Synthesis, Cytotoxicity, and Cell Death Profile of Polyaminoanthraquinones as Antitumor Agents


Corresponding authors: Chaojie Wang,; Jin Zhao,


Investigation on designed polyaminoanthraquinones revealed that the anthraquinones bearing triamine motifs are generally more potent than their counterparts with diamine or tetramine motifs. Compared with the reference drug mitoxantrone (MTX), 9b and 9c exhibited better inhibitory activity on cancerous HepG2 cells and preferable cell selectivity in the further screen of normal QSG7701 cells, although they were not assimilated by cancer cells via the polyamine transporter. The presence of polyamine motifs elevated the interaction of compounds 9b and 9c with lysosomes and resulted in distinct mode of cell death. 9c and MTX could cause caspases-dependent HepG2 cell apoptotic death involving in mitochondrial membrane potential (MMP) loss, cytochrome c release, and caspase-3 activation. However, 9b, which contained only one less methylene group in the polyamine tail, produced cytotoxicity by necrosis. In conclusion, the modification of anthraquinones with polyamines may furnish potent anticancer drug candidates against hepatocellular carcinoma undergoing distinct cell death mechanisms.

Naturally occurring molecules can be served as useful scaffolds in the course of drug discovery. Owing to the chemical diversity, the structural complexity, and the inherent biologic activity, natural molecules may exert potent antitumor activities (1–5). In this respect, the conjugation of two natural scaffolds with great anticancer potential should yield better hit in cancer chemotherapy.

Natural polyamines, which sophisticatedly maintain homeostasis under physiological conditions, play a crucial role in modulating cell growth. Interestingly, the existing polyamine scaffolds could reinforce the interaction of polyamine derivatives with biological targets (6,7). Numerous cytotoxic drugs such as phthalocyanine (8), paclitaxel (9), and artemisinin (10) have been coupled with polyamines to achieve improved antitumor potency and cell selectivity. It is worthy of attention that F14512, an epipodophyllotoxin–spermine derivative recently reported, has reached phase I clinical trials (11). More recently, we found that the presence of polyamine motif may endow some intriguing features to the parent drugs (12).

Natural and synthetic anthraquinones have a wide spectrum of antitumor activity. For example, mitoxantrone (MTX), a commercial anthraquinone drug in clinical use, plays an important role in cancer chemotherapy by interfering with DNA, act as DNA topoisomerase poisons (13). However, because of the severe side-effects, that is, myelosuppression, cardiotoxicity, and extravasation, much attention has been paid to enhance the curative effect and attenuate the side-effect. In addition to the modification on the parent anthraquinone backbones (14), variations in side chains can also result in biological improvements (15). As a kind of lead compounds structurally similar to anthraquinones, anthracene–polyamine conjugates showed potent tumor cell growth inhibition in vitro and could be preferentially transported into cancer cells with upregulated polyamine transporter (PAT) (16–19). Recently, the naphthoquinone–polyamine conjugate reported by Bolognesi and coworkers was found to inhibit EGF binding to its receptor (EGFR) and induces apoptotic cell death without the PAT information (20,21). To date, there are only limited reports about polyamine-containing anthraquinone derivatives as antitumor agents (22). In this article, we planned to attach polyamine skeletons to the 9, 10-anthraquinone frameworks, with the expectation that such hybridization would generate synergistic effects in cytotoxicity. In addition, the antitumor activity and cell death profile of anthraquinone–polyamine conjugates were further investigated.

Methods and Materials


Purifications by column chromatography were carried out over silica gel (230–400 mesh). Melting points were determined on a WRS-1B melting point apparatus and were uncorrected. NMR spectra were recorded on Bruker AV-400 spectrometer (400 MHz for 1H NMR and 100 MHz for 13C NMR spectra). Mass spectra were determined by APCI recorded on a Esquire 3000 LC-MS mass instrument. Elemental analyses were performed by the Vario Elemental CHSN-O microanalyzer, and the purity of tested compounds was >95%. Compounds 5ab and 6ab were prepared according to the procedures as previously reported in the literature (23,24).

General procedure for the synthesis of compounds 2ac

1-Chloro-9,10-anthraquinone or 1,8-bischloro-9,10-anthraquinone (4.12 mmol) reacted with N,N-dimethylamino)-ethylamine or 3-aminopropanol (8.24 mmol) in ethylene glycol (30 mL) at 130 °C for 12 h. Cold water (50 mL) was added and solid filtered. The residue was purified by flash column chromatography to give the intermediate. The intermediates 2ab were directly transformed into hydrochloride salts without further purification.

