Eighteen 5,7-Dihalo-8-quinolinol and 2,2'-Bipyridine Co(II) Complexes as a New Class of Promising Anticancer Agents.

Here we first report the design of a series of bis-chelate Co(II) 5,7-dihalo-8-quinolinol-phenanthroline derivative complexes, [Co(py)(QL1)2] (Co1), [Co(py)(QL2)2] (Co2), [Co(Phen)(QL1)2] (Co3), [Co(Phen)(QL2)2] (Co4), [Co(DPQ)(QL1)2]·(CH3OH)4 (Co5), [Co(DPQ)(QL2)2] (Co6), [Co(DPPZ)(QL1)2]·CH3OH (Co7), [Co(MDP)(QL1)2]·3H2O (Co8), [Co(ODP)(QL1)2]·CH3OH (Co9), [Co(PPT)(QL1)2]·CH3OH (Co10), [Co(ClPT)(QL1)2] (Co11), [Co(dpy)(QL3)2] (Co12), [Co(mpy)(QL1)2] (Co13), [Co(Phen)(QL4)2] (Co14), [Co(ODP)(QL4)2] (Co15), [Co(mpy)(QL4)2]I (Co16), [Co(ClPT)(QL4)2] (Co17), and [Co(ClPT)(QL5)2] (Co18), with 5,7-dihalo-8-quinolinol and 2,2'-bipyridine mixed ligands. The antitumor activity of Co1-Co18 has been evaluated against human HeLa (cervical) cancer cells in vitro (IC50 values = 0.8 nM-11.88 μM), as well as in vivo against HeLa xenograft tumor growth (TIR = 43.7%, p < 0.05). Importantly, Co7 exhibited high safety in vivo and was more effective in inhibiting HeLa tumor xenograft growth (43.7%) than cisplatin (35.2%) under the same conditions (2.0 mg/kg). In contrast, the H-QL1 and DPPZ ligands greatly enhanced the activity and selectivity of Co7 in comparison to Co1-Co6, Co8-Co18, and previously reported cobalt(II) compounds. In addition, Co7 (0.8 nM) inhibited telomerase activity, caused G2/M phase arrest, and induced mitochondrial dysfunction at a concentration 5662.5 times lower than Co1 (4.53 μM) in related assays. Taken together, Co7 showed low toxicity, and the combination could be a novel Co(II) antitumor compound candidate.

Pt-based drugs were extensively used to treat a large number of tumors in the clinic.[1−10] However, cisplatin (cis-[PtCl2(NH3)2]) and its derivatives were limited by drug resistance and some severe side effects,[1114] and consequently, the Ru, Ti, Au, Co, Ir, Os, Rh, Fe, Cu, etc., complexes have been designed and attracted attention.[16−38] In addition, cobalt has emerged as a key of vitamin B12 (cobalamin) metabolism and its metal complexes have been reported as DNA cleavage agents, antiviral, antifungal, antitumor antiproliferative, antioxidant, and have shown antimicrobial activity, such as oxoisoaporphine,[15] shydroxamic acids prodrugs,[16,17] 2-benzimidazole derivatives,[18] acetylenehexacarbonyldicobalt,[19] 2-acetylpyridine and malonic acid dihydrazide,[20] antiulcer drug famotidine,[21] sparfloxacin,[23] valine-derived Schiff bases,[24]N-benzoyl-N′-dialkylthiourea derivatives,[25] nonsteroidal anti-inflammatory drugs,[28] nonsteroidal anti-inflammatory drug tolfenamic acid,[30] 2-acetoxy-(2-propynyl)benzoate]hexacarbonyldicobalt,[31] a fluorescent coumarin,[32] and tetradentate phenolate-based ligand[35] cobalt(II/III) complexes.

Recently, many novel hydroxyquinoline metal compounds, such as Zn, Cu, Ni, Ir, Os, Rh, Au, Co, Fe, Sn, Ru, Pt, Pd, and Ln metal complexes,[39−60] have proved to be promising anticancer drugs in vitro and in vivo. Among them, a small amount of copper(II) complexes of hydroxyquinolines were designed and preliminarily identified as anticancer drugs.[50−60] However, to date, Co(II) complexes bearing 5,7-dihalo-2-methyl-8-quinolinol and o-phenanthroline derivative mixed ligands have not been reported.

To gain the mixed chelating cobalt(II) complexes with high anticancer activity in vitro and in vivo, we first designed 18 novel Co(II) complexes Co1–Co18 with 2,2′-bipyridine (py), 1,10-phenanthroline (Phen), 5,7-dichloro-2-methyl-8-quinolinol (H-QL1), dipyridoquinoxaline (DPQ), 5,7-dibromo-2-methyl-8-quinolinol (H-QL2), dipyridophenazine (DPPZ), 5,6-dimethyl-10-phenanthroline (MDP), 4,4′-dimethoxy-2,2′-bipyridyl (ODP), 4,7-diphenyl-10-phenanthroline (PPT), 4,7-dichloro-1,10-phenanthroline (ClPT), 5-chloro-7-iodo-8-hydroxy-quinoline (H-QL3), 5,5-dimethyl-2,2-dipyridine (dpy), 5,7-diiodo-8-hydroxyquinoline (H-QL4), 4,4′-dimethyl-2′-bipyridine (mpy), and 5,7-dibromo-8-quinolinol (H-QL5). Additionally, the biological properties of Co1–Co18 have been evaluated.

First, the mononuclear complexes Co1–Co18 were prepared by CH3CN–CH3OH (3.5 mL/1.5 mL) reflux of H-QL1 (or H-QL2, H-QL3, H-QL4, H-QL5) and py, Phen, DPQ, DPPZ, MDP, ODP, PPT, ClPT, dpy, and mpy with cobalt(II) acetate (2:1:1) at 80 °C for 24 h, respectively (Scheme ). These 18 new Co(II) complexes were structurally fully characterized (Figures and S1–S45).

Scheme 1

General Synthetic Pathway for Co1–Co18

Figure 1

Crystal structures of Co1 and Co7.

The Co(II) atoms in Co1–Co18 were six-coordinated and surrounded by two deprotonated QL ligands (N∧O-ligand) and one second ligand molecule (N∧N-ligand) showing a distorted octahedral geometry (Figures and S1–S16). In addition, a diagram of Co1–Co18 is shown in Figures and S1–S16, and selected bond distances (Å) and angles (deg) are listed in Tables S1–S54, and the bond lengths (Å) of Co1–Co18 remained normal.

MTT assay was carried out to gauge the in vitro anticancer activity of Co1–C o18, H-QL1, H-QL2, py, Phen, DPQ, DPPZ, MDP, ODP, PPT, ClPT, H-QL3, dpy, H-QL4, CoCl2·6H2O, mpy, and H-QL5 using BEL-7404 (hepatocellular), Hep-G2 (hepatocellular), HeLa (cervical), MCF-7 (breast) cancer cells, and normal HL-7702 (hepatocyte) cells. As a general observation (Table S55), Co7 was more active (IC50 values = 0.80 nM) than the Co1–Co6, Co8–Co18, cisplatin, H-QL1, H-QL2, py, Phen, DPQ, DPPZ, MDP, ODP, PPT, ClPT, H-QL3, dpy, H-QL4, CoCl2·6H2O, mpy, and H-QL5 in all tested cells, and the cytotoxicity of Co1–Co18 against HeLa cells followed the order Co7 > Co5 > Co8 > Co11 > Co3 > Co10> Co9 > Co13 > Co1 > Co6 > Co4> Co2 > Co12 > Co18 > cisplatin > Co15 > Co17 > Co14 > Co16. The higher in vitro anticancer activity for Co7 may be due to the H-QL1 and DPPZ ligands. Such observed, different antitumor effects may be due to the electronic effect of the methyl group and the halogenated and more extended planar ligand of H-QL1 and DPPZ. Interestingly, Co1–Co18 showed low toxicity (IC50 > 60.0 μM) to normal HL-7702 (hepatocyte) cells. Compared with previously reported 8-hydroxyquinoline metal complexes (IC50 ≥ 1.00 nM),[39−60]Co7 exhibited higher cytotoxicity against HeLa cells (IC50 = 0.80 ± 0.21 nM).

Thus, ICP-MS assay showed that the Co(II) concentrations of Co7 ((22.96 ± 0.15 nmol of Co)/106 cells) and Co1 ((18.06 ± 0.05 nmol of Co)/106 cells) were significantly above those of control groups and cisplatin ((4.11 ± 0.59 nmol of Pt)/106 cells),[15] with Co7 (0.80 nM) showing the highest cell accumulation of Co ((22.96 ± 0.15 nmol of Co)/106 cells) and a high extent in HeLa nuclear fraction (Table S56). Thus, Co7 (0.80 nM) showed higher toxicity on the HeLa (cervical) cancer cells possibly due to its better cellular uptake.

Thus, the induction of the level of c-myc, hTERT, and thus, telomerase in HeLa cells by Co1 (4.53 μM) and Co7 (0.80 nM) was investigated using a TRAP-silver staining assay and Western blot. As shown in Figure , Co7 (0.80 nM) showed a more inhibitory effect on c-myc, hTERT, and telomerase activity than that of Co1 (4.53 μM), suggesting that Co1 (4.53 μM) and Co7 (0.80 nM) inhibited c-myc and hTERT, and thus, telomerase levels were related to a variety of malignant cancers.[61−65] Importantly, inhibition of telomerase in Co7 (0.80 nM)-treated cells was 44.10%, while that caused by Co1 (4.53 μM) only reached 7.13%.

Figure 2

Level of telomerase (A) and related factors (B,C) in HeLa cells induced by Co1 (4.53 μM) and Co7 (0.80 nM) at 24 h.

