Synthesis, Structure Characterization and Antitumor Activity Study of a New Iron(III) Complex of 5-Nitro-8-hydroxylquinoline (HNOQ)

Abstract

A new iron(III) complex (1) of 5-nitro-8-hydroxylquinoline (HNOQ) was synthesized and structurally characterized in its solid state and solution state by IR, UV-Vis, electrospray ionization (ESI)-MS, elemental analysis, conductivity and X-ray single crystal diffraction analysis. The DNA binding study suggested that complex 1 interacted with calf thymus (ct)-DNA mainly via an intercalative binding mode. By 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, the in vitro cytotoxicity of complex 1, comparing with HNOQ and cisplatin, was screened towards a series of tumor cell lines as well as the normal liver cell line HL-7702. Complex 1 showed higher cytotoxicity towards the tested tumor cell lines but lower cytotoxicity towards HL-7702 than HNOQ, in which the T-24 was the most sensitive cell line for 1. Complex 1 caused G2 phase cell cycle arrest and induced cell apoptosis in T-24 cells in a dose-dependent mode, evidenced by changes in cell morphology. Targeting the mitochondrial pathway due to the redox potential of Fe(III)/Fe(II), the apoptotic mechanism in T-24 cells treated by 1 was investigated by reactive oxygen species (ROS) detection, intracellular [Ca²⁺] measurement and caspase-9 and caspase-3 activity assay. It suggested that complex 1 induced cell apoptosis by triggering the caspase-9 and caspase-3 activation via a mitochondrion-mediated pathway.

Cancer, which remains an ever-increasing threat to the people’s health worldwide, has become the second most common cause of death nowadays.^1,2) Considerable efforts have been made by a growing number of researchers to design new therapeutic anticancer agents.³⁾ Cisplatin, fortuitously discovered by Rosenberg et al. in 1965, is one of the clinical successes of metal-based anticancer drugs and is widely used in the treatment of several human carcinomas.^4,5) Today, the series of platinum-based drugs, such as cisplatin, carboplatin, oxaliplatin, etc., are routinely used in clinic or in the combined chemotherapy, to treat various malignancies (ovarian, lung, testicular, colorectal, head and neck cancer, etc.).⁶⁾ However, the severe side effects and drug resistance in the therapeutic process of the platinum drugs have motivated inorganic chemists to find new platinum complexes with higher efficacy, reduced toxicity and lower incidences of drug resistance.⁷⁾ This provides us with the motivation to focus on developing new non-platinum metal complexes with minimal side effects and maximal curative potentials.

Iron plays an important role in many crucial biological systems, such as DNA synthesis, cell growth and proliferation.^8,9) It is regarded as one of the most studied DNA-damaging metals and this feature can be used to synthesize anti-cancer agents.¹⁰⁾ For instance, the iron complex of the glycopeptide bleomycin, which is clinically used in cancer treatment, can cause double-strand breaks in DNA.¹¹⁾ This successful application of an iron complex has motivated investigations of new iron complexes for cancer treatment. During the last decade, a number of iron complexes, involving hydroxyferrocifen iron complexes,¹²⁾ salophen iron complexes¹³⁾ and dipyridophenazine iron complexes,¹⁴⁾ were synthesized and studied for their antitumor activities.

On the other hand, in the past years, 8-hydroxyquinoline and its derivatives have raised medicinal chemists’ interests and have been investigated for various biological and medical applications, such as potential antitumor/antineoplastic agents.^15,16) For example, Chan et al. reported the design and synthesis of a series of 8-hydroxyquinoline derivatives, in which 8-hydroxy-2-quinolinecarbaldehyde showed good in vitro cytotoxicity against the human carcinoma cell lines.¹⁷⁾ Barilli et al. reported the cytotoxic properties of the transition metal complexes of 8-hydroxyquinoline derivatives on HeLa cells.¹⁸⁾ Tardito et al. reported the cytotoxicity of the copper complex of 5-Cl-7-I-8-hydroxyquinoline and found that it showed potential anticancer activities with reasonably minor adverse side effects.¹⁹⁾ From these studies, we were inspired to explore more metal complexes of 8-hydroxyquinoline ligands with other rational substitutive groups.

5-Nitro-8-hydroxylquinoline (HNOQ) is a typical 8-hydroxyquinoline derivative. Although quite some metal complexes of HNOQ have been reported,^20,21) their biological activities, especially antitumor activities, have not yet been fully explored.^22,23) As to the iron complexes of HNOQ, only their coordination chemistry was studied in the early years.²⁴⁾ In this work, considering the potential antitumor activity of HNOQ and the anti-proliferative effect of iron complexes, a new iron(III) complex of HNOQ was synthesized and structurally characterized. With the aim of exploring new antitumor agents, its DNA binding property was primarily studied and its in vitro cytotoxicity against a series of tumor cell lines was screened. Aiming for the most sensitive tumor cell line T-24, the intracellular apoptotic pathway under the treatment of the iron(III) complex was further investigated and discussed for better understanding its possible antitumor mechanism.

Experimental

Chemicals

All chemicals including iron(III) salt were analytical grade and used without purification. HNOQ was purchased from Alfa-Aesar. RPMI 1640 and fetal bovine serum (FBS) were obtained from Hyclone (U.S.A.). 3-(4,5-Dimathylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), ribonuclease (RNase) A, propidium iodide (PI), Hoechst 33258, acridine orange/ethidium bromide (AO/EB), rhodamine-123 (Rh123), 2-7′-dichlorodihydrofluorescein diacetate (DCFH-DA), Fluo-3 AM were purchased from Sigma Chemicals Co. (U.S.A.). The CasPGLOW™ Fluorescein Activite Caspase-9 and Caspase-3 Staining kit were purchased from BioVision. Calf thymus (ct)-DNA was from Sigma-Aldrich. PBR322 DNA was from Thermo Scientific (U.S.A.). Tris(hydroxymethyl)aminomethane (Tris buffer) was prepared using twice distilled water. All solvent used for spectroscopic studies were of HPLC grade.

Instrumentation

Electrospray ionization (ESI)-MS and IR (in KBr pellets) were recorded on a Bruker HCT Electrospray Ionization Mass Spectrometer and a PerkinElmer, Inc. FT-IR Spectrometer, respectively. Elemental analysis (C, H, N) was measured by a PerkinElmer, Inc. Series II CHNS/O 2400 elemental analyzer. UV-Vis absorption spectra were collected on a PerkinElmer, Inc. Lambda 45 UV-Visible Spectrophotometer. Conductivity was measured on a DDS-307 conductivity meter. Fluorescence measurements were obtained on a Shimadzu RF-5301/PC spectrofluorophotometer. Flow cytometry (FCM) was recorded on FACS Aria II Flow Cytometer (BD Biosciences, U.S.A.). Fluorescence images were obtained by a Nikon TE2000 microscopy system (Japan).

