Diterpenylquinone Hybrids: Synthesis and Assessment of Gastroprotective Mechanisms of Action in Human Cells

A modern approach in the search for new bioactive molecules is the synthesis of novel chemical entities combining molecules of different biosynthetic origin presenting biological effects as single compounds. Gastroprotective compounds from South American medicinal plants, namely quinones and diterpenes, were used as building blocks to obtain hybrid diterpenylquinones. Starting from the labdane diterpene junicedric acid and two isomers, as well as from three quinones, including lapachol, 18 hybrid molecules were synthesized. Six of them are described for the first time. The potential gastroprotective mechanisms of action of the compounds were assessed in dose-response experiments using human gastric epithelial cells (AGS) and human lung fibroblasts (MRC-5). The following studies were carried out: stimulation of cell proliferation, cytoprotection against sodium taurocholate (NaT)-induced damage, synthesis of PGE2 and total reduced sulfhydryl (GSH) content. The antioxidant capacity of the compounds was determined on the inhibition of the lipoperoxidation in human erythrocyte membranes. Hybrid compounds presented activities different from those shown by the starting compounds, supporting the potential of this approach in the search for new bioactive molecules. The effects might be modulated by selective modification in the terpene or quinone moieties of the new molecules. Structure-activity relationships are discussed.


Introduction
South American traditional medicine uses plants in therapeutic practice. Several herbal crude drugs have been shown to contain bioactive constituents that can be clearly related to their beneficial effects. The resin of the "araucaria" tree (Araucaria araucana (Mol.) K. Koch, Araucariaceae) was used by the Mapuche amerindians in Chile to treat gastric ulcers and to relieve stomach pain [1]. The stem bark of the crude drug "lapacho" or "taheebo" (Tabebuia spp., Bignoniaceae) was formerly used by the Guarani as an anti-inflammatory agent and then incorporated into Paraguayan folk medicine, to treat cancer and wounds [2]. The naphthoquinone lapachol from Tabebuia species [3] and labdane diterpenes from A. araucana, display gastroprotective activity in vivo [4] supporting the ethnobotanical information. The bioactive compounds from those crude drugs (quinones and diterpenes) exert their effect as single chemical entities. A new approach in drug design is to combine two different molecules with individual intrinsic effects into a single new hybrid compound [5]. Little has been done on the synthesis of products combining naturally occurring moieties arising from different biosynthetic pathways into such hybrid molecules. Recent studies presented the synthesis, gastroprotective effect and cytotoxicity of diterpenylnaphthoquinones [6] and the potential gastroprotective effect of novel sesquiterpene quinone derivatives in human cells [7]. Terpenylquinones have been described from some marine organisms and display several biological activities, including cytotoxicity and anti-inflammatory effect [8]. New hybrid compounds containing terpene and quinone moieties have been prepared and some of them present relevant biological effects including cytotoxicity [9,10] and antifungal effect [11].
The study of gastroprotective activity of compounds has been traditionally carried out using laboratory animals. However, due to social and economic pressures, the international trend has been a gradual reduction in the use of experimental animals. The use of primary cell culture as well as immortalized cell lines allows in vitro testing and studies of possible mechanisms of action. There are several advantages in using cell cultures as biological models, namely: the amount of compound is reduced to a minimum, variables are much better controlled than when working with entire animals and the costs are substantially reduced. More advanced techniques consider the use of gastric epithelial cells [7,12,13].
The aim of this work was to prepare new hybrid compounds using as starting molecules gastroprotective diterpenes with the labdane skeleton and quinones. Human cell cultures were used to assess their potential mechanisms of gastroprotective activity and compare it to that of the starting substances. To draw some structure-activity trends, the hybrids included variations both in the quinone as well as in the diterpene moieties.

