1,10-phenanthroline promotes copper complexes into tumor cells and induces apoptosis by inhibiting the proteasome activity

Abstract

Indole-3-acetic acid and indole-3-propionic acid, two potent natural plant growth hormones, have attracted attention as promising prodrugs in cancer therapy. Copper is known to be a cofactor essential for tumor angiogenesis. We have previously reported that taurine, l-glutamine, and quinoline-2-carboxaldehyde Schiff base copper complexes inhibit cell proliferation and proteasome activity in human cancer cells. In the current study, we synthesized two types of copper complexes, dinuclear complexes and ternary complexes, to investigate whether a certain structure could easily carry copper into cancer cells and consequently inhibit tumor proteasome activity and induce apoptosis. We observed that ternary complexes binding with 1,10-phenanthroline are more potent proteasome inhibitors and apoptosis inducers than dinuclear complexes in PC-3 human prostate cancer cells. Furthermore, the ternary complexes potently inhibit proteasome activity before induction of apoptosis in MDA-MB-231 human breast cancer cells, but not in nontumorigenic MCF-10A cells. Our results suggest that copper complexes binding with 1,10-phenanthroline as the third ligand could serve as potent, selective proteasome inhibitors and apoptosis inducers in tumor cells, and that the ternary complexes may be good potential anticancer drugs.

Introduction

Indole-3-acetic acid (IAA) and indole-3-propionic acid (IPA), two potent natural plant growth hormones, have been studied for many years in the agricultural field. IAA is able to regulate plant growth, cell division, and differentiation [1, 2]. In recent years, IAA and IPA have also attracted attention as promising prodrugs in cancer therapy. IAA induces apoptosis in PC-3 prostate cancer cells in combination with UVB irradiation [3]. IAA and horseradish peroxidase in combination can lead to apoptosis in TCCSUP human urinary bladder carcinoma and G361 human melanoma cells via both death-receptor-mediated and mitochondrial apoptotic pathways [4, 5]. IPA may also be a potential therapeutic agent against carcinogenesis by protecting against free-radical damage [6].

Copper is an important cofactor in many proteins and enzymes. All animals require copper for survival and normal physiological function. The participation and activation of copper allows many key enzymes to affect metabolic processes in organisms [7–10]. One key pathway in which copper is important is angiogenesis [11]. Angiogenesis plays a key role in growth, development, and wound healing processes. At the same time, angiogenesis also aids in tumor cell proliferation and metastasis [12–15]. Reducing the copper content in vivo, resulting in inhibition of angiogenesis without destruction of normal cellular function, has been a hot topic for the treatment of cancer. However, the mechanism of angiogenic sensitivity to copper is still not fully understood.

Ubiquitin-mediated proteasomal proteolysis is the major mechanism of degradation of cellular proteins in human cells [16–18]. The 20S proteasome, which is the proteolytic core of the multicatalytic 26S proteasome complex, has several proteolytic activities, including chymotrypsin-like, trypsin-like, peptidylglutamyl peptide hydrolyzing, small neutral amino acid preferring, and branched chain amino acid preferring activities [19–22]. However, it has been shown that only inhibition of the chymotrypsin-like activity is tightly associated with induction of tumor cell death programs [22–25]. At the same time, an associated accumulation of target proteins (e.g., IκB-α) will occur [20, 26].

In the current study, we used the plant growth hormones IAA (1 in Fig. 1) and IPA (2 in Fig. 1) as copper-binding ligands to synthesize and characterize their copper complexes (3–6 in Fig. 1). We then investigated the abilities of 1–6 to inhibit cell proliferation in PC-3 human prostate cancer cells and chymotrypsin-like activity of purified 20S proteasome. We found that 1 and 2 had almost no inhibitory effect on cell proliferation or chymotrypsin-like activity of purified 20S proteasome, whereas 3–6 had differing effects on cell proliferation but a similar inhibitory effect on chymotrypsin-like activity of purified 20S proteasome, at different concentrations. Furthermore, we studied the proteasome-inhibitory and apoptosis-inducing activities of these compounds in the PC-3 prostate cancer cells. Similar to purified proteasome activity and cellular proliferation, 1 and 2 could neither inhibit the chymotrypsin-like activity of the proteasome in human prostate cancer cell cultures nor induce cancer cell death. Moderate inhibitory effects were observed with 3 and 4, whereas 5 and 6 were the most potent. Finally, we compared the effects of 5 and 6 in MDA-MB-231 breast cancer cells with the effects in nontumorigenic MCF-10A cells. Our new findings suggest that (1) copper binding with 1,10-phenanthroline as the third ligand could promote tumor cells to take up copper, resulting in potent proteasome inhibition and apoptosis induction in cancer cells, and (2) MDA-MB-231 cells are more sensitive to the novel candidates 5 and 6 than nontumorigenic cells, suggesting tumor-selective targeting.

Fig. 1.

