Terpenoids from Euphorbia pedroi as Multidrug-Resistance Reversers

Article pubs.acs.org/jnp

Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX

Terpenoids from Euphorbia pedroi as Multidrug-Resistance Reversers Ricardo J. Ferreira,†,⊥ Annamaŕ ia Kincses,‡ Maŕ ió Gajdać s,‡ Gabriella Spengler,‡ Daniel J. V. A. dos Santos,†,§ Joseph Molnaŕ ,‡ and Maria-Jose ́ U. Ferreira*,†

J. Nat. Prod. Downloaded from pubs.acs.org by KAOHSIUNG MEDICAL UNIV on 09/10/18. For personal use only.

†

iMed.ULisboa (Research Institute for Medicines), Faculty of Pharmacy, Universidade de Lisboa, Avenida Prof. Gama Pinto, 1649-003 Lisboa, Portugal ‡ Department of Medical Microbiology and Immunobiology, Faculty of Medicine, University of Szeged, Dóm tér 10, H-6720 Szeged, Hungary § LAQV@REQUIMTE/Department of Chemistry and Biochemistry, Faculty of Sciences, University of Porto, Rua do Campo Alegre, 4169-007 Porto, Portugal S Supporting Information *

ABSTRACT: The phytochemical study of Euphorbia pedroi led to the isolation of a new tetracyclic triterpenoid with an unusual spiro scaffold, spiropedroxodiol (1), along with seven known terpenoids (2−8). Aiming at obtaining compounds with improved multidrug-resistance (MDR) reversal activity, compound 8, an ent-abietane diterpene, was derivatized by introducing nitrogen-containing and aromatic moieties, yielding compounds 9−14. The structures of compounds were characterized by detailed spectroscopic analysis, including 2D NMR experiments (COSY, HMQC/HSQC, HMBC, and NOESY). Compounds 1−14 were evaluated for their MDR-reversing activity on human ABCB1 gene transfected mouse lymphoma cells (L5178YMDR) through a combination of functional and chemosensitivity assays. The natural compounds 1−8 were further evaluated on resistant human colon adenocarcinoma cells (Colo320), and, additionally, their cytotoxicity was assessed on noncancerous mouse (NIH/3T3) and human (MRC-5) embryonic fibroblast cell lines. While spiropedroxodiol (1) was found to be a very strong MDR reversal agent in both L5178Y-MDR and Colo320 cells, the chemical modifications of helioscopinolide E (8) at C-3 positively contributed to increase the MDR reversal activity of compounds 10, 12, and 13. Furthermore, in combination assays, compounds 1 and 7−14 enhanced synergistically the cytotoxicity of doxorubicin. Finally, by means of molecular docking, the key residues and binding modes by which compounds 1−14 may interact with a murine P-glycoprotein model were identified, allowing additional insights on the efflux modulation mechanism of these compounds.

M

(BCRP, ABCG2) are the most frequently implied in the MDR phenotype.2 Regarding P-gp, several studies have reported that it is able to transport numerous substrates such as β-amyloid peptides, phospholipids, steroids, and hormones.3 Physiologically, P-gp is found at the apical surface of kidney proximal tubule cells, the canalicular membranes of hepatocytes, villous intestinal cells, and blood−tissue barriers, as in the brain, placenta, and testis, where it is thought to participate in cellular detoxification pathways. However, in cancer cells, the overexpression of this

ultidrug resistance (MDR) to anticancer agents is becoming an ever-greater worldwide health concern. Despite the establishment of several mechanisms implied in MDR, one of the most widely investigated is the overexpression of proteins on the surface of cancer cells, belonging to the ABC transporter superfamily. These act as efflux pumps by actively transporting chemotherapeutic drugs from the intracellular compartment and impairing their cytotoxic effects by lowering their intracellular concentration.1 From the known 49 transporters identified so far in human cells, P-glycoprotein (P-gp, ABCB1), multidrug resistance protein 1 (MRP1, ABCC1), and breast cancer resistance protein © XXXX American Chemical Society and American Society of Pharmacognosy

Received: April 26, 2018

A

DOI: 10.1021/acs.jnatprod.8b00326 J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products

Article

acetate soluble fraction of the methanol extract, this yielded a new spirotriterpenoid, named spiropedroxodiol (1). Also obtained were the seven known terpenoids 7,11-dioxoobtusifoliol (2),19 β-sitostenone (3),20 cycloart-23-ene-3β,25diol (4),21 cycloart-25-ene-3β,24-diol (5),22 and oleanolic acid (6)23 and the diterpene lactones helioscopinolides B (7)24 and E (8),25 identified by comparison of their spectroscopic data to those reported in the literature (Figure 1 and Supporting Information).

particular transporter is also correlated with the decrease of the intracellular concentration of anticancer drugs, impairing the efficacy of chemotherapeutic regimens.4 From the discovery by Tsuruo et al. that verapamil could reverse MDR in P388 leukemia cells and in its multidrugresistant P300/VCR counterpart, by acting as a P-gp efflux modulator,5 three generations of inhibitors have been designed so far. While the first-generation included pharmacologically active compounds such as verapamil, cyclosporine, and quinidine, the second-generation modulators comprised related compounds such as dexverapamil, dexniguldipine, and valspodar, but lacked intrinsic pharmacological activity. More recently, the thirdgeneration modulators were obtained by quantitative structure− activity relationships (QSAR) and high-throughput screening/ combinatorial chemistry and are increasingly more selective and potent toward P-gp.6 However, to date, no effective MDR efflux modulator is clinically available for coadministration with anticancer drugs, mainly due to the lack of in vivo efficacy of third-generation modulators in phase III trials.7 Toward that end, recent studies have identified a possible answer for the clinical failure of previously developed generations of P-gp modulators: the polyspecificity of the drugbinding pocket of this particular efflux pump, mainly due to the existence of multiple drug-binding sites.8−10 Nonetheless, the investigation of novel small molecules able to act as MDR reversal agents is still considered as a key approach in future drug development toward MDR reversal and has been recommended by the U.S. FDA as one of the most promising approaches to deal with P-gp efflux pump activity (which also occurs in cancer cells).11 Along with high-throughput screening of compound databases, the isolation of compounds from natural sources is considered one of the best approaches for the identification of new scaffolds with MDR reversal properties.12 Our search for effective MDR reversers13−17 has led to the phytochemical study of Euphorbia pedroi Molero & Rovira (Euphorbiaceae), an endemic plant of the western sea cliffs in Portugal.18 A new spirotriterpenoid (1) was isolated from the methanolic extract, along with seven known terpenoids (2−8). Helioscopinolide E (8), an ent-abietane diterpene, was derivatized, yielding compounds 9−14. Both the isolated compounds (1−8) and the synthesized derivatives (9−14) were evaluated for their activity as P-gp modulators through the rhodamine-123 (R123) accumulation assay, by flow cytometry, using the ABCB1-transfected mouse T-lymphoma L5178Y cell line as model. The isolated compounds 1−8 were assessed further on human colon adenocarcinoma cells (Colo320). All the compounds (1−14) were also evaluated for their cytotoxic and antiproliferative activities in both parental (PAR) and resistant (MDR) cell lines mentioned above. Additionally, the cytotoxicity of compounds 1−8 was also assessed on normal cell lines (NIH/3T3 and MRC-5). A drug combination assay was performed for compounds 1 and 7−14 on the L5178Y-MDR cell line, thus evaluating their effect in combination with the known anticancer drug doxorubicin. Finally, molecular docking studies were conducted to better understand how compounds 1−14 may act as P-glycoprotein efflux modulators.

Figure 1. Key COSY, HMBC, and main NOESY correlations for spiropedroxodiol (1).