Synthesis and Characterization

Complex 1 was prepared by treating HNOQ (0.1 mmol, 0.019 g) with FeCl₃·6H₂O (0.1 mmol, 0.016 g) in ethanol–pyridine–water (9 : 1 : 1) under solvothermal conditions at 80°C for 3 d. Brown block crystals of 1 were harvested and suitable crystals were selected for X-ray diffraction analysis. Yield (0.029 g, 95%). ESI-MS (in dimethyl sulfoxide (DMSO) containing aqueous solution): m/z 309.0 [Fe^II(NOQ)+2CH₃OH]⁺, 512.1 [Fe^III(NOQ)₂+DMSO]⁺ and 544.9 [Fe(NOQ)₂+DMSO+CH₃OH]⁺. IR (KBr, cm⁻¹): 3411 (m, v(OH)), 1568 (s, v(C=N)), 1499 (s, v(C=C)), 1291 (s, v(C–N)), 1191 (s, v(C–O)), 574 (s, v(Fe–N)), 478 (s, v(Fe–O)). UV-Vis: ε_254 nm=1.14×10⁵ L·mol⁻¹·cm⁻¹. Anal. Calcd for C₂₃H₁₅Cl₂FeN₅O₆: C 47.29; H 2.59; N 11.99%. Found: C 47.28; H 2.63; N 11.82% (for the synthetic route, see Chart 1).

Chart 1. The Synthetic Route of Complex 1

X-Ray Crystallography

The collection of single crystal X-ray diffraction data for complex 1 was obtained by a Bruker Smart Apex II CCD diffractometer equipped with graphite monochromated MoKα radiation (λ=0.71073 Å). Crysalis RED was used for cell refinement, data reduction and absorption correction. The structure was solved by direct methods with SHELXL-97 programs.²⁵⁾ The full-matrix least squares methods were used for the refinement of non-hydrogen atoms. All hydrogen atoms were added theoretically, riding on the concerned atoms. DIAMOND was used for the drawing presentation of the complex.

Conductivity Experiment

An aqueous solution of compex 1 at 1.0 µM was prepared by adding a certain amount of single crystals of compex 1 into ultra pure water, according to the molecular weight of [Py+H]·[FeCl₂(NOQ)₂]·1.5H₂O. The conductivity of the solution of complex 1 was tested under the residence time ranging from 0 to 12 h at 25°C. All experiments were repeated for three times and taken the average value.

DNA Binding Experiments

All the spectral analysis for DNA binding studies were tested in 0.1 M of Tris buffer saline (TBS; pH=7.34). The stock solution of ct-DNA (2.0 mM) was stored at 4°C no more than 5 d before use. The 2.0 mM stock solution of complex 1, according to the molecular weight of [Py+H]·[FeCl₂(NOQ)₂]·1.5H₂O, was prepared by dissolving it in DMSO and diluting suitably with TBS to required concentrations for all the experiments. For UV-Vis spectral analysis, a 3.0 mL constant concentration of complex 1 (2.0×10⁻⁵ M) was prepared and then the stock solution of DNA was gradually added. The absorption spectra were recorded after the mixture solution being reacted enough. The electrostatic interaction experiment for 1 was also carried out using UV-Vis absorption spectrum. A fixed solution (3.0 mL) containing 2.0×10⁻⁵ M of complex 1 was titrated by successive additions of sodium dodecyl sulfate (SDS) solution and the corresponding absorption intensity was recorded. The DNA competitive binding study between EB and complex 1 was performed using fluorescence emission spectra by maintaining a constant concentration of ct-DNA (2.0×10⁻⁵ M) and EB (2.0×10⁻⁶ M) and varying the complex 1 concentration. In plasmid DNA unwinding assay, the stock solution (2.0 mM) of complex 1 was diluted to gradient concentrations (5, 10, 20 µM) by TBE buffer (TBE: Tris-Boric acid-ethylenediaminetetraacetic acid (EDTA) buffer solution). Various concentrations of complex 1 was mixed with 0.5 µg pBR322 DNA and made up to a total 25 mL by TBE buffer. All samples were incubated at 37.5°C in dark for 2 h. Then 10 mL of each sample mixed with 1 mL DNA loading buffer was electrophoresed in the 0.8% agarose gel immersed in the TBE buffer solution for 4 h under the 70 V electric field. Finally, the gel was stained with 1.27 µM EB for 30 min. The DNA bands were visualized by placing the gel on UV illuminator and photographed.

Cell Culture and MTT Assay

The T-24, SK-OV-3, BEL-7404, HeLa, MGC-803 tumor cell lines and HL-7702 human normal liver cell were purchased from the Shanghai Cell Bank of the Chinese Academy of Sciences. Those cell lines were cultured at 37°C in a humidified atmosphere of 5% CO₂–95% air in RPMI-1640 medium supplemented with 100 units/mL penicillin, 100 µg/mL streptomycin, and 10% FBS. The stock solution of complex 1 (2.0 mM) was diluted by phosphate buffered saline (PBS) to the required concentration immediately before use. Cisplatin was selected as a reference metallodrug for evaluating the potency of complex 1.²⁶⁾ The in vitro cytotoxic activity evaluation was performed by MTT assay, about 1×10⁴ cells well⁻¹ were seeded in 180 µL of supplemented culture medium in 96-well micro plates, and incubated at 37°C in a humidified atmosphere of 5% CO₂ for 24 h to make sure cell adhesion. Then appropriate concentrations (2.5, 5, 10, 20, 40 µM, respectively) of HNOQ, complex 1 and cisplatin were added. After that, cells were further incubated for 48 h. Upon completion of the incubation, 10 µL of MTT (5 mg/mL) in PBS (pH 7.40) was added to each well, the plates were continued to incubate for another 4 h in cell culture incubator. Then the medium was replaced with 100 µL of DMSO to dissolve formed formazan crystals. Finally, the absorbance was read at enzyme labelling instrument with 570/630 nm double wavelength measurement by an enzyme-linked immunosorbent assay (ELISA) reader. The cytotoxic activity evaluation was obtained by the IC₅₀ values, which were calculated by the Bliss method (n=5).