Chemical structures of 1–6

Materials and methods

Materials

Compounds 1 and 2 were purchased from J&K Scientific (Beijing, China). Complexes 3–6 were synthesized in the Key Laboratory of Marine Chemistry Engineering and Technology (Ocean University of China, Qingdao, China). Bisbenzimide, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), dimethyl sulfoxide (DMSO), and other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA). All compounds were dissolved in DMSO at stock concentrations of 50 mM and stored at 4°C. RPMI 1640 medium and penicillin/streptomycin were purchased from Invitrogen (Carlsbad, CA, USA). Fetal bovine serum was from Aleken Biologicals (Nash, TX, USA). Purified human 20S proteasome was from Boston Biochem (Boston, MA, USA). Rabbit polyclonal antibody against human poly(ADP-ribose) polymerase (PARP; H-250), mouse monoclonal antibodies against IκB-α (H-4), goat polyclonal antibody against β-actin (C-11), and secondary antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Elemental analysis was performed with a PerkinElmer model 2400 analyzer. IR spectra were recorded as KBr pellets with a Nicolet 170SX spectrophotometer in the 4,000–400-cm⁻¹ region. ¹H NMR spectra were recorded with a Bruker AVANCE III (600- MHz) spectrometer. ¹³C NMR spectra were recorded with a Bruker AV-600 (600-MHz) spectrometer.

Cell culture and whole-cell extract preparation

PC-3 human prostate cancer and MDA-MB-231 human breast cancer cells were purchased from American Type Culture Collection (Manassas, VA, USA) and grown in RPMI 1640 or 1:1 Dulbecco’s modified Eagle’s medium / F-12 medium, respectively, supplemented with 10 % fetal bovine serum, 100 U mL⁻¹ penicillin, and 100 µg mL⁻¹ streptomycin (Life Technologies, Carlsbad, CA USA). MCF-10A cells (immortalized but nontumorigenic) were provided by Fred Miller (Karmanos Cancer Institute, Detroit, MI, USA) and were grown in 1:1 Dulbecco’s modified Eagle’s medium/F12 medium supplemented with 5% (v/v) horse serum, 0.029 mol L⁻¹ sodium bicarbonate, 10 mM N-(2-hydroxyethyl)piperazine-N′-ethanesulfonic acid buffer solution, 100 U mL⁻¹ penicillin, 5 mg insulin, 10 µg epidermal growth factor, and 250 µg hydrocortisone. Cells were cultured in an atmosphere containing 5 % CO₂ at 37°C. Whole-cell extracts were prepared as described previously [27]. Briefly, cells were harvested, washed with 1× phosphate-buffered saline, homogenized in a lysis buffer [50 mM tris(hydroxymethyl)aminomethane (Tris)– HCl, pH 8.0, 150 mM NaCl, 0.5 % NP40], vortexed for 30 min at 4°C, and centrifuged at 12,000g for 12 min. The supernatants were collected as whole-cell extracts and used for measurement of chymotrypsin-like activity and Western blot analysis.

Cell proliferation assays

The MTT assay was used to measure the effects of compounds 1 and 2 and their complexes 3–6 on prostate cancer cell proliferation. Cells were seeded in triplicate in 96-well plates and grown to 70–80 % confluency, followed by treatment with 1–6 at the concentrations indicated (1–4, 40 µM; 5 and 6, 5, 10, 20, and 40 µM). After 16 h incubation at 37°C, inhibition of cell proliferation was measured as previously described [28].

Purified 20S proteasome activity assay

Purified 20S human proteasome (35 ng) was incubated with different concentrations of 1–6 and 20 µM fluorogenic peptide substrate N-succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (Suc-LLVY-AMC) in 100 mL assay buffer (20 mM Tris–HCl, pH 7.5) for 2 h at 37°C. Then, hydrolysis of fluorogenic substrates was measured with a Wallac Victor 3 multilabel plate reader (PerkinElmer; Waltham, MA, USA) with an excitation filter of 365 nm and an emission filter of 460 nm.

Cellular morphological analysis

Cellular changes were observed using an Axiovert 25 microscope (Carl Zeiss Microscopy, Thornwood, FL USA). Rounded and detached cells were considered apoptotic cells.

Analysis of proteasomal chymotrypsin-like activity in cell extracts

Whole-cell extracts (10 µg) of human prostate cancer cells, treated as indicated, were incubated for 2 h at 37°C in 100 µL assay buffer (20 mM Tris–HCl, pH 7.5) with 20 µM fluorogenic substrate Suc-LLVY-AMC (Anaspec; Fremont, CA, USA), followed by measurement of production of hydrolyzed amino-4-methylcoumarin groups, as previously described [28].

Western blot analysis

Prostate cancer cells were treated as indicated in the figure legends, harvested, and lysed. Protein concentrations of whole-cell lysates were determined using the Bradford protein assay (Bio-Rad, Hercules, CA USA). Cell lysates (40 µg) were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. Western blot analysis was done using specific antibodies against IκB-α, PARP, and β-actin, followed by visualization with enhanced chemiluminescence reagent (Denville Scientific, Metuchen, NJ, USA) as previously described [29].

Syntheses of copper complexes

Complexes 3 and 4

Synthesis of these complexes was done by following a previously described procedure [30].

