (8a, a gift from Dr. Annaliese K. Franz (University of California, Davis)) (Scheme 1) [37]. The CD spectra of (R)-8 (EtOH, 23 °C) revealed Cotton effects (λ 208 nm (Δε +13.1), 232 (−10.7), 259 (+3.5)) similar to dioxindoles recently investigated [38]. Hydrogenolysis of geobarretin A (1) (Scheme 1) gave debromodihydrogeobarrettin A (1a) with a CD spectrum identical to that of (R)-8 ( Figure 3). Given the identical CD spectra of 1a and (R)-8, whose absolute configuration was established by X-ray crystallography, and the similarity of their Cotton effects with those of (R)-5-methyldioxindole [39] and the natural product (+)-trikentramide I [38], the C-3 configuration of compound 1 was assigned as R, unambiguously.
Determination of amino acid configuration using Marfey's method and chiral derivatizing agents (CDAs) such as L-fluoro-2-4-dinitrophenyl-5-L-alanine amide (FDAA, Marfey's reagent) or the homolog, L-fluoro-2,4-dinitrophenyl-5-L-leucineamide (FDLA), are popular and effective [40]. Compound 1a was hydrolyzed (6 M HCl, 110 • C) and the product was derivatized with the new CDA, L-FDTA [41], to produce 9. Comparison with standards D-and L-Arg-DTA derivatives was then achieved using HPLC ( Figure S32) and UPLC-MS analyses. D-and L-Arg-DTA derivatives were observed with integral values of 19:81 suggesting either natural 1 is partially racemic at C-12, or that Arg underwent partial racemization under the conditions of hydrolysis. The latter possibility is ruled out by precedence; Arg essentially is stable to the conditions of hydrolysis. In a careful systematic study Kaiser and Benner (2005) demonstrated that the amount of D-Arg produced during acid hydrolysis of L-Arg (110 • C, 6 M HCl, 20 h) is not more than 1.0 ± 0% of the total for free amino acid, or 2.8 ± 0.4% when bonded in lysozyme [42]. We conclude that compound 1 is a cryptic mixture of diastereomers with stereofidelity at C-3, but partial epimerization at C-12. Therefore, the configuration of geobarrettin A (1) is 3R, 8Z, 12S.
The sole chiral amino acid residue in barettin (4), first reported in 1986 from a deep-water specimen of G. barretti collected in Sweden, presumably from Koster fjord [12], was assigned as Pro, but the configuration was not determined at that time. A second report of 4 in 2004 from a Norwegian deep-water sample of G. barretti [11] corrected the structure to an Arg-containing DKP, but also omitted the specific rotation. Finally, the configuration of 4 was established as L-Arg by comparison of optical rotations of natural and synthetic 4 [42]. Although they had similar signs, the specific rotations of synthetic and natural 4 were different ([α] 26 D −32.5 (c 2, MeOH) [18]; natural 4 [α] D −25 (c 3, MeOH)) [9]. In the present study, the rotation of 4 was found to be even higher ([α] 25 D −84 (c 0.5, MeOH)). In order to resolve this paradox, the L-FDTA method was applied to the hydrolysate as described above for 1. HPLC analysis detected the presence of D-Arg-DTA and L-Arg-DTA in a ratio of 19:81 suggesting L-Arg was predominant, but-as with 3-our sample of 4 was also partially racemic. Assuming no racemization occurred during the synthesis of 4 [42], our finding would be consistent with partially racemic natural 4 in the Swedish sample. The hydrolysis was repeated on debromo-8,9-dihydrobarettin (5a) (Figure 1), obtained by hydrogenolysis of 4, and the product was again converted to L-DTA derivatives and analyzed by HPLC as before to give D-Arg-DTA and L-Arg-DTA (19:81).