1-(3-Hydroxyl-propylamino)-9, 10-anthraquinone 2c

Red solid; yield 89%, m.p. 183–184 °C; 1H NMR (400 MHz, DMSO-d6): δ 9.64–9.60 (m, 1H, NH), 8.17–8.03 (m, 2H, ArH), 7.85–7.74 (m, 2H, ArH), 7.58–7.53 (m, 1H, ArH), 7.35–7.32 (m, 1H, ArH), 7.18–7.14 (t, 1H, = 8.0 Hz, ArH), 3.57–3.52 (m, 2H, CH2), 3.38–3.35 (m, 2H, CH2), 1.81–1.74 (m, 2H, CH2); 13C NMR (100 MHz, DMSO-d6): δ 182.2, 149.8, 134.4, 133.7, 132.5, 130.8, 118.0, 116.1, 109.9, 60.2, 41.2, 32.1; Anal. Calcd. for C17H15NO3·0.22H2O: C, 71.58; H, 5.46; N, 4.91. Found: C, 71.96; H, 5.28; N, 4.87; APCI-MS: m/z 282.4 [M + H]+.

General procedure for the synthesis of compounds 3ab

The intermediate (2ab) was dissolved in ethanol (5 mL), and hydrochloric acid (4 N) was slowly added at 0 °C. The precipitate was collected by filtration and washed with absolute ethanol to give the target compounds 3ab.

1-[(2-N, N-Bismethylamino)-ethylamino]-9, 10-anthraquinone hydrochloride 3a

Red solid; yield 81%, m.p. 187–188 °C; 1H NMR (400 MHz, D2O): δ 9.82 (br s, 1H, NH), 8.29 (d, 1H, = 8.0 Hz, ArH), 8.22 (d, 1H, = 8.0 Hz, ArH), 7.74–7.71 (m, 2H, ArH), 7.59–7.51 (m, 2H, ArH), 7.04 (d, 1H, = 8.0 Hz, ArH), 3.44–3.39 (m, 2H, CH2), 2.67 (t, 2H, = 6.0 Hz, CH2), 2.35 (s, 6H, 2CH3); 13C NMR (100 MHz, D2O): δ 185.0, 184.0, 151.6, 135.4, 135.1, 134.8, 134.1, 133.1, 133.0, 126.9, 126.8, 118.0, 115.7, 113.2, 58.12, 5.72, 41.1 (2C); Anal. Calcd. for C18H19N2ClO2: C, 65.35; H, 5.79; N, 8.47. Found: C, 65.09; H, 5.72; N, 8.32; APCI-MS: m/z 295.3 [M + H]+.

1-Chloro-8-[(2-N,N-bismethylamino)-ethylamino]-9,10-anthraquinone hydrochloride 3b

Red solid; yield 76%, m.p. 193–194 °C; 1H NMR (400 MHz, D2O): δ 7.84 (m, 1H, ArH), 7.72–7.69 (m, 2H, ArH), 7.55–7.54 (m, 3H, ArH), 4.40 (m, 2H, CH2), 3.67 (m, 2H, CH2), 3.04 (s, 6H, 2CH3); 13C NMR (100 MHz, D2O): δ 182.9 (2C), 136.9 (2C), 135.8 (2C), 135.2, 135.1, 132.8, 131.8, 128.0 (2C), 126.9, 126.8, 52.4, 50.7, 43.5, 41.8; Anal. Calcd. for C18H18N2Cl2O2·H2O: C, 56.41; H, 5.26; N, 7.31. Found: C, 56.65; H, 5.28; N, 7.19; APCI-MS: m/z 329.7 [M + H]+.

Synthesis of compound 7

To a solution of 1-amino-9, 10-anthraquinone (1.0 g, 4.4 mmol) in dichloromethane (100 mL), ferric trichloride (0.4 g, 2.0 mmol), and epichlorohydrin (9.6 mL, 126.0 mmol) were slowly added. The mixture was stirred at room temperature for 12 h, then water (200 mL) was added, and the reaction solution was extracted with dichloromethane and dried with anhydrous magnesium sulfate. The solvent was removed, and the residue was recrystallized in the mixture of n-hexane/ethanol (v/v, 5/1) to give compound 7 as red solid.

Yield 86%, m.p. 162–163 °C. 1H NMR (400 MHz, DMSO-d6): δ 9.62 (m, 1H, NH), 8.11–8.09 (m, 1H, ArH), 8.05–8.03 (m, 1H, ArH), 7.83–7.80 (m, 1H, ArH), 7.79–7.76 (m, 1H, ArH), 7.57–7.53 (m, 1H, ArH), 7.34 (d, 1H, = 6.8 Hz, ArH), 7.15 (d, 1H, = 8.4 Hz, ArH), 4.66 (t, 1H, = 5.0 Hz, CH), 3.57–3.53 (m, 2H, CH2), 3.38–3.34 (m, 1H, OH), 1.82–1.75 (m, 2H, CH2); APCI-MS: m/z 316.8 [M + H]+.