Complex 3: Yield: 83 %; Anal. calc. for 3 (%, [Cu₂(C₁₀ H₈O₂N)₄(H₂O)₂]·2H₂O, formula weight 895.86 g mol⁻¹); C, 53.63; H, 4.50; N, 6.25; O, 21.44; Cu, 14.18. Found (%): C, 53.97; H, 4.42; N, 6.18; O, 20.88; Cu, 14.55. UV: λ_max (nm): 213, 228, 280. IR (KBr, cm⁻¹): 3,404.75, ν(−NH−); 1,610.90, ν_as(COO⁻); 1,420.42, ν_s(COO⁻); 453.99, ν(M−O). ¹H NMR (DMSO, 600 MHz): δ (ppm) 11.362 (1H, s, −NH−); 7.503 [1H, d, C(7)H]; 7.337 [1H, d, C(10)H]; 7.225 [1H, d, C(3)H]; 7.079 [1H, t, C(8)H]; 6.985 [1H, t, C(9)H]; 3.631 [2H, s, C(2)H]. Thermogravimetric analysis: lost 4.57 % (calculated 4.02 %, 2H₂O) in the first step at 25–140 °C; residue 22.04 % (calculated 21.43 %, CuO). Molar conductivity Λ_m (S cm² mol⁻¹): 8.74.

Complex 4: Yield: 80 %; Anal. calc. for 4 (%, [Cu₂(C₁₁ H₁₀O₂N)₄(H₂O)₂]·2H₂O, formula weight 951.96 g mol⁻¹); C, 55.51; H, 5.08; N, 5.88; O, 20.19; Cu, 13.34. Found (%): C, 56.27; H, 4.80; N, 6.06; O, 19.26; Cu, 13.61. UV: λ_max (nm): 232, 285. IR (KBr, cm⁻¹): 3,408.16, ν(−NH−); 1,588.13, ν_as (COO⁻); 1,414.59, ν_s(COO⁻); 501.32, ν(M−O). ¹H NMR (DMSO, 600 MHz): δ (ppm) 10.189 (1H, s, −NH−); 7.858 [1H, d, C(8)H]; 7.627 [1H, d, C(11)H]; 7.321 [1H, d, C(4)H]; 7.139 [1H, t, C(9)H]; 6.885 [1H, t, C(10)H]; 3.797 [2H, s, C(2)H]; 3.338 [2H, t, C(3)H]. ¹³C NMR (DMSO): δ (ppm) 182.64 (−COO); 138.27, 125.51, 123.07, 120.39, 116.73, 115.24, 114.43, and 112.95 (indole ring carbons); 33.67 and 27.36 (−CH₂). Thermogravimetric analysis: lost 3.84 % (calculated 3.78 %, 2H₂O) in the first step at 25–125 °C; residue 21.82 % (calculated 20.12 %, CuO). Λ_m (S cm mol⁻¹): 6.59.

Compounds 5 and 6

Ligand 1 (0.175 g, 1 mM) or ligand 2 (0.189 g, 1 mM) was dissolved in 5 mL water to give a solution. Cu(CH₃-COO)₂·nH₂O (0.100 g, 0.5 mM), dissolved in 10 mL anhydrous ethanol, was added dropwise to the above-mentioned solution with stirring and was allowed to react for 3 h, followed by addition of C₁₂H₈N₂ (0.5 mM) dissolved in 5 mL anhydrous ethanol, and reaction for 3 h gave a green precipitate, which was filtered off to give the complexes.

Complex 5: Yield: 71 %; Anal. calc. for 5 (%, [Cu(C₁₀ H₈O₂N)₂(C₁₂H₈N₂)], formula weight 592.1 g mol⁻¹); C, 64.91; H, 4.09; N, 9.48; O, 10.82; Cu, 10.72. Found (%): C, 65.39; H, 4.32; N, 9.24; O, 10.53; Cu, 10.51. IR (KBr, cm⁻¹): 3,390.36, ν(−NH−); 1,548.05, ν_as(COO⁻); 1,414.59, ν_s(COO⁻); 498.15, ν(M−O); 533.30, ν(M−N). ¹H NMR (DMSO, 600 MHz): δ (ppm) 11.000 (s, −NH−, indole ring); 8.763 (d, H-2, 9, 1,10-phenanthroline); 7.230 (s, H-5, 6, 1,10-phenanthroline); 7.360, 7.596 (d, H-7, 10, indole ring); 6.962, 7.063 (t, H-8, 9, indole ring); 7.126 (d, H-3, indole ring); 3.465 (s, −CH₂−indole ring); ¹³C NMR (DMSO): δ (ppm) 181.12 (−COO); 150.49, 141.27, 130.56, 124.11 (1,10-phenanthroline carbons); 136.31, 126.54, 123.20, 120.25, 118.75, 118.17, 115.81, and 111.13 (indole ring carbons); 18.46 (−CH₂). Thermogravimetric analysis: residue 16.73 % (calculated 15.84 %, CuO). Λ_m (S cm² mol⁻¹): 10.67.

