Antioxidant potential and LC/MS metabolic profile of anticancer fractions from Echium angustifolium Mill. aerial parts

Plant material

The Echium angustifolium Mill. aerial parts were collected in April 2017 from the Northwestern coast of Egypt (Agiba region, Mersa Matruh Governorate). The plant material was authenticated by Professor Dr. Azza El-Hadidy, Professor of Plant Taxonomy at Cairo University, Egypt. A voucher sample (Ea-2017-A) was kept at the Herbarium of the Faculty of Science, Cairo University, Egypt.

Extraction and fractionation of the E. angustifolium aerial parts

The aerial parts of E. angustifolium were air-dried and then powdered. The powder (2.0 kg) was extracted with 80% ethanol (80% EtOH) by maceration at room temperature (3 × 6 l, each 48 hours). The combined extracts were concentrated under reduced pressure at 45°C till dryness to give 146 g of dried extract. The total extract was subjected to a biological screening (in vitro anticancer assay), followed by bioguided fractionation in order to isolate the active fractions by suspending them in 70% aqueous MeOH (600 ml) in a separating funnel using liquid–liquid partition fractionation with n-hexane, yielding the fatted n-hexane fraction (Ea-F, 33.8 g) and the defatted fraction (EaDf, 105 g). The EaDf was subjected to in vitro cytotoxic assay, and the results indicated that the fraction has a significant activity. Hence, it was further subjected to fractionation on polyamide 6S column chromatography using a gradient elution of H₂O–EtOH (100:0–0:100). All the obtained fractions (500 ml per fraction) were subjected to analytical normal phase TLC plates (silica gel 60 GF254 Normal phase-thin layer chromatography (NP-TLC); purchased from Merck, Darmstadt, Germany). These plates were seen under UV (254 and 365 nm) and immersed in 5% sulfuric acid in methanol followed by heating at 105°C for 3 minutes. The effluents characterized by similar TLC chromatographic patterns were combined and dried to give two promising subfractions (Ea-DfD, 0.94 g, and Ea-DfE, 0.46 g), followed by activity evaluation of these resultant fractions using in vitro cytotoxic and antioxidant assays. Finally, these two fractions were subjected to Liquid chromatography-electrospray ionization-mass spectrometry (LC-ESI-MS)/MS to identify their chemical constituents.

LC-ESI-MS/MS profiling of the two subfractions (Ea-DfD and Ea-DfE)

LC-ESI-MS/MS analysis was assessed for the two subfractions (Ea-DfD and Ea-DfE) from the total extract of the E. angustifolium aerial parts via ACQUITY UPLC™. This apparatus is equipped with a BEH-C18 column of dimensions 2.1 × 50 mm and a 1.7 μm particle size (Waters, Milford, CT). The mobile phase is composed of acidified H₂O with 0.1% formic acid (A) and acetonitrile containing 0.1% formic acid (B). The gradient system with a flow rate 0.2 ml/minute was as follows: 0–2 minutes (A:B 90:10, v/v), 2–5 minutes (70:30), 5–15 minutes (30:70), 15–22 minutes (10:90), 22–25 minutes (10:90), 26–29 minutes (0:100), and 29–32 minutes (90:10). The detector (ESI) in both modes was used within the m/z range of 100–1,000 (Al-Madhagy et al., 2019). Parameters were as follows: capillary voltage 3 kV, cone voltage 30 eV, dissolution gas flow 600 l/hour, cone gas flow 50 l/hour, and source and dissolution temperatures 150°C and 400°C, respectively. MassLynx 4.1 software was used for mass spectra processing, and the peaks of the compounds were annotated by comparing their molecular and fragmentation ions to the literature.