General procedure for the synthesis of compounds 9ae

Compound 7 (0.6 g, 1.9 mmol) and respective N-protected amine (8.3 mmol) and anhydrous potassium carbonate (0.8 g, 5.6 mmol) were dissolved in acetonitrile (30 mL). The mixture was refluxed for 4 h. The solvent was removed under vacuum, and the residue was dissolved in chloroform, washed with water, and dried over anhydrous sodium sulfate. The solvent was removed, and the residue was purified by flash column chromatography to give the Boc-protected intermediates 8ae.

The intermediates 8ae were dissolved in ethanol (5 mL), and 4 N hydrochloric acid was slowly added at 0 °C. The precipitation was collected by filtration and washed with absolute ethanol to give the target compounds 9ae.

1-[3-(2-Amino-ethylamino)-2-hydroxyl-propylamino)]-9,10-anthraquinone hydrochloride 9a

Red solid; yield 68%, m.p. 193–194 °C; 1H NMR (400 MHz, D2O): δ 7.21–7.17 (m, 2H, ArH), 7.07–7.06 (d, 1H, = 6.4 Hz, ArH), 6.97–6.96 (d, 1H, = 6.4 Hz, ArH), 6.64 (m, 1H, ArH), 6.29–6.27 (d, 1H, = 6.4 Hz, ArH), 6.16–6.14 (d, 1H, = 8.4 Hz, ArH), 3.97 (m, 1H, CH), 3.51–3.46 (m, 4H, 2CH2), 3.27–3.24 (d, 1H, = 11.2 Hz, CH), 3.13–3.11 (d, 1H, = 11.0 Hz, CH), 2.88–2.76 (m, 2H, CH2). 13C NMR (100 MHz, D2O): δ 183.1, 183.1, 150.1, 150.0, 135.1, 134.3, 133.0, 132.8, 131.8, 125.8, 125.7, 118.6, 116.3, 110.6, 65.2, 51.0, 45.2, 44.5, 35.5; Anal. Calcd. for C19H23Cl2N3O3·H2O: C, 53.03; H, 5.86; N, 9.76. Found: C, 53.12; H, 5.88; N, 9.54; APCI-MS: m/z 340.4 [M + H]+.

1-[3-(3-Amino-propylamino)-2-hydroxyl-propylamino)]-9,10-anthraquinone hydrochloride 9b

Red solid; yield 68%, m.p. 194–195 °C; 1H NMR (400 MHz, D2O): δ 7.26–7.22 (m, 2H, ArH), 7.15 (d, 1H, = 7.2 Hz, ArH), 7.05 (d, 1H, = 7.2 Hz, ArH), 6.68 (d, 1H, = 7.2Hz, ArH), 6.37 (d, 1H, = 7.2 Hz, ArH), 6.122 (d, 1H, = 8.8 Hz, ArH), 3.97–3.95 (m, 1H, CH), 3.19 (t, 2H, = 8.2 Hz, CH2), 3.08 (t, 2H, = 8.0 Hz, CH2), 3.03–3.00 (m, 2H, CH2), 2.88–2.81 (m, 2H, CH2), 2.13–2.09 (m, 2H, CH2); 13C NMR (100 MHz, D2O) δ 183.1, 183.1, 150.1, 135.1, 134.3, 133.0, 132.8, 131.8, 130.7, 125.8, 125.7, 118.6, 118.6, 116.3, 110.6, 65.2, 51.0, 45.2, 44.5, 35.5; Anal. Calcd. for C20H25Cl2N3O2·H2O: C, 54.06; H, 6.12; N, 9.46. Found: C, 54.26; H, 5.94; N, 9.53; APCI-MS: m/z 354.2 [M + H]+.

1-[3-(4-Amino-butylamino)-2-hydroxyl-propylamino)]-9,10-anthraquinone hydrochloride 9c

Red solid; yield 62%, m.p. 198–199 °C; 1H NMR (400 MHz, D2O): δ 7.28–7.24 (m, 2H, ArH), 7.15 (d, 1H, = 6.8 Hz, ArH), 7.05 (d, 1H, = 6.8 Hz, ArH), 6.71 (t, 1H, = 7.6 Hz, ArH), 6.36 (d, 1H, = 7.2 Hz, ArH), 6.23 (d, 1H, = 8.4 Hz, ArH), 3.99–3.98 (m, 1H, CH), 3.20–3.16 (m, 3H, CH+ CH), 3.08–3.05 (m, 3H, CH+ CH), 2.90–2.83 (m, 2H, CH2), 1.84–1.78 (m, 4H, 2CH2); 13C NMR (100 MHz, D2O): δ 183.2, 183.2, 150.1, 135.1, 134.3, 133.1, 132.9, 131.9, 130.9, 125.8, 125.7, 118.7, 116.3, 110.7, 65.3, 50.5, 47.2, 45.3, 38.8, 24.0, 22.7; Anal. Calcd. for C21H27Cl2N3O2·0.9H2O: C, 55.24; H, 6.36; N, 9.20. Found: C, 55.48; H, 6.01; N, 8.87; APCI-MS: m/z 368.3 [M + H]+.