Complex 6: Yield: 67 %; Anal. calc. for 6 (%, [Cu(C₁₁H₁₀O₂N)₂(C₁₂H₈N₂)], formula weight 620.16 g mol⁻¹); C, 65.85; H, 4.55; N, 9.03; O, 10.33; Cu, 10.24. Found (%): C, 65.75; H, 4.50; N, 9.08; O, 10.40; Cu, 10.27. IR (KBr, cm⁻¹): 3,402.16, ν(−NH−); 1,589.32, ν_as(COO⁻); 1,414.56, ν_s(COO⁻); 500.32, ν(M−O); 550.21, ν(M−N). ¹H NMR (DMSO, 600 MHz): δ (ppm) 10.834 (s, −NH−, indole ring); 8.835 (d, H-2, 9, 1,10-phenanthroline); 7.309 (s, H-5, 6, 1,10-phenanthroline); 7.646, 7.478 (d, H-8, 11, indole ring); 6.981, 7.102 (t, H-9, 10, indole ring); 7.244 (d, H-3, indole ring); 4.461 (s, −CH₂−indole ring); 3.570 (t, −CH₂−indole ring). ¹³C NMR (DMSO): δ (ppm) 178.37 (−COO); 154.51, 143.43, 138.36, 130.50 (1,10-phenanthroline carbons); 136.35, 128.44, 124.66, 120.65, 118.19, 118.04, 114.36, and 111.27 (indole ring carbons); 27.51, 25.37 (−CH₂). Thermogravimetric analysis: residue 16.99 % (calculated 15.48 %, CuO). Λ_m (S cm² mol⁻¹): 7.92.

Results

Synthesis and characterization of copper complexes

We previously demonstrated that some copper complexes can selectively inhibit proliferation and induce apoptosis in cancer cells [31–34]. We found that the chemical structure plays an important role in the activity of these complexes. In the current study, we synthesized two types of copper complexes, dinuclear complexes (3 and 4) and ternary complexes (5 and 6), of which 4–6 were new and novel. All these complexes were characterized by elemental analysis, IR spectroscopy, ¹H NMR spectroscopy, ¹³C NMR spectroscopy, UV–vis spectroscopy, and thermogravimetric analysis. Because these compounds were not crytallized, X-ray diffraction was not performed. The carbon, hydrogen, and nitrogen contents in each complex were measured by an elemental analyzer. The experimental data were comparable to the calculated values, supporting the composition of the complexes. According to the IR spectra of the synthesized metal complexes (3–6), all of the strong −NH− absorption peaks appeared in the range from 3,390.36 to 3,412.61 cm⁻¹ without any obvious shift compared with the ligands (data not shown), suggesting that the nitrogen in the indole ring did not form a coordination bond with the metal. In comparison with the IR spectrum of the ligand, the metal complexes exhibited two new peaks at 1,536.68–1,588.13 and 1,414.59–1,459.09 cm⁻¹, respectively, which could be attributed to ν_as(COO⁻) and ν_s(COO⁻). The magnitude of ν_as(COO⁻) − ν_s(COO⁻) was less than 200 cm⁻¹ in the complexes, indicating that the oxygen in the – COO⁻ group was coordinated to the metal ion in a bidentate fashion [35]. A new peak appeared at 453.99 or 501.32 cm⁻¹ in the two complexes, and could be attributed to the vibration of Cu–O. Compared with the dinuclear complexes, a new peak at 533.30 or 550.21 cm⁻¹ appeared in the ternary complexes, suggesting the new chemical bond M–N was formed.

The UV–vis absorption spectra of 3 and 4 were recorded in the 200–500-nm range in DMSO. λ_max at 228 and 280 nm for 3 and λ_max at 232 and 285 nm for 4 are attributed to π–π* and n–π* transitions of ligands, respectively. The absorption bands are shifted, which is ascribed to the metal-to-ligand charge transfer transitions.

The assignments for the ¹H NMR and ¹³C NMR spectra are presented in “Materials and methods.” The ¹H NMR spectra of 3 and 4 showed there was coordination of the ligand with the Cu(II) ion; the signals showed small shifts due to the electronic redistribution of the ligands, which was attributed to interaction of the ligands with the Cu(II) ion. In addition, the hydrogen atom of −NH− still existed and the hydrogen atom of −COOH was displaced by the metal ion. The ¹H NMR spectra are further supported by ¹³C NMR spectra. Compared with free ligands 1 and 2, the ¹³C NMR signal values are downfield shifted upon coordination to the Cu(II) ion (¹H NMR and ¹³C NMR spectral data of ligands 1 and 2 not shown). Consequently, the new complexes were formed by coordination of the copper ion, which is also supported by the IR spectra. A comparison between the ¹H and ¹³C NMR spectra of 1–6 clearly indicated the presence of 1,10-phenanthroline.

In the range from 25 to 800 °C, complexes 3 and 4 decomposed mainly in two steps. In the first step, they lost 4.57 and 3.84 % (calculated 4.02 and 3.78 %) of their mass, respectively, which was attributed to the two molecules of crystal water. The residue rates of 3 and 4 were 22.04 and 20.92 %, which were consistent with the calculated values (21.43 and 20.12 %). Complexes 5 and 6 mainly underwent only one step of decomposition and the residue rates were 16.73 and 16.99 %, respectively, which coincided with the calculated values (15.84 and 15.48 %).

All the compounds are soluble in dimethylformamide and DMSO and stable in air. Moreover, the molar conductivities (Λ_m) of 3–6 in DMSO were 8.74, 6.95, 10.67, and 7.92 S cm² mol⁻¹, respectively, which are less than 35 S cm² mol⁻¹ [36]; hence, these complexes were considered to be nonelectrolytes and were quite stable in culture media.

