Diiron Hexacarbonyl Complex Induces Site-Specific Release of Carbon Monoxide in Cancer Cells Triggered by Endogenous Glutathione.

In this study, we have evaluated a water-soluble, nontarget reagent and a carrier-free diiron hexacarbonyl complex, [Fe2{μ-SCH2CH(OH)CH2(OH)}2(CO)6] (TG-FeCORM), that can induce the site-specific release of carbon monoxide (CO) in cancer cells triggered by endogenous glutathione (GSH). The releasing rate of CO was dependent on the amount of endogenous GSH. Being the amount of endogenous GSH higher in cancer cells than in normal cells, the CO-releasing rate resulted faster in cancer cells. Moreover, the anti-inflammatory properties related to the intracellular CO release of TG-FeCORM were also confirmed in the living HeLa cells.

Introduction

Carbon monoxide (CO) as one of the byproducts of heme catabolism through heme oxygenase-1 has been admitted as an endogenous vital messenger molecule in mammals.[1−3] CO plays many important roles in protecting tissue via its antiproliferative and anti-inflammatory effects.[4−6] Thus, CO appears to have great potentials in therapeutic applications. Unfortunately, CO is also a highly toxic gas because of its high binding capability with hemoglobin and myoglobin.[1,2] The precise control of CO location is one of the critical factors for useful therapeutic responses. CO-releasing molecules (CORMs) have been used to simulate the therapeutic effect of CO in a few biological conditions.[7] Organometallic carbonyl complexes are well-suited to be candidates as CORMs.[8−10] Among organometallic carbonyl complexes, CORM-2 and water-soluble CORM-3 are currently the most frequently used CORMs for investigating physiological functions of CO release both in vitro and in vivo.[8] CO release from CORMs in tissues for treatment requires safe conditions. Therefore, site-specific CORMs as therapeutic agents represent a crucial point to be evaluated. To date, organic photo-CORMs,[11] micelles-CORMs,[12] nanocarrier CORMs,[13] and bovine serum albumin[14] have been explored for the site-specific CO release. However, only several nontarget reagents and carrier-free CORMs for site-specific CO release have yet been reported.[10d,10p] In addition, many CORMs have a fast CO-releasing rate in physiological buffer, with a short half-life (t1/2).[8] Stable CORMs with longer t1/2 are required to improve the functions and bioavailability. Therefore, novel transition metal carbonyl complexes that are suitable for site-specific CO release with high biocompatibility, excellent bioavailability, and low cost need to be further investigated.

[Fe2{μ-SCH2CH(OH)CH2(OH)}2(CO)6] (TG-FeCORM) is a water-soluble organometallic diiron hexacarbonyl compound, which has been reported in our previous work.[15]TG-FeCORM releases CO through substitution reaction by thiol groups of cysteamine (CysA). Because iron is an essential element in human life, TG-FeCORM was confirmed to be biocompatible.[15] It is known that glutathione (GSH) has a thiol group and its concentration is different in normal cells (∼2 mM) and cancer cells (∼10 mM).[16−18] Herein, we reported that GSH could stimulate TG-FeCORM to release CO in a concentration-dependent way. Moreover, we observed for the first time that the CO-releasing rate of TG-FeCORM triggered by GSH was significantly different in normal and cancer cells. As a transition metal carbonyl complex, TG-FeCORM is able to achieve site-specific release of CO in cancer cells (Scheme ).

Scheme 1

Schematic Illustration of TG-FeCORM Synthesis and Its Proposal Site-Specific CO Release in Cancer Cells Triggered by Endogenous GSH

Results and Discussion

The kinetic analysis of CO release from TG-FeCORM was monitored by UV–vis spectroscopy using the myoglobin assay. The high-intensity absorbance Soret band, nearly 420 nm, was also observed here.[19] Linear variations were not apparent for the analysis of CO release. Therefore, the release of CO from TG-FeCORM triggered by GSH was investigated by monitoring the absorption wavelength between 500 and 600 nm, and two new absorption peaks at 540 and 577 nm were observed that belong to the Q band region of MbCO (Figure S1).[10e,10n,20] Without GSH, TG-FeCORM completely released CO in 760 min, and the t1/2 was 480 min (Figure a, Table S1, Supporting Information). The result was caused by a reducing agent, sodium dithionite, which was used to reduce deoxymyoglobin.[10e,10n] In our previous study, TG-FeCORM did not decompose until 150 min in the physiological saline detected by infrared spectroscopy.[15b] Therefore, TG-FeCORM in physiological conditions is very stable. Upon the addition of 2 mM GSH, CO was released completely in 255 min with a t1/2 of 125 min (Figure b, Table S1, Supporting Information). With 10 mM GSH, CO was completely released in 105 min with a t1/2 of about 53 min (Figure c, Table S1, Supporting Information). Overall, the CO-releasing rate depended on the amount of GSH (Figure d). The intracellular endogenous GSH concentration is ∼2 mM in normal cells.[17] Whereas the concentration of endogenous GSH in cancer cells is 4 times that of normal cells.[18] As indicated by our results, it is inferred that differences in GSH concentrations could lead to significant differences of CO release from TG-FeCORM.

Figure 1

Time-dependent UV–vis absorption spectra of TG-FeCORM in the absence of GSH (a), and in the presence of 2 mM GSH (b), and 10 mM GSH (c). CO release by TG-FeCORM (20 μM) triggered by different concentrations of GSH vs time in reduced deoxymyoglobin (d). All of the myoglobin assays were carried out at 37 °C in phosphate buffer solution.