Cell Cycle Experiment

T-24 cells were treated with complex 1 at 5.0, 10.0, 20.0 µM for 24 h. Then cells were collected and fixed with ice-cold 70% ethanol at −20°C overnight. Before testing, fixed T-24 cells were resuspended in 500 µL of PBS (containing 50 µg/mL PI and 100 µg/mL RNase A) and stained for 45 min at 37°C in the dark. The cell cycle distribution was recorded on a FACS AriaII flow cytometer (BD) and calculated using MFLT32 LT software.

Cell Apoptosis Experiment

T-24 cells were plated in 6-well plates at the concentration of 1×10⁵ cells/mL, incubated with complete medium only (control) and medium containing 5.0, 10.0, 20.0 µM concentrations of complex 1 for 8 h, respectively. Then cells were trypsinized and harvested, added 100 µL 1× binding buffer, stained with 5 µL of annexin V-fluorescein isothiocyanate (FITC) and 5 µL of PI at 4°C for 20 min, and finally resuspended with 400 µL 1× binding buffer. Cell apoptosis was measured by flow cytometry.

Hoechst 33258 and AO/EB Experiment

The tumor cell’s morphological changes were detected by Hoechst 33258 and AO/EB double staining for cellular apoptosis. For Hoechst 33258 assay, T-24 cells adhered on a coverslip were treated with 5.0, 10.0, 20.0 µM of complex 1 for 8 h, respectively. Upon completion of the incubation, the cells were added 0.5 mL of stationary liquid to fix cells on coverslip for 10 min, stained at 37°C by 0.5 mL Hoechst 33258 fluorescent dye in dark for 5 min after rinsing twice with PBS. Then washed twice with PBS again, a blob of anti-fluorescence quenching liquid was placed on a slide, which was next covered by a coverslip. The images of T-24 cells were captured by a CarlZeiss LSM710 confocal microscope. For AO/EB double staining assay, the T-24 cells were treated with 5.0, 10.0, 20.0 µM of complex 1 for 8 h. Then the treated cells were trypsinized and collected, suspended in PBS, stained suspension cells with AO-EB working solution (AO: 100 µg/mL; EB: 100 µg/mL) for 5 min at 37°C to detect apoptosis. The treated cells were then observed immediately in a fluorescence microscope (Nikon TE2000, Japan).

Measurement of Mitochondrial Membrane Potential

Mitochondrial membrane potential (Δψ) was determined flow cytometrically using Rh 123 staining. For flow cytometric analysis, T-24 cells treated with 10.0 µM of complex 1 for 8 h were harvested, washed twice with PBS, stained with 10 µg/mL Rh 123 for 30 min at 37°C, washed twice with culture medium before monitoring using flow cytometry.

Measurement of Reactive Oxygen Species (ROS) Production

The intracellular ROS production level was measured using the stain of DCFH-DA by flow cytometry analysis. T-24 cells were cultured in 6-well plates at a cell density of 2×10⁵ cells/well. The cells were then being exposed to 10.0 µM of complex 1 for 8 h at 37°C. Then the treated cells were loaded for 30 min with 100 µM DCFH-DA at 37°C. After that, the loaded cells were washed twice with serum-free cell culture medium, and then maintained in 500 mL serum-free culture medium. The level of ROS generation was detected immediately by flow cytometry.

Measurement of Cytoplasmic Calcium Concentration ([Ca²⁺]_c)

The intracellular Ca²⁺ concentration was tested by flow cytometry using Fluo-3 AM staining. Firstly, T-24 cells were exposed to 10.0 µM of complex 1. Next, the cells were then being stained by 5 µM Fluo-3 AM for 30 min at 37°C and was then washed with PBS. Finally, the cells were measured by flow cytometry analysis immediately.

Determination of Caspase-9 and Caspase-3 Activity by Flow Cytometric Analysis

The assessments of caspase-9 and caspase-3 activity were performed on a flow cytometry using FITC-LEHD-FMK (for caspase-9) or FITC-DEVD-FMK (for caspase-3) staining. The treated T-24 tumour cells were exposed to 10.0 µM of complex 1 for 8 h and the controlled cells were harvested, centrifuged and washed twice with PBS, then added 300 µL buffer and 1 µL of FITC-LEHD-FMK (for caspase-9) or FITC-DEVD-FMK (for caspase-3) consequently and incubated for 1.0 h in the dark at 37°C. Then, the T-24 cells were monitored by a fluorescent activated cell sorter (FACS) AriaII flow cytometer. The analysis results were described as the percent change on the activity comparing with the untreated control.

Results and Discussion

Structural Characterization

The UV-Vis absorption spectrum of complex 1 was shown in Fig. S1, in which three characteristic absorption bands were found. The most intensive bands at ca. 254 and 409 nm were attributed to the π–π* electron transition of the aromatic structure and n–π* electron transition of the nitro structure of HNOQ ligand, respectively. The less intense band at ca. 312 nm is typical of charge transfer from metal to ligand (MLCT).²⁷⁾ The FT-IR spectra for HNOQ and complex 1 were shown in Figs. S2 and S3. The IR spectrum of the HNOQ ligand showed a broad band at 3203 cm⁻¹ attributed to the O–H stretch (v(O–H)) of 8-hydroxyl. The C–O stretch (v(C–O)) is observed at 1200 cm⁻¹. The bands at 1571 and 1508 cm⁻¹ were attributed to the ring stretching frequencies of v(C=N) and v(C=C) in the quinoline ligand, respectively. The bands at 1320 and 1288 cm⁻¹ were assigned to the C–N stretch v(C–N). Comparatively, obvious absorption changes could be found in the IR spectrum of complex 1. The strong absorption band at ca. 3203 cm⁻¹ (v(O–H)) disappeared, indicating the de-protonation of the 8-hydroxyl of HNOQ after coordinating to Fe(III), while the emerging band at 3411 cm⁻¹ was attributed to the O–H stretch (v(O–H)) of water in the crystal lattice. The ring stretching frequencies of v(C=N) and v(C=C) are shifted to 1568 cm⁻¹ (Δv=−4 cm⁻¹) and 1499 cm⁻¹ (Δv=−9 cm⁻¹), respectively. The C–O stretch (v(C–O)) is shifted to 1191 cm⁻¹ (Δv=−9 cm⁻¹), and two C–N stretches v(C–N) are shifted to 1291 cm⁻¹ (Δv=−29 cm⁻¹) and 1236 cm⁻¹ (Δv=−52 cm⁻¹), respectively. The magnitude of the shift Δv obviously indicates that the O and N coordinated with metal ion. This is further supported by bands in the far infrared region, 574 cm⁻¹, corresponding to Fe–N and 478 cm⁻¹ to Fe–O bond stretches, respectively.²⁸⁾