Compounds 1–6 inhibit proliferation of PC-3 human prostate cancer cells

We first investigated the antiproliferative ability of 1–6 (Fig. 1). Highly metastatic PC-3 human prostate cancer cells were treated with 5, 10, 20, or 40 µM 5 and 6 and 40 µM 1–4 for 16 h, followed by analysis by MTT assay. Cells treated with DMSO were used as a control. We found that 5 and 6 had similar growth-inhibitory activity, resulting in 55 and 81 % inhibition at 5 µM (Fig. 2), 73 and 92 % inhibition at 10 µM, and more than 98 % inhibition at 20 and 40 µM, respectively. However, 3 and 4 induced less than 50 % inhibition (Fig. 2) and 1 and 2 had no effect at 40 µM after 16 h of treatment (Fig. 2).

Fig. 2.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay of PC-3 cells treated with 1–6. PC-3 cells were treated with each compound for 16 h at various concentrations as indicated. After 16 h, the medium was removed, and the cells were treated with MTT solution, as described in “Materials and methods.” Dimethyl sulfoxide (DMSO) was used as a control

Compounds 3–6, but not compounds 1 and 2, inhibit chymotrypsin-like activity of the purified 20S proteasome in vitro

To investigate whether these compounds can inhibit proteasome activity and have similar effects as on cell proliferation, we incubated 1–6 at various concentrations with a purified human 20S proteasome for 2 h, with DMSO treatment as a control. We found that 3–6 had similar potency toward proteasomal chymotrypsin-like activity, with IC₅₀ values of 3, 1.3, 4.2, and 3.3 µM, respectively (Fig. 3), whereas their ligands 1 and 2 had little effect (Fig. 3). Combined with the MTT data (Fig. 2), these results suggest that copper is the essential component in these complexes and that the proteasome-inhibitory abilities of these copper complexes in cancer cells are related to their chemical structures, as the presence of 1,10-phenanthroline in the complexes probably creates a more potent anticancer agent (see later).

Fig. 3.

Inhibition of 20S proteasome. Purified human proteasome (35 ng) was incubated with DMSO or various concentrations of 1–6 for 2 h, followed by proteasomal chymotrypsin-like activity assay. CT chymotrypsin

Dose-dependent tumor cellular proteasome inhibition and apoptosis induction by compounds 1, 3, and 5

To determine whether the chemical structure is crucial for compounds to inhibit tumor cellular proteasome activity and induce apoptosis, PC-3 cells were treated with various concentrations of compounds 1, 3, and 5 or DMSO as a control for 16 h and the cellular proteasomal chymotrypsin-like activity was measured. We found that 5 was the most potent, inhibiting more than 97 % activity (Fig. 4a) at 20 and 30 µM. Compound 3 had a weaker effect, inducing only 20 and 35 % inhibition at 20 and 30 µM, respectively (Fig. 4a), and 1 had no effect (Fig. 4a). We have reported that accumulation of target proteins is associated with inhibition of proteasome activity [37]. The proteasome target protein IκB-α was analyzed by Western blotting. As expected, we observed accumulation of IκB-α mainly in cells treated with 5, whereas little accumulation occurred in cells treated with 3 and 1 (Fig. 4b). These results suggest that 5 is a potent inhibitor of the cellular proteasome in PC-3 cells, whereas 3 and 1 are much weaker in their effect, with the order of inhibition being 5 > 3 > 1.

Fig. 4.

Dose–response experiment in PC-3 cells. PC-3 cells were treated with either DMSO or the indicated concentrations of 1, 3, and 5 for 16 h. This was followed by measuring inhibition of the proteasomal chymotrypsin-like activity using the fluorescent N-succinyl-Leu-Leu-Val-Tyr-60-amino-4-methylcoumarin (Suc-LLVY-AMC) (a), Western blot analysis using antibodies against IκB-α, poly(ADP-ribose) polymerase (PARP) and β-actin (as loading control) (b), and observation of morphological changes (c). CT chymotrypsin

It has been reported that inhibition of tumor cellular proteasome activity is associated with induction of apoptosis [25, 28]. To assess the ability of these compounds to induce apoptosis, we studied morphological changes and apoptosis-associated PARP cleavage in the same experiment. At 16 h, treatment with 5 at 10 µM resulted in generation of PARP cleavage fragment p85 (Fig. 4b). Both full-length p110 PARP and cleaved p85 PARP disappeared at 20 and 30 µM because of extensive protein degradation caused by apoptosis (Fig. 4b). Compound 3 weakly induced PARP cleavage at 20 and 30 µM (Fig. 4b), whereas 1 induced no cleavage. At the same time, apoptosis-associated cellular morphology changes (rounding and shrinkage) were also observed in cells treated with 5, at concentrations as low as 10 µM (Fig. 4c). These changes did not occur in cells treated with 1 and 3, even at the highest concentration. Our results suggest that apoptosis induced by compounds 1, 3, and 5 correlates well with inhibition of the proteasome, and the rank is 5 > 3 > 1.

