Erythroid induction activity of Aquilegia fragrans and Aquilegia pubiflora and identification of compounds using liquid chromatography-tandem mass spectrometry

2.1 Sample collections

A. fragrans was gathered from village Lawat of Athmaqam tehsil in Neelum district (34°41′18.5 ∼ N 73°58′12.3 ∼ E′) while A. pubiflora was obtained from Makra village, Patekha tehsil in Muzaffarabad district (34°27′14.1 ∼ N 73°32′56.2 ∼ E′) of Azad Kashmir, Pakistan. Plants were identified by the Taxonomist from the Herbarium, University of Karachi, Pakistan and voucher numbers were assigned after deposit of plant material. Voucher no. 95,314 was given for A. fragrans while voucher no. 95314B was given for A. pubiflora.

2.2 Reagents and chemicals

For the mobile phase, methanol (HPLC grade) was acquired from “Merck” KGaA, (Darmstadt, Germany). Formic acid was bought from “DaeJung Chemical and Metals”, (Korea) and type I water was obtained from Barnstead™ GenPure (USA) water purification assembly. Reagents for erythroid induction activity are Fetal Bovine Serum (Gibco), RPMI-1640 (Gibco), Pen/Strip (Gibco), DMSO (Sigma Aldrich, France), Benzidine hydrochloride (Perkin Chemicals, China), Hydrogen Peroxide-30% (Sigma Aldrich, Germany), Glacial acetic acid (Lab Scan), Trypan Blue 4% (Sigma Aldrich, UK), Hydroxyurea (Sigma Aldrich USA).

2.3 Sample preparation

Shade dried whole plants were ground using liquid nitrogen into homogenized powder. Exactly 1 g of each sample was weighed and extracted with 10 mL methanol in a 15 mL falcon tube using a ultra-sonicator for 30 min which were then centrifugated for 15 min at room temperature at 12,000 RPM in order to remove solid particles. 1 mL of upper layer from each sample was filtered through a PTFE syringe-driven filter (0.22 µm) and transferred to 1.5 mL centrifuge tubes for LC-MS analysis. Filtered sample extracts were further diluted by taking 50 µL of sample and diluted with 950 µL (20 times) of methanol in HPLC vial.

2.4 Erythroid induction activity

K562 cells (CCL243, ATCC,), were cultured in RPMI-1460 complete medium containing 10% FBS, and 1% penicillin/streptomycin solution, in a 5% CO₂ at 37 °C under humid conditions. 10 mg/mL stock solutions of methanolic extracts of each plant were prepared in sterile deionized water as a solvent to prepare a stock solution. To determine the erythroid induction, the cells were seeded at density of 4x10⁴ cells/mL in culture and treated in concentrated depend manner range from 250 µg to 15.625 µg of each extract. The cells were then placed in 5% CO₂ incubator at 37 °C for five days.

For erythroid induction benzidine-hydrogen peroxide was used. Briefly, 0.2% benzidine solution was prepared in 0.5 M acetic acid, and 20 µL of hydrogen peroxide was added per 1 mL of ice-chilled 0.2% benzidine solution. On day five, the cells were washed with phosphate buffer saline (PBS) and resuspended in it. Equal volume of benzidine-hydrogen peroxide solution was added to cell suspension in each well and plate was placed in the dark for about 3 min at room temperature (Theodorou et al., 2016). Then the cells were observed under an inverted compound microscope to determine percent benzidine positive cells.

2.5 Analysis of cell viability

The cell viability was determined by using alamarBlue fluorescence assay. This assay utilizes the non-toxic resazurin dye, in healthy cells resazurin undergoes reduction by mitochondrial enzymes and produces a highly fluorescent pink color. This fluorescence intensity of resorufin has been reported as a direct indication of the number of healthy living cells in the medium (Rampersad, 2012).

To determine the viability, cells were seeded (1 × 10⁴ cells/mL) in 96 wells plate and treated with extracts in a dose-dependent manner and incubated for periods of up to 48 h at 37 °C. After incubation, 10 µL alamarBlue dye (Sigma, USA) (25 mg/100 mL PBS) was added to each well and was incubated in dark for 4 h. Cell viability and growth kinetics were determined by measuring the fluorescence intensity at 560 nm (excitation wavelength) and 570 nm (emission wavelength) with SpectraMax M5e Multi-Mode Microplate Reader (Molecular Devices, USA) using SoftMax Pro software (Molecular Devices USA). The cell viability was calculated using the following formula. $$\%\mathrm{C}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{V}\mathrm{i}\mathrm{a}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}=\left(Ft-Fb\right)/\left(Fc-Fb\right)\times 100$$ where Fb, Fc, and Ft describes for fluorescence of blank (only medium), untreated cells, and treated cells, respectively. The data were analyzed and presented as mean ± SEM (Error for Mean) using Graph Pad Prism 5.

2.6 LC-ESI-QTOF-MS/MS analysis

The samples were analyzed using LC-ESI-QTOF-MS/MS mass spectrometer (Bruker MaXis-II, Bremen, Germany), coupled to Thermo Fisher Scientific™ Ultimate™ 3000 series UPLC system, (USA), attached with an autosampler, solvent degasser, column thermostat. Chromatography was optimized using five different columns including columns Zorbax (Eclipse XDB-C18), Zorbax (Eclipse XDB-phenyl), Zorbax (Eclipse XDB-CN), Poroshell (120 EC-C18), Nucleodur Gravity column (RP-C18). For efficient chromatographic separation, the Macherey-Nagel RP-C18 Nucleodur Gravity column (3.0 × 100.0 mm, 1.8 µm particle size) was used, at 45 °C. Mobile phase A was an aqueous solution (0.1% formic acid) while methanol (0.1% formic acid) was used as mobile phase B with 0.70 mL/min flow rate. The solvent gradient system was started with 20% B, raised to 70% from 0.0 to 3.0 mins, and kept constant for 3.00 min and then again raised to 90% from 6.0 to 8.0 mins then lowered back to 20% at 9.0 min with 1 min equilibrium.

Tandem mass spectrometric analysis was performed using electrospray ionization through positive and negative ionization modes separately. The ion source parameter were set as voltage at 4500 V, drying gas at 270 °C temperature and 12.0 L/min flow rate, nebulizer gas with 45.0 psi pressure, mass range at 100 to 2000 m/z while scan speed at 5 Hz for MS, and 12 Hz for MS/MS analysis. Acquisition software, ‘Bruker Compass TargetAnalysis’ was used for quick screening of metabolites.

2.7 Identification of compounds

Compounds identification was based on various features including exact mass matching, error > 5 ppm, MS/MS spectral data was compared with available ESI-MS libraries. An in-house library was prepared from Dictionary of Natural Products on DVD (DNP ver. 26.2) based on previously identified compounds of Aquilegia species. Then these identified compounds were screened against high-resolution mass spectra using Target Analysis version 1.3 (Bruker Daltonics, Bremen Germany). MS/MS-based identification was performed using different available ESI mass spectral libraries including MassBank of North America, MassBank of Europe, mzCloud and NIST-MS/MS library through comparison of exact masses, isotopic pattern and MS/MS fragmentation pattern.

3.1 Erythroid induction assay

Both, A. fragrans and A. pubiflora, investigated in this study showed erythroid induction activity in a concentration-dependent manner. The results of Erythroid induction are given in Table 1 and Fig. 1. Both extracts have shown substantial erythroid induction activity at indicated concentration as compared to untreated control group. The extracts have shown more erythroid induction at lower concentrations which decreases with the increase in extract concentration. However, the A. fragrans extract was found more potent at all the concentrations as compared to A. pubiflora. The maximum activity was observed as 27.15 ± 5.1 and 31.30 ± 1.0 percent benzidine positive cells by A. fragrans and A. pubiflora extracts, respectively at 15.625 (µg/mL) concentration.

Close

Fig. 1

Erythroid induction activity of Aquilegia pubiflora and Aquilegia fragrans compared with standard drug hydroxy urea and control. The results are represented as mean %benzidine positive cells at different concentrations.

Export to PPT

3.2 Evaluation of cell viability

In the cell viability experiment, AF and AP extracts have shown concentration-dependent response. Cell viabilities were represented as percent of control and data is expressed as mean ± SEM for experiments (n = 3). The gradual increase in cell viability was observed with a decrease in the concentration of treatment along. AP extract showed 6.3% cell viability (93.7% cytotoxicity) and 57.6% cell viability (42.6% cytotoxicity) at 250 µg/ml and 125 µg/ml concentrations respectively, but no cytotoxic effects were observed at concentrations ≤ 62.25 µg/ml against K562 cell line. Similarly, 94.5% cell viability (5.5% cytotoxicity) was observed at 250 µg/ml AF extract, but it had no cytotoxic effects at concentrations ≤ 125 µg/ml. Many bioactive plants are toxic on higher concentrations and their dose determines their therapeutic significance (Deshpande, 2002). Both AP and AF extracts showed pro-proliferative effects on K562 cells at lower concentrations (Fig. 2). Our findings suggest that the potent erythroid differentiation of K562 cells by methanolic extracts of A. fragrans and A. pubiflora, signify these plants as potential sources of HbF inducers for the treatment and efficient management of β-thalassemia.

Close

Fig. 2

Cell viability of K562 cells against AP, AF and hydroxy urea at different concentration levels, represented as percent viability of control.

Export to PPT

3.3 Chemical fingerprinting of Aquilegia samples through identification of compounds

Plant samples were analyzed using high-resolution LC-ESI-QTOF tandem mass spectrometry. LC-MS method was optimized using five columns with different chemistry and dimensions. The best-resolved chromatogram (Fig. 3) was observed using RP-C18, column (3.0 × 100.0 mm, 1.8 µm) and thus selected for LC-MS analysis. Further optimizations of run-time, mobile phase gradient and flowrate were performed on the selected column. Based on exact masses and MS/MS matching, total thirty chemical compounds were identified including nineteen compounds through positive ionization mode and thirteen through negative ionization mode. For identification of compounds, every m/z values were checked for mass error (less than 5 ppm) and mSigma values (less than 50 units) which shows the matching of observed isotopic pattern with a calculated isotopic pattern of that molecular formula. The list of identified compounds through both positive and negative ionization modes are given in Table 2. Fig. 3 shows the Base Peak Chromatograms of both species through negative and positive modes of ionization. These chromatograms have peaks labeled with the codes of identified compounds which were found in each of these peaks. Except three compounds, all the compounds were identified in the retention time between 2 min and 4 min.

Close

Fig. 3

Base Peak Chromatograms of: AF in positive mode, AF in negative mode, AP in positive mode and AP in negative mode, respectively. Peaks are labeled with the codes of compounds identified in it.

Export to PPT

Close

Fig. 3

Base Peak Chromatograms of: AF in positive mode, AF in negative mode, AP in positive mode and AP in negative mode, respectively. Peaks are labeled with the codes of compounds identified in it.

Export to PPT

The reported biological activities of identified compounds were also tabulated such as cardioprotective, prevent endothelial dysfunction, reduce overexpression, inhibition of α-glucosidase enzyme, antidiabetic, antidepressant, anti-inflammatory, antimicrobial, antiviral, antifungal and protozoal, antineoplastic, antiplatelet and anticancer activity (Table 2).

3.4 Chemometric analysis of identified compounds

To visualize the distribution of identified compounds between both species, chemometric analysis were performed. Perseus software (version 1.6.2.1) was utilized to generate heatmap clusters. Heatmap from compounds identified using positive ionization mode revealed that compound 6 was found with high intensities in both plants and could be responsible for erythroid induction activity of these plants. Compounds 1 and 5 were found in higher intensity in AF sample while lower intensity in AP. Similarly, compound 4 exhibited lower intensity in AF than AP. Compounds 9, 12 and 13 were found with good intensities in AF while completely absent in AP. Compounds 8, 14, 16 and 19 were absent in AF sample but found in AP with low intensity. Compound 10 showed the highest intensity compared to all the compounds but present in AF only. Heatmap from the positive mode of ionization is given in Fig. 4.

Close

Fig. 4

Heatmap and hierarchical clustering based on peak areas of identified compounds in A. pubiflora and A. fragrans through positive ion mode.

Export to PPT

Through negative ionization mode, compounds 2, 6, 9, 10, 12 and 14 were present in both species. Compound 1 was found with higher intensity in AF, while compound 9 was at a lower intensity in AP. Compound 14 exhibited higher intensity in AP and lower intensity in AF. Compound 6 was present with higher intensity in AP while lower in AF. Compound 12 exhibited a very high intensity in AP as compared to AF sample. Compounds 1, 3, 4, 5, 7, 11, 13 were not found in AP while compounds 1 and 3 were present with higher intensities in AF. Heatmap from the negative mode of ionization is illustrated in Fig. 5.

Close

Fig. 5

Heatmap and hierarchical clustering based on peak areas of identified compounds in A. pubiflora and A. fragrans through negative ion mode.

Export to PPT

Comparative studies of both plant species showed that in positive ion mode out of 19 identified compounds, 16 compounds were confined to one species while 5 compounds are similar in both species. Similarly, in negative ion mode, total thirteen compounds were identified among these 8 compounds were different and 6 compounds were similar in both species (Table 2). The previous phytochemical studies have reported cycloartane glycosides from Aquilegia fragrans while no phytochemical studies were performed on Aquilegia pubiflora. However, these compounds were not seen in the present study. The compounds identified here are mainly belonged to flavonoids class of natural products while some steroids, lignans, terpenoids, benzofuran and coumarins were also found. None of these identified compounds were previously reported to have erythroid induction activity.