Hybrid Compounds
A series of hybrid compounds was synthesized using as building blocks diterpenes and quinones. The quinones included lapachol (1) and its hydrogenation products 2 and 3. The quinone part of the molecule included an isoprenyl side chain, hydrogenated side chains with aromatic rings in the naphthoquinone moiety or hydrogenated side chains and aromatic rings. The diterpene junicedric acid (4) with an exocyclic double bond, its isomer 5 with an Δ8(9) double bond and the hydrogenation product 6 were used as the terpene part of the hybrid molecules. Eighteen hybrids, including six new compounds, were synthesized (Scheme 1), differing in some structural features while maintaining common characteristics to evaluate structure-activity trends. All compounds were characterized by spectroscopic and spectrometric means and are in agreement with the proposed structures.

Cytotoxicity
The cytotoxicity of all compounds was evaluated on confluent cultures of human lung fibroblasts (MRC-5) and human gastric epithelial cells (AGS) to determine the working concentrations in the subsequent mechanism of action experiments. Cell viability was determined by means of the neutral red uptake assay and results are expressed as IC 50 values (µM) in Table 1. Hydrogenation increases cytotoxicity for the lapachol derivatives 2 and 3. From the starting diterpenes, Δ8(9) junicedric acid (5) presented lower toxicity towards MRC-5 and AGS cells, with IC 50 values of 834 and 666 µM, respectively. The diterpene with higher cytotoxicity was compound 6 (IC 50 values of 163 and 283 µM for MRC-5 and AGS cells, respectively).
Among the six hybrid compounds derived from diterpene 4 (compounds 7-12), the carboxylic acid derivative 7 resulted more cytotoxic than the parent quinone and diterpene 1 and 4, respectively. A similar trend was observed for hybrids 9 and 11, prepared from quinones 2 and 3, and diterpene 4, respectively. Methylation of 7, 9 and 11 led to the less cytotoxic derivatives of the series 8, 10 and 12, indicating that the free C-19 carboxylic acid function is relevant for the cytotoxicity of this series. The highest cytotoxicity value was found for compound 11, in which both, the prenylated side chain and the benzene ring of the lapachol were reduced by hydrogenation.
The diterpene 5 differs from 4 in the position of the double bond, leading to different configurations for both compounds and their hybrid derivatives. Hybrid compounds containing diterpene 5 (compounds 19-24), showed the same trend than derivatives from diterpene 4 when presenting a free C-19 COOH function. Therefore, compounds 19, 21 and 23 were more toxic than their corresponding C-19 COOMe derivatives 20, 22 and 24. As for the series from diterpene 4, the most cytotoxic compound for the derivative series of 5 resulted compound 23, with the prenylated side chain and the benzene ring of the lapachol completely reduced.
Among the hybrid compounds obtained from diterpene 6 (compounds 13-18), it was observed that hybrids 13 and 17 bearing a carboxylic acid function at C-19, were more toxic than the corresponding methylester derivatives 14 and 18. However, when only hydrogenation of the prenylated side chain of the lapachol was present in the hybrid compounds (15 and 16), the ester derivative 16 showed higher cytotoxicity that the corresponding carboxylic acid derivative 15.

MRC-5 Fibroblast Proliferation
A key factor in the renewal and repair of gastric mucosa is the proliferative capacity of fibroblasts [14]. The ability of the compounds to accelerate cell proliferation and hence gastric wound healing was determined using MRC-5 fibroblasts. The compounds were evaluated at six concentrations lower than the IC 50 values, being IC 50 /2 the highest concentration tested. The positive results are presented in Figure 4. Percentual proliferative increase compared to untreated cells is shown in parentheses. Stimulation of fibroblast proliferation was observed for the diterpene 5 (3.6%), the hybrid 11 (3.6%), obtained by the combination of 4 and 3 and the derivatives from diterpene 6: 13 (7.3%), 17 (5.5 to 14.5%) and 18 (3.6%). The active compounds 11, 17 and 18 have in common the quinone 3 in their structures. These compounds differ from the inactive derivatives 23 and 24 (also containing the quinone 3) by the more planar configuration of the diterpene moiety of compounds 23 and 24. None of the hybrid derivatives that showed activity contain the diterpene moiety 5.