Anti-Inflammatory Activity
To evaluate the potential anti-inflammatory activity of compounds 1-7, their effects on DC secretion of the pro-inflammatory cytokine IL-12p40 and the anti-inflammatory cytokine IL-10 was determined. The tested compounds did not show cytotoxic effects on the DCs, with the cell viability being more than 90% in the presence of the compounds at 10 µg/mL (data not shown). Compound 4 inhibited by more than 50% secretion of both IL-12p40 and IL-10. Compounds 2 and 6 decreased DC secretion of IL-12p40 by 29 and 32%, respectively, without affecting secretion of IL-10, which suggests an overall anti-inflammatory activity of these compounds. Compound 3 slightly decreased DC secretion of IL-12p40 (13%) but increased secretion of the anti-inflammatory cytokine IL-10 by 40%, which also suggests an anti-inflammatory activity. Compounds 1, 5, and 7 did not affect DC secretion of IL-12p40 or IL-10 ( Figure 4).  [9]. In the present study, the rotation of 4 was found to be even higher ([α] 25 D −84 (c 0.5, MeOH)). In order to resolve this paradox, the L-FDTA method was applied to the hydrolysate as described above for 1. HPLC analysis detected the presence of D-Arg-DTA and L-Arg-DTA in a ratio of 19:81 suggesting L-Arg was predominant, but-as with 3-our sample of 4 was also partially racemic. Assuming no racemization occurred during the synthesis of 4 [42], our finding would be consistent with partially racemic natural 4 in the Swedish sample. The hydrolysis was repeated on debromo-8,9-dihydrobarettin (5a) (Figure 1), obtained by hydrogenolysis of 4, and the product was again converted to L-DTA derivatives and analyzed by HPLC as before to give D-Arg-DTA and L-Arg-DTA (19:81).

Anti-Inflammatory Activity
To evaluate the potential anti-inflammatory activity of compounds 1-7, their effects on DC secretion of the pro-inflammatory cytokine IL-12p40 and the anti-inflammatory cytokine IL-10 was determined. The tested compounds did not show cytotoxic effects on the DCs, with the cell viability being more than 90% in the presence of the compounds at 10 μg/mL (data not shown). Compound 4 inhibited by more than 50% secretion of both IL-12p40 and IL-10. Compounds 2 and 6 decreased DC secretion of IL-12p40 by 29 and 32%, respectively, without affecting secretion of IL-10, which suggests an overall anti-inflammatory activity of these compounds. Compound 3 slightly decreased DC secretion of IL-12p40 (13%) but increased secretion of the anti-inflammatory cytokine IL-10 by 40%, which also suggests an anti-inflammatory activity. Compounds 1, 5, and 7 did not affect DC secretion of IL-12p40 or IL-10 ( Figure 4).  We next examined whether compounds 2, 3, and 4 affected cytokine secretion by the DCs in a dose-dependent manner. The effects of compounds 2 and 3 on IL-12p40 and IL-10 secretion by DCs were not dose-dependent although for compound 2 there was a tendency towards an increasing effect on IL-12p40 secretion at higher concentrations. Compound 4 inhibited secretion of IL-12p40 and IL-10 in a dose-dependent manner with IC 50 being 21.04 µM for IL-12p40 and 11.80 µM for IL-10 ( Figure 5), with IL-12p40 and IL-10 secretion decreasing at the lowest concentrations tested, i.e., 5.96 µM (2.5 µg/mL) and 11.93 µM (5 µg/mL), respectively.
Mar. Drugs 2018, 16, x FOR PEER REVIEW 9 of 17 5), with IL-12p40 and IL-10 secretion decreasing at the lowest concentrations tested, i.e., 5.96 μM (2.5 μg/mL) and 11.93 μM (5 μg/mL), respectively. To further elucidate the anti-inflammatory activity of compounds 2-4, DCs matured in their presence were co-cultured with allogeneic CD4 + T cells and the differentiation of the T cells investigated by determining secretion of the cytokines IL-10, IL-17, and IFN-γ. T cells co-cultured with DCs matured in the presence of compounds 2 or 3 secreted less IL-10 and IFN-γ than T cells cocultured with DCs matured in the absence of the compounds, but maturing DCs in the presence of compounds 2 or 3 did not affect T cell secretion of IL-17 ( Figure 6). Co-culturing T cells with DCs matured in the presence of compound 4 resulted in a substantial decrease in secretion of IL-10 with no effect on secretion of either IL-17 or IFN-γ.  To further elucidate the anti-inflammatory activity of compounds 2-4, DCs matured in their presence were co-cultured with allogeneic CD4 + T cells and the differentiation of the T cells investigated by determining secretion of the cytokines IL-10, IL-17, and IFN-γ. T cells co-cultured with DCs matured in the presence of compounds 2 or 3 secreted less IL-10 and IFN-γ than T cells co-cultured with DCs matured in the absence of the compounds, but maturing DCs in the presence of compounds 2 or 3 did not affect T cell secretion of IL-17 ( Figure 6). Co-culturing T cells with DCs matured in the presence of compound 4 resulted in a substantial decrease in secretion of IL-10 with no effect on secretion of either IL-17 or IFN-γ.
Mar. Drugs 2018, 16, x FOR PEER REVIEW 9 of 17 5), with IL-12p40 and IL-10 secretion decreasing at the lowest concentrations tested, i.e., 5.96 μM (2.5 μg/mL) and 11.93 μM (5 μg/mL), respectively. To further elucidate the anti-inflammatory activity of compounds 2-4, DCs matured in their presence were co-cultured with allogeneic CD4 + T cells and the differentiation of the T cells investigated by determining secretion of the cytokines IL-10, IL-17, and IFN-γ. T cells co-cultured with DCs matured in the presence of compounds 2 or 3 secreted less IL-10 and IFN-γ than T cells cocultured with DCs matured in the absence of the compounds, but maturing DCs in the presence of compounds 2 or 3 did not affect T cell secretion of IL-17 ( Figure 6). Co-culturing T cells with DCs matured in the presence of compound 4 resulted in a substantial decrease in secretion of IL-10 with no effect on secretion of either IL-17 or IFN-γ.

Discussion
Three new 6-bromoindole derivatives were isolated from G. barretti, including a dioxindole featuring a DKP-type cyclic dipeptide, geobarrettin A (1); a 6-bromoindole possessing DKP system, geobarrettin B (2); and a new 6-bromoindole alkaloid, geobarrettin C (3). Compounds 2 and 3 inhibited IL-12p40 production by DCs and DCs treated with compounds 2 and 3 reduced IFN-γ secretion by co-cultured T cells, hence reduced Th1 responses, which are linked to inflammatory disorders and many chronic inflammatory diseases [48,49]. As IL-12 is the main inducer of Th1 polarization of T cells with subsequent IFN-γ secretion [50], the down-regulation of IFN-γ observed in the co-culture of T cells with DCs is most likely resulting from a reduced ability of the DCs matured in the presence of compounds 2 and 3 to secrete IL-12p40 (one of the two chains that form the IL-12 molecule). Compound 2 did not affect IL-10 secretion by DCs but compound 3 increased IL-10 secretion by DCs. Therefore, the decreased concentration of IL-10 in co-cultures of T cells and DCs matured in the presence of compounds 2 and 3 was unexpected. The reduced IL-10 levels observed in the co-cultures were most likely the result of reduced secretion by the T cells but not the DCs, but whether it is so needs to be confirmed, e.g., by intracellular staining for IL-10 on T cells and DCs in the co-culture experiments. Although the decreased secretion of IL-10 in the co-cultures may hamper the anti-inflammatory effect of compounds 2 and 3, the downregulation of IFN-γ secretion is strongly suggestive of inhibition of inflammatory Th1 response and subsequently that compounds 2 and 3 may have the potential of being a starting point for development of new anti-inflammatory drugs.
The anti-inflammatory effect of compound 4, shown previously in a monocytic cell line [6], was confirmed in the present study as it downregulated secretion of IL-12p40 by the DCs. However, compound 4 also downregulated IL-10 production, which could interfere with the anti-inflammatory effect observed by reduced IL-12p40 secretion. This seemed to be the case as, when DCs treated with compound 4 were co-cultured with T cells, the effect of compound 4 was not anti-inflammatory as neither Th1 nor Th17 cytokines were affected. These results were unexpected and suggest that the effect of compound 4 to decrease IL-10 secretion by the DCs seems to be the determining factor, overriding the effect of downregulation of IL-12p40 and subsequently leading to a reduction in IL-10 in the co-culture of the two cell types.
Despite structural similarities of compounds 1, 2, 4, and 5, which all possess the structure of a DKP-type cyclic dipeptide, there were remarkable differences in their anti-inflammatory effects. Oxidation of the indole ring in compound 1 as compared with compound 4 caused the disappearance of the anti-inflammatory activity, indicating that the indole skeleton is important in inhibition of cytokine secretion by the DCs. Both the number and the position of double bonds on the DKP-type cyclic dipeptide may affect the anti-inflammatory activity (2 vs. 4 vs. 5). The disappearance of the double bond at C-8 could decrease the anti-inflammatory activity, as compound 5 did not affect cytokine secretion by the DCs whilst compounds 4 and 2 did, suggesting that the double bond at C-8 is required for the activity. However, the double bond at C-12 may be responsible for the reduction of anti-inflammatory activity when comparing compound 2 with compound 4. Considering these observations, the bromotryptophan containing the double bond at C-8 may be important for the activity. The double bond at the N,N,N-trimethylethanaminium group increased the suppression of IL-12p40 production and increased the IL-10 secretion (6 vs. 7). Collectively, 6-bromoindole derivatives may have anti-inflammatory activity that depends on the bromotryptophan nucleus (1, 2, 4, and 5) or the side chain at C-3 position of the 6-bromoindole (3, 6, and 7), suggesting that there may be more than one potential target site or mode of action.
The observations described above indicate that G. barretti is a prolific source of 6-bromoindoles with potential anti-inflammatory activities, which may be a starting point for the development of new drug s with a potential for being used in the treatment of inflammatory diseases in the future.

General Produres
Optical rotations were measured on a P-2000 polarimeter (Jasco, OK, USA) equipped with a 10 mm pathlength cell. UV spectra were recorded on a NanoVue TM spectrophotometer (GE Healthcare Life Science, Little Chalfont, UK). ECD spectra were measured on a JASCO J-810 spectropolarimeter in quartz cells (1 or 5 mm pathlength) at 23 • C. IR spectra were measured on a Spectrum Two TM FTIR spectrometer (Perkin Elmer, Waltham, MA, USA). NMR spectra were recorded on a Bruker AM-400 spectrometer (proton frequency 400.13 MHz and carbon frequency 100.62 MHz, respectively) for compounds 4-7 (in CD 3 OD and/or DMSO-d 6 ) using TMS as an internal standard or a Bruker Avance 600 spectrometer (proton frequency 600.13 MHz and carbon frequency 150.76 MHz, respectively) for compounds 1 (in CD 3 OD and DMSO-d 6 ), 2-3 (in CD 3 OD), and 7 (in DMSO-d 6 ). The 1 H NMR spectrum of compound 8 (in CDCl 3 ) was recorded on a Varian Mercury 400 spectrometer and the 13 C NMR spectrum was measured on a Varian Xsens 500 spectrometer equipped with a 13 C{ 1 H} cryoprobe at 125 MHz. High-resolution mass spectra were obtained on a Waters G1 Synapt qTOF mass spectrometer. An UHPLC system (ACQUITY UPLC Waters) was coupled in line with a qTOF mass spectrometer (Synapt G1, Waters) operating in the positive mode. HPLC was performed on a Dionex 3000 HPLC system equipped with a G1310A isopump, a G1322A degasser, a G1314A VWD detector (210 nm), a 250 × 21.2 mm Phenomenex Luna C18(2) column (5 µm), and a 250 × 4.6 mm Phenomenex Gemini-NX C18 column (5 µm). Alkaloids were detected by TLC on Merck silica gel F254 plates by immersing the plates in Dragendorff's reagent. VLC chromatography was performed on C 18 adsorbent (LiChroprep RP-18, 40-63 µm, Merck Inc., Darmstadt, Germany). All organic solvents were purchased from Sigma-Aldrich and were HPLC grade or the highest degree of purity.

Animal Materials
The sponge material Geodia barretti was collected in the west of Iceland (65 • 27.6 N-30 • 46.6 W) at 388 m depth in September 2010. The sponge was identified by Dr. Hans Tore Rapp, University of Bergen (Norway). A voucher specimen was deposited at the Department of Natural Products Chemistry, Faculty of Pharmaceutical Sciences, University of Iceland. The collected specimens (six in total) (wet weight, 1.8 kg) were immediately frozen and stored at −20 • C.

Extraction and Isolation
The frozen sponge material was cut into small pieces and lyophilized prior to extraction with CH 2 Cl 2 :CH 3 OH (v/v, 1:1) mixture (3 × 20 L, each for 24 h) at room temperature. The combined extracts were concentrated under vacuum to yield a dark gum and stored at −20 • C. The crude extract (1.8 g) was submitted to a modified Kupchan partition to yield five subextracts, namely hexane (fraction A), chloroform (fraction B), dichloromethane (fraction C), n-butanol (fraction D), and water (fraction E). Each fraction was analyzed by UPLC-qTOF-MS before preparative-scale isolation work commenced and the data were processed and analyzed by MassLynx programme and compared to available references in ChemSpider database [51] and SciFinder Scholar (Chemical Abstracts Service, Columbus, OH, USA). The fractions B and C showed similar patterns on TLC and the analysis of qTOF-MS data revealed similar chemical compositions. Thus, they were combined and fractionated by VLC on a C18 reversed-phase column using gradient elution of MeOH-

Acid Hydrolysis of Debromodihydrogeobarrettin B (1a)
Approximately 0.31 mg of compound 1a was separately hydrolyzed with 6 M HCl (0.8 mL) for 15 h at 110 • C, dried under a stream of N 2 followed by high vacuum to remove volatiles, and the resulting material subjected to further derivatization (see below).

Absolute Configuration of the Amino Acid of Geobarrettin A (1)
The determination of absolute configuration of the amino acids was performed as previously described [41]. After hydrolysis of 1a, the residue was dissolved in acetone (50 µL) and treated with L-FDTA (25 mM in acetone, 50 µL) and 1 M NaHCO 3 (50 µL) and heated to 80 • C for 20 min.

Maturation and Activation of DCs
DCs were developed and matured from human monocytes as previously described [52,53]. CD14 + monocytes were isolated from peripheral blood mononuclear cells obtained from healthy human donors using CD14 Microbeads (Miltenyi Biotech, Bergisch Gladbach, Germany). Immature DCs were obtained by culturing CD14 + monocytes at 5 × 10 5 cells/mL for 7 days in the presence of IL-4 at 12.5 ng/mL and GM-CSF at 25 ng/mL (both from R&D Systems, Abingdon, England), with fresh medium and cytokines added at day 3. The immature DCs were matured and activated by culturing them at 2.5 × 10 5 cells/mL for 24 h with IL-1β at 10 ng/mL, TNF-α at 50 ng/mL (both from R&D Systems), and lipopolysaccharide (LPS) at 500 ng/mL (Sigma-Aldrich, Munich, Germany). Pure compounds were dissolved in DMSO at the concentration of 10 µg/mL and added to the DCs at the same time as the cytokines and LPS. The final concentration of DMSO in the medium of DCs cultured with the pure compounds was 0.002% and the same concentration of DMSO was used as solvent control. In order to determine whether the effects of compounds 2-4 were dose-dependent, the concentrations of 2.5, 5, 7.5, 10 µg/mL were used. After 24 h the mature and activated DCs were harvested and the effects of the pure compounds on DC maturation and activation determined by measuring cytokine concentration in the culture medium by ELISA. Cell viability was determined by counting cells following staining with trypan blue and calculating the percentage of live cells.

Co-Culture of DCs and Allogeneic CD4 + T Cells
DCs matured in the presence/absence of a pure compound at 10 µg/mL or solvent control only were co-cultured with allogeneic CD4 + T cells at DC:T cell ratio of 1:10 (2 × 10 5 DCs/mL: 2 × 10 6 T cells/mL) for 6 days. CD4 + T cells were obtained from PBMCs using CD4 Microbeads (Miltenyi Biotec). The effects of the pure compounds on the ability of the DCs to differentiate the CD4 + T cells were determined by measuring cytokine concentrations in the co-culture supernatants by ELISA.

Determination of Cytokine Concentration by ELISA
The concentrations of IL-12p40 and IL-10 in culture supernatants from DCs and of IFN-γ, IL-17 and IL-10 in supernatants from co-cultures of DCs and allogeneic CD4 + T cells were measured by sandwich ELISA using DuoSets from R&D Systems according to the manufacturer's protocol. The results were expressed as secretion index (SI), which was calculated by dividing the cytokine concentration in supernatants from DCs cultured with pure compound or co-cultures of these DCs with allogeneic CD4 + T cells by the cytokine concentration in supernatants of DCs cultured with solvent only or co-cultures of these DCs with allogeneic CD4 + T cells.

Statistical Analysis
Data are presented as the mean values ± standard error of the mean (SEM). As the data were not normally distributed, Mann-Whitney U test or Kruskal Wallis one-way ANOVA with Tukey's post-hoc test were used to determine statistical differences between treatments (SigmaStat 3.1, Systat Software, San Jose, CA, USA) and p < 0.05 was considered as statistically significant.

Conclusions
In conclusion, three new 6-bromoindole derivatives, geobarrettin A-C (1-3), and four known ones (4-7) were obtained from the marine sponge G. barretti collected from west of Iceland. Compounds 2, 3 and 4 showed anti-inflammatory properties by inhibiting DC secretion of IL-12p40 with varying effects on IL-10, and the anti-inflammatory effect of compounds 2 and 3 was confirmed by inhibition of IFN-γ secretion in co-cultures of T cells and DCs matured in the presence of the compounds. It is increasingly being recognized that low-grade, subclinical inflammation is a significant pathogenic factor in many chronic diseases that have not, until now, been considered inflammatory in nature. Importantly, this includes most diseases that today are the main cause of morbidity and mortality in Western countries, such as atherosclerotic diseases, cancers, chronic pain disorders, and Alzheimer's disease [54]. Therefore, the discovery of the two new 6-bromoindole derivatives with anti-inflammatory effects is important as they could be used in the development of treatments for diseases with inflammatory components. Acknowledgments: The authors would like to thank Hans Tore Rapp at the University of Bergen for the identification of animal material, Nathalie Kringlstein for technical assistance, Annaliese Franz at University of California, Davis, for the generous gift of 8a, Sigridur Jonsdottir at University of Iceland for running NMR spectra on a Bruker AM-400 spectrometer, Finnur Freyr Eiriksson and Margret Thorsteinsdottir for running UPLC-qTOF-MS and Kare Telnes for giving the permission to use the sponge picture in the Graphical abstract.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.