The structure of complex 1 in its solid state was further determined by single crystal X-ray diffraction analysis, as shown in Fig. 1. It indicated that complex 1 crystallized in a triclinic crystal system with space group P-1. The Fe(1) atom was six coordinated by two NOQ ligands via the heterocyclic N atom and the de-protonated hydroxyl O atom, respectively, as well as two Cl atoms, to form a distorted octahedral coordination geometry. Since the two NOQ and two Cl⁻ ligands coordinating to Fe(III) formed in a negative charged species, [Fe^IIICl₂(NOQ)₂]⁻, a protonated pyridine, [Py+H]⁺, existed as a counter ion for each [Fe^IIICl₂(NOQ)₂]⁻ to maintain electrical neutrality of complex 1. In each crystal unit, there existed two [Fe^IIICl₂(NOQ)₂]⁻ and two [Py+H]⁺, as well as three H₂O as solvent molecules, which were shown in Fig. S4. The details of crystallographic data and refinement parameters were depicted in Table 1, and the selected coordinating bond lengths and bond angles of complex 1 were depicted in Table S1. As shown by the detailed structural data listed in Table S1, the Fe–O bond lengths ranged from 1.982(4) to 1.998(4) Å, the Fe–N bond lengths ranged from 2.205(5) to 2.237(5) Å and the Fe–Cl bond lengths ranged from 2.272(2) to 2.325(2) Å. The chelating bond angles of N(1)–Fe(1)–O(1) and N(3)–Fe(1)–O(4) were 77.51(17)° and 76.12(17)°, respectively. While the distortion on the octahedral coordination geometry of complex 1 could be indicated by the following angles, such as N(1)–Fe(1)–N(3) (84.34(17)°), N(3)–Fe(1)–O(1) (86.51(17)°) and N(3)–Fe(1)–Cl(2) (88.40(14)°).

Fig. 1. The ORTEP Presentation of the Crystal Structure of Complex 1 with Thermal Ellipsoids at the 30% Probability

Table 1. Crystal Data and Structure Refinement Parameters of the Complex 1

Empirical formula	C23H15Cl2FeNO6
Formula weight	584.15
Temperature/K	293(2)
Crystal system	Triclinic
Space group	P–1
a/Å, b/Å, c/Å	8.4700(17), 16.574(3), 18.072(4)
α/°, β/°, γ/°	94.99(3), 92.88(3), 93.99(3)
Volume/Å3	2517.2(9)
Z	2
ρcalc/mg mm−3	1.607
μ/mm−1	0.868
F(000)	1236
2θ range for data collection	4.54 to 52.78°
Index ranges	−10≤h≤10, −20≤k≤20, −22≤l≤22
Reflections collected	23200
Independent reflections	10249 [R(int)=0.1162]
Data/restraints/parameters	10249/0/679
Goodness-of-fit on F2	0.917
Final R indexes [I>2σ(I)]	R1=0.0693, wR2=0.1172
Final R indexes [all data]	R1=0.2231, wR2=0.1638
Largest diff. peak/hole/eÅ−3	0.328/−0.350

Moreover, the existing species of complex 1 in aqueous solution was investigated by ESI-MS analysis, to analyze the exact species of the metal complex existing in solution state, in order to better understand its antitumor activity as well as the potential structure–activity relationship (SAR). As demonstrated in Fig. S5, it was found that the major abundances of ESI-MS for 1 peaked at 309.0, 512.1 and 544.9 in the positive ion mode, corresponding to 309.0 [Fe^II(NOQ)+2CH₃OH]⁺, 512.1 [Fe^III(NOQ)₂+DMSO]⁺ and 544.9 [Fe^III(NOQ)₂+DMSO+CH₃OH]⁺, respectively. It strongly suggested that complex 1 tended to maintain a mono-nuclear coordinating structure in aqueous solution, with the NOQ/Fe(III) ratio from 1 : 1 to 2 : 1. Notably, it implied that there also existed the variable valence in the Fe centre from +3 to +2. This is closely related to the redox activity of Fe(III)/Fe(II), which may contribute to the potential ROS generation in the cellular level.^29–31) In addition, according to the ESI-MS analysis for 1, both chlorides might dissociate as leaving groups in aqueous solution, which led to the changes on the charged status of 1. The conductivity experiment was thus carried out for further assessment. As shown in Fig. 2, after the stock solution of complex 1 was diluted by the ultra pure water and was monitored immediately, the conductivity of complex 1 was observed to increase rapidly. Finally, the conductivity value of the diluted solution of 1 boosted about 3.8 times in 11.0 h, which further supported the analysis results of ESI-MS.

Fig. 2. The Variations on the Conductivity of Complex 1 (1 µM) in Aqueous Solution at 25°C, Shown as the Average Value in a Triplicate Experiment, Which Indicated That It Would Be Dissociated in Aqueous Solution to Give Ionic Species, as Suggested by the Results of ESI-MS

In Vitro Antitumor Activity Evaluation by MTT Assay

The cytotoxicity of complex 1 against five typical tumor cell lines (SK-OV-3, BEL-7404, HeLa, T-24, MGC80-3) were evaluated by MTT assay, comparing with HNOQ and cisplatin, and the human normal liver cell line HL-7702 was also tested for the selectivity study. The IC₅₀ values of the tested compounds were listed in Table S2, and the histogram for IC₅₀ values was shown in Fig. 3. It was found that the in vitro cytotoxicity of complex 1 was higher than that of HNOQ towards most of the tumor cell lines, except for SK-OV-3, in which the T-24 was the most sensitive tumor cell line for 1, with the IC₅₀ value of 9.24±0.07 µM. Against T-24, complex 1 was even more cytotoxic than cisplatin (IC₅₀=14.05±0.03 µM), even though cisplatin was still much more cytotoxic than 1 towards the other four tumor cell lines. Although the complex 1 was suggested by ESI-MS to partly dissociate into both the species of [Fe^II(NOQ)]⁺ and [Fe^III(NOQ)₂]⁺, which led the uncertainty of the molar ratio for [complex 1]/[HNOQ], regarding the HNOQ, fell in the interval of 1 : 1 to 2 : 1 in MTT assay, the higher in vitro cytotoxicity of complex 1 than HNOQ against most of the tumor cell lines should not be denied.

Fig. 3. IC₅₀ Values (µM) of HNOQ, Complex 1 and Cisplatin towards Five Selected Tumor Cell Lines as Well as the Normal Liver Cell Line HL-7702 after Incubations for 48 h

On the other hand, it should be emphasized that complex 1 also showed much lower cytotoxicity towards the human normal liver cell line HL-7702, with the IC₅₀ value of only 95.41±0.06 µM. It indicated that the cytotoxicity of 1 towards HL-7702 was the lowest among all these six tested cell lines, suggesting the satisfying cytotoxic selectivity of 1 on the tumor cells. Furthermore, towards HL-7702, the cytotoxicity of 1 was also much lower than both of the HNOQ (IC₅₀=25.61±0.05 µM) and cisplatin (IC₅₀=5.04±0.06 µM), implying its potential for the better therapeutic effect on the liver cancer.

DNA Binding Studies

DNA is generally regarded as the primary antitumor target. DNA replication, in tumor cells, can be blocked by the intercalations between the small molecule and the base pairs of DNA.³²⁾ Thus, the characterization of the interaction between metal complex and DNA in vitro is of great importance due to the contribution for insights into its binding mechanism and the toxicological effect of metal based chemotherapeutic drug.³³⁾ To better understand the antitumor activity of complex 1, the DNA binding property of 1 was studied by UV-Vis absorption, fluorescence spectrum and agarose gel electrophoresis.

UV-Vis Absorption Spectral Analysis

UV-Vis absorption spectroscopy is an effective method to examine the binding mode of small molecules with DNA. Thus, with the increasing addition of SDS, electrostatic interaction of complex 1 with ct-DNA was primarily studied by UV-Vis absorption spectrum. SDS is a kind of probe for understanding whether the electrostatic interaction existed due to the aggregated SDS anions acting as appropriate substitute for DNA polyanionic backbone, which led to spectral changes.³⁴⁾ As indicated in Fig. S6, the addition of increasing amounts of SDS didn’t cause obvious hypochromicity on the maximum absorbance of complex 1 at ca. 254 nm, which suggested that the electrostatic interaction did not exist between complex 1 and DNA, most probably due to their same negative charged structure in the solutions.

To ascertain the possible binding mode between complex 1 and DNA, the interaction of complex 1 with ct-DNA was also investigated by UV-Vis absorption spectrum. Generally, hypochromic effect and red shift were observed in the absorption spectra of small molecule if it intercalated between DNA base pairs.³⁵⁾ The absorption spectra of 1 in the absence and presence of ct-DNA were depicted in Fig. 4. As shown in Fig. 4, in the absence of ct-DNA, the maximum absorbance (A_max=2.280) of 1 appeared at 254 nm, which could be attributed to the π–π* electron transition of the aromatic structure of HNOQ. After the gradient increase of ct-DNA till the [DNA]/[1] ratio reached 7 : 1, the maximum absorbance (A_max) of 1 gradually red-shifted to 292 nm and gradually decreased to 0.711, with a total hypochromic ratio of 69% for 1. It strongly revealed that complex 1 interacted with ct-DNA in an intercalative binding mode.

Fig. 4. The UV-Vis Absorption Spectra of Complex 1 at 2.0×10⁻⁵ M in the Absence (Dashed Line) and Presence (Solid Lines) of Increasing Amounts of ct-DNA from 1 : 1 to 7 : 1

Fluorescence Spectral Analysis

The DNA binding property of complex 1 was also discussed by fluorescent spectral analysis, based on the competitive binding between EB and 1 with ct-DNA. In the presence of DNA, the fluorescent emission intensity of EB, one of the most sensitive fluorescent probes for DNA, can be greatly enhanced because of its strong intercalation between the adjacent DNA base pairs.³⁶⁾ If the EB was replaced by other small molecules, its fluorescent emission will be significantly quenched. It can be used to distinguish the intercalative and non-intercalative compounds.³⁷⁾ As shown in Fig. 5, in the absence of complex 1, the EB molecules bound by ct-DNA gave strong fluorescence emission with maxium emission intensity at 582.8 nm. Under the gradient addition of complex 1, the fluorescent intensity was gradually quenched. It indicated that there existed competitive binding between complex 1 and EB, which suggested an intercalation binding mode of 1 with DNA. The fluorescence quenching constant (K_q) was calculated according to Stern–Volmer equation³⁸⁾:   

(1)

where I₀ is the fluorescence intensity of EB bound with DNA in the absence of complex 1, I is the fluorescence intensity EB in the presence of complex 1, and [Q] is the concentration of complex 1. As indicated by the inset plot of Fig. 5, K_q was calculated to be 1.16×10⁴ M⁻¹. It suggested that the intercalation binding mode of complex 1 with ct-DNA was considerably strong, which is consistent with the above results of the UV-Vis spectral analysis.

Fig. 5. Fluorescence Emission Spectra of EB Bound with ct-DNA ([DNA]=2.0×10⁻⁵ M, [EB]=2.0×10⁻⁶ M) in the Absence and Presence of Complex 1 with [Complex 1]/[EB] Ratios Range from 1 : 1 to 9 : 1

Agarose Gel Electrophoresis Assay

The DNA binding mode of complex 1 was further ascertained by agarose gel electrophoresis assay. pBR322 DNA usually exhibits three forms of DNA with different electrophoretic mobility, containing supercoiled DNA (Form I), nicked DNA (Form II) and linear DNA (Form III). When small molecules interact with DNA in an intercalation mode, the migration rate of supercoiled DNA decreases.³⁹⁾ As shown in Fig. 6(a), upon the gradual addition of complex 1 to DNA, the migration rate of supercoiled DNA (Form I) gradually reduced. The relative optical densitiy of Form I also gradually decreased from 100 to 96.36, 96.08 and 8.32% after pBR322 DNA being treated with increasing concentrations of 5.0, 10.0 and 20.0 µM complex 1, respectively (as shown in Fig. 6(b)). It revealed that complex 1 can bind with DNA by intercalative binding mode.