1-{3-[3-(4-Amino-butylamino)-propylamino-2-hydroxyl-propylamino)]-9,10-anthraquinone hydrochloride 9d

Red solid; yield 65%, m.p. 202–203 °C; 1H NMR (400 MHz, D2O): δ 7.33–7.28 (m, 2H, ArH), 7.19 (d, 1H, = 6.8 Hz, ArH), 7.11 (d, 1H, = 7.2 Hz, ArH), 6.77 (t, 1H, = 7.8 Hz, ArH), 6.41 (d, 1H, = 7.2 Hz, ArH), 6.29 (d, 1H, = 8.8 Hz, ArH), 4.05–4.03 (m, 1H, CH), 3.27–3.07 (m, 12H, 6CH2,), 2.96–2.89 (m, 2H, CH2), 2.18–2.13 (m, 2H, CH2), 1.90–1.88 (t, 2H, CH2); 13C NMR (100 MHz, D2O): δ 183.3, 183.2, 150.2, 135.2, 134.4, 133.1, 133.0, 132.0, 130.9, 125.9, 125.8, 118.7, 116.3, 110.8, 65.4, 50.6, 47.3, 47.1, 45.4, 44.7, 36.7, 23.8, 22.9, 22.8; Anal. Calcd. for C24H35Cl3N4O3·1.4H2O: C, 51.55; H, 6.81; N, 10.02. Found: C, 51.37; H, 7.14; N, 10.41; APCI-MS: m/z 425.3 [M + H]+.

1-{3-[3-(4-Amino-butylamino)-butylamino-2-hydroxyl-propylamino)]-9,10-anthraquinone hydrochloride 9e

Red solid; yield 63%, m.p. 201–202 °C; 1H NMR (400 MHz, D2O): δ 7.45–7.42 (m, 2H, ArH), 7.40 (d, 1H, = 4.0 Hz, ArH), 7.31 (d, 1H, = 7.2 Hz, ArH), 6.92 (t, 1H, = 8.0Hz, ArH), 6.63 (d, 1H, = 7.2Hz, ArH), 6.46 (d, 1H, = 8.8 Hz, ArH), 4.10–4.07 (m, 1H, CH), 3.27–2.17 (m, 3H, CH+ CH), 3.13–3.08 (m, 6H, 3CH2), 3.04–3.01 (m, 3H, CH+ CH), 1.83–1.82 (m, 4H, 2CH2), 1.76–1.75 (m, 4H, 2CH2); 13C NMR (100 MHz, D2O): δ 183.8, 183.7, 150.5, 135.4, 134.5, 134.5, 133.3, 132.4, 131.2, 126.0, 126.0, 118.8, 116.4, 111.1, 65.4, 50.5, 47.2, 46.9, 46.9, 45.4, 38.8, 23.9, 22.9, 22.8, 22.8; Anal. Calcd for C25H37Cl3N4O3·1.3H2O: C, 52.55; H, 6.99; N, 9.81. Found: C, 52.25; H, 7.28; N, 9.59; APCI-MS: m/z 439.3 [M + H]+.


The reagents used were of commercial origin and were employed without further purification. Solvents were purified and dried by standard procedures. Chemicals used in bioassay were purchased from Sigma, unless otherwise indicated. RPMI1640 and fetal calf serum (FCS) were ordered from Gibco. Primary antibodies against caspase-3 and cytochrome c as well as Cy5-conjugated goat antimouse secondary antibody were purchased from Santa Cruz biotechnology. Stock solution (10 mm) was prepared in DMSO and diluted to various concentrations with serum-free culture medium.

DNA interaction assay

Fluorescence titration was performed at a fixed concentration of DNA-EB system ([DNA] = 4.15 × 10−5 m, [EB] = 1.52 × 10−5 m) in sodium phosphate buffer (20 mm sodium phosphate, 20 mm NaCl, pH 6.3). Small aliquots of 9b were added into the solution at final concentrations from 0 to 0.8 × 10−5 m. Fluorescence intensity was measured at 602.9 nm.

Cell lines and culture conditions

Cell lines, Hela, K562, HepG2, HCT116 and QSG7701 were obtained from American Type Culture Collection (ATCC). Cells were cultured in RPMI1640, supplemented with 10% heat-inactivated FCS, antibiotics (penicillin, 100 units/mL; streptomycin sulfate, 100 μg/mL) at 37 °C, in an atmosphere of 95% air and 5% CO2 under humidified conditions. Aminoguanidine (1 mm) was added as an inhibitor of amine oxidase derived from FCS and had no effect on the various parameters of the cell measured in this study.

