Leaves of O. ferruginea, i.e., Olea europaea var. ferruginea (Royle), were collected from Tarnab, District Peshawar, Pakistan, during the summer season. The leaves were identified as O. ferruginea leaves by Dr. Muhammad Ilayas (University of Swabi). The voucher specimen number UOS/Bot104 was deposited in the herbarium of the Department of Botany of the University of Swabi, Pakistan. The extraction was carried out through modification of a protocol reported previously [6]. The leaves of O. ferruginea were dried in the shade at room temperature and then ground into a fine powder using a grinding machine; 5 kg of powder was extracted with 5 L of a H${}_2$O/methanol (3:7) mixture for 15 days at room temperature. The resulting extract was filtered and concentrated using a rotavapor. This process was repeated twice more to achieve complete extraction; 1.5 kg of crude extract was then suspended in 2 L of distilled water for extraction with n-hexane, dichloromethane, and ethyl acetate. The ethyl acetate fraction, after evaporation, which provided 170 g of extract, was loaded into a silica gel chromatographic column. The elution was performed with a gradient of dichloromethane and methanol (from 100:0 to 60:40), and 6 fractions were collected. Fraction 4 was re-chromatographed with n-hexane and ethyl acetate as the mobile phase, increasing solvent polarity from 100:0 to 20:80, giving 5 subfractions. Subfraction 3 provided cycloolivil ((1) 60 mg) and ferruginan ((2) 15 mg), which were obtained as white solids after further purification by column chromatography with n-hexane and ethyl acetate (100:0–30:70). On the other hand, ferruginan A ((3) 35 mg) was obtained from subfraction 4, which was processed similarly. The compounds were fully characterized, and analytical profiles were in agreement with previously reported data [7, 9].

The compounds extracted from O. ferruginea were evaluated for their antiproliferative activity on the U-87 MG malignant glioma cells. We profoundly acknowledge Prof. Norah Defamie, Pôle Biologie Sante Université de Poitiers, France, for kindly providing the cells that were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and that are identified by the ATCC code HTB-14. Cells were cultured in the Cancer Cell Culture & Precision Oncomedicine Lab, Institute of Basic Medical Sciences, Khyber Medical University Peshawar, Pakistan, and the cell line was used in agreement with previous studies from our group [15, 16]. Authentication of cell lines was performed by the authors, and PCR analysis was carried out [15, 17]. The cells were cultured in DMEM media (Gibco, USA) supplemented with 10% FBS (Cegrogen biotech, Germany) and 1% penicillin-streptomycin (Gibco, USA) at 37 °C in a 5% CO${}_2$ humidified incubator (Humanlablne, Korea). Upon reaching confluency, the cells were seeded in a 96-well plate, and a growth inhibition assay was performed following a procedure reported previously [18]. Briefly, 5000 cells were seeded into each well of a 96-well plate and allowed to grow overnight at 37 °C with 5% CO${}_2$ supply. Stock solutions of cycloolivil (1), ferruginan (2), and ferruginan A (3) were prepared in DMSO (3 mg/mL). The cells were then treated in triplicate with different dilutions of the compounds (400, 200, 100, 50, 25, 12.5, and 6.25 µM), and an equivalent volume of vehicle was used as a negative control. In addition, cisplatin was used as the positive control, utilizing the same dilutions as the test compounds. Then, the cells were incubated in a CO${}_2$ incubator at 37 °C for 24 hours. Afterward, cells were fixed with 4% formalin (Scharlab, Spain) for 10 minutes at room temperature. Following fixation, 100 µL of 0.1% crystal violet (Daejung, Korea) was added to each well, and then a 10 min incubation at room temperature was performed. Cells were washed three times with PBS (Sigma Aldrich, St. Louis, MO, USA), and the plates were allowed to dry overnight. Finally, 200 µL of acetic acid (Merck, Germany) was added to each well to solubilize the crystal violet stain, and absorption spectra were recorded at 590 nm using a spectrophotometer (Biotek-ELISA, Santa Clara, CA, USA). IC${}_{50}$ values were calculated for each compound based on linear regression analysis.

Molecular docking studies were performed using Molecular Operating Environment (MOE 2016.0802). X-ray crystal structure of duplex DNA containing a cisplatin intrastrand cross-link was obtained from the Protein Data Bank (PDB ID 3LPV) [19, 20], as also the crystal structure of EGFR (PDB ID 3POZ). The downloaded crystal structures were prepared and 3D protonated by using the “Prepare” module of MOE [21, 22]. For docking simulations, docking grids were determined within 10 Å from the binding site. In the case of EGFR, the docking procedure was validated by using the re-docking method. Cognate ligand 03P, also known as TAK-285, was re-docked in the binding site. The pose showed a root-mean-square deviation (RMSD) of 2.0 Å with respect to the cognate, and interactions with key amino acid residues were maintained. For all the ligands, ten docking poses were generated, and the top-ranked conformations based on docking score were selected. Furthermore, we also performed docking studies by using AutoDock 4.2 [23] to validate the results using the same parameters. The docking studies were carried out following our previously reported method [21]. Ligand interaction and visualization studies were carried out using Discovery Studio Visualizer (BIOVIA, Dassault Systèmes, Waltham, MA, USA), while 3D artworks were processed using UCSF Chimera [24].

Concerning the studies focused on the antiproliferative activity of compounds from O. ferruginea, only a few reports can be retrieved in the literature. Ferruginan (2) was previously described as a cytotoxic agent in the MCF-7 breast cancer cell line (IC${}_{50}$ = 10.41 µg/mL) [9]. Hashmi et al. [10] tested quercetin, $\beta$-amyrin, oleuropein, and ligstroside isolated from O. ferruginea against HeLa and Vero cells, while Liaqat et al. [11] evaluated the cytotoxicity on HepG2 human liver cancer cell line of hexane, chloroform, ethanol, methanol, and water extracts of O. ferruginea barks, where the ethanol extract was highlighted as the most promising.

The treatment of gliomas, tumors affecting the CNS, is very challenging. Despite new surgery protocols and advances in treatments and radiation techniques, the survival rate struggles to improve, mainly due to recurrence [25]. Thus, the search for novel therapeutic tools of natural and synthetic origin is wide open [26, 27]. In this work, to evaluate the anticancer effects of cycloolivil (1), ferruginan (2), and ferruginan A (3), U-87 MG malignant glioma cells were treated with the compounds (6.25–400 µM), and the concentration range was selected in agreement with previous studies [9, 10, 11]. Cell survival was plotted against drug concentration, and IC${}_{50}$ values were calculated. The result indicated that the cell survival decreased upon the increase of compound concentration (Supplementary Fig. 2), and the IC${}_{50}$ values are reported in Table 1. Overall, the compounds showed mild cytotoxic activity, with IC${}_{50}$ values in the µM range. In particular, in the current experimental model, ferruginan A (3) was highlighted as the most promising compound of the set, as an IC${}_{50}$ value of 216.4 $\pm$ 0.1 µM was calculated for this molecule against U-87 MG cells. For reference, cisplatin was tested, and an IC${}_{50}$ value of 124.3 $\pm$ 0.1 µM was obtained for this compound under the mentioned experimental conditions.

Computational drug discovery tools were enrolled in the current study to assist in the interpretation of experimental data. Double stranded DNA (dsDNA) is one of the potential targets of compounds developed for the treatment of glioma, and it also represents the macromolecular interactor of cisplatin. In particular, cisplatin binds the N7-purine bases, resulting in cell division block and apoptotic cell death, and several recent studies report the efficacy of therapies based on the combination of cisplatin and other cytotoxic drugs [12]. Thus, the first molecular modeling simulation was performed by docking the isolated compounds to the structure of a dsDNA possessing a cisplatin 1,2-d(GpG) intrastrand cross-link (PDB ID 3LPV). The results of the docking study are depicted in Fig. 2, in which the 2D interaction maps, as well as the 3D models, are reported for all the compounds. Cycloolivil (1) interacts through hydrogen bonds with adenine 17 and guanine 16 and with guanine 14 through a hydrophobic interaction. Ferruginan (2) binds thymine 8, adenine 17, and adenine 15 through hydrogen bonds thanks to the hydroxyl groups. The interaction pattern is slightly different for ferruginan A (3). Hydrogen bonds are formed with adenine 15, cytosine 17, and cytosine 18, but $\pi$-$\pi$ interactions with adenine 15 and guanine 17 can also be observed. The calculated binding energy values computed for cycloolivil (1), ferruginan (2), and ferruginan A (3) are –6.8, –5.9, and –7.0 kcal/mol, respectively. The compounds showed the same trend when AutoDock was used. In this case, the values computed for cycloolivil (1), ferruginan (2), and ferruginan A (3) are –6.1, –5.6, and –6.6 kcal/mol, respectively.

Fig. 2.

Molecular docking study for the interaction of cycloolivil (1, a), ferruginan (2, b), and ferruginan A (3, c) with dsDNA. 2D interaction maps and 3D models are depicted for the tested compounds.

The compounds could potentially exert their antiproliferative activity through several molecular mechanisms. Among these, it must be noted that EGFR is a target of primary relevance in the context of the treatment of glioma, as amplification and mutation of the EGFR gene were observed in these tumors [13, 14]. The EGFR gene encodes for a transmembrane glycoprotein from the protein kinase superfamily. This protein works as a receptor for the epidermal growth factor (EGF): on the cell surface, EGFR binds EGF, promoting receptor dimerization and tyrosine autophosphorylation, resulting in enhanced cell growth [13]. Thus, in view of a potential multi-target mechanism involving the isolated compounds in their antiproliferative activity against glioma cells, another molecular docking study was performed considering the 3D structure of EGFR (PDB ID 3POZ). The results of the docking study are depicted in Fig. 3, in which the 2D interaction maps, as well as the 3D models, are reported for all the compounds. Cycloolivil (1) interacts with Lys745, Asp885, Cys797, and Asn842 through hydrogen bonds. Similarly, ferruginan (2) binds Arg841, Met793, Leu788, and Gln791, and interactions are mainly mediated by hydroxyl groups. Ferruginan A (3) interacts with the protein through a pattern that is intermediate between the two previous compounds. More specifically, Cys797 and Lys745 are targeted like in the case of cycloolivil (1), while the compound also binds Gln791 as in the case of ferruginan (2). Overall, in this case, the docking study suggests that the -OH groups represent the main groups mediating the interaction with the macromolecular target. The calculated binding energy values computed for cycloolivil (1), ferruginan (2), and ferruginan A (3) are –7.0, –8.0, and –7.4 kcal/mol, respectively. Also, in this case, the compounds showed the same trend when AutoDock was used. In this case, the values computed for cycloolivil (1), ferruginan (2), and ferruginan A (3) are –7.5, –8.3, and –7.8 kcal/mol, respectively. Additionally, the cognate ligand was re-docked for reference with both software and values of –10.4 kcal/mol and –10.9 kcal/mol were obtained by using MOE and AutoDock, respectively.

Fig. 3.

Molecular docking study for the interaction of cycloolivil (1, a), ferruginan (2, b), and ferruginan A (3, c) with EGFR. 2D interaction maps and 3D models are depicted for the tested compounds.