Fig. 6. (a) Gel Electrophoresis Mobility Shift Assay of pBR322 DNA Treated with Increasing Concentrations of 5.0, 10.0 and 20.0 µM of Complex 1 at 25°C after 2 h of Incubation in TBE Buffer; (b) Quantitative Evaluation on the Relative Optical Densitiy of the Supercoiled Form DNA in the Gel

Cell Cycle Arrest

Since complex 1 performed the most sensitive anti-proliferative effect on T-24 human tumor cell line, the cell cycle progression of the T-24 cells treated by 1 was primarily examined by using flow cytometric analysis, to explore the intracellular mechanism for its growth inhibition. As shown in Fig. 7, after the end of the treatment of 0, 5.0, 10.0, and 20.0 µM complex 1 for 24 h, T-24 cells were observed a significant accumulation in the proportion of cells at the G2 phase in a concentration-dependent manner. The G2 phase population increased from 14.81% (for control) to 35.47%, while the population of G1 phase decreased from 60.73% (for control) to 40.44%. It suggested that complex 1 induced cell cycle arrest in T-24 cells at G2 phase.⁴⁰⁾ Since the G2/M cell cycle checkpoint is one of the crucial important responses to the DNA damage, it further suggested that DNA was an important intracellular target for complex 1.

Fig. 7. Percentage of Cell Populations in Various Phases (G1, S, G2 Phase) of Cell Cycle Distribution in T-24 Cells after Treatment for 24 h with 5.0, 10.0, 20.0 µM of Complex 1

Cell Apoptosis Assay

Considering the essential role of cell cycle arrest in tumor cells apoptosis,⁴¹⁾ and the apoptosis induction being regarded as an anticancer approach for many bioactive agents,⁴²⁾ the cell apoptosis assay was studied using annexin V-FITC/PI staining by flow cytometry for detecting whether the anti-proliferative effect on T-24 cells was due to apoptosis induction.⁴³⁾ As shown in Fig. 8, the populations of apoptotic cells, including late apoptotic cells (see Q2 zone) and early apoptotic cells (see Q3 zone) by complex 1, were increased as compared to the control. The percentages of apoptotic T-24 cells were enhanced from 5.8 to 7.8, 10.8, 14.5%, respectively, after the end of the treatment of 5.0, 10.0, and 20.0 µM complex 1 for 8 h. The significant increments on the apoptotic tumor cells indicated that complex 1 effectively inhibited the proliferation and induced apoptosis of T-24 cells in a dose-dependent mode.⁴⁴⁾

Fig. 8. The Cell Apoptosis Induction in the T-24 Cells Treated by 5.0, 10.0, and 20.0 µM of Complex 1 for 8 h Was Examined by Flow Cytometry Using PI and Annexin V-FITC Double Staining Assay

Cell Morphological Observation by Hoechst 33258 and AO/EB Double Staining

To further confirm the existing of apoptosis in T-24 cells induced by complex 1, morphological changes were detected using Hoechst 33258 staining and AO/EB double staining. During the process of cell apoptosis, morphological changes such as cell shrinkage, nuclear fragmentation, chromatin condensation and the formation of apoptotic bodies may be observed.⁴⁵⁾ As shown in Fig. 9A, after treatment of complex 1 at gradient concentrations (5.0, 10.0, 20.0 µM) for 8h, more and more tumor cells with apoptotic features, such as cell shrinkage and nucleus fragmentation, could be observed, comparing with the living cells for control. While as shown in Fig. 9B, in the untreated control group, green fluorescence was emitted in the living cells. Comparatively, the tumor cells treated by complex 1 showed typical apoptotic characteristic. The cells emitting bright green fluorescence in the nuclear chromatin suggested the apoptotic status of early-stage, while the cells emitting orange-red fluorescence suggested that of late-stage.^46,47) The apoptotic features were more and more obvious with the increase of complex 1 concentration. Thus, viewed from the characteristic cellular morphology in the T-24 cells by Hoechst 33258 and AO/EB staining, respectively, complex 1 could effectively induce cell apoptosis in the T-24 tumor cells in a concentration-dependent manner.

Fig. 9. Morphological Observation on the Apoptosis of the T-24 Cells Incubated with Complex 1 of Gradient Concentrations (0, 5.0, 10.0, 20.0 µM) for 8 h by Using the Hoechst 33258 Staining (A) and AO/EB Staining (B), Respectively

The Mitochondrial Membrane Potential Assay for Cell Apoptosis

Apoptotic stimuli leads to the convergence of many signals at mitochondria and many of these stimuli trigger a change of the mitochondrial membrane permeability. The loss of mitochondrial membrane potential (Δψ) was considered as a characteristic of cell apoptosis in the early stage. The decrease of Δψ causes depolarization of mitochondrial membrane which leads to less uptake of Rh123, a fluorescent dye by mitochondria. Hence, the alterations of Δψ T-24 cells were monitored by flow cytometry using Rh123 staining. As shown in Fig. 10, left marker was regarded as mitochondria with low membrane potential, i.e., depolarized mitochondria and right marker was regarded as mitochondria with high membrane potential, i.e., polarized mitochondria.⁴⁸⁾ Comparing to the untreated control cells, the mitochondrial polarization was decreased from 52.1 to 46.1% after the end of the treatment of 10.0 µM complex 1 for 8 h. The loss of mitochondrial membrane potential in T-24 cells intimated that the T-24 cell apoptosis induced by complex 1 might be an intrinsic mitochondria-mediated pathway.

Fig. 10. The T-24 Cells Treated with 10.0 µM of Complex 1 for 8 h and Stained by Rh123 Indicated the Loss of the Mitochondrial Membrane Potential (Δψ) by Flow Cytometry

Measurement on the ROS Generation

To study the upstream regulatory mechanisms leading to complex 1-induced mitochondrial dysfunction, the effect of complex 1 on intracellular ROS levels was examined. T-24 cells were treated with 10.0 µM of complex 1 for 8 h and measured by flow cytometry analysis using the DCFH-DA staining, since DCFH-DA can change to the fluorescent DCF when bound with ROS. As shown in Fig. 11, compared with the control, the level of ROS generation in the tumor cells significantly enhanced from 51.8 to 65.2% under the treatment of 9.0 µM of complex 1 for 8 h. So the cell apoptosis induced by 1 was suggested to be closely related to the ROS-mediated mitochondrial dysfunction pathway, especially in the early-stage apoptosis, in which the redox activity of Fe(III)/Fe(II) as the centre of complex 1 should play the key role.^29–31)

Fig. 11. The Increment on the ROS Generation in the T-24 Cells When Treated with 10.0 µM of Complex 1 for 8 h

Measurement of the Intracellular Concentration of Ca²⁺

Considering the close relationship between ROS and intracellular calcium levels ([Ca²⁺]_c),⁴⁹⁾ as well as the enhanced intracellular [Ca²⁺] being regarded as one of the major contributing factors towards the cell apoptosis by regulating the enzyme activity and disrupting the mitochondrial function,⁵⁰⁾ the measurement of [Ca²⁺]_c of T-24 cells was detected by flow cytometry analysis with fluorescent probe Fluo-3/AM. As depicted in Fig. 12, obvious elevation of [Ca²⁺]_c was found after T-24 cells being exposed by 10.0 µM complex 1 for 8 h as comparing with the control, it was raised from 52.3 to 59.5%. These results proposed the mitochondrial apoptotic pathway induced by 1 via the mitochondria dysfunction triggered by ROS generation, which could be proven by the enhancement of the intracellular [Ca²⁺]_c levels.

Fig. 12. The Detection on the Intracellular Level of [Ca²⁺]_c in the T-24 Cells Treated with 10.0 µM of Complex 1 for 8 h by Flow Cytometry Using Fluo-3/AM as the Fluorescent Probe

Assessment on the Caspase-9 and Caspase-3 Activation for Cell Apoptosis

Caspase-9 and caspase-3 played an essential role as an executor of cell apoptosis. To further confirm the intrinsic mitochondrion-mediated apoptotic pathway induced by complex 1, the activation of caspase-9 and -3 was assessed using FITC-LEHD-FMK (for caspase-9) and FITC-DEVD-FMK (for caspase-3) as probe molecules, respectively. As shown in Fig. 13, after exposing to 10.0 µM of complex 1 for 8 h, the level of activated caspase-9 cells were found promotion from 1.9 to 7.8% in T-24 cells treated by complex 1, it was increased about four times as comparing with the control. Subsequently, the activated caspase-3 cells were also found significantly elevation after treatment with 10.0 µM complex 1. It was elevated from 3.0 to 10.8%. It strongly suggested that complex 1 could effectively trigger the activation of caspase-9 as the initiator of the caspase cascade, and then activate the caspase-3 as the downstream protein of the caspase cascade.

Fig. 13. The Activation of Caspase-9 and Caspase-3 in T-24 Cells When Treated with 10.0 µM Complex 1 for 8 h

Conclusion

In this study, a new iron(III) complex of HNOQ was synthesized and structurally characterized. This complex formed as a mono-nuclear Fe(III) complex both in the crystallized state and in the aqueous solution state, implying the stability of this complex to maintain the coordinated state of NOQ to Fe(III). This complex most probably bound with DNA via an intercalative binding mode, and it was more cytotoxic to a series of typical human tumor cell lines than HNOQ, in which the T-24 cell line was the most sensitive one. More importantly, complex 1 showed much lower cytotoxicity towards HL-7702 cells compared with those of the free HNOQ and cisplatin, suggesting the good cytotoxic selectivity of complex 1 towards the tested tumor cells. Its antitumor mechanism was investigated on the cellular level targeting the mitochondrial pathway for cell apoptosis, due to the detected ROS generation and the intracellular Ca²⁺ release. From this study, it strongly suggested that the presence of this complex in the T-24 cells induced the mitochondria-mediated apoptotic pathway by activating the caspase cascade. In this pathway, the ROS generation and the enhancement of intracellular [Ca²⁺]_c level were obviously observed, which was closely related to the mitochondria dysfunction. The redox activity of Fe(III)/Fe(II) as the centre of this complex should play the key role for ROS generation. However, the network of the apoptotic pathway in tumor cells is so complicated that other possible apoptotic pathways attributed to this complex may not be excluded.

Acknowledgments

This work was financially supported by the National Natural Science Foundations of China (Nos. 21271051, 21561005), the Doctoral Construction Program and State Key Laboratory Foundation of Guangxi Normal University (CMEMR2013-C05, CMEMR2012-B01/2015-A10). We also thank Dr. Anna Renfrew from the University of Sydney for English syntax check.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

The online version of this article contains supplementary materials. CCDC No. 1429321 for complex 1 contains the supplementary crystallographic data for this paper. The data can be obtained free of charge via http://www.ccdc.cam.ac.uk, or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB21EZ, U.K. (Fax: [phone]; e-mail: [e-mail]).