In Vitro cytotoxic assay

The human colorectal [Human Colorectal Carcinoma cell line (HCT116)] and hepatocellular [Human hepatocellular carcinoma G2 (HEPG2)] carcinoma cell lines were acquired from the American Type Culture Collection and kept at −180°C in liquid nitrogen and then subcultured in the National Cancer Institute, Egypt. These cancer cells were suspended in a RPMI 1640 medium (Sigma-Aldrich, USA) in 1% L-glutamine (Lonza, Belgium), 10% fetal bovine serum (Sigma-Aldrich, USA), and 1% from the mixture of antibiotic (10,000 U/ml streptomycin sulfate and 10,000 U/ml K-penicillin) and antimycotic (25 μg/ml amphotericin B), which was purchased from Lonza, Belgium. The total extract of the aerial parts of E. angustifolium (Ea), the EaDf, and the two subfractions (Ea-DfD and Ea-DfE) were evaluated for cytotoxicity potential according to the method previously described by Skehan et al. (1990). Cells (10⁴ cells/well) were put in a microplate (96 wells) for 24 hours. Triplicates of concentrations of each sample (50, 25, 12.5, and 5 μg/ml) beside the negative control were added to the wells and incubated at 37°C in 5% CO₂ for 48 hours. Then, cells were stained with sulforhodamine B stain. Acetic acid was used to remove the excess stain; the fixed stain was dissolved in Tris-Ethylenediamine tetraacetic acid (EDTA) buffer. An Enzyme-linked immunosorbent assay (ELISA) reader was used to detect the color intensity. The viable cells showed the relation between surviving fraction and extract concentration. The IC₅₀ value of the tested extracts was deduced and expressed as IC₅₀ ± SD (triplicate).

Antioxidant activity

The antioxidant activity of the two subfractions (Ea-DfD and Ea-DfE) was assessed using four different in vitro methods: 2,2′-diphenyl-1-picryl-hydrazyl hydrate (DPPH), 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulphonate) (ABTS), ferric-reducing ability power (FRAP), and oxygen radical absorbance capacity (ORAC). All the results were recorded using a microplate reader FLUOstar Omega (BMG LABTECH, Germany).

DPPH free radical assay

An initial screening step was carried out to identify the range of IC₅₀. Solutions of 1,000 and 100 μg/ml in alcohol from both Ea-DfD and Ea-DfE samples were prepared and tested. Then, the concentrations of 60, 50, 40, 30, and 20 μg/ml were prepared. A Trolox standard was prepared as a stock solution of 100 μM in methanol, and then seven concentrations were prepared from it: 50, 40, 30, 20, 15, 10, and 5 μM. In a 96-well plate, 100 μl of a freshly prepared 0.1% DPPH reagent in methanol was mixed with 100 μl of the tested extracts (n = 3). Then, the reaction was incubated for 30 minutes at room temperature in the dark (Boly et al., 2016), and the reduction in DPPH color intensity was detected at 540 nm. Data and IC₅₀ value calculation was analyzed by Microsoft Excel® and GraphPad Prism 5® and represented as means ± SD according to the following equation:

% inhibition = [(Ac–As) / Ac] × 100.

ABTS assay

1 mM of Trolox in methanol was prepared as a stock solution, and then five serial dilutions (600, 500, 400, 300, 200, 100, and 50 μM) were prepared, whereas 0.1 mg/ml from both Ea-DfD and Ea-DfE samples in alcohol was prepared. The assay was carried out in microplates with minor modifications to the method of Arnao et al. (2001); briefly, 192 mg of ABTS was dissolved in a 50 ml volumetric flask with distilled water. After that, 1 ml of this solution was added to 17 μl from 140 mM of potassium persulfate and left for 24 hours in the dark. 1 ml of the reaction mixture was diluted with methanol to the volume of 50 ml to obtain the final ABTS reagent used in the assay. 190 μl of this freshly prepared ABTS reagent was added to 10 μl of each sample (n = 4) and left in the dark for 120 minutes. Finally, the decrease in ABTS color intensity was analyzed at 734 nm, and means ± SD were calculated according to the equation described above. The results are expressed as μM TE/mg sample using the linear regression equation extracted from the linear dose-inhibition calibration curve of Trolox.

FRAP assay

5 mM in methanol of Trolox was prepared as a stock solution, and 10 serial dilutions (4,000, 3,000, 2,000, 1,000, 800, 600, 400, 200, 100, and 50 μM) were prepared. A concentration of 0.1 mg/ml from the two samples was prepared in 70% ethanol. The FRAP assay was evaluated in microplates by Benzi and Strain (1996), with minor changes; briefly, a fresh 2,4,6-tripyridyl-s-triazine (TPTZ) reagent was prepared from 10 mM TPTZ in HCl (40 mM), 300 mM acetate buffer of pH 3.6, and 20 mM FeCl₃ in 1:10:1 v/v/v, respectively. In a 96-well plate, 190 μl from TPTZ was added to 10 μl of the sample (n = 3), incubated at the same conditions, and measured at λ = 593 nm. The ferric-reducing ability of the samples is expressed as μM TE/mg sample by the linear regression equation extracted from the linear dose–response calibration curve of Trolox.

ORAC assay

1 mM of Trolox in methanol was prepared as a stock solution, and nine serial dilutions (400, 300, 200, 150, 100, 75, 50, 25, and 12.5 μM) were prepared. The tested samples (20 μg/ml) were prepared in methanol. The ORAC assay was evaluated based on Liang et al. (2014), with minor modifications; briefly, 12.5 μl of the tested samples was mixed with 75 μl fluoresceine (10 nM) and incubated for 30 minutes at 37°C. Fluorescence measurement (485 EX, 520 EM) was carried out for three cycles (cycle time, 90 seconds) for background measurement. Then, 12.5 μl from 240 mM of 2,2-azobis(2-amidinopropane) was prepared and added immediately to each well, and fluorescence measurement was continued for 2.5 hours (100 cycles, each 90 seconds). Data are represented as means (n = 3) ± SD, and the results are presented as μM TE/mg sample using the linear regression equation extracted from the linear dose-inhibition calibration curve of Trolox.

In vitro anticancer activity

To assess the anticancer effect of the E. angustifolium plant, the total alcoholic extract was screened against liver and colon cancer cell lines (HEPG2 and HCT116, respectively), and it possessed a potential anticancer effect. The extract inhibited the growth of both tested cell lines HEPG2 and HCT116 with IC₅₀ of 22 ± 0.6 and 15 ± 1.1 μg/ml, respectively, as listed in Table 1. These results suggested bioguided fractionation by hexane to yield fatted and defatted fractions that were also screened against the same previous cell lines. The EaDf showed good growth inhibition for both liver and colon cancer cells with IC₅₀ 20 ± 0.6 and 19.7 ± 1.1 μg/ml, respectively. Consequently, fractionation of the defatted part on the polyamide column was carried out to yield several fractions that were also biologically screened against the same previous cell lines. It was found that only two fractions from the polyamide column (Ea-DfD and Ea-DfE) possessed potent anticancer activity. Ea-DfD induced IC₅₀ 10 ± 0.2 and 4 ± 0.8 μg/ml, while Ea-DfE induced IC₅₀ 4 ± 0.5 and 11 ± 0.7 μg/ml against HEPG2 and HCT116, respectively (Table 1). These values of IC₅₀ were comparable to those induced by doxorubicin (4.57 ± 0.5 and 3.73 ± 0.6 μg/ml against HEPG2 and HCT116, respectively). The antitumor activity of two previously isolated lignans from E. angustifolium against breast and liver cell cancer lines (MCF7 and HepG2) was proved, and they displayed promising in vitro antitumor activity lignans (El-Rokh et al., 2018b). This previous report was found to be consistent with the above results, as different lignans were detected in the tested fraction (Ea-DfD) by the LC/MS technique. Additionally, the presence of different phenolic acids’ derivatives with hydroxyl group substitution which were identified in both tested fractions (Ea-DfD and Ea-DfE) by LC/MS was known to induce anticancer effect via their antioxidant role (Godlewska-Zy?kiewicz et al., 2020; Teixeira et al., 2013). Therefore, it can be assumed that the possessed anticancer effect of Ea-DfD and Ea-DfE on liver and colon cell lines is due to their enrichment with lignans and hydroxyl phenolic acids’ derivatives.

Antioxidant activity

Oxidative stress is a well-known cause of cancer, and antioxidants’ role in cancer inhibition and treatment is clear. Additionally, most plants are good sources of antioxidants, and the free radical scavenging activity of E. angustifolium polar fraction was previously evidenced by the DPPH method (Bahmani et al., 2017; Al-Rimawia et al., 2020). Considering that Ea-DfD and Ea-DfE possessed good anticancer activities, it was of value to study also their antioxidant potential as a possible mechanism of action. To assess antioxidant activity, four different in vitro methods (DPPH, ABTS, FRAP, and ORAC) were applied for each of Ea-DfD and Ea-DfE active fractions and repeated thrice to verify the reproducibility of these tests. According to the mechanism of reaction, all of them are categorized as electron transfer methods except ORAC that is a hydrogen-transfer method (Dontha, 2016).

Herein, all the applied methods showed the strong antioxidant activity of both fractions. Regarding the DPPH assay, it was found that fraction Ea-DfE highly inhibited the DPPH free radical with IC₅₀ 25.06 μg/ml that is more reactive than Ea-DfD with IC₅₀ 42.61 μg/ml and very close to Trolox (reference drug) that exhibited IC₅₀ of 24.42 μg/ml as represented in Table 2 and shown in Figure 1. Additionally, the ABTS, FRAP, and ORAC results were also similar to those found in the DPPH and confirmed the antioxidant potential of the tested fractions as both Ea-DfD and Ea-DfE gave high values of Trolox equivalent antioxidant capacity, whereas fraction Ea-DfE possessed 3,501.98, 2,967.85, and 8,673.35 μM TE/mg dry wt. that is a stronger antioxidant than fraction Ea-DfD which showed 3,552.52, 2,089.28, and 5,147.16 μM TE/mg dry wt. for ABTS, ORAC (Fig. 2), and FRAP assays, respectively (Table 2). Considering that Ea-DfD and Ea-DfE fractions are rich in phenolic acid derivatives that are well-known antioxidants, the ability to quench free radicals and consequently to prevent cancer can be explained by the presence of hydroxyl groups attached to aromatic rings (Golezar et al., 2017; Piazzon et al., 2012).

LC-ESI-MS/MS profiling of two active subfractions (EaDf-D and EaDf-E) resulting from the total extract of the E. angustifolium aerial parts

To explore the chemical profile of the two active fractions (EaDf-D and EaDf-E), LC-ESI-MS/MS was used in both ionization modes (negative and positive ion modes). The identification of the secondary metabolites was based on analysis of MS/MS data (i.e., molecular and fragments ions), retention times (R_t), and exploratory MS data of each compound in the mass bank, as well as the comparison with the published data in the literature. All the suggested secondary metabolites from the two analyzed fractions (EaDf-D and EaDf-E) and their fragmentations were summarized in Tables 3 and 4, respectively, and it was noticed that the peak number is the same as the compound number only in Table 3. The basic structures of the tentatively identified compounds from Ea-DfD and Ea-DfE are shown in Figure 3.

Basically, the LC-ESI-MS base peak chromatogram of the first active fraction (EaDf-D) in the negative and positive ion modes (Fig. 4) represented the tentative identification of nine phenolic compounds including six phenolic acid derivatives and three lignans (Table 3). Compound 1 was detected only in the negative ion mode and showed [M−H]⁻ at m/z 197.1 and the fragment ion 153.1 [M−H−CO₂]⁻, with a neutral loss of 44 amu. This agreed with that of syringic acid (Purnamasari et al., 2021). Compound 2 revealed [M−H]⁻ at m/z 179.1, 181.1 corresponding to [M+H]⁺, and a fragment peak at 163 [M+H−H₂O]⁺. Therefore, it was tentatively identified as caffeic acid (Purnamasari et al., 2021; Wei et al., 2010). Compound 9 presented molecular ions: [M+H]⁺ at m/z 209.1, [M−H]⁻ at 207.1, and a fragment pattern at 163. Thus, this compound might be associated with dimethylcaffeic acid (Mohamadi et al., 2015; Purnamasari et al., 2021). Compound 3 indicated [M−H]⁻ and [M+H]⁺ at m/z 436.3 and 438.3, respectively. Also, it showed fragmentations at m/z 292.2 [M+H−146]⁺ and 147 [M+H−291]⁺ consistent with the loss of coumaroyl and coumaroyl-spermidine moieties, respectively. These confirmed the identification of this phytochemical as N,N-dicoumaroyl spermidine (A?alar et al., 2018; Collison et al., 2015). Similarly, the MS data of compounds 4 and 5 indicated that these two compounds could be identified as phenolic acid spermidine derivatives as well, whereas compound 4 gave molecular ions at m/z 468.3 [M+H]⁺ and 466.2 [M−H]⁻. Additionally, fragmentation patterns at 436.3 [M+H-31]⁺, 292.2 [M+H−176]⁺, and 147 [M+H−321]⁺ supported the presence of feruloyl and coumaroyl moieties attached to spermidine in the structure. By comparing our results with the literature (A?alar et al., 2018; Cho et al., 2013; Wu et al., 2016), this compound was tentatively identified as N-coumaroyl-N′-feruloyl spermidine, while compound 5 was suggested to be N,N-diferuloyl spermidine that showed [M+H]⁺ at m/z 498.3, [M−H]⁻ at 496.2, and fragments at m/z 463.2, 322.3 [M+H−176]⁺, 177.1 [M+H−321]⁺, and 353.3 [M−H−143]⁻ and strengthened the attachment of two feruloyl moieties with spermidine constituting the compound structure. Our findings were in agreement with the published data (A?alar et al., 2018; Wu et al., 2016). The three latter compounds (6, 7, and 8) were matched and tended to be as echiumin A, echiumin D, and echiumin B, respectively, by comparing their MS findings with the previous data (El-Rokh et al., 2018b) which reported the isolation and identification of these sucrose diesters of aryldihydronaphthalene-type lignans from our investigated plant (E. angustifolium) growing in Egypt. Our results showed that echiumin A (6) detected a molecular ion peak at m/z 1013.5 [M−H]⁻, echiumin D (7) gave molecular ions at m/z 755.4 [M+H]⁺ and 753.4 [M−H]^-, and finally echiumin B (8) represented molecular ions at m/z 1,007.4 [M+Na]⁺, 985.5 [M+H]⁺, and 983.5 [M−H]⁻. The existence of these lignan compounds in the EaDf-D fraction was reconfirmed by the TLC profile by the presence of some green fluorescence spots under long-wavelength UV (365 nm), which is in agreement with El-Rokh et al., 2018b.

Figure 1. DPPH inhibitory action of Ea-DfD and Ea-DfE subfractions from the defatted extract of E. angustifolium and Trolox. [Click here to view]

Table 1. Evaluation of the IC50 of the E. angustifolium aerial parts’ extracts against the HEPG2 and HCT116 cancer cell lines. [Click here to view]

Table 2. Antioxidant activities of Ea-DfD and Ea-DfE subfractions from the defatted extract of the E. angustifolium aerial parts. [Click here to view]

Table 3. Phytochemical compounds detected and suggested in the Ea-DfD subfraction resulting from the total extract of E. angustifolium aerial parts using LC-ESI-MS/MS in positive and negative ionization modes. [Click here to view]

Figure 2. Antioxidant activities by ORAC assay. (A) Signal curves of different Trolox concentrations and blank indicating the decay of fluoresceine with different concentrations of Trolox. (B) Signal curves of Ea-DfD and Ea-DfE subfractions from the defatted extract of E. angustifolium and blank indicating the decay of fluoresceine upon applying the samples. (C) Blank corrected linear regression curve of Trolox. [Click here to view]

Figure 3. Basic structures of the tentatively identified compounds from Ea-DfD and Ea-DfE fractions: (A) spermidine phenolic acids, (B) syringic acid, (C) other phenolic acid derivatives, (D) echiumin D, and (E) echiumins A and B. [Click here to view]

Figure 4. LC-ESI-MS base peak chromatogram of Ea-DfD fraction of the extract from the E. angustifolium aerial parts (A) in the negative mode and (B) in the positive mode. [Click here to view]

Table 4. Peak assignment of tentatively identified metabolites in the Ea-DfE subfraction obtained from the total extract of the E. angustifolium aerial parts using LC-ESI-MS/MS in positive and negative ionization modes. [Click here to view]

Figure 5. LC-ESI-MS base peak chromatogram of Ea-DfE fraction from the extract of the E. angustifolium aerial parts (A) in the negative mode and (B) in the positive mode. [Click here to view]

On the other hand, LC-ESI-MS of the second active fraction (EaDf-D) tentatively displayed assignment of four phenolic acid derivatives (Table 4) as shown in its base peak chromatogram in both ionization modes (Fig. 5). The first identified peak in this chromatogram showed the same mass spectrum as compound 3. Thus, it was tentatively identified as N,N-dicoumaroyl spermidine (A?alar et al., 2018; Collison et al., 2015). The mass spectrum of peak (3) indicated molecular ions [M+H]⁺ at m/z 355, [M−H]⁻ at 353, and [M+Na]⁺ at 377 and fragmentation ions at m/z 208.9 and 177.9 in the positive ion mode which revealed that this peak could be tentatively identified as chlorogenic acid, 11 (Chiang et al., 2004; Purnamasari et al., 2021). It was noticed that peak (2) represented just a molecular ion at m/z 369.1 [M−H]⁻, while peak (4) showed [M−H]⁻ and [M+H]⁺ at m/z 383 and 385, respectively. Moreover, fragments at m/z 163 and 209 [M+H−176]⁺ corroborate a neutral loss of hexuronyl moiety (glucuronyl moiety), and by comparing these recent MS findings with the mass spectral data of compound 9 and the exploratory MS data of the suggested compound in the mass bank as well as by reviewing the literature (Piazzon et al., 2012), it can be suggested that peak (4) is thought to be tentatively identified as dimethylcaffeoyl glucuronide, 12, and consequently, peak (2) might be identified as methylcaffeoyl glucuronide, 10, with the absence of one methyl group compared to compound 12.