Cell proliferation and cell viability MTT assay

Chemosensitivity was assessed using the 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl tetrazolium bromide (MTT) assay. Briefly, exponentially growing K562 cells were seeded into 96-well plates at 4000 cells per well and treated with indicated concentrations of samples for 48 h, and then 10 μL of MTT (10 mg/mL) was added. After incubation for 4 h at 37 °C, the purple formazan crystals (i.e., a reduced form of MTT) generated from viable cells were dissolved by adding 100 μL of 10% SDS (sodium dodecyl sulfate) in each well. The absorbance of each well was then read at 570 nm.

In addition, exponentially growing Hela, HepG2, HCT116, or QSG7701 cells were seeded into 96-well plates at 5000 cells/well and allowed to attach overnight. The cells were treated with the indicated concentration (0.1, 1.0, 10, 30, 50 μm) of samples for 48 h, and then 100 μL of MTT (1 mg/mL) was added. After incubation for 4 h at 37 °C, the MTT solution was removed, and the remaining formazan crystals were dissolved with 150 μL of DMSO in each well. The absorbance of each well was then read at 570 nm.

Polyamine transport experiment

The cells were treated with the increasing concentrations of synthetic polyamine conjugates in the absence and the presence of DFMO (5 mm) or SPD (500 μm) for 48 h, and 100 μL MTT (1 mg/mL) was added to each well. After incubation at 37 °C for 4 h, the MTT solution was removed and the crystals of the viable cells were dissolved with DMSO. The absorbance of each well was read at 570 nm. The inhibition rate was calculated from plotted results using untreated cells as 100%.

Lysosomal membrane stability assay

Untreated or treated with all novel conjugates (10 μm) for 48 h, HepG2 cells, both adherent and floating, were collected and stained with Lyso-Tracker Red (50 nm) for 60 min at room temperature. After washed, the cells were stained by Hoechst 33342 (2 μg/mL) for 15 min at room temperature. Lysosomal membrane stability was determined by assessing red fluorescence of Lyso-Tracker Red with high-content screening (HCS) (Thermo Scientific Cellomics ArrayScan Vti, Cellomics, Inc., Pittsburgh, PA).

Apoptosis and necrosis assay

HepG2 cells were seeded into six-well plates and treated with conjugates (10 μm). After 48 h, the cells were washed twice with cold PBS and then stained with annexin V-FITC (5 μL annexin V-FITC was added in 400 μL of binding buffer) or propidium iodide (PI) (100 μg/mL) at 4 °C in the dark for 15 min. After washing twice with cold PBS, the cells were counterstained with Hoechst 33342 (2 μg/mL) for 15 min and then were determined using HCS.

Measurement of mitochondrial membrane potential

Mitochondrial membrane potential (MMP) was assessed by the retention of Rh123, a membrane-permeable fluorescent cationic dye. The uptake of Rh123 by mitochondria is proportional to the MMP. Briefly, cells were incubated with Rh123 (0.1 μg/mL) in the dark at room temperature for 20 min. After being washed with PBS, the HepG2 cells were analyzed by HCS.

Double immunocytochemical staining

HepG2 cells were untreated or treated with conjugates (10 μm) for 48 h. After being washed with PBS, the cells were fixed with 10% formaldehyde at room temperature for 20 min, permeabilized with 90% methanol at −4 °C for 10 min, and blocked with 10% defatted milk at room temperature for 1 h. Cells were washed three times with PBS after fixation and permeabilization and with TBST after blocking. Thereafter, the cells were incubated with rabbit anticytochrome c or caspase-3 (diluted 1:500; Santa Cruz, CA, USA) for 1 h at room temperature, respectively, and thereafter with Cy5 goat anti-rabbit IgG (diluted 1:100) for 1 h at room temperature. Antibody was removed by washing three times with TBST, after each antibody incubation step. Finally, cells were counterstained with Hoechst 33342 (2 μg/mL) for 15 min and then washed three times with TBST. The fluorescence intensity change of cytochrome c or caspase-3 was analyzed by HCS.

Data analysis

All data are presented as mean ± SD and analyzed using students t-test or analysis of variance (anova) followed by q-test.

Results and Discussion


We decided to conjugate 1-chloro-, 1,8-bischloro-, or 1-amino-substituted 9, 10 anthraquinone with alkylamines that are either commercially available or easy to be synthesized. Initially, we synthesized three compounds (2c, 3a, and 3b) containing short alkylamine chains (Scheme 1). The synthesis was performed by amination of 1-chloro- or 1,8-bischloro-9,10-anthraquinone. The target compound 2c was prepared via nucleophilic substitution of 3-aminopropanol with quinone aromatic nucleus. Because of the low reactivity of 1-chloro-anthraquinone, the transformation was achieved in high-boiling-point solvent ethylene glycol, and the yield was comparatively approving (yield 92%). In a similar procedure, N, N-dimethyl-ethanediamine was smoothly converted to compounds 2ab. Following flash silica gel column chromatography, the intermediates 2ab were treated with hydrochloric acid to form its hydrochloride salts 3a and 3b with the yield of two-step 81% and 76%, respectively.

                Scheme 1:

 Synthesis of target compounds 2c, 3a, and 3b. Reagents and conditions: (a) N, N-dimethyl-ethanediamine or 3-aminopropanol, ethylene glycol mono-ethyl ether, reflux, 12 h; (b) 4 m hydrochloric acid, ethanol, 0 °C, overnight.

To investigate the effect of the alkylamine side chains on biological activity, the target compounds 9ae were decorated with various polyamine skeletons differing in chain length and number of nitrogen atoms within the carbon chains. As an attempt, one hydroxy was introduced adjacent to aromatic ring with expectation of better pharmacology.

For the synthesis of the target molecules with one terminal amine function, the alkylamine chains must be selected, protected. The Boc-protected polyamine chains were prepared through the classic Gabriel procedure as previously reported (23,24). As outlined in Scheme 2, the N-bromosuccinimides 4ab reacted with Boc-protected diamines to give compounds 5ab, and subsequent deprotection of the intermediates in hydrazine hydrate provides the Boc-protected triamines 6ab.

                Scheme 2:

 Preparation of polyamine skeletons. Reagents and conditions: (a) N-protected butanediamine, K2CO3, acetonitrile, 40 °C, overnight; BOC2O, methanol, r.t., overnight; (b) 80% hydrazine hydrate, ethanol, r.t., overnight.

Scheme 3 shows the synthetic route of the target compounds 9ae. To introduce the hydroxy functional group into the side chain, epichlorohydrin was used as starting material and was reacted with 1-amino-9, 10-anthraquinone in the presence of catalytic amount of FeCl3 to give the hydroxy-containing compound 7. The attachment of respective Boc-protected alkylamines to compound 7 was performed through nucleophilic substitution. Finally, deprotection of the Boc groups in the intermediates 8ae was accomplished by treatment with hydrochloric acid in ethanol to yield the target compounds 9ae as hydrochloride salts.

                Scheme 3:

 Synthetic route of target compounds 9ae. Reagents and conditions: (a) epichlorohydrin, FeCl3, CH2Cl2, r.t., 15 h; (b) respective N-protected amines, K2CO3, acetonitrile, reflux, 4 h; (c) 4 m hydrochloric acid, ethanol, 0 °C, overnight.


The antiproliferative properties of the newly synthesized polyamino-anthraquinone derivatives, 2c, 3ab, and 9ae, were evaluated in three types of cancer cell lines: K562 (human myelogenous leukemia cell line), Hela (human cervical carcinoma cell line), HepG2 (human hepatoma cell line), and one normal cell line: QSG7701 (normal hepatic cell line) for liver cell selectivity. Mitoxantrone was used as a reference compound, and the IC50 values (the concentration required for 50% cell growth inhibition) were summarized in Table 1.

Table 1. In vitro cytotoxic results of polyaminoanthraquinone derivatives against K562, Hela, HCT116, HepG2, and QSG7701 cells
  1. aIC50 values were from three independent experiments, and the standard deviations were <10%. IC50 values were given only if they were <50 μm, which was the maximum concentration tested.

  2. bThe data (above 50 μm) were extrapolated ones.

  3. cN demonstrated the IC50 value no. tested.

2c >50bNc>50>50
3a 6.63 ± 0.459.10 ± 0.847.50 ± 0.0139.57 ± 1.04
3b 27.90 ± 0.283.40 ± 0.5618.20 ± 0.3848.61 ± 1.62
9a 5.63 ± 0.4915.37 ± 0.766.59 ± 0.1337.44 ± 0.35
9b 10.24 ± 0.302.74 ± 0.462.67 ± 0.5824.03 ± 0.15
9c 12.90 ± 0.494.18 ± 0.574.81 ± 0.1920.49 ± 0.15
9d 16.68 ± 0.174.55 ± 0.136.04 ± 0.1922.82 ± 0.31
9e 22.59 ± 0.119.80 ± 0.2124.14 ± 0.8376.77 ± 0.87
MTX4.55 ± 1.952.35 ± 0.0518.60 ± 0.540.62 ± 0.31

As shown in Table 1, the compounds 3ab, 9ae, bearing an alkylamine side chain adjacent to the quinone nucleus, exhibited significant anticancer activities in three tested tumor cell lines with the IC50 values ranging from 2.67 to 27.90 μm, while 2c, which has a hydroxyalkyl side chain, was found to be inactive against the growth of all cell lines tested, with the IC50 value above 50 μm. This suggested that the alkylamine side chain probably accounts for the difference. Generally speaking, the anthraquinones with triamine motifs showed superior inhibitory activity than their diamine or tetramine counterparts.