Dose-dependent tumor cellular proteasome inhibition and apoptosis induction by compounds 2, 4, and 6

If the presence of 1,10-phenanthroline in 5 is critical for proteasome inhibition and apoptosis induction, we should observe similar effects in PC-3 cells treated with 6. To test this idea, we performed another set of experiments with compounds 2, 4, and 6, which have structures similar to those of 1, 3, and 5, respectively. We performed PC-3 cellular proteasomal chymotrypsin-like activity assay and Western blot analysis, and observed the morphological changes. The results indicated that 6 is the most potent, inducing 78, 90, and 97 % inhibition of chymotrypsin-like activity at 10, 20, and 30 µM, respectively (Fig. 5a). Compound 4 induced only approximately 20 % inhibition at 20 and 30 µM and 2 induced less than 10 % inhibition (Fig. 5a). Consistently, accumulation of IκB-α (Fig. 5b) was observed in cells treated with 6 and 4, but not in cells treated with 2 (Fig. 5b). Furthermore, we observed PARP cleavage (Fig. 5b) after treatment with 10 and 20 µM 6, in line with the morphological changes (rounded and shrunken) (Fig. 5c). PARP cleavage (Fig. 5b) and morphological changes (Fig. 5c) were inconspicuous in cells treated with 2 and 4 even at 30 µM. These data demonstrate that the rank of proteasome inhibition and apoptosis induction is 6 > 4 > 2. The results further support the conclusion that the presence of 1,10-phenanthroline in the copper complex is important for the proteasome-inhibitory and apoptosis-inducing activities of these compounds.

Fig. 5.

Dose–response experiment in PC-3 cells. PC-3 cells were treated with either DMSO or the indicated concentrations of 2, 4, and 6 for 16 h. This was followed by measuring inhibition of the proteasomal chymotrypsin-like activity using the fluorescent SucLLVY-AMC (a), Western blot analysis using antibodies against IκB-α, PARP, and β-actin (as a loading control) (b), and observation of morphological changes (c). CT chymotrypsin

Time-dependent tumor cellular proteasome inhibition and apoptosis induction by compounds 5 and 6

If inhibition of the proteasome is responsible for induction of apoptosis, inhibition of proteasome activity should occur prior to cell death [31, 38]. To test this hypothesis, we treated PC-3 cells with 15 µM 5 and 6 for 3, 9, or 24 h, followed by measurement of proteasome activity (Fig. 6). The results show that proteasome inhibition by 5 and 6 started as early as 3 h, with only 58 and 34 % proteasome activity remaining, respectively (Fig. 6a). The chymotrypsin-like activity was further decreased at later times (Fig. 6a). Consistently, the levels of proteasome target protein IκB-α increased gradually starting at 3 and 9 h treatment with 5 and 6, respectively (Fig. 6b). In the same kinetic experiment, however, cell death was not detected until 9 h (Fig. 6c). Also, PARP cleavage appeared mainly at 9 and 24 h treatment with 5 and 6, respectively (Fig. 6b). These data demonstrate that the apoptosis induced by 5 and 6 is a consequence of proteasome inhibition.

Fig. 6.

Kinetic effects of 5 and 6 on PC-3 cells. PC-3 cells were treated with 15 µM 5 for the times indicated. This was followed by chymotrypsin-like activity assay using extracts (a), Western blot analysis with antibodies against IκB-α, PARP, and β-actin (b), and observation of morphological changes (c). CT chymotrypsin

Compounds 5 and 6 exhibit more biological activity in MDA-MB-231 breast cancer cells than in nontumorigenic MCF-10A cells

A novel anticancer drug should selectively induce apoptosis in tumor cells but not in noncancer cells. To investigate whether 5 and 6 exhibit selectivity toward tumor cells, we treated MDA-MB-231 human breast cancer cells and immortalized human breast, nontumorigenic MCF-10A cells with 5 and 6 at 5 and 10 µM for 24 h, followed by MTT assay, measurement of proteasomal chymotrypsin-like activity, Western blotting, and observation of morphological changes. As shown in Fig. 7a, 5 and 6 at 10 µM exhibited dramatically different effects on cell proliferation, causing 80–90 % inhibition in MDA-MB-231 cells but little inhibition in MCF-10A cells. Also, when these compounds were used at 5 µM, no inhibition of cell proliferation was observed in MCF-10A cells, whereas nearly 50 % inhibition occurred in MDA-MB-231 cells (Fig. 7a). Therefore, MDA-MB-231 cells are much more sensitive to inhibition by 5 and 6 than nontumorigenic MCF-10A cells.

Fig. 7.