Furthermore, a G2/M population of 22.89% was observed in Co7 (0.80 nM)-treated cells, while the other corresponding G2/M populations of 20.27% and 12.35% were observed in the Co1 (4.53 μM) treated and control cells (Figure ), suggesting that Co1 (4.53 μM) and Co7 (0.80 nM) caused G2/M cycle arrest. In addition, Co7 (0.80 nM) and Co1 (4.53 μM) could inhibit the expression of cyclin B1 and CDK1 in HeLa (cervical) cells (Figure S46), mainly due to the G2/M phase arrest (Figure ) that they cause and their ability to inhibit telomerase.[66−68] Further, immunofluorescence (Figure S47) and Western blot assays (Figure S46) were carried out. Clearly, Co7 (0.80 nM) and Co1 (4.53 μM) could up-regulate the H2A.X and cleaved-PARP levels, indicating that Co7 (0.80 nM) and Co1 (4.53 μM) remarkably induced DNA damage (Figures S46 and S47) and caused G2/M phase arrest following the order of Co7 > Co1.

Figure 3

Co7 (0.80 nM) and Co1 (4.53 μM) caused G2/M phase arrest for 24 h.

In addition, Co7 (0.80 nM) mainly accumulated in a nuclear fraction (Table S56) but also was distributed in the mitochondria. Thus, Co7 (0.80 nM) and Co1 (4.53 μM) caused obviously up-regulated reactive oxygen species (ROS, Figure S48), intracellular [Ca2+] (Figure ), and caspase-3/9 levels (Figure S49 and S50), and down-regulated mitochondrial membrane potential (ΔΨm) (Figure S51) in HeLa cells, illustrating that the ΔΨm, ROS generation, intracellular [Ca2+], and caspase-3/9 played a key role in cancer mitochondrial function damage and apoptosis.[15,69−72]

Figure 4

Effects of Co7 (0.80 nM) and Co1 (4.53 μM) on the [Ca2+] level in HeLa cells at 24 h.

Therefore, to clarify the antimigration effects of Co7 (0.80 nM) and Co1 (4.53 μM) in HeLa cells (Figure a–d), the trans-well migration assay was carried out. It was found that Co7 (0.80 nM) could significantly induce cell migration at 0.80 nM than that of Co1 (4.53 μM).

Figure 5

Antimigration (a–d) and apoptosis (e–g) effect of Co7 (0.80 nM) and Co1 (4.53 μM) on HeLa cells for 24 h.

For this, to further investigate the ability of Co7 (0.80 nM) and Co1 (4.53 μM) to induce HeLa cell apoptotic by flow cytometry (FCM). As shown in Figure e–g, the percentages of apoptotic cells treated with Co7 (0.80 nM) and Co1 (4.53 μM) were 95.68% and 34.42%, respectively, suggesting that Co complexes could cause cell death at higher rates than other 8-quinolinate metal complexes.[39−60]

Furthermore, treatment of HeLa xenografts by Co7 (2.0 mg/kg/q2d) was related to significant reduction (TIR = 43.7%, p < 0.05) in tumor growth (Figure and Tables S57–S59), which represents approximately 1.3-fold reduction compared with cisplatin-treated groups (IR = 35.2 ± 5.8%, p < 0.05).[51,55,56,72−74] In addition, Co7 (2.0 mg/kg/q2d)-treated mice displayed no obvious signs of toxicity (Figure and Tables S57–S59) as indicated by a relatively stable mouse body weight (mend= 20.5 ± 1.3 g) compared to solvent control (mend= 20.7 ± 1.4 g).

Figure 6

Tumor volume (A, mm3 ± SD) and images (B) of the HeLa tumors of Co7 (n = 6) for 21.0 days.

In conclusion, we have shown Co1–Co18 complexes containing mixed 5,7-dihalo-8-quinolinol (H-QL1–H-QL5) and 2,2′-bipyridine derivative-based ligands as potential Co(II) complexes with superior cytotoxicity compared with previously reported Co compounds. MTT studies have demonstrated that Co7 was ca. 300.6 times more cytotoxic than cisplatin (15.03 ± 1.05 μM), with IC50 values of 0.80 ± 0.21 nM while maintaining a high extent in HeLa nuclear fraction and targeting telomerase; thus, it was considerably less cytotoxic to normal HL-7702 (hepatocyte) cells. In addition, the antitumor activity of Co7 has been evaluated in vivo against HeLa xenograft growth (TIR = 43.7%, p < 0.05). It also inhibited telomerase activity, caused G2/M phase arrest, and induced mitochondrial dysfunction at a 104.6-fold lower concentration than Co1 (4.53 μM) in related assays. Thus, the superior cytotoxicity (IC50 = 0.80 ± 0.21 nM) and selectivity index of Co7 in comparison to cisplatin could make it a novel antitumor cobalt(II)-based drug.