Compounds 3ab and 9ae showed comparable antitumor potency, with IC50 values generally being in the same order of magnitude with MTX. Of the tested compounds, 9b and 9c stand out as the most efficacious compounds on account of their remarkable performance in solid tumors (especially for HepG2 cells). However, previous reports revealed that the presence of polyamines may result in cancer cell selectivity (19). The growth inhibitory activity against cancerous HepG2 and normal QSG-7701 cells supported that the synthesized compounds were generally more selective than MTX, which is unexpectedly toxic to QSG-7701 cells. The elevated toxicity to HepG2 cells and downregulated toxicity to normal QSG-7701 indicated that polyaminoanthraquinones were at least well worthy of further investigation as antihepatocellular carcinoma agents.

Although anthracene–polyamine conjugates could be ferried by cells via PAT (18), whether polyaminoanthraquinone derivatives could harness PAT is unknown. Interestingly, our results of the MTT assay with DFMO-(adifluoromethylornithine) or Spd (spermidine)-treated HepG2 cells demonstrated that the addition of DFMO or Spd had no effects on the IC50 values of these polyaminoanthraquinone compounds, indicating the cell uptake instead of PAT (data not shown). The entrance of polyaminoanthraquinone derivatives into cells may involve different mechanisms, such as passive diffusion and utilization of another transporter on the cell membrane, as recently observed in dinaphthalimide–polyamine conjugates (25).

The anthraquinone frameworks usually behaved as efficient DNA intercalators. As expected, the newly synthesized polyaminoanthraquinones also exhibited significant interactions with DNA. As shown in Figure 1, 9b could reduce the fluorescent intensity of the DNA-EB complex in a concentration-dependent manner, that is, 9b could replace EB from the DNA-EB complex. Such result obviously indicated that the anthraquinone derivative 9b also possessed the ability as a DNA intercalator.

Figure 1.

 The interaction of 9b on DNA-EB system. [DNA] = 4.15 × 10−5 m, [EB] = 1.52 × 10−5 m; 1–7: [9b] = 0.00, 0.2, 0.3, 0.4, 0.5, 0.6, 0.8 × 10−5 m.

It is known that anthraquinone derivatives may be multi-targeted drugs. To date, although they were generally viewed as DNA-targeted agents, there were few reports discussing intracellular targets on cell level. Previous report indicated that the presence of polyamines in naphthalimides may furnish lysosomes as a new target compared with parent drug amonafide (26). Lysosomes are membrane-bound organelles, which contain an arsenal of different hydrolases and act as crucial role in living cells (27). Destabilization of lysosomal membrane and release of lysosomal content into the cytosol have assuredly been found to influence cell death fate: apoptosis or necrosis (28). The induction of lysosomal membrane permeabilization (LMP) can be analyzed by fluorescence spectrum assay using Lyso-Tracker Red as the probe. Lyso-Tracker Red prefers to accumulate in normal lysosomes, and the fluorescence intensity of Lyso-Tracker Red will decrease in case of lysosomal rupture. As shown in Figure 2, the fluorescence intensity of Lyso-Tracker Red significantly degraded in compounds 9b- and 9c-treated HepG2 cells at 10 μm, indicating the induction of LMP. Furthermore, the intensity of LMP induced by 9b and 9c is more powerful than the reference MTX, which is in concordance with their antiproliferative properties in HepG2 cells. It is speculated that the obviously upregulated impact on LMP by the introduction of polyamine to anthraquinones may be related to their improved cell selectivity.

Figure 2.

 Lysosomal membrane stability was detected in HepG2 cells after 9bc (10 μm) and MTX (20 μm) treatment for 48 h using high-content screening (HCS). (A) The fluorescence intensity of Lyso-Tracker Red was determined using HCS, mitoxantrone as a positive control. Each point represents the means ± SD from four independent experiments. (B) Typical lysosomal membrane permeabilization induction of 9bc and MTX in HepG2 cells using Lyso-Tracker Red and Hoechst 33342 double staining (Magnification, 20×).

Because of the intriguing biological activities described previously, the pharmacology profile of compounds 9b and 9c was further investigated in HepG2 cells. On the basis of the apoptotic properties of anthracene- and naphthoquinone- polyamine conjugates, the apoptosis assay was implemented through the fluorescence staining using annexin V-FITC and propidium iodide (PI). PI is a DNA-binding fluorochrome that intercalates in the double helix. Generally, PI could not permeate into the living cell membrane and result in nucleus staining, whereas it could permeate into the necrotic cells because of its damaged membrane and thus stain nucleus. With regard to the apoptotic cells, the phosphatidylserine (PS) will evaginate on the cell surface, and annexin V-FITC will bind with PS. As shown in Figure 3, the cell number of annexin V-FITC staining significantly increased after 9c treatment (green). Meanwhile, PI staining is not obvious in HepG2 cells (red). And the typical characteristics of apoptotic cell, such as chromatin condensation, nuclear fragmentation, and formation of apoptotic bodies, were also observed (Figure 3). In contrast, after 9b treatment, HepG2 cells were obviously stained by PI but not by annexin V-FITC. These results suggested that the cell death of 9c-treated HepG2 cells was mainly because of apoptosis, while 9b, necrosis. And that the observed alternative cell death profile of tested compounds may result from the downstream signaling pathway of LMP.

Figure 3.

 Apoptosis and necrosis were detected in HepG2 cells after 9bc (10 μm) and MTX (20 μm) treatment for 48 h using high-content screening (HCS). Up: Apoptosis was detected by annexin V-FITC and Hoechst 33342 double staining. Down: Necrosis was detected by PI and Hoechst 33342 double staining (Magnification, 20×).

It has been widely accepted that both extrinsic and intrinsic pathways in programmed cell death converge at caspase-3, and caspase-3 activation is normally considered as a key mediator in cell apoptosis (29). On the basis of the induction of compounds 9bc on LMP, the cell death pathway ought to be investigated for a better comprehension of the selected compounds on the mode of action. With compounds 9b, 9c in hand, we further investigated their effect on caspase-3 activation. As shown in Figure 4, the great difference was observed between 9b and 9c. Compared with mitoxantrone, 9b did not exhibit caspase-3 activation property with none of the fluorescence detected in 9b-treated HepG2 cells. However, compound 9c and MTX could significantly induce caspase-3 activation (the strong fluorescence intensity showed in Figure 4). These results obviously indicated that cell treatments by 9b and 9c result in different cell death pathways. Similar behaviors of naphthalimides (30) and naphthoquinones (20) were also reported recently.

Figure 4.

 Caspase-3 activation was detected in HepG2 cells after 9bc (10 μm) and MTX (20 μm) treatment for 48 h using high-content screening (HCS). Compared with control *p < 0.05, **p < 0.01.

To further understand how the compound 9c induced the apoptotic signaling pathway, the changes in mitochondrial membrane permeability were initially detected. Mitochondria are considered as one of the crucial organelle in cell apoptosis, and major consequences of its changes could be characterized as the mitochondrial membrane potential (MMP) loss and release of cytochrome c. Initiated by the loss of integrity of the outer mitochondrial membrane, cytochrome c and other pro-apoptotic proteins were released into cytosol and thus trigger apoptotic caspase cascade (i.e., caspase-3 activating) (31). So in this study, the induction of mitochondrial membrane potential loss and release of cytochrome c induced by 9c in HepG2 cells were performed using double immunocytochemical staining. As shown in Figure 5, the fluorescence intensity of Rh123 (a membrane-permeable fluorescent cationic dye) significantly decreased, and meanwhile, the immunofluorescence of cytochrome c obviously increased in 9c-treated HepG2 cells. This suggested that 9c could induce MMP loss and trigger cytochrome c release. Particularly, these effects were in positive correlation with its caspase-3 activation effect.

Figure 5.

 The change in MMP and Cyto c release of 9c (10 μm) and MTX (20 μm) was detected using Rh123 staining and immunofluorescence assay by HCS, respectively. Each point represents the means ± SD from four independent experiments. Compared with control *p < 0.05, **p < 0.01.


In conclusion, the in vitro biological assay on polyaminoanthraquinones indicated that the triamine-modified anthraquinones possess comparable or better cytotoxic activity toward solid human cancerous Hela and HepG2 cells compared with MTX. Notably, the increased cytotoxicity of 9b and 9c to HepG2 and less toxic on normal hepatic QSG7701 cells corroborated that the existing polyamine motif will improve the tumor-specific capability of the polyamine-containing drugs. Interestingly, mechanistic studies revealed that 9b, 9c, and MTX target lysosomes to trigger HepG2 cell death. However, the models of cell death initiated by 9b and 9c were unexpectedly different. Compound 9c, which contained a terminal aminobutyl scaffold in the side chain, was confirmed to orchestrate a mitochondria-mediated cell apoptotic death involving in MMP loss, cytochrome c release from mitochondria to cytoplasm, and caspase-3 activation, whereas 9b, which had a terminal aminopropyl scaffold, induced the cell necrosis. Because structurally similar drug candidates with diverse cell death mechanisms may have potential therapeutic applications, such as preventing chemoresistance, both detailed molecular mechanisms of 9bc and polyamine-containing anthraquinones as promising formworks for the development of antihepatocellular carcinoma agents are worthy of further investigation, which is under way in our laboratory.


The authors gratefully thank the National Natural Science Foundation of China (Nos. 21272056, 20872027), the Key Scientific and Technological Projects of Henan Province (Nos. 092102310025, 102102210075, and 102102310195), and the Co-construction Projects of Henan University (No. SBGJ090518) for financial support of this research.

Conflict of Interest

The authors declared that there is no conflict of interests.