The different effects of 5 and 6 in normal breast cells and breast cancer cells. Normal, immortalized MCF-10A human breast cells and MDA-MB-231 human breast cancer cells were treated with 5 and 10 µM 5 and 6 for 24 h. DMSO was used as a control. a Inhibition of cell proliferation in MDA-MB-231 cells, but not in normal MCF-10A cells. b Chymotrypsin-like activity assay. c Western blot analysis with antibodies against IκB-α and β-actin. d Morphological changes. CT chymotrypsin

Additionally, to further determine the differential effects of 5 and 6 on MDA-MB-231 cells and nontumorigenic MCF-10A cells, we measured proteasomal chymotrypsinlike activity in these two cell lines after 24 h treatment with 5 and 6 at 5 and 10 µM. We found that 5 and 6 at 5 µM induced 51 and 35 % inhibition, respectively, in MDA-MB-231 cells, but only 14 and 8 % inhibition in MCF-10A cells (Fig. 7b). Even when used at 10 µM, the compounds were more potent in MDA-MB-231 cells, reaching 90 and 62 % inhibition of chymotrypsin-like activity (Fig. 7b), whereas in MCF-10A cells, the inhibition reached only 34 and 24 %. Consistent with these results, accumulation of IκB-α (Fig. 7c) was observed in MDA-MB-231 cells treated with 5 and 6 at 5 and 10 µM, whereas notably no accumulation was observed in MCF-10A cells treated with 5 µM 5 and 6 or 10 µM 6, and very little occurred in MCF-10A cells treated with 10 µM 5 (Fig. 7c). In the same experiment, morphological changes were observed to determine whether the compounds would induce less apoptosis in nontumorigenic MCF-10A cells compared with cancerous MDA-MB-231 cells. As in PC-3 prostate cancer cells, 5 and 6 at 5 and 10 µM caused MDA-MB-231 cells to round and shrink, whereas nearly no morphological changes were observed in MCF-10A cells (Fig. 7d). Taken together, these results demonstrate that MDA-MB-231 cells are more sensitive to the novel compounds 5 and 6 than nontumorigenic cells, suggesting tumor selectivity of these novel copper compounds.

Discussion

Decreasing the copper content in vivo to inhibit angiogenesis has long been considered a selective mechanism against cancer cells and tissues [14, 15, 39, 40]. Recently, many copper complexes have been used experimentally as prodrugs to inhibit the proteasome. We previously reported that disulfiram copper, pyrrolidine dithiocarbamate copper, clioquinol copper, l-glutamine Schiff base copper, and taurine Schiff base copper complexes could inhibit proteasome activity and cell proliferation, as well as induce apoptosis in breast cancer and prostate cancer cells [28, 31–34]. One overarching hypothesis in our study is that (1) a synthetic copper complex could deliver the copper to the proteasome, causing proteasome inhibition via direct interaction and tumor cell death, (2) a copper-binding ligand could reduce cellular copper levels, ultimately blocking the binding of copper to its cellular targets, (3) a copper-binding ligand could interact and complex with cellular copper and deliver it to the proteasome, leading to direct proteasome inhibition, and (4) the delivered Cu(II) could be reduced to Cu(I) [14, 15], followed by generation of reactive oxygen species (ROS) that triggers oxidation and deactivation of the cellular proteasome. The synthetic copper complexes should not be different from the mixture of ligands and copper as well as the proposed ligand– copper complexes formed in tumor cells. It should also be noted that proteasome inhibition is only one of several possible mechanisms of action for these compounds. It is not the only mechanism. Other potential mechanisms of action must be explored in the future.

Not all copper complexes are able to induce apoptosis, and only those complexes that can carry copper ions into tumor cells and can prevent copper from interacting with many nonspecific proteins could induce apoptosis. In this study, we synthesized and chose two kinds of copper complexes, dinuclear complexes (3 and 4) and ternary complexes (5 and 6). We first measured the effects of these complexes on cell proliferation by MTT assay and the ability of the complexes to inhibit the activity of purified 20S proteasome. We found that the ternary copper complexes, 5 and 6, with 1,10-phenanthroline as the third ligand possessed strong cell-proliferation-inhibitory abilities (Fig. 2), followed by dinuclear copper complexes 3 and 4 (Fig. 2), but ligands 1 and 2 had no inhibitory effect (Fig. 2). Inhibition of chymotrypsin-like activity of purified 20S proteasome by complexes 3–6 was extremely similar (Fig. 3). Again, 1 and 2 had no inhibitory effect (Fig. 3). These results suggest that copper is a necessary factor for the inhibition of cancer cell proliferation by these complexes and 1,10-phenanthroline may play a key role in carrying copper into cells.

To further study this hypothesis, we compared the proteasome-inhibitory and apoptosis-inducing abilities of these compounds in PC-3 prostate cancer cells. We found that treatment with 5 and 6 significantly reduced proteasomal chymotrypsin-like activity (Figs. 4a, 5a) and resulted in the accumulation of proteasome target protein IκB-α (Figs. 4b, 5b). Compounds 3 and 4 caused little inhibition of chymotrypsin-like activity (Figs. 4a, 5a), and only 4 resulted in some accumulation of IκB-α (Fig. 5b). Ligands 1 and 2 had no inhibitory effect (Figs. 4a, b, 5a, b). Compounds 5 and 6 also induced apoptosis-related cleavage of PARP (Figs. 4b, 5b) and resulted in cellular morphological changes (Figs. 4c, 5c). Compound 3, but not compound 4, induced little PARP cleavage (Fig. 4b). Generally, actin can be used as an internal control, but actin and many other housekeeping proteins are degraded during apoptosis, and cannot be used as an internal control in dead cells. When human cancer cells were treated with higher concentrations, many cells underwent apoptosis (Fig 4b, 20 and 30 µM 5; Fig. 5b, 30 µM 6), and many proteins including actin are degraded. The IκB-α expressed may fully prove that protein loading is correct. These results were highly consistent with those of the MTT assay. With respect to proteasome-inhibitory activity, 5 and 6 had similarly high activities toward not only purified 20S proteasome, but also tumor cellular proteasome. In contrast, 3 and 4 were potent against purified 20S proteasome, but were only weak inhibitors of cellular proteasome activity. However, it is important to note that the effects of these compounds on proteasome activity in purified 20S proteasome (Fig. 3), although similar, were not identical in cultured cells (Fig. 4a). In general, it is not uncommon to see increased potency under purified conditions, because there is no competition from other proteins for binding of the compounds. Thus, more compounds are needed to inhibit the cellular proteasome to the same degree as for the purified proteasome. Therefore, the critical factor is most likely the structures of the copper complexes allowing easy transport of the copper into cancer cells, with 1,10-phenanthroline being the best choice for a third ligand.