References

• 1) Koiri R. K., Trigun S. K., Dubey S. K., Singh S., Mishra L., Biometals, 21, 117–126 (2008).
• 2) Fei B. L., Xu W. S., Gao W. L., Zhang J., Zhao Y., Long J. Y., Anson C. E., Powell A. K., J. Photochem. Photobiol. B, 142, 77–85 (2015).
• 3) Pradeepa S. M., Naik H. S. B., Kumar B. V., Priyadarsini K. I., Barik A., Jayakumar S., Inorg. Chim. Acta, 428, 138–146 (2015).
• 4) Spiegel K., Magistrato A., Org. Biomol. Chem., 4, 2507–2517 (2006).
• 5) Rosenberg B., Vancamp L., Trosko J. E., Mansour V. H., Nature (London), 222, 385–386 (1969).
• 6) Munteanu C. R., Suntharalingam K., Dalton Trans., 44, 13796–13808 (2015).
• 7) Yan Y. K., Melchart M., Habtemariam A., Sadler P. J., Chem. Commun., 2005, 4764–4776 (2005).
• 8) Torti S. V., Torti F. M., Nat. Rev. Cancer, 13, 342–355 (2013).
• 9) Chen J., Luo Z., Zhao Z., Xie L., Zheng W., Chen T., Biomaterials, 71, 168–177 (2015).
• 10) Angelé-Martínez C., Goodman C., Brumaghim J., Metallomics, 6, 1358–1381 (2014).
• 11) Absalon M. J., Wu W., Kozarich J. W., Stubbe J., Biochemistry, 34, 2076–2086 (1995).
• 12) Hillard E., Vessieres A., Thouin L., Jaouen G., Amatore C., Angew. Chem. Int. Ed., 45, 285–290 (2006).
• 13) Würtenberger I., Follia V., Lerch F., Cwikla C., Fahrner N., Kalchschmidt C., Flögel B., Kircher B., Gust R., J. Med. Chem., 58, 588–597 (2015).
• 14) Saha S., Majumdar R., Roy M., Dighe R. R., Chakravarty A. R., Inorg. Chem., 48, 2652–2663 (2009).
• 15) Shen A. Y., Wu S. N., Chiu C. T., J. Pharm. Pharmacol., 51, 543–548 (1999).
• 16) Mao X., Li X., Sprangers R., Wang X., Venugopal A., Wood T., Zhang Y., Kuntz D. A., Coe E., Trudel S., Rose D., Batey R. A., Kay L. E., Schimmer A. D., Leukemia, 23, 585–590 (2009).
• 17) Chan S. H., Chui C. H., Chan S. W., Kok S. H., Chan D., Tsoi M. Y. T., Leung P. H. M., Lam A. K. Y., Chan A. S. C., Lam K. H., Tang J. C. O., Med. Chem. Lett., 4, 170–174 (2013).
• 18) Barilli A., Atzeri C., Bassanetti I., Ingoglia F., Dall’Asta V., Bussolati O., Maffini M., Mucchino C., Marchiò L., Mol. Pharm., 11, 1151–1163 (2014).
• 19) Tardito S., Barilli A., Bassanetti I., Tegoni M., Bussolati O., Franchi-Gazzola R., Mucchino C., Marchiò L., J. Med. Chem., 55, 10448–10459 (2012).
• 20) Gao H. L., Jiang S. X., Hu Y. M., Li F. F., Zhang Q. Q., Shi X. Y., Cui J. Z., Inorg. Chem. Commun., 44, 58–62 (2014).
• 21) Shoja M., Gershon H., Bray D., Clarke D. D., Monatsh. Chem., 129, 843–853 (1998).
• 22) Heidary D. K., Howerton B. S., Glazer E. C., J. Med. Chem., 57, 8936–8946 (2014).
• 23) Tang Q., Ni W. X., Leung C. F., Man W. L., Lau K. K. K., Liang Y., Lam Y. W., Wong W. Y., Peng S. M., Liu G. J., Lau T. C., Chem. Commun., 49, 9980–9982 (2013).
• 24) Steger H. F., Corsini A., J. Inorg. Nucl. Chem., 35, 1621–1636 (1973).
• 25) Sheldrick G. M., SHELXS-97, Program for the Solution of Crystal Structures, University of Göttingen, Germany, 1997.
• 26) Qin Q. P., Chen Z. F., Shen W. Y., Jiang Y. H., Cao D., Li Y. L., Xu Q. M., Liu Y. C., Huang K. B., Liang H., Eur. J. Med. Chem., 89, 77–87 (2015).
• 27) Zhao X. S., Chen Y. X., Luo J. S., Wang H., Li S. L., Zhou H. P., Wu J. Y., Tian Y. P., Spectrochim. Acta A Mol. Biomol. Spectrosc., 123, 30–36 (2014).
• 28) Dharmaraja J., Subbaraj P., Esakkidurai T., Shobana S., J. Coord. Chem., 68, 4314–4344 (2015).
• 29) Sarkar T., Banerjee S., Hussain A., RSC Adv., 5, 29276–29284 (2015).
• 30) Marzenell P., Hagen H., Sellner L., Zenz T., Grinyte R., Pavlov V., Daum S., Mokhir A., J. Med. Chem., 56, 6935–6944 (2013).
• 31) Schikora M., Reznikov A., Chaykovskaya L., Sachinska O., Polyakova L., Mokhir A., Bioorg. Med. Chem. Lett., 25, 3447–3450 (2015).
• 32) Dutta S., Abe H., Aoyagi S., Kibayashi C., Gates K. S., J. Am. Chem. Soc., 127, 15004–15005 (2005).
• 33) Tabassum S., Asim A., Arjmand F., Afzal M., Bagchi V., Eur. J. Med. Chem., 58, 308–316 (2012).
• 34) Ma D. L., Che C. M., Chemistry, 9, 6133–6144 (2003).
• 35) Ramakrishnan S., Suresh E., Riyasdeen A., Akbarsha M. A., Palaniandavar M., Dalton Trans., 40, 3524–3536 (2011).
• 36) Li P., Niu M. J., Hong M., Cheng S., Dou J. M., J. Inorg. Biochem., 137, 101–108 (2014).
• 37) Wen J. H., Li C. Y., Geng Z. R., Ma X. Y., Wang Z. L., Chem. Commun., 47, 11330–11332 (2011).
• 38) Lakowicz J. R., “Principles of Fluorescence Spectroscopy,” 3rd ed., Springer-Verlag, Germany, 2006.
• 39) Dunham S. U., Chifotides H. T., Mikulski S., Burr A. E., Dunbar K. R., Biochemistry, 44, 996–1003 (2005).
• 40) Shao K. P., Zhang X. Y., Chen P. J., Xue D. Q., He P., Ma L. Y., Zheng J. X., Zhang Q. R., Liu H. M., Bioorg. Med. Chem. Lett., 24, 3877–3881 (2014).
• 41) Chou C. C., Yang J. S., Lu H. S., Ip S. W., Lo C., Wu C. C., Lin J. P., Tang N. Y., Chung J. G., Chou M. J., Teng Y. H., Chen D. R., Arch. Pharm. Res., 33, 1181–1191 (2010).
• 42) Fan L., Ma Y., Liu Y., Zheng D., Huang G., Eur. J. Pharmacol., 743, 79–88 (2014).
• 43) Guichard A., Humbert P., Tissot M., Muret P., Courderot-Masuyer C., Viennet C., Int. J. Pharm., 479, 422–429 (2015).
• 44) Ma C. G., Song M. M., Zhang Y., Yan M. Q., Zhang M., Bi H., Toxicol. Rep., 1, 114–121 (2014).
• 45) Janicke R. U., Sprengart M. L., Wati M. R., Porter A. G., J. Biol. Chem., 273, 9357–9360 (1998).
• 46) Huang H. L., Li Z. Z., Liang Z. H., Yao J. H., Liu Y. J., Eur. J. Med. Chem., 46, 3282–3290 (2011).
• 47) Liu Z., Xiang Q., Du L., Song G., Wang Y., Liu X., Food Chem., 141, 289–296 (2013).
• 48) Singh B. K., Tripathi M., Pandey P. K., Kakkar P., Toxicology, 275, 1–9 (2010).
• 49) Girish K. S., Paul M., Thushara R. M., Hemshekhar M., Sundaram M. S., Rangappa K. S., Kemparaju K., Biochem. Biophys. Res. Commun., 438, 198–204 (2013).
• 50) Xie Z., Zhang Y., Li A., Li P., Ji W., Huang D., Toxicol. Lett., 192, 115–118 (2010).