Compound 1, named spiropedroxodiol, was isolated as a colorless oil with [α]20 D +12.0 (c 0.1, CHCl3). From its IR spectrum, characteristic absorption bands corresponding to hydroxy (3383 cm−1) and carbonyl (1691 cm−1) groups were observed. From the low-resolution ESIMS data a protonated molecular ion at m/z 459 [M + H]+ was identified. Furthermore, a sodium adduct ion peak at m/z 481.3655 [M + Na]+ (calcd for C30H50NaO3, 481.3652) was observed in the HRESIMS, from which a molecular formula of C30H50O3 and a degree of unsaturation of six were inferred. The 1H NMR spectrum of 1 revealed the presence of seven methyl groups, three tertiary (δH 0.66, 1.20, 1.47) and four secondary at δH 0.94 (J = 5.5 Hz), 0.96 (J = 5.5 Hz), 1.01 (J = 6.8 Hz), and 1.03 (J = 6.8 Hz). Furthermore, a terminal double bond (δH 4.65 and 4.72) and two oxymethines (δH 3.04 and 4.34) were additionally found. From the 13C NMR and DEPT experiments, 30 carbons could be proposed including seven methyl groups, 10 methylene

■

RESULTS AND DISCUSSION Characterization of the Isolated Compounds. The airdried powdered aerial parts of E. pedroi were extracted exhaustively with methanol. Following subsequent chromatographic fractionation and further purification of the ethyl B



Article

units (including one sp2 at δC 106.3), seven methine groups (including two oxygenated methines at δC 77.4 and 80.6), and six quaternary carbons including one carbonyl (δC 215.2) and one olefinic (δC 156.7). Although pointing to a tetracyclic skeleton, the 13C NMR spectrum suggested the presence of a spiro rearrangement from the presence of a quaternary carbon signal at δC 64.2, similar to the resonance value reported for C-9 of spiroterpenoids such as spiroinonotsuoxodiol and 3,7-dihydroxy4,14-dimethyl-7-(8→9)-abeo-cholestan-8-one, isolated from Inonotis obliquos and Euphorbia of f icinalis, respectively.26,27 Further structural details of compound 1 were obtained from 2D NMR experiments (COSY, HMBC, HMQC/HSQC, and NOESY), which allowed an unambiguous assignment of all carbon signals (Table 1, Figure 1). The 1H−1H COSY experiment enabled the identification of four key fragments (A−D), which were additionally connected by 2JC−H and 3JC−H HMBC correlations that allowed a determination of the locations of the functional groups in 1. In the 1 H−1H COSY spectrum, H-7 correlated with H-6 diastereotopic protons, while in an HMBC experiment 3JC−H correlations between Me-19 and the carbon signals at C-1, C-5, and C-9 and the heterocorrelations between H-6a, H-6b and C-5, C-7, and C-9 corroborated the presence of a quaternary carbon at C-9. Moreover, correlations of Me-18 with C-12, C-13, C-14, and C-17 and of Me-29 with the two quaternary carbons C-13 and C-14, the carbonyl group at C-8, and the methylene group at C-15 clarified the position of the carbonyl group and further corroborated the presence of a spiro skeleton. The position of the hydroxy group at C-7 (δC 80.6) agreed with the data reported in the literature for spiroinonotsuoxodiol (δC 80.6, C-7) and 3,7-dihydroxy-4,14-dimethyl-7-(8→ 9)-abeo-cholestan-8-one (δC 80.4, C-7). The relative configurations of the tetrahedral stereocenters of 1 were determined by a NOESY experiment (Figure 1). While cross-peaks between Me-18/Me-19/H-1a suggested a β-orientation of both methyl groups, NOE correlations between H-6a/H-7/Me-29/H-17 and H-3/Me-30/H-5/H-6a corroborated the α-orientation of these protons and the β-orientation of the hydroxy groups at C-3 and C-7. The 20R stereochemistry (H-20) was inferred from the chemical shift of the doublet (δH 0.94, Me-21) observed in the 1 H NMR spectrum of 1, which was slightly downfield (ΔδH ∼0.10 ppm) than the signal occurring in the spectra of other compounds in which the configuration of C-20 is inverted.26,27 Thus, compound 1 was assigned as 3,7-dihydroxy-7-(8→9)abeo-ergost-24(28)-en-8-one. The lowest energy structure (Figure 2), obtained by using the PM3 Hamiltonian in the semiempirical quantum chemistry program MOPAC (included in the MOE package), further supported the proposed assignments and further agreed with the NOESY data determined experimentally. It is also worth noting that the spiro scaffold described herein is highly unusual, but nonetheless a possible biosynthetic pathway has been suggested for similar compounds.28,29 Preparation of Helioscopinolide E (8) Derivatives. As helioscopinolide E (8) was obtained in a reasonable amount, a set of ent-abietane derivatives was prepared by chemical derivatization of the carbonyl moiety. According to the literature, the presence of nitrogen atoms and aromatic rings is characterized as key features for the P-gp efflux modulation.30,31 Thus, the rationale for the chemical derivatization of compound 8 was the introduction of these structural features, in order to potentiate the MDR reversal capabilities of its derivatives.

Table 1. NMR Spectroscopic Data (300 MHz, CDCl3) for Spiropedroxodiol (1)a position

δC (ppm)

1a 1b 2a 2b 3 4 5 6a 6b 7 8 9 10 11a 11b 12a 12b 13 14 15a 15b 16a 16b 17 18 19 20 21 22a 22b 23a 23b 24 25 26 27 28a 28b 29 30

30.3 28.8 77.4 38.3 48.3 37.9 80.6 215.2 65.2 47.8 30.3 30.9 48.3 61.1 29.7 27.2 50.3 16.6 18.6 35.8 18.9 34.9 31.4 156.7 33.9 22.1 22.0 106.3 19.7 16.3

δH [mult, J (Hz)] 1.60 2.00 1.41 1.98 3.04 1.66 1.22 2.39 1.39 4.34

(1H, (1H, (1H, (1H, (1H, (1H, (1H, (1H, (1H, (1H,

m)b m)b m)b m)b ddd, J = 12.0, 9.5, 5.3) m)b m)b dt, J = 13.3, 7.6) dd, J = 13.3, 3.4) dd, J = 8.0, 3.4)

1.60 2.02 1.74 2.07

(1H, (1H, (1H, (1H,

m)b m)b m) m)

1.25 1.88 1.33 1.96 1.67 0.66 1.47 1.40 0.94 1.15 1.58 1.92 2.13

(1H, (1H, (1H, (1H, (1H, (3H, (3H, (1H, (3H, (1H, (1H, (1H, (1H,

m)b m)b m)b m)b m)b s) s) m)b d, J = 5.5) m) m)b m) m)

2.22 1.03 1.01 4.65 4.72 1.20 0.96

(1H, (3H, (3H, (1H, (1H, (3H, (3H,

m) d, J = 6.8) d, J = 6.8) s) s) s) d, J = 5.5)

a

Signals assigned using 2D COSY, HMQC/HSQC, and HMBC NMR data and ref 28. bOverlapping signals.

Figure 2. Energy-minimized structure for spiropedroxodiol (1).

Herein, while the condensation of compound 8 and hydroxylamine hydrochloride (Scheme 1, i) yielded the corresponding C



Article

R123 accumulation assay, by flow cytometry. The natural compounds 1−8 were further evaluated on multidrug-resistant human colon adenocarcinoma cells (Colo320). The compounds were tested at 2.0 and 20 μM, and compound 1 was also tested at 0.2 μM. Verapamil was used as positive control (20 μM). The results are summarized in Tables 2 and 3.

oxime (9) (CN−OH), its subsequent acylation with acyl anhydride or chlorides (Scheme 1, ii) yielded five new derivatives Scheme 1. Preparation of the Helioscopinolide E (8) Derivatives 9−14a

Table 2. Effect of Compounds 1−8 on P-gp-Mediated R123 Efflux, in ABCB1-Transfected Mouse T-Lymphoma (L5178Y-MDR) and Human Colon Adenocarcinoma (Colo320) Cells fluorescence activity ratio compound

concentration (μM)

L5178Y-MDR

Colo320

0.2 2.0 20 2.0 20 2.0 20 2.0 20 2.0 20 2.0 20 2.0 20 2.0 20 20 2%

43.75 54.38 96.02 6.14 30.68 2.51 22.56 10.68 67.05 5.91 43.58 1.06 1.70 1.70 8.18 1.84 10.23 9.66 1.01

1.68 2.27 8.00 1.09 1.84 0.97 1.34 1.26 3.45 0.77 5.91 2.10 4.71 1.73 6.78 0.44 0.98 4.10 0.79

1

2

a

Reagents and conditions: (i) NH2OH·HCl (5 equiv) in pyridine, rt, overnight; (ii) 3.0 equiv anhydride and acyl chlorides, rt, 6−12 h.

3

(10−14). The imine function at C-3 in compound 9 was confirmed by the analysis of the 1H and 13C NMR spectra, in which a strong diamagnetic effect at this carbon (δC 215.7 → 166.3, ΔδC = −48.9) and diamagnetic effects at C-2 (δC 34.5 → 17.7, ΔδC = −16.8) and C-4 (δC 47.6 → 41.4, ΔδC = −6.2) corroborated the substitution of the carbonyl moiety by the corresponding imine. Similarly, a diamagnetic effect on the protons H-2 was also observed, together with deshielding effects of proton and carbon signals of the geminal methyl groups Me-18 and Me-19. Following the reaction of the oxime (9) with acetic anhydride, butyryl, benzoyl, furan-2-carbonyl, and tiophene-2carbonyl chlorides, this yielded derivatives 9−14. In the 13C NMR spectra of compounds 9−14, the new acyl moieties were corroborated structurally by observing the appearance of a new signal corresponding to the carbonyl of the ester linkage (ΔδC ≈ 171.0−175.0) along with small paramagnetic shifts (ΔδC = +3.2−8.8) of the imine carbon at C-3 (see Experimental Section). Antiproliferative and Cytotoxic Activities. The antiproliferative effects of all compounds obtained were assessed in ABCB1-transfected mouse T-lymphoma cells (L5178Y-MDR) and the corresponding parental cell line (L5178Y-PAR) using the thiazolyl blue tetrazolium bromide (MTT) assay.32,33 Additionally, the cytotoxicity of the isolated compounds 1−8 was also assessed in human colon adenocarcinoma cells (Colo320) and the sensitive cell line (Colo205) and in mouse (NIH/3T3) and human (MRC-5) embryonic fibroblast cell lines (Tables S1 and S2, Supporting Information). Herein, no significant antiproliferative activity and cytotoxicity were found for compounds 1−14 (IC50 > 10 μM). However, it is worth noting that while compound 6 showed an increased antiproliferative activity in resistant mouse lymphoma cells (Table S1, Supporting Information), compound 3 also displayed increased cytotoxic activity toward resistant human colon adenocarcinoma cells when compared with the corresponding sensitive cells and human embryonic fibroblast cell line (Table S2, Supporting Information). Modulation of the P-Glycoprotein Efflux. The ability of compounds 1−14 to modulate P-gp efflux was assessed on the parental mouse T-lymphoma cell line (PAR) and its transfected MDR subline (MDR), through the standard functional

4 5 6 7 8 verapamil DMSO

Table 3. Effect of Compounds 9−14 on P-gp-Mediated R123 Efflux in ABCB1-Transfected Mouse T-Lymphoma (L5178Y-MDR) Cells compound 9 10 11 12 13 14 verapamil DMSO

concentration (μM) fluorescence activity ratio L5178Y-MDR 2.0 20 2.0 20 2.0 20 2.0 20 2.0 20 2.0 20 20 2%

0.96 4.88 3.02 38.64 1.36 10.38 5.78 56.37 2.10 40.69 1.09 7.09 9.66 1.01

The P-gp-modulating potential of the test compounds was assessed by determining fluorescence activity ratio (FAR) values in the R123 efflux assay. To this end, FAR values are related to the cytoplasmic accumulation ratio of R123 between L5178Y-MDR and L5178Y-PAR cells. If P-gp modulation occurred, the FAR value for a given compound would be greater than 1, and when this ratio was higher than 10, compounds could be classified as strong modulators. During sample preparation and the flow cytometry assay, toxicity effects D



Article

Figure 3. Flow cytometry histograms in the R123 accumulation assay for spiropedroxodiol (1) at 0.2, 2.0, and 20 μM in the chemoresistant L5178Y-MDR and parental L5178Y-PAR (left) and in the human adenocarcinoma resistant Colo320 and parental Colo205 (right) cell lines.

were excluded due to the absence of significant changes in size (FSC) and granularity (SSC) of the cell population (data not shown). As shown in Table 2, spiropedroxodiol (1) was the most active compound, displaying considerable P-gp efflux modulation even when tested at lower concentrations such as 0.2 μM in both the L5178Y-MDR (FAR = 43.75, 54.38, and 96.02, at 0.2, 2.0, and 20 μM, respectively) and Colo320 (FAR = 1.68, 2.27, and 8.00 at 0.2, 2.0, and 20 μM, respectively) cell lines (Figure 3). It is notable that 7,11-dioxo-obtusifoliol (2) and the two cycloartane derivatives (4 and 5) also showed good profiles for P-gp efflux modulation at both 2.0 μM (FAR = 6.14, 10.68, and 5.91, respectively) and 20 μM (FAR = 30.68, 67.05, and 43.58, respectively), in agreement with previous results.34,35 Interestingly, oleanolic acid (6) was also found to be a good efflux modulator in Colo320 cells (FAR = 2.10 at 2.0 μM). When comparing the ability of compounds 9−14 as P-gp modulators (Table 3) with that of the parent compound, helioscopinolide E (8) (FAR = 1.84 and 10.23 at 2.0 and 20 μM, respectively), a significant improvement was observed for compounds 10, 12, and 13 (FAR values ranging from 38.64 to 56.37, at 20 μM), with compound 12, possessing a benzoyl moiety, being the most active. Conversely, the activity of compound 11, bearing a butyryl moiety, was similar to the parent compound, and introduction of the oxime function in compound 9 decreased the activity. Drug Combination Assay. Aiming at evaluating the type of in vitro interactions between the compounds and the anticancer drug doxorubicin, a known P-gp substrate, the isolated compounds 1, 7, and 8 and derivatives 9−14 were also assessed in a combination chemotherapy model using MDR mouse T-lymphoma cells (Table 4). The extent of interaction between doxorubicin and a given compound was calculated by the combination index (CI) as suggested by Chou36 and expressed using the CI for mutually exclusive drugs, with a CI ≈ 1 indicating additivity, CI < 1 defining synergisms, and CI > 1 being related to antagonism. As shown in Table 4, all tested compounds enhanced synergistically the cytotoxicity of doxorubicin, displaying in the case of compound 12 strong synergism (CI, 0.152). Molecular Docking Studies. As it was possible to verify that compounds 1−14 were predicted to interact with P-gp (using previously developed pharmacophoric models)37 (data not shown), molecular docking studies were performed to obtain further insights on how they can act as efflux modulators. Accordingly, a previously refined murine P-gp structure,

Table 4. Drug Combination Assays between Selected Compounds 1, 7, and 8 and Derivatives 9−14 with Doxorubicin Using Mouse T-Lymphoma (L5178Y-MDR) Cells compound

best ratioa

1 7 8 9 10 11 12 13 14

1:6 1:12.5 1:25 1:6 1:6 1:25 1:25 1:25 1:50

CI at ED50 (±SD)b

interaction

± ± ± ± ± ± ± ± ±

slight synergism synergism synergism synergism synergism synergism strong synergism synergism synergism

0.872 0.652 0.367 0.477 0.501 0.387 0.152 0.454 0.324

0.182 0.099 0.051 0.252 0.102 0.156 0.156 0.129 0.022

a

Data are shown as the best combination ratio between doxorubicin and compounds. bCombination index (CI) values are represented as the mean of three CI values calculated from different drug ratios ± standard deviation (SD) of the mean, for an inhibitory concentration of 50% (ED50). 0.1 < CI ≤ 0.3, strong synergism; 0.3 < CI ≤ 0.85, synergism; 0.85 < CI ≤ 0.9, slight synergism.

having 100% residue identity with human P-gp for the residues inside the drug-binding pocket,38 was used to identify the key residues, binding modes, and cross-interaction capability with P-gp, as previously reported. In the following paragraphs, the molecular docking results have been used to classify each compound as nonsubstrate, substrate, nontransported substrate, or modulator, unless stated otherwise.8 The results showed that while only compounds 1, 2, and 8−12 displayed similar binding preferences for the M- and R-sites (ΔΔG ≤ 0.2 kcal·mol−1, Table S3, Supporting Information), the binding energies for compounds 4−9 and 14 were all lower at the substrate binding R-site when compared with the other two drug-binding sites. Furthermore, as compound 13 was the only one with all docking poses at the R-site, this result suggests that this derivative acts mainly by competing with natural substrates (as a high-affinity substrate). Cycloart-23-ene-3β,25diol (4), oleanolic acid (6), and compounds 12−14 were the only compounds with energies below −10.0 kcal·mol−1. For helioscopinolide E (8) and derivatives 9−14, while the addition of aromatic moieties in compounds 12−14 improved binding affinities toward P-gp (ΔG from −10.1 to −10.6 kcal·mol−1), its relative binding preference toward the M-site remained unaffected (Table S3, Supporting Information), with compounds 13 and 14 preferring the substrate-binding R-site instead. E



Article

on a Shimadzu Affinity-1 FTIR spectrometer using NaCl pellets. NMR spectra were recorded on Bruker 400 (1H 400 MHz, 13C 101 MHz) and Bruker 300 (1H 300 MHz, 13C 75 MHz) Ultra-Shield instruments, with all 1H and 13C NMR chemical shifts expressed in δ (ppm) referenced to the solvent used and the proton coupling constants J in hertz (Hz). Spectra were assigned using appropriate COSY, DEPT, HMQC/HSQC, and HMBC sequences. ESIMS was performed on a Triple Quadrupole mass spectrometer (Micromass Quattro Micro API, Waters), and HRESIMS was obtained on a FTICR-MS Apex Ultra 7T instrument (Brüker Daltonics). Column and preparative thin-layer chromatography were performed using silica gel 60 (Merck Millipore 109385, 0.040−0.063 mm) and GF254 (Merck Millipore 105554), respectively. All reactions were monitored by TLC by using TLC silica 60 F254 coated plates, visualized under UV light and by spraying with sulfuric acid−methanol (1:1) followed by heating. The purity of all compounds (≥95%) was determined on a HPLC-UV system (Merck-Hitachi) using a Merck LiChrospher 100 RP-18 column (5 μm, 125 × 4 mm) and MeOH−H2O mixtures as mobile phase. Plant Material. Euphorbia pedroi was collected with authorization of the Instituto para a Conservaçaõ da Natureza e das Florestas (ICNF) at the Parque Natural da Arrábida (June 2011) and identified by the plant taxonomist Dr. Teresa Vasconcelos, of Instituto Superior de Agronomia, Universidade de Lisboa. A voucher specimen (no. LISI-06/2011) has been deposited at the herbarium of Instituto Superior de Agronomia. Extraction and Isolation. The air-dried powdered aerial parts of E. pedroi (17.9 kg) were exhaustively extracted with methanol (10 × 12.5 L) at room temperature. The solvent was evaporated under vacuum (at 40 °C) to yield a residue of 3.33 kg. Next, the residue was resuspended in MeOH−H2O (1:1) and extracted sequentially with n-hexane (3 × 0.3 L) and ethyl acetate (EtOAc, 10 × 12.5 L). The EtOAc-soluble fraction was then dried over Na2SO4 and evaporated under vacuum to obtain a residue (901 g) that was chromatographed over silica gel (6.5 kg), using as eluents mixtures of n-hexane−EtOAc (1:0 to 0:1) and EtOAc−MeOH (1:0 to 5:1). According to differences in composition (as indicated by TLC), 10 crude fractions were obtained (fractions A−L). Fractions D−G were selected for further workup due to their TLC profiles. Fraction D (2.71 g) was processed further through column chromatography (110 g of silica gel) with n-hexane−EtOAc (1:0 to 7:3, used in increased gradients of 10%, 500 mL each) and CH2Cl2− MeOH (1:0 to 9:1, 200 mL each), providing nine pooled fractions (D1−D9). Fraction D4 (338 mg) was again chromatographed by column chromatography (15 g of silica gel) with n-hexane−EtOAc (1:0 to 4:1, 500 mL each) to obtain five fractions (D4A−D4E). β-Sitostenone (3, 15 mg) was obtained by preparative TLC from fraction D4B, using CH2Cl2 as eluent. Fraction D5 was separated through column chromatography (60 g of silica gel), using n-hexane− EtOAc as mobile phase, from 1:0 to 7:3 with increasing gradients of 10%, to obtain four fractions (D5A−D5D). Cycloart-25-ene-3β,24diol (5, 56 mg) was isolated from fraction D5C through column chromatography (190 mg, 15 g silica gel) using CH2Cl2−acetone (1:0 to 4:1) as mobile phase. Cycloart-23-ene-3β,25-diol (4, 13 mg) was isolated from fraction D5D using a CombiFlash system, with standard columns (501 mg, 12 g silica gel prepacked column, RediSepRf, TeleDyne Isco) and CH2Cl2−acetone as eluent (1:0 to 4:1, 18 mL·min−1 with an increasing gradient of 2%) followed by preparative TLC using 1% CH2Cl2−MeOH as mobile phase. Fraction F was processed through column chromatography (9.2 g, 500 g of silica gel) using n-hexane−EtOAc (1:0 to 3:7, gradient of 10% in 500 mL portions) as eluents, leading to 12 pooled fractions (F1−F12). 7,11-Dioxo-obtusifoliol (2) was obtained from fraction F5 through three sequential purification steps. First, using the Combiflash system (587 mg, 43 g prepacked C18 column, RediSepRf, Teledyne Isco, with MeOH−H2O as mobile phase with an increasing gradient of 10% from 0:1 to 1:0, 10 mL·min−1), five fractions (F5A−F5E) were obtained. Next, fraction F5D was further separated using the Combiflash system (137 mg, 12 g prepacked silica gel column, RediSepRf, Teledyne Isco) with a CH2Cl−MeOH gradient (1:0 to 4:1,

Regarding residue interaction, it is worth noting that while compounds 1−5, 8, 9, and 12 revealed an increased preference for interaction with aromatic residues when at the M-site (Tables S4 and S5, Supporting Information), oleanolic acid (6) and helioscopinolide B (7) were the only compounds that displayed increased interactions with hydrophobic (30% and 42%) and polar (50% and 26%) residues, instead of aromatic residues (20% and 32%). As previous studies characterized the interaction with aromatic side-chains at the modulator M-site as crucial for the efflux inhibition by P-gp modulators (through π−π stacking or CH−π interactions),8,37,38 these results are in agreement with the experimentally observed results herein. Thus, compound 6 had the lowest FAR value for both cell lines, while compounds 1−5, showing a high percentage of interactions with aromatic residues (50% to 67%), consistently showed higher activities at both concentrations tested (Table 3). Furthermore, and although the same pattern can be observed for compounds 9−14, in which the presence of an aromatic moiety had in general a positive effect on their MDR reversal activities (Tables S3−S5, Supporting Information), the presence of the additional nitrogen and oxygen/sulfur atoms can also be expected to increase the preference for the molecule to bind to more polar pockets as the H- and R-sites. A closer inspection revealed an interaction pattern common to most compounds, involving residues Phe39, Leu296, Thr297, Phe696, and Ser697 (Tables S4 and S5, Supporting Information). However, while in compounds 1−5 and 12 additional interactions with the aromatic side chains of residues Tyr271, Phe300, and Phe696 were found, for compound 6 the interactions occurred mainly with the residues Gln294, Ile304, Gln689, and L939. Interestingly, an increased interaction with glutamine residues had already been suggested8,38 as more frequent in substrates: as can be seen in Figure S74 (Supporting Information), poor modulators show a reduced number of interactions with aromatic side chains (phenylalanines and tyrosines), being replaced by contacts with polar residues such as glutamines, serines, and threonines. The cross-interaction capability (CIc)8 also assessed the potential impact of compounds 1−14 in impairing conformational changes leading to efflux. Thus, compound 1 was observed to have one of the highest estimated CIc values (nontransported substrate, Figure S74A, Supporting Information), only behind 8 (nontransported substrate) and 6 (substrate), followed by compounds 12 (modulator) and 9 (nontransported substrate). These results suggest that while compounds 1−3, 8, and 12 may impair conformational changes when bound at the M-site due to its high affinity and good CIc, other compounds such as 4, 5, and 10−14 could act as competitive modulators when at the R-site or as weak to moderate efflux modulators when bound to the M-site. For oleanolic acid (6, Figure S74A, Supporting Information) and the oxime 9, the above results suggest that both compounds are unable to impair efflux due to the reduced interactions with aromatic residues (at the M-site, 6) and/or to the increased preference for binding at the R-site (9).

■

EXPERIMENTAL SECTION

General Experimental Procedures. All purchased solvents were used after distillation. Optical rotations were obtained in a PerkinElmer 241 polarimeter, with a 10 cm path length quartz cell at a temperature of 20 °C, and all samples were dissolved in spectroscopic grade CHCl3 or MeOH. Infrared spectra were determined F



Article

2% increments with an 8 mL·min−1 flow) and by sequential preparative TLC (n-hexane−acetone, 3:2; n-hexane−EtOAc, 4:1) to obtain 6 mg of compound 2 as a white powder. From fraction F6, oleanolic acid (6, 17 mg) was obtained as a white precipitate from n-hexane− EtOAc. Fraction F7 (412 mg) was submitted to reversed-phase chromatography using the Combiflash system (43 g prepacked C18 column, RediSepRf) with MeOH−H2O as mobile phase (0:1 to 1:0, increasing gradient of 10%, 10 mL·min−1), producing 11 pooled fractions; helioscopinolide B (7, 51 mg) was obtained from F7G by preparative TLC, using n-hexane−acetone (4:1) as eluent. In turn, helioscopinolide E (8, 239 mg) was obtained from fraction F8 by crystallization from n-hexane−EtOAc, with the mother liquor further separated by chromatography using the Combiflash system (1.33 g, 24 g prepacked silica gel column) with n-hexane−EtOAc as mobile phase (1:0 to 0:1). This afforded four subfractions (F8A−F8D); from fraction F8D, standard column chromatography (100 mg, 10 g silica gel) was performed in order to purify and obtain the compound spiropedroxodiol (1, 28 mg). Spiropedroxodiol (3,7-Dihydroxy-7(8→9)-abeo-ergost-24(28)en-8-one) (1): colorless oil; [α]20 D + 12.0 (c 0.1, CHCl3); IR (NaCl) νmax 3383, 1691, 1658, 1456, 1379 cm−1; 1H NMR (CDCl3, 300 MHz), δ 4.72 (1H, s, H-28b), 4.65 (1H, s, H-28a), 4.34 (1H, dd, J = 8.0, 3.4 Hz, H-7), 3.04 (1H, ddd, J = 12.0, 9.5, 5.3 Hz, H-3), 2.39 (1H, dt, J = 13.3, 7.6 Hz, H-6a), 2.22 (1H, m, H-25), 2.13 (1H, m, H-23b), 2.07 (1H, m, H-12b), 2.02 (1H, m, H-11b), 2.00−1.98 (2H, m, H-1b, H-2b), 1.96 (1H, m, H- 16b), 1.92 (1H, m, H-23a), 1.88 (1H, m, H-15b), 1.74 (1H, m, H-12a), 1.67 (1H, m, H-17), 1.66 (1H, m, H-4), 1.60 (2H, m, H1-a, H-11a), 1.58 (1H, m, H-22b), 1.47 (3H, s, H-19), 1.41 (2H, m, H-2a), 1.40 (1H, m, H-20), 1.39 (1H, dd, J = 13.3, 3.4 Hz, H-6b), 1.33 (1H, m, H- 16a), 1.25 (1H, m, H-15a), 1.22 (1H, m, H-5), 1.20 (3H, s, H-29), 1.15 (1H, m, H-22a), 1.03 (3H, d, J = 6.8 Hz, H-26), 1.01 (3H, d, J = 6.8 Hz, H-27), 0.96 (3H, d, J = 5.5 Hz, H-30), 0.94 (3H, d, J = 5.5 Hz, H-21), 0.66 (3H, s, H-18); 13 C NMR (75 MHz, CDCl3) δ 215.2, 156.7, 106.3, 80.6, 77.4, 64.2, 61.1, 50.3, 48.3, 48.3, 47.8, 38.3, 37.9, 35.8, 34.9, 33.9, 31.4, 30.9, 30.3, 30.3, 29.7, 28.8, 27.2, 22.1, 22.0, 19.7, 18.9, 18.6, 16.6, 16.3; ESIMS m/z 459 [M + H]+ (100); HREIMS m/z 481.36547 (calcd for C30H50NaO3, 481.36522). Preparation of 3-Hydroxyiminohelioscopinolide (9). To 200 mg (1 equiv) of helioscopinolide E (8), dissolved in dry pyridine (2.5 mL), was added 5 equiv of hydroxylamine hydrochloride. The mixture was then stirred at room temperature overnight, until the reaction was complete (as seen by analytical TLC, mobile phase n-hexane−EtOAc, 7:3). The solvent was removed under reduced pressure, and the residue was purified by column chromatography, using n-hexane−EtOAc (1:0 to 1:1), to afford 200 mg (0.61 mmol, yield 95%) of compound 9. 3-Hydroxyiminohelioscopinolide (9): amorphous, white powder; 1 H NMR (400 MHz, CDCl3) δ 6.31 (1H, s, H-14), 4.88 (1H, dd, J = 12.7, 5.5 Hz, H-12), 3.26 (1H, ddd, J = 14.8, 4.6, 3.3 Hz, H-2a), 2.57 (1H, d, J = 5.7 Hz, H-7a), 2.53 (1H, d, J = 5.9 Hz, H-7b), 2.26−2.12 (2H, H-1a, H-11a), 2.20 (1H, d, J = 8.4 Hz, H-9), 2.08 (1H, m, H-2b), 1.95−1.83 (1H, m, H-6b), 1.83 (3H, d, J = 1.6 Hz, H-17), 1.56−1.52 (1H, m, H-5), 1.52−1.43 (1H, m, H-6b, H-11b), 1.33 (1H, dt, J = 12.7, 4.6 Hz, H-1b), 1.20 (3H, ş H-19), 1.09 (3H, s, H-18), 1.04 (3H, s, H-20); 13C NMR (101 MHz, CDCl3) δ 175.2, 166.3, 155.9, 150.9, 117.0, 114.6, 75.9, 55.2, 51.2, 41.4, 40.6, 37.4, 36.8, 27.8, 27.6, 24.2, 23.1, 17.7, 16.6, 8.5; ESIMS m/z (positive mode) 330 [M + H]+. General Procedure for the Preparation of Derivatives 10−14. To 40 mg (1 equiv) of compound 9, dissolved in dry pyridine (2 mL), was added 3 equiv of the suitable anhydride/acyl chloride. The mixture was then stirred at room temperature until the reaction was complete, as controlled by analytical TLC (mobile phase n-hexane−EtOAc, 7:3). The solvent was evaporated, under reduced pressure (40 °C), and the residue was purified by column chromatography, using n-hexane−EtOAc (1:0 to 1:1). 3-Acetoxyiminohelioscopinolide (10): white, amorphous powder, from the reaction of compound 9 with acetic anhydride (34 mg, 75% yield); 1H NMR (400 MHz, CDCl3) δ 6.31 (1H, s, H-14), 4.86

(1H, dd, J = 13.5, 6.0 Hz, H-12), 3.13 (1H, dt, J = 15.1, 4.3 Hz, H-2a), 2.58−2.53 (2H, m, H-7a, H-7b), 2.51 (1H, dd, J = 13.2, 5.5 Hz, H-11a), 2.31 (1H, ddd, J = 15.1, 13.2, 5.3 Hz, H-2b), 2.20 (1H, d, J = 8.4 Hz, H-9), 2.18 (3H, s, H-2′), 2.07 (1H, ddd, J = 13.6, 4.4 Hz, H-1a), 1.94−1.86 (1H, m, H-6a), 1.83 (3H, d, J = 1.6 Hz, H-17), 1.55 (1H, m, H-5), 1.53−1.47 (1H, m, H-6b, H-11b), 1.39 (1H, dt, J = 13.1, 4.9 Hz, H-1b), 1.29 (3H, ş H-19), 1.15 (3H, s, H-18), 1.04 (3H, s, H-20); 13C NMR (101 MHz, CDCl3) δ 175.2, 171.7, 169.5, 155.7, 150.4, 117.1, 114.8, 75.8, 55.2, 51.0, 41.6, 41.2, 37.6, 36.7, 27.8, 27.5, 24.1, 20.1, 20.0, 19.9, 16.7, 8.5; ESIMS m/z (positive mode) 372 [M + H]+. 3-(Butyryloxy)iminohelioscopinolide (11): amorphous, white powder, from the reaction of compound 9 with butyryl chloride (28 mg; 58% yield); 1H NMR (400 MHz, CDCl3) δ 6.31 (1H, s, H-14), 4.86 (1H, ddd, J = 13.4, 6.2, 1.8 Hz, H-12), 3.11 (1H, dt, J = 15.1, 4.5 Hz, H-2a), 2.58−2.53 (2H, m, H-7a, H-7b), 2.51 (1H, dd, J = 13.6, 6.0 Hz, H-2b), 2.41 (2H, t, J = 7.4 Hz, H-2′), 2.31 (1H, ddd, J = 15.1, 13.0, 5.3 Hz, H-11b), 2.21 (1H, d, J = 8.1 Hz, H-9), 2.06 (1H, ddd, J = 13.1, 4.7 Hz, H-1a), 1.94 (1H, m, H-6a), 1.83 (3H, d, J = 1.6 Hz, H-17), 1.72 (2H, hex, J = 7.4 Hz, H-3′), 1.60 (1H, m, H-5), 1.62−1.53 (2H, m, H-6b, H-11b), 1.45 (1H, dt, J = 13.4, 4.1 Hz, H-1b), 1.30 (3H, ş H-19), 1.15 (3H, s, H-18), 1.04 (3H, s, H-20), 0.99 (3H, t, J = 7.4 Hz, H-4′); 13C NMR (101 MHz, CDCl3) δ 175.2, 173.7, 171.7, 155.7, 150.5, 117.1, 114.8, 75.8, 55.2, 51.0, 41.8, 41.2, 37.6, 36., 35.1, 7.8, 27.6, 24.1, 23.4, 20.4, 18.5, 16.7, 13.9, 8.5; ESIMS m/z (positive mode) 400 [M + H]+. 3-(Benzoyloxy)iminohelioscopinolide (12): amorphous, white powder obtained by reacting compound 9 with benzoyl chloride (37 mg; 70% yield); 1H NMR (400 MHz, CDCl3) δ 8.07 (2H, d, J = 7.1, H-3′), 7.61 (1H, t, J = 7.4 Hz, H-5′), 7.49 (2H, t, J = 7.5 Hz, H-4′), 6.32 (1H, s, H-14), 4.89 (1H, ddd, J = 13.6, 6.2, 1.8 Hz, H-12), 3.27 (1H, dt, J = 15.2, 4.4 Hz, H-2a), 2.63−2.54 (2H, m, H-7a, H-7b), 2.64−2.47 (1H, m, H-11a), 2.47 (1H, ddd, J = 15.1, 12.8, 5.7 Hz, H-2b), 2.25 (1H, br d, J = 8.1 Hz, H-9), 2.13 (1H, ddd, J = 13.1, 4.8 Hz, H-1a), 2.02−1.89 (1H, m, H-6a), 1.86 (3H, d, J = 1.6 Hz, H-17), 1.60 (1H, m, H-5), 1.60−1.54 (2H, m, H-6b, H-11b), 1.45 (1H, dt, J = 13.4, 4.3 Hz, H-1b), 1.41 (3H, ş H-19), 1.24 (3H, s, H-18), 1.09 (3H, s, H-20); 13C NMR (101 MHz, CDCl3) δ 175.2, 175.0, 164.3, 155.7, 150.5, 133.3, 130.2, 129.5, 128.7, 117.1, 114.8, 75.8, 55.2, 51.0, 41.9, 41.2, 37.6, 36.7, 27.8, 27.6, 24.1, 23.5, 20.5, 16.7, 8.5; ESIMS m/z (positive mode) 434 [M + H]+. 3-(Furan-2-carbonyl)oxyiminohelioscopinolide (13): amorphous, white powder from the reaction of 9 with furan-2-carbonyl chloride (32 mg; 61% yield); 1H NMR (400 MHz, CDCl3) δ 7.62 (1H, dd, J = 1.8, 0.8 Hz, H-5′), 7.25 (1H, dd, J = 3.5, 0.9 Hz, H-3′), 6.54 (1H, dd, J = 3.5, 1.7 Hz, H-4′), 6.32 (1H, s, H-14), 4.88 (1H, ddd, J = 13.3, 6.2, 1.8 Hz, H-12), 3.24 (1H, dt, J = 15.1, 4.5 Hz, H-2a), 2.60−2.56 (2H, m, H-7a, H-7b), 2.52−2.48 (1H, m, H-11a), 2.43 (1H, ddd, J = 15.1, 13.2, 5.6 Hz, H-2b), 2.23 (1H, br d, J = 8.2 Hz, H-9), 2.11 (1H, ddd, J = 12.9, 4.6, 0.6 Hz, H-1a), 1.97−1.89 (1H, m, H-6a), 1.84 (3H, d, J = 1.6 Hz, H-17), 1.60 (1H, dd, J = 8.6, 6.9 Hz, H-5), 1.62−1.54 (2H, m, H-6b, H-11b), 1.45 (1H, dt, J = 13.3, 5.2 Hz, H-1b), 1.37 (3H, s, H-19), 1.21 (3H, s, H-18), 1.07 (3H, s, H-20); 13C NMR (101 MHz, CDCl3) δ 175.2, 175.1, 166.7, 155.7, 150.4, 134.0, 132.8, 132.2, 128.0, 117.1, 114.8, 75.8, 55.2, 51.0, 41.9, 41.2, 37.6, 36.6, 27.8, 27.6, 24.1, 23.5, 20.4, 16.7, 8.5 ppm; ESIMS m/z (positive mode) 424 [M + H]+. 3-(Thiophene-2-carbonyl)oxyiminohelioscopinolide (14): amorphous, white powder from the reaction of compound 9 with thiophene2-carbonyl chloride (38 mg; 72% yield); 1H NMR (400 MHz, CDCl3) δ 7.88 (1H, dd, J = 3.8, 1.3 Hz, H-5′), 7.60 (1H, dd, J = 5.0, 1.3 Hz, H-3′), 7.15 (1H, dd, J = 5.0, 3.7 Hz, H-4′), 6.33 (1H, s, H-14), 4.87 (1H, dd, J = 13.1, 5.2 Hz, H-12), 3.23 (1H, dt, J = 15.1, 5.3 Hz, H-2a), 2.58−2.54 (2H, m, H-7a, H-7b), 2.55−2.47 (1H, m, H-11a), 2.45 (1H, ddd, J = 15.1, 13.1, 5.3 Hz, H-2b), 2.23 (1H, br d, J = 8.1 Hz, H-9), 2.13 (1H, ddd, J = 13.1, 4.6 Hz, H-1a), 1.96−1.84 (1H, m, H-6a), 1.83 (3H, s, H-17), 1.60 (1H, dd, J = 8.0, 5.5 Hz, H-5), 1.60−1.54 (2H, m, H-6b, H-11b), 1.45 (1H, dt, J = 13.4, 5.9 Hz, H-1b), 1.37 (3H, ş H-19), 1.22 (3H, s, H-18), 1.07 (3H, s, H-20); 13C NMR (101 MHz, CDCl3) δ 175.2, 174.8, 160.0, 155.7, G



Article

150.4, 134.0, 132.8, 132.2, 128.0, 117.1, 114.8, 75.8, 55.2, 51.0, 41.8, 41.2, 37.6, 36.6, 27.8, 27.6, 24.0, 23.5, 20.4, 16.7, 8.5; ESIMS m/z (positive mode) 440 [M + H]+. Cell Lines and Cultures. Parental and ABCB1-transfected L5178Y mouse T-lymphoma cells (ECACC cat. no. 87111908, U.S. FDA, Silver Springs, MD, USA)39 and human colon adenocarcinoma cell lines (Colo 205 parent and Colo 320/MDR-LRP expressing ABCB1), ATCC-CCL-220.1 (Colo320) and ATCC-CCL-222 (Colo205) were used as described in previous papers by our group.13−15 Mouse fibroblasts (NIH/3T3, ATCC CRL-1658TM) and human fibroblasts (MRC-5, ATCC CCL-171TM) were cultivated in DMEM (Gibco 52100-039) or EMEM (ATCC 30−2003), respectively, and supplemented with 10% heat-inactivated fetal bovine serum (Biowest, VWR International Kft, Debrecen, Hungary), 100 U/L of L-glutamine, 1% Na pyruvate, 1% penicillin/streptomycin (Sigma-Aldrich Chemie GmbH), and 0.1% nystatin (8.3 g/L in ethylene glycol). The adherent cells were detached using a combination of 0.25% trypsin and 0.02% EDTA for 5 min at 37 °C. Before each cytotoxicity assay using this cell line, cells were seeded in untreated 96-well flat-bottom microtiter plates, following a 4 h incubation period in a humidified atmosphere (5% CO2, 95% air) at 37 °C. All cell lines were incubated in a humidified atmosphere (5% CO2, 95% air) at 37 °C. Antiproliferative and Cytotoxicity Assays. The antiproliferative activity and cytotoxicity of the compounds were assessed with a range of decreasing concentrations (2-fold dilutions) in ABCB1transfected mouse T-lymphoma cell lines (L5178Y-PAR and L5178YMDR) (compounds 1−14). The cytotoxicity of isolated compounds 1−8 was additionally assessed in human colon adenocarcinoma (Colo205 and Colo320), mouse embryonic fibroblasts (NIH/3T3), and human lung embryonic fibroblasts (MRC-5). An antiproliferative assay was also performed in Colo205 and Colo320 cells for compounds 1−8. As previously described,13−15 test compounds were added to cells distributed into 96-well flat bottom microtiter plates at concentrations of 6 × 103 (antiproliferative) or 1 × 104 (cytotoxicity) and initially incubated for 72 h (antiproliferative) or 24 h (cytotoxicity), after which an MTT solution in phosphate-buffered saline (PBS) was added to each well and incubated for another 4 h. Following this, 100 μL of sodium dodecyl sulfate (10% in a 0.01 M HCl solution) was added to each well and incubated overnight at 37 °C. Cell growth was determined in quadruplicate by measuring optical density (OD) at 550 nm (reference 630 nm) in a Multiscan EX ELISA reader (Thermo Labsystems, Cheshire, WA, USA). The percentage of inhibition of cell growth was determined according to the equation ÅÄÅ OD ÑÉ Å ÑÑ sample − ODmedium control Ñ 100 − ÅÅÅÅ ÑÑ × 100 ÅÅÇ ODcell control − ODmedium control ÑÑÑÖ

FAR =

FL‐1MDRtreated /FL‐1MDRuntreated FL‐1PARtreated /FL‐1PARuntreated

Drug Combination Assay. The combination studies were designed using a fixed ratio of compounds across a concentration gradient. The dilution of doxorubicin was made in a horizontal direction (14.7−0.1 μM), while the dilutions of the MDR-reversal agents (at 2-fold of their IC50 values) were made vertically in a microtiter plate, to a final volume of 200 μL of medium per well. The cells were distributed in 100 μL aliquots into wells at a concentration of 2 × 105/mL, and the plates were incubated for 48 h, under standard conditions. Cell growth was determined after MTT staining, as above-described. Drug interactions were assessed according to Chou using the CalcuSyn v2.2 software.36 Each dose−response curve (individual agents as well as combinations) was fitted to a linear model using the median effect equation to obtain the median effect value (thus corresponding to the IC50) and slope (m). Goodness-of-fit was assessed using the linear correlation coefficient r, and only data with r > 0.90 was considered. Combination index values are represented as the means of three CI values calculated from different drug ratios ± standard deviation (SD) of the mean, for an inhibitory concentration of 50% (IC50). The extent of interaction between test compounds was expressed using CI for mutually exclusive substances. A CI close to 1 indicates additivity, CI < 1 defines synergism, and CI > 1 is related to antagonism. Molecular Docking. Molecular docking was performed using a previously published P-glycoprotein, derived from the original crystallographic data, comprising 100% identity between mouse and human structures for the residues inside the drug-binding pocket. MarvinSketch 17.24.042 was used for drawing structures. All ligands were exported to Molecular Operating Environment (MOE),43 minimized with the MMFF94x44 force field (adjusting hydrogen and lone pairs by default), and exported again as mol2 files in order to generate PDBQT files with AutoDockTools v1.5.6rc45 for utilization in AutoDock VINA 1.1.246 docking software. The binding location was defined by a docking box including the whole internal cavity defined by Aller et al.47 and in agreement with a previously published study,8 centered at the DBP and with dimensions xyz of 35.25, 25.50, and 45.25 Å, respectively (xy corresponds to the membrane plane). Due to the large search space volume (over 40 000 Å3), the “exhaustiveness” parameter was set to 50. Visual inspection of the docking poses was made in MOE to allow the identification of individual docking zones. Identification of binding residues, hydrogen bonding, and cross-interaction capability of compounds were evaluated using Ligplot48,49 and in-house python scripts.

■

ASSOCIATED CONTENT

S Supporting Information *

The results were expressed as the mean ± SD, and the IC50 values were obtained by best fitting the dose-dependent inhibition curves independently in Libreoffice Calc and in GraphPad Prism 5.03 for Windows software.40,41 Rhodamine-123 Accumulation Assay. As referred to in previous publications,13−15 cells were adjusted initially to a density of 2 × 106/mL, resuspended in a serum-free McCoy’s 5A medium (L5178Y cells) or RPM1 1640 (Colo cells), and distributed in 500 μL aliquots in Eppendorf centrifuge tubes. All test compounds were added at 2 and 20 μM with verapamil (positive control, EGIS Pharmaceuticals PLC, Budapest, Hungary) at 20 μM, and DMSO at 2% as solvent control. After 10 min of incubation at room temperature, 10 μL (5.2 μM final concentration) of R123 was added and further incubated for 20 min at 37 °C. The samples were washed twice, resuspended in 1 mL of PBS, and analyzed by flow cytometry (Partec CyFlow Space instrument, Partec GmbH, Münster, Germany). Histograms were evaluated regarding mean fluorescence intensity (FL-1), standard deviation, and both FSC and SSC parameters. The fluorescence activity ratio was calculated as the quotient between FL-1 of treated/untreated resistant cell line (L5178Y-MDR, Colo320) over treated/untreated sensitive cell line (L5178Y-PAR, Colo205), according to the following equation:

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jnatprod.8b00326.

■

Additional information (PDF)

AUTHOR INFORMATION

Corresponding Author

*Tel: +351 217946475. Fax: +351 217946470. E-mail: mjuferreira@ff.ulisboa.pt. ORCID

Ricardo J. Ferreira: 0000-0003-2590-8229 Maria-José U. Ferreira: 0000-0002-8742-1486 Present Address ⊥

Molecular Biophysics, Department of Cell and Molecular Biology, Uppsala Biomedicinska Centrum BMC, Husargatan 3, 751 24 Uppsala, Sweden. Notes

The authors declare no competing financial interest. H



■

Article

(27) Ferreira, M. J. U.; Ascenso, J. R.; Tavares, O. S. J. Nat. Prod. 1995, 58, 275−279. (28) Daoubi, M.; Marquez, N.; Mazoir, N.; Benharref, A.; Hernández-Galán, R.; Muñoz, E.; Collado, I. G. Bioorg. Med. Chem. 2007, 15, 4577−4584. (29) Handa, N.; Yamada, T.; Tanaka, R. Phytochemistry 2010, 71, 1774−1779. (30) Suzuki, T.; Fukazawa, N.; San-Nohe, K. J. Med. Chem. 1997, 40, 2047−2052. (31) Zamora, J. M.; Pearce, H. L.; Beck, W. T. Mol. Pharmacol. 1988, 33, 454−462. (32) Berridge, M.; Tan, A.; McCoy, K.; Wang, R. Biochemica 1996, 4, 4−9. (33) Mosmann, T. J. Immunol. Methods 1983, 65, 55−63. (34) Molnár, J.; Gyémánt, N.; Tanaka, M.; Hohmann, J.; BergmannLeitner, E.; Molnár, P.; Deli, J.; Didiziapetris, R.; Ferreira, M. J. U. Curr. Pharm. Des. 2006, 12, 287−311. (35) Ferreira, M. J. U.; Gyémánt, N.; Madureira, A. M.; Molnár, J. Anticancer Res. 2005, 25, 3259−3262. (36) Chou, T.-C. Cancer Res. 2010, 70, 440−446. (37) Ferreira, R. J.; dos Santos, D. J. V. A.; Ferreira, M. J. U.; Guedes, R. C. J. Chem. Inf. Model. 2011, 51, 1315−1324. (38) Ferreira, R. J.; Ferreira, M. J. U.; dos Santos, D. J. V. A. J. Chem. Theory Comput. 2012, 8, 1853−1864. (39) Pastan, I.; Gottesman, M. M.; Ueda, K.; Lovelace, E.; Rutherford, A. V.; Willingham, M. C. Proc. Natl. Acad. Sci. U. S. A. 1988, 85, 4486−4490. (40) Motulsky, H.; Hristopoulos, A. Fitting Models to Biological Data Using Linear and Nonlinear Regression: A Practical Guide to Curve Fitting; Oxford University Press: New York, 2004. (41) GraphPad Prism Software v.5.03; Graphpad: La Jolla, CA, 2015. (42) MarvinSketch v.17.2.20; ChemAxon: Cambridge, MA, 2017. (43) Molecular Operating Environment (MOE) v.2016.08; Chemical Computing Group Inc.: Montreal, QC, 2016. (44) Halgren, T. A. J. Comput. Chem. 1999, 20, 720−729. (45) Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew, R. K.; Goodsell, D. S.; Olson, A. J. J. Comput. Chem. 2009, 30, 2785− 2791. (46) Trott, O.; Olson, A. J. J. Comput. Chem. 2010, 31, 455−461. (47) Aller, S. G.; Yu, J.; Ward, A.; Weng, Y.; Chittaboina, S.; Zhuo, R.; Harrell, P. M.; Trinh, Y. T.; Zhang, Q.; Urbatsch, I. L.; Chang, G. Science 2009, 323, 1718−1722. (48) Wallace, A. C.; Laskowski, R. A.; Thornton, J. M. Protein Eng., Des. Sel. 1995, 8, 127−134. (49) Laskowski, R. A.; Swindells, M. B. J. Chem. Inf. Model. 2011, 51, 2778−2786.

ACKNOWLEDGMENTS This project received funding from European Structural & Investment Funds through the COMPETE Programme and from National Funds through FCT, Fundaçaõ para a Ciência e a Tecnologia, under the Programme grants PTDC/ QEQMED/0905/2012, UID/DTP/04138/2013, SAICTPAC/0019/2015, and PTDC/MED-QUI/30591/2017. R.J.F. acknowledges FCT for the Ph.D. grant SFRH/BD/84285/ 2012. The study was also supported by the project GINOP2.3.2-15-2016-00012. We also acknowledge Dr. T. Vasconcelos, ISA, Universidade de Lisboa, for plant material identification.

■

REFERENCES

(1) Szakács, G.; Paterson, J. K.; Ludwig, J. A.; Booth-Genthe, C.; Gottesman, M. M. Nat. Rev. Drug Discovery 2006, 5, 219−234. (2) Eckford, P. D. W.; Sharom, F. J. Chem. Rev. 2009, 109, 2989− 3011. (3) Borges-Walmsley, M. I.; McKeegan, K. S.; Walmsley, A. R. Biochem. J. 2003, 376, 313−338. (4) Gottesman, M. M.; Lavi, O.; Hall, M. D.; Gillet, J.-P. Annu. Rev. Pharmacol. Toxicol. 2016, 56, 85−102. (5) Tsuruo, T.; Lida, H.; Tsukagoshi, S.; Sakurai, Y. Cancer Res. 1981, 41, 1967−1972. (6) Palmeira, A.; Sousa, E.; Vasconcelos, M. H.; Pinto, M. Curr. Med. Chem. 2012, 19, 1946−2025. (7) Colabufo, N. A.; Berardi, F.; Cantore, M.; Contino, M.; Inglese, C.; Niso, M.; Perrone, R. J. Med. Chem. 2009, 53, 1883−1897. (8) Ferreira, R. J.; Ferreira, M. J. U.; dos Santos, D. J. V. A. J. Chem. Inf. Model. 2013, 53, 1747−1760. (9) Ferreira, R. J.; Ferreira, M. J. U.; dos Santos, D. J. V. A. WIREs Comput. Mol. Sci. 2015, 5, 27−55. (10) Ferreira, R. J.; Ferreira, M. J. U.; dos Santos, D. J. V. A. J. Chem. Theory Comput. 2015, 11, 4525−4529. (11) Food and Drug Administration. In Vitro Metabolism- and Transporter-Mediated Drug-Drug Interaction Studies; 2017; pp 1−47. (12) Barnes, E. C.; Kumar, R.; Davis, R. A. Nat. Prod. Rep. 2016, 33, 372−381. (13) Reis, M.; Paterna, A.; Mónico, A.; Molnar, J.; Lage, H.; Ferreira, M. J. U. Planta Med. 2014, 80, 1739−1745. (14) Reis, M. A.; Paterna, A.; Ferreira, R. J.; Lage, H.; Ferreira, M. J. U. Bioorg. Med. Chem. 2014, 22, 3696−3702. (15) Vieira, C.; Duarte, N.; Reis, M. A.; Spengler, G.; Madureira, A. M.; Molnár, J.; Ferreira, M. J. U. Bioorg. Med. Chem. 2014, 22, 6392− 6400. (16) Paterna, A.; Gomes, S. E.; Borralho, P. M.; Mulhovo, S.; Rodrigues, C. M. P.; Ferreira, M. J. U. J. Ethnopharmacol. 2016, 194, 236−244. (17) Ferreira, M. J. U.; Duarte, N.; Reis, M.; Madureira, A. M.; Molnár, J. Phytochem. Rev. 2014, 13, 915−935. (18) Rovira, A.; Molero, J. An. Jard. Bot. Madrid 1997, 55, 189−200. (19) Tanaka, R.; Kasubuchi, K.; Kita, S.; Matsunaga, S. Phytochemistry 1999, 51, 457−463. (20) Gopalakrishnan, M.; Narayanan, C. S.; Grenz, M. J. Agric. Food Chem. 1990, 38, 2133−2136. (21) de Pascual Teresa, J.; Urones, J. G.; Marcos, I. S.; Basabe, P.; Cuadrado, M. S.; Fernandez Moro, R. Phytochemistry 1987, 26, 1767−1776. (22) Anjaneyulu, V.; Rao, G. S.; Connolly, J. D. Phytochemistry 1985, 24, 1610−1612. (23) Huang, K. F.; Liou, L. E. Chin. Pharm. J. 1997, 49, 305−314. (24) Crespi-Perellino, N.; Garofano, L.; Arlandini, E.; Pinciroli, V.; Minghetti, A.; Vincieri, F. F.; Danieli, B. J. Nat. Prod. 1996, 59, 773− 776. (25) Borghi, D.; Baumer, L.; Ballabio, M.; Arlandini, E.; Perellino, N. C.; Minghetti, A.; Vincieri, F. F. J. Nat. Prod. 1991, 54, 1503−1508. (26) Nes, W. R.; Thankamma, E. V.; Krevitz, K. J. Am. Chem. Soc. 1977, 99, 260−262. I


Terpenoids from Euphorbia pedroi as Multidrug-Resistance Reversers

Recommend Documents