Sodium Taurocholate (NaT)-Induced Damage
The model of AGS cells damaged by NaT (bile salt) was used to assess the gastroprotective effect of compounds against the gastric ulcer induced by bile reflux [15]. Results are presented as a reduction in cell viability. Treatment of the cells with 10 mM NaT during 30 min caused a 48% reduction of cell viability compared with untreated controls. Cells were pre-treated during 60 min with the compounds at 1/2, 1/4 and 1/8 of the respective IC 50 values and then 10 mM NaT was added to all wells. The reference compound sucralfate at 4 mg/mL (580 µM) showed a reduction of 35% in cell viability. The positive results are shown in Figures 5, 6 and 7. Figure 5. Effect of pre-treatment during 60 min with the reference compound sucralfate (Suc), the quinones 1 and 2 and the diterpenes 5 and 6 followed by an incubation during 30 min with 10 mM NaT on the viability of AGS cells determined by the neutral red uptake assay. Each value represents the mean ± SD of three different experiments in quadruplicate. ANOVA followed by Dunnett´s test. * p < 0.05 compared to NaT group.  Figure 6. Effect of pre-treatment during 60 min with the reference compound sucralfate (Suc) and derivatives 13, 15, 16 and 17 from 17β-dihydrojunicedric acid, followed by an incubation during 30 min with 10 mM NaT on the viability of AGS cells determined by the neutral red uptake assay. Each value represents the mean ± SD of three different experiments in quadruplicate. ANOVA followed by Dunnett's test. * p < 0.05 compared to NaT group.  . Effect of pre-treatment during 60 min with the reference compound sucralfate (Suc) and Δ8(9) junicedric acid derivatives 19, 21 and 23 followed by an incubation during 30 min with 10 mM NaT on the viability of AGS cells determined by the neutral red uptake assay. Each value represents the mean ± SD of three different experiments in quadruplicate. ANOVA followed by Dunnett's test. * p < 0.05 compared to NaT group. The results showed significant cytoprotective effect (p < 0.05) for the starting quinones 1 (25,22 and 29% cell protection at 285, 143 and 71 µM, respectively) and 2 (27,23 and 23%, at 146, 73 and 37 µM respectively), the diterpenes 5 (35, 25 and 19% reduction in cell viability at 333, 167 and 83 µM, respectively) and 6 (25% cell viability reduction at 71 and 35 µM) ( Figure 5).
Cytoprotection was observed for the diterpene 6 hybrids 13, 15-17 ( Figure 6). The effect of compound 17 at 30, 15 and 7.5 µM was in the same range as sucralfate (34%-36%), while the activity of 13 at 89 and 45 µM was 38% and the effect of 16 at 81 µM was 44%, compared to NaT controls. Best effect, even higher than that of sucralfate, was found for the hybrid 15 at 500, 250 and 125 µM, reducing cell viability by 21, 27 and 27%, respectively.
The hybrids of the diterpene 5 with 1, 2 and 3 (compounds 19, 21 and 23) were active ( Figure 7). However, the corresponding methyl esters were devoid of effect, suggesting that the free COOH at C-19 is needed for cytoprotection. Even if COOH is needed for cytoprotection, the carboxylic acid derivatives were the most toxic compounds. The cytoprotection effect might be due to the implication of OH residues in radical mechanisms with ROS species, and that might be the reason why the methyl ester derivatives, lacking of OH, do not present cytoprotective effects. Compound Cytoprotective compounds may protect against NaT-induced damage forming a physical barrier or binding bile salts, interacting with cell membranes or modifying the expression of trefoil factor family 2 mRNA and c-fos protein [16,17].

Effect on Prostaglandin E 2 (PGE 2 ) Content
A relevant gastroprotective mechanism involves the prostaglandins that accelerate ulcer healing, epithelial cell proliferation, synthesis of growth factors and inhibition of inflammatory cell infiltration [18]. The prostaglandin E 2 (PGE 2 ) is considered to be the major humoral factor providing resistance to the gastric mucosa [19]. Some terpenes and their derivatives exert their gastroprotective effect increasing the gastric PGE 2 synthesis in vitro and in vivo [7,20]. Experiments were carried out using post-confluent AGS cells. The results are presented in Table 2. Among the starting quinones and  The dihydroprenyl lapachol hybrids 9-10, 15, 21-22, bearing the same quinone moiety but differing in the identity of the terpene, increased PGE 2 synthesis as follows: 9 (12.

Effect on Cellular Reduced Glutathione (GSH) Content
Generation of gastric ulcers has been related to the effect of free radicals. This fact explains the ulcerogenic effect of ethanol on the gastric mucosa. Intracellular GSH content is a determining factor that contributes to the protection of the gastric mucosa against oxidative stress [21]. Experiments were carried out using post-confluent AGS cells. The known stimulant of GSH synthesis, N-acetyl-Lcysteine (NAC) was used as positive control. Under our experimental conditions, NAC (750 µM) increased GSH content by 30% compared to untreated controls. Results are summarized in Table 3. Among the starting diterpenes, junicedric acid (4) and 17β-dihydrojunicedric acid (6) significantly (p < 0.05 and p < 0.01) stimulated GSH production. Best activity was observed for compound 6 (60 and 43% increase at 71 and 37 µM, respectively). Among the hybrid compounds, a good effect was found for the junicedric acid-lapachol derivatives 7 (23 and 18% at 36 and 18 µM, respectively) and 8 (24 and 19% at 250 and 125 µM, respectively) as well as for dihydroprenyl lapachol junicedrate 9 (16 and 15% at 12 and 6 µM, respectively).

Inhibition of Lipoperoxidation
Antioxidant effect of the compounds was assessed using human erythrocyte membranes. The products were tested at 100 µg/mL and results are presented in Table 4. Among the starting quinones, the hydrogenation products of lapachol 2 and 3 were more active than lapachol (1), with IC 50 values of 25.4 and 33.3 µg/mL, respectively. From the diterpene moieties, only Δ8(9)-junicedric acid (5) was active with an IC 50 of 43.8 µg/mL. The inhibition of lipoperoxidation for the hybrids prepared from the diterpene 4 (compounds 7-12) was moderate. Better effects were found for the diterpene 4 derivative with lapachol and a COOCH 3 at C-19 (8, IC 50 : 81.4 µg/mL), and the hybrids containing 4 and the quinone 2, either with a free C-19 carboxylic acid (9, IC 50 : 40.1 µg/mL) or as the methyl ester (10, IC 50 : 99.4 µg/mL). Among the hybrids of diterpene 6 (compounds 13-18), best antioxidant activity was found for: 13, 14 and 15 with IC 50 values of 61.1, 19.7 and 25.3 µg/mL, respectively. The three compounds contain the quinone 1 or 2 in the hybrid structure. The compounds 19-24 were prepared using Δ8(9) junicedric acid (5) as the diterpene moiety and presented moderate antioxidant effect, best activity was displayed by 20 (IC 50 : 58.1 µg/mL) and 23 (IC 50 : 59.5 µg/mL). In the same assay, the reference compound catechin presented an IC 50 value of 75.4 µg/mL. Structural diversity is a requirement to draw structure-activity relationships/trends in compounds series. The mechanisms of action of the starting diterpenes and quinones were compared with those of the new hybrids. The starting compounds showed activity in one or two (exceptionally three) out of the mechanisms of action investigated. Diterpene 5 stimulates cell proliferation, while none of its hybrids presented effect. From the four hybrids of diterpenes 4 and 6 with the quinone 3, three were active.
The quinones 1-2 and diterpenes 5-6 were active in the NaT assay. Hybrids from diterpenes 5 and 6 with quinones show effects on NaT when the COOH group at C-19 of the diterpene is free (compounds 13, 15, 17, 19, 21 and 23). The corresponding methyl esters (except compound 16) were devoid of effect. The products prepared from the diterpene 4, with a 8(17) double bond were inactive.
Among the parent molecules, only lapachol increased PGE 2 synthesis. Relevant changes in PGE 2 levels were observed for hybrid compounds, with several products stimulating synthesis. Regardless of the diterpene moiety, hybrids with dihydroprenyl lapachol 2 and two methyl esters from the hybrids with 3 were active.
In a previous report, part of the compounds presented in this study, were evaluated for gastroprotective effect by the HCl-EtOH-induced gastric lesions model in mice [6].The compounds were administered orally at a single dose of 5 mg/kg. Compounds 9-10 and 21-22 displayed significant activity in vivo [6]. In the present study, these compounds stimulated the PGE 2 synthesis in AGS cells. Our results show that compounds 8 and 20, active in vivo, increased GSH content in AGS cells and protected erythrocyte membranes against induced lipoperoxidation. These in vitro findings sustain the in vivo observations suggesting that different and complementary mechanisms may be underlying the gastroprotective effect of the compounds.
Gastric lesions are a consequence of an imbalance between defensive and aggressive factors in the gastric mucosa. Increase in the defensive factors, namely: restitution of gastric epithelium, protection against damage induced by bile salts and free radicals, increase in PGE 2 levels, GSH, bicarbonate and others, may reduce the appearance of gastric lesions in animals [7,16,17]. It has been shown that gastroprotective compounds exert their effects mainly stimulating the defensive factors in the gastric mucosa rather than inhibiting the aggressive agents (gastric acid or pepsin) [16,17,22].
In this work the possible mechanisms of gastroprotective effect of diterpenylquinones were assessed using human lung fibroblasts (MRC-5) and gastric epithelial cells (AGS). The evaluation of the compound activity on some defensive factors depending on gastric epithelial cells has been carried out using AGS cells. This cell line has been successfully used in the study of different aspects related to the gastric mucosa [7,23,24]. The hybrid compounds investigated stimulated one or more of the defensive factors described as relevant to protect the gastric mucosa. To obtain hybrid molecules with better activity than the parent compounds, structural modifications should be undertaken, followed by a suitable combination of substituents placed at the appropriate positions of hybrid compounds [25]. To understand the relationship between in vitro and in vivo activity, further research should be carried out, since factors such as absorption, bioavailability, location of the targets in different tissues or organs, receptor and enzyme densities will influence in effect [26].

General Experimental Procedures
Optical rotations were obtained for solutions in CHCl 3 (concentrations expressed in g/100 mL) on a Jasco DIP 370 polarimeter (Jasco Analytical Instruments, Easton, MD, USA). IR spectra were recorded on a Nicolet Nexus FT-IR instrument (Thermo Electron Corporation, Waltham, MA, USA). 1 H-NMR spectra were recorded at 400 MHz and 13 C-NMR data were obtained at 100 MHz on a Bruker Avance spectrometer (Bruker, Rheinstetten, Germany). Chemical shifts are given in δ (ppm) with TMS as the internal standard. Mass spectra were measured in a LC/MSD-TOF (Agilent Technologies) by the electrospray technique. Positive or negative ion mode was used according to the samples. As internal reference in the ESI (+) mode, purine (m/z 121.0509) and HP-0921 (m/z 922.0098) were used. In the negative mode: ESI (−), purine (m/z 112.9856) and HP-0921 (m/z 1033.9881) were used as internal references. Silica gel 60 (Merck, 63-200 µm particle size) was used for column chromatography, precoated silica gel plates (Merck, Kieselgel 60 F254, 0.25 mm) were used for thin layer chromatography (TLC). TLC spots were visualized by spraying the chromatograms with p-anisaldehyde-ethanol-acetic acid-H 2 SO 4 (2:170:20:10 v/v) and heating at 110 °C for 3 min. 1,3-Dicyclohexylcarbodiimide (DCC) and dimethylaminopyridine (DMAP) were from Merck (Schuchardt, Germany). Oxalyl chloride was from Sigma-Aldrich (St. Louis, MO, USA).

Quinones
The naphthoquinone lapachol (1) was isolated from the wood of "lapacho" tree (Tabebuia sp). The hydrogenation products of lapachol (compounds 2 and 3) were prepared trating (1) in EtOAc with Pd/C [6]. The synthetic scheme for the synthesis is presented in [6].

Diterpenes
The diterpene 4 was obtained from the resin of Araucaria araucana as previously reported [6]. The isomers 5 and 6 were prepared from 4. Briefly, treatment of 4 with HBr in acetic acid afforded Δ8(9) junicedric acid (5) while catalytic hydrogenation of 4 in ethyl acetate with 10% Pd/C in a 1:10 molar ratio yielded the hydrogenation product 6. The synthetic scheme for the synthesis is shown in [6].

Hybrid Compounds
The compounds 7-14 and 19-22 were synthesized according to [6]. The new compounds 15 and 17 were prepared by treating the diacid 6 (1 mEq) in dry CH 2 Cl 2 (DCM) with 1,3-dicyclohexylcarbodiimide (DCC) (1 mEq) at room temperature under constant stirring. After 10 min, the quinones 2 and 3 (1 mEq) dissolved in dry DCM, were added together with a catalytic amount of dimethylaminopyridine (DMAP). After 2-4 h, the reaction was stopped by adding water and extracted with DCM. Compound 23 was synthesized from the diacid 5 under an inert (N 2 ) atmosphere. Briefly, the diterpene 5 was dissolved in dry CH 2 Cl 2 (DCM) and ice cooled under nitrogen flow. To this solution, oxalyl chloride in dry DCM was added drop wise with stirring. The diterpene: oxalyl chloride molar ratio was 1:40. The mixture was stirred at room temperature overnight, then the DCM was evaporated and the residue vacuum dried. The dry residue was dissolved in dry DCM and the quinone 3 was added as well as triethylamine (TEA) under constant N 2 flow. The diterpene acyl chloride-quinone-TEA molar ratio was 1:1:2. The reaction mixture was left at room temperature under stirring and inert atmosphere for two days. Then, the mixture was washed two times with water, and the aqueous phase extracted with DCM to obtain the crude reaction mixture.

Cytotoxicity Assay
Cytotoxicity values were required as a reference to determine the working concentrations in the subsequent experiments. Briefly, confluent cultures of MRC-5 and AGS cells were treated during 24 h with medium containing the compounds at concentrations ranging from 0 up to 1,000 µM. The products were first dissolved in DMSO (1% final concentration) and then in the corresponding culture medium supplemented with 2% FBS. Untreated cells were used as controls. Cell viability was determined at the end of the incubation by means of the neutral red uptake (NRU) assay [27]. Lansoprazole (Sigma-Aldrich Co., min. 98% by TLC) was used as reference compound.

MRC-5 Fibroblast Proliferation Assay
One day after seeding, cells were treated during 4 days with medium supplemented with 10% FBS and the studied compounds at concentrations ranging from 1/64 up to 1/2 of the respective IC 50 values. Untreated cells were used as controls. Cell viability was determined at the end of the incubation by means of the NRU assay [27]. starting compounds, supporting the potential of this approach in the search for new bioactive molecules. These effects were modulated by selective modification in the terpene or quinone moieties of the hybrids. The changes in effect associated to the stereochemistry of the diterpenes (i.e., a more flexible configuration for diterpenes with a 8 (17) double bond or the 17-β-dihydroderivative vs. the Δ8(9) isomer) or with modifications in the planar configuration of the napththoquinones, suggest that the different mechanisms involved might include interactions with active sites of enzymes and/or receptors. Molecular modeling can be a useful tool for better understanding the clues behind the observed effects.