It is known that proteasome inhibition is an effective means of causing cancer cell death [41, 42]. Thus, we performed a kinetic experiment to test whether 5 and 6 induce cancer cell death by the same mechanism. It is clear that the proteasomal chymotrypsin-like activity was inhibited as early as 3 h by 5 and 6 (Fig. 6a), consistent with the accumulation of target protein IκB-α (for 5 this started at 3 h and for 6 it started at 9 h) (Fig. 6b). Consistent with proteasome-inhibition-induced apoptosis, PARP cleavage was not observed until after 9 and 24 h treatment with 5 and 6, respectively (Fig. 6b). The cellular morphological changes (detached and rounded) were visible starting at 9 h. Our results indicate that cancer cell death induced by 5 and 6 occurs after inhibition of proteasome activity.

One of the most important indicators for novel cancer drugs should be effects in cancer cells but not in noncancer cells [43]. We thus examined the possibility that 5 and 6 could selectively inhibit cell proliferation and proteasome activity and induce cell death in breast cancer cells, but not in noncancer cells. We found that 5 and 6 had little effect on nontumorigenic MCF-10A cells, in contrast to their effect on MDA-MB-231 breast cancer cells (Fig. 7). We observed much more proteasome inhibition in MDA-MB-231 cells than in MCF-10A cells (Fig. 7b), and this was associated with the accumulation of target protein IκB-α (Fig. 7c). The MDA-MB-231 cells detached and rounded when treated with 5 and 6, but nearly no effects on nontumorigenic MCF-10A cells were observed (Fig. 7d). These results suggest that 5 and 6 are toxic to cancer cells, but less toxic to noncancer cells, owing to their selective inhibition of proteasome activity.

Copper homeostasis in humans is tightly regulated. Copper uptake by cells occurs through copper transporter 1 in the cell membrane. Most of the cellular copper is found in the form of Cu(I) [44]. Copper is a redox-active metal, and the chemical processes that trigger its activity have usually been associated with the generation of ROS [45, 46]. It has been reported that oxidative damage plays a role in carcinogenesis and oxidative stress is an important mechanism of action for a number of anticancer drugs [47]. Cancer cells exist in an elevated oxidation state and, therefore, are more vulnerable to increased levels of ROS [48, 49]. It is known that 1,10-phenanthroline possesses good coplanarity. The complexes with 1,10-phenanthroline as the third ligand have a more planar structure than the other complexes.

We previously showed with the use of inductively coupled plasma optical emission spectroscopy that compounds with the ability to bind copper easily transport copper into breast cancer cells [50], and that when leukemia cells are pretreated with copper, posttreatment with a copper-binding compound causes inhibition of the cellular proteasome and induction of apoptosis in these cancer cells [33]. We have also shown that treatment of breast cancer cells with Au(III) compounds produced significant levels of ROS, associated with proteasome inhibition and apoptosis, and that inhibition of ROS induction by N-acetyl-l-cysteine results in inhibition of Au(III)-induced proteasome inhibition and cell death [51, 52]. Thus, on the basis of the results of the current study and those of previous studies performed in our laboratory, we hypothesize that these Cu(II) complexes could transport copper into the cell, wherein Cu(II) could be reduced to Cu(I) [14, 15], followed by subsequent ROS generation, ultimately resulting in inhibition of the cellular proteasome. However, there may be other possibilities in addition to the reduction of copper and its mechanism of action. Further studies are needed in order to understand the detailed molecular mechanisms.

In summary, these data support organic copper complexes binding with 1,10-phenanthroline as the third ligand as novel proteasome inhibitors and apoptosis inducers in human cancer cells. The possible mechanisms involved could be that ternary copper complexes with 1,10-phenanthroline as the third ligand are more easily able to transfer copper into the cells than dinuclear complexes, leading to direct proteasome interaction/inhibition and/or through oxidation of the proteasome by copper, which consequently causes deactivation of the proteasome. Therefore this species of copper complex has great potential to be developed into anticancer drugs. Further studies should also investigate the effects of 5 and 6 on inducing tumor death in animal models.

Abbreviations

DMSO

Dimethyl sulfoxide

IAA

Indole-3-acetic acid

IPA

Indole-3-propionic acid

MTT

3-(4,5-Dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide

PARP

Poly(ADP-ribose) polymerase

ROS

Reactive oxygen species

Suc-LLVY-AMC

N-Succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin