Phytoconstituents with cytotoxic activity from Ulmus pumila L

Farouk R. Melek1*, Soheir M. El Zalabani2, Neveen S. Ghaly1, Omar M. Sabry2, Walid Fayad3, Ann G. Boulis1 1Chemistry of Natural Compounds Department, National Research Centre, 60014618, Dokki, Giza, 12622, Egypt 2Pharmacognosy Department, Faculty of Pharmacy, Cairo University, Cairo, 11562, Egypt 3Drug Bioassay-Cell Culture Laboratory, Pharmacognosy Department, National Research Centre, Dokki, Giza, 12622, Egypt

Ulmus pumila L., renowned as Asiatic elm, Chinese elm, and dwarf elm, is a deciduous tree belonging to central Asia. In folk medicine, its leaf and stem bark extracts are employed as diuretic, demulcent, antipyretic, and laxative remedies (Duke and Ayensu, 1985). Ulmus pumila L. was reported to possess large amounts of phenols and flavonoids with potent antioxidant activities (Kim et al., 2010). In addition, previous studies on the constituents of the root bark of this species led to the characterization of two potentially cytotoxic sesquiterpenoids, namely, mansonones E and F (Wang et al., 2004), as well as various bioactive triterpenoids (Wang et al., 2006). Moreover, four triterpenoids, namely, oleanolic acid, friedelin, maslinic acid, and arjunolic acid, were also isolated from the methanol extract of U. pumila L. (Ghosh et al., 2012). Furthermore, a recent phytochemical study on the stem bark extract of U. pumila L. led to separation of Icariside E 4 which strongly prohibited nitric oxide generation in LPSactivated macrophages (Joo et al., 2014). As an extension to our interest in exploring bioactive compounds from natural sources, we described in this report the isolation and identification of 13 compounds from U. pumila L. stem bark and leafy branches. The cytotoxic activity of the stem bark methanol extract and some of the isolated triterpenoids against five human carcinoma cell lines was also reported.

Plant material
The stem bark and leafy branches of U. pumila L. were gathered from Orman Botanical Garden, Giza, Egypt, in January and February 2018. A voucher specimen, encoded M131, was submitted to the National Research Centre herbarium, Giza, Egypt.

General experimental procedures
Vacuum liquid chromatography (VLC) was achieved with silica gel H 60 (E-Merck, Darmstadt, Germany) and polyamide 11 (E-Merck, Darmstadt, Germany). Preparative and analytical thin layer chromatography were carried out using silica gel (E-Merck, Darmstadt, Germany). Chromatograms were first visualized under ultraviolet (UV) light and then sprayed with 20% sulfuric acid in methanol or ferric chloride reagent. Column chromatography was performed using Sephadex LH-20 (Sigma-Aldrich, St. Louis, MO). Infrared (IR) spectra were run on a JASCO FT/IR-6100 Fourier Transform IR Spectrometer (Oklahoma, USA). Mass spectra (MS) were acquired by means of a Thermo ISQ Single Quadrupole Mass Spectrometer (THERMO Scientific Corporation, USA). UV spectra were displayed on a Shimadzu Double Beam Spectrophotometer UV-1650 (Shimadzu, Japan). Nuclear magnetic resonance (NMR) spectra were obtained via a Bruker High Performance Digital FT-NMR-Spectrophotometer Avance III HD ( 1 H-NMR: 400 MHz, 13 C-NMR: 100 MHz, Bremen, Germany). Chemical shifts were expressed on the δ scale and tetramethylsilane was used as an internal standard.

Extraction and isolation of stem bark and leafy branch constituents
Air-dried powdered stem bark and leafy branches (1 and 1.2 kg, resp.) were separately extracted with methanol (5 l × 3) at room temperature. Upon vacuum evaporation, the stem bark extract yielded a reddish-brown residue (50 g) and the leafy branch extract yielded a dark green residue (110 g). A portion of each dried extract (45 g of stem bark extract and 100 g of leafy branch extract) was individually suspended in distilled water (500 ml) and then partitioned with dichloromethane (250 ml × 5), ethyl acetate (250 ml × 5), and water-saturated n-butanol (300 ml × 4), in succession. The solvent-free dichloromethane fraction (19 g) from the stem bark extract and ethyl acetate fraction from leafy branch extract (4.2 g) were subjected to VLC (silica gel 500 g, and polyamide 11 250 g, resp.).
Elution of the polyamide 11 bed was started with H 2 O and then by 10% increments of MeOH up to 80%. Eighty fractions, 100 ml each, were collected and monitored by TLC (solvent system, EtOAc-MeOH-H 2 O 30:5:4 v/v/v). Spots were detected in visible and UV (365 nm) lights, before and after exposure to ammonia vapor or spraying with ferric chloride. Fractions eluted with 10, 20, and 30% MeOH, with compound 10 as the major component, were mixed. After evaporating the solvent, the residue was subjected to column chromatography (Sephadex LH-20; eluent, H 2 O-MeOH 1:1 v/v) to yield compound 10 (11.5 mg). Fractions eluted with 50% and 60% MeOH were combined based on TLC analysis. After solvent evaporation, the residue was chromatographed on a Sephadex LH-20 column (eluent, H 2 O-MeOH 1:1 v/v) to yield a mixture of compounds 11 and 12 (41.1 mg) together with compound 13 (10.5 mg).

Identification of the isolated compounds
Compounds 1-13 shown in Figure 1 were identified based on the following spectral data: Friedelin (

Acid hydrolysis of glycosides
Aliquots (3 mg, each) of the isolated glycosides were separately subjected to acid hydrolysis. The sample was dissolved in 3 ml of 2N hydrochloric acid-methanol mixture (1:1 v/v) and heated under reflux on a water bath for 2 hours. The reaction mixture was further evaporated under vacuum to dryness, and the residue was suspended in distilled water (10 ml) and then repeatedly extracted with ethyl acetate. The ethyl acetate layer was subjected to TLC (solvent system, CHCl 3 -MeOH 5:1 v/v) alongside reference aglycones. Meanwhile, the aqueous layer was diluted with methanol and evaporated to dryness and the residue obtained was investigated by PC and TLC-(solvent systems: n-butanolacetic acid-water 4:1:5 v/v/v, upper layer; and isopropanol:water 7:1 v/v, resp.) for detection of sugar moieties (Mabry et al., 1970). Hydrolysis of compound 10 afforded D-glucose and L-rhamnose, while the mixture of compounds 11 and 12 yielded D-glucose and D-galactose.
Cell culture was carried out under sterile conditions using a laminar airflow cabinet biosafety class II level. The culture was maintained in McCoy's 5a medium in case of HCT-116 cell line, Dulbecco's Modified Eagle Medium: Nutrient Mixture F12 in case of A549 and RPE-1 cell lines, and Eagle's Minimum Essential Medium in case of MCF7, HepG2, and HOS cell lines. The culture media were supplied with 1% antibiotic-antimycotic mixture (10,000 U/ml potassium penicillin, 10,000 µg/ml streptomycin sulfate, and 25 µg/ml amphotericin B), 1% L-glutamine, and 10% heat-inactivated fetal bovine serum. Cisplatin was used as a positive control and 0.5% DMSO solution as a negative control (Thabrew et al., 1997).

Cell viability assay
The cells were seeded at concentrations of 10,000 cells/ well in case of MCF-7, HepG2, A549, and HOS cell lines and 20,000 cells/well in case of HCT-116 and RPE-1 cell lines, using 96-well microtiter plastic plates at 37°C for 24 hours under 5% CO 2 in a carbon dioxide incubator. Stock solutions of the test isolates were prepared at concentrations of 20, 10, 5, and 2.5 mg/ ml for the stem bark methanol extract and 20, 10, 5, and 2.5 mM for each isolate. Culture media were aspirated from the cell culture plates, and four different concentrations of the test isolates were prepared, in triplicates. This was carried out by adding an aliquot (1 μl) of each stock solution of the test isolates to fresh medium with cells (199 μl) in each well to reach final concentrations of 100, 50, 25, and 12.5 μg/ml for the crude extract and 100, 50, 25, and 12.5 μM for each isolate. After incubation for 72 hours, the media were aspirated and 40 µl of 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide (MTT) salt (2.5 μg/ml) was added to each well followed by 4 h. incubation at 37ºC under 5% CO 2 . In order to stop the reaction and dissolve the formed crystals, 200 µl of 10% solution of sodium dodecyl sulfate in deionized water was added to each well, followed by incubation overnight, at 37°C. The absorbance was measured using a microplate multiwell reader at 595 nm. Cell viability was determined using a modified procedure of MTT assay, based on mitochondrial-dependent reduction of the yellow MTT to purple formazan (Mosmann, 1983). The cytotoxicity percentage was calculated as follows: % Cytotoxicity = [1− (Avg X /Avg NC )] × 100, where Avg indicates average; X indicates absorbance of sample, and NC indicates absorbance of the negative control.
The half maximal inhibitory concentration (IC 50 ) values of the tested samples were calculated using GraphPad Prism software (version 5.0.), and the selectivity indices (SI) of the cytotoxic samples were deduced from the following equation (Pritchett et al., 2014): SI = IC 50 of sample against normal cell line/IC 50 of sample against cancer cell line.
The IC 50 values and SI of the methanol extract of the stem bark, triterpenoidal isolates, and positive control (cisplatin) against the tested cell lines are recorded in Tables 1 and 2.
The cytotoxic efficiency of the stem bark methanol extract (Table 1) might be ascribed to its triterpenoidal components as many of these constituents play an essential role in the upregulation and downregulation of several important genes that influence the apoptotic effects (Prabhu et al., 2011). The isolated betulinic acid was moderately active against MCF-7, HCT-116, and A549 cell lines (respective IC 50 values: 22.39 ± 0.09, 22.29 ± 0.05, and 42.33 ± 0.06 μM); nevertheless, it was insufficiently selective to MCF-7 and HCT-116 cell lines (SI = 1.4) and lacked selectivity to A549 cell line (SI = 0.7). The cytotoxic potential of betulinic acid was previously explained by its ability to induce apoptosis by directly affecting the mitochondria leading to cleavage of caspase-9 and caspase-3 and activation of nuclear factor-kappa-B (NF-kappa-B), which is a key regulator of stress-induced transcriptional activation (Tripathi et al., 2009). Furthermore, betulinic acid was found to inhibit angiogenesis and metastatic activity through inhibition of aminopeptidase N enzyme (Melzig and Bormann, 1998). Studying the structure-activity relationship of betulinic acid revealed that the skeleton composed of rings A, B, and C as well as the carboxylic acid function at C-28 was essential for eliciting its cytotoxic activity (Mukherjee et al., 2006).

CONCLUSION
The variability in triterpenoidal composition between locally acclimatized U. pumila L. samples and those obtained from plants growing abroad could be attributed to environmental conditions. Moreover, the established cytotoxic efficiency of betulinic acid suggests its use as a lead compound for synthesizing potential cytotoxic agents.

AUTHORS' CONTRIBUTIONS
All authors have contributed to gathering literature data, carrying out the chemical and spectral analyses, interpreting the  results, and drafting of this manuscript. All authors have read and approved the final manuscript.