TG-FeCORM had excellent cell compatibility when its concentration was less than 80 μmol/L in vitro (Figures S2–S7, Supporting Information). The intracellular concentrations of GSH in normal cells (HL-7702) and cancer cells (BEL-7402 and HeLa) were detected according to a commercial GSH assay kit.[21] Intracellular GSH concentrations of HL-7702, BEL-7402, and HeLa cells were determined to be approximately 3 ± 1, 14 ± 2, and 12 ± 1 mM, respectively. The measured concentrations of endogenous GSH in cells are consistent with the data detected by the Yang’s group.[21] The results confirmed that there are about fourfold or fivefold higher concentrations of endogenous GSH in BEL-7402 and HeLa cells over HL-7702 cells. The concentration difference of endogenous GSH between normal cells and cancer cells is in accordance with the results of the Kuppusamy’s group.[18] CO release from TG-FeCORM triggered by endogenous GSH was further evaluated in vitro. Intracellular CO-releasing profiles were measured using COP-1, a CO-responsive turn-on fluorescent probe, with a limit of detection of 1 μM (∼28 ppb).[22] In contrast to the negative control (Figure , top row), clear green fluorescent signals were observed when HeLa cells were treated with the positive control CORM-3 (Figure , middle row), as well as TG-FeCORM (Figure , bottom row), confirming that both CORM-3 and TG-FeCORM could release CO in the living HeLa cells. The time-dependent mean value of total fluorescence intensity of living HeLa cells was integrated at the same condition to test the CO-releasing process (Figure S8, Supporting Information). CO release from CORM-3 or TG-FeCORM was fast in the first 5 min. The nonsignificant change of fluorescence intensity indicates that the CO release from CORM-3 did not change during 10–60 min (Figures and S8, Supporting Information). By contrast, the evident increase of fluorescence intensity suggests that CO release from TG-FeCORM is continuous within 60 min (Figures and S8, Supporting Information). Thus, the releasing time of TG-FeCORM is more durable than CORM-3 in living HeLa cells. With 10 mM GSH, CO was completely released in 105 min, according to the myoglobin assay. The detected endogenous GSH concentration in HeLa cells is about 12 mM. Consequently, CO may be completely released from TG-FeCORM in about 100 min combined with the above results.

Figure 2

Time-dependent confocal microscopy images of CO site-specific release by TG-FeCORM in living HeLa cells. HeLa cells were preincubated with nothing (top row), 20 μM CORM-3 (middle row), and 20 μM TG-FeCORM (bottom row) for 30 min. All samples were incubated with 1 μM COP-1 for 30 min. Images were taken every 10 min. Scale bars: 20 μm.

The performances of CO release from TG-FeCORM were further investigated with fluorescence imaging experiments, carried out in vitro using normal cells (HL-7702) and cancer cells (BEL-7402 and HeLa). As shown in Figure a, green fluorescence was very weak in HL-7702 treated with COP-1 to detect CO release, indicating that the amount of CO release by TG-FeCORM triggered by endogenous GSH was very low. In contrast, the green fluorescence was remarkably enhanced in BEL-7402 and HeLa cells, indicating that released CO in cancer cells was clearly higher compared to that in normal cells. These results demonstrated that the differences in the concentration of endogenous GSH could lead to a site-specific CO release from TG-FeCORM in cancer cells. We also evaluated the effect of CORM-3 in vitro. The fluorescence signals were observed in HL-7702, BEL-7402, and HeLa cells (Figure b), showing that CORM-3 could not achieve site-specific CO release in cancer cells. In contrast, TG-FeCORM could lead to the site-specific CO release triggered by endogenous GSH in cancer cells. For HL-7702 cells, fluorescence induced by TG-FeCORM is weak, as shown in Figure a, whereas that induced by CORM-3 is very strong, as shown in Figure b. Thus, compared to CORM-3, TG-FeCORM is relatively stable in normal cells.

Figure 3

Confocal microscopy images for cellular CO release from TG-FeCORM (a), CORM-3 (b) in living HL-7702, BEL-7402, and HeLa cells, respectively. Images were taken after the treatment with 20 μM TG-FeCORM (a), 20 μM CORM-3 (b), and Hoechst 33342 for 30 min, followed by 1 μM COP-1 for 30 min. In each panel, left pictures show the nucleus stained using Hoechst 33342 (blue), middle pictures show COP-1 detected by CO with turn-on signal (green), and right pictures show merged images. Scale bars: 20 μm.

To confirm the anti-inflammatory effect of CO release from TG-FeCORM, the expression levels of inflammatory cytokines were detected with the enzyme-linked immunosorbent assay (ELISA). The proinflammatory cytokines IL-6, IL-1α, an anti-inflammatory cytokine IL-10, and a tumor necrosis factor TNF-α in HeLa cells were stimulated by lipopolysaccharides (LPS) in vitro.[9,13b,23] LPS had no distinct cytotoxicity toward HeLa cells in the range of 0–100 μmol/L (Figure S9, Supporting Information). TG-FeCORM had excellent cell compatibility to HeLa cells when its concentration is in the range of 0–40 μmol/L (Figure S10, Supporting Information). As shown in Figure , HeLa cells displayed a positive inflammatory response with the LPS stimulated, as the concentrations of IL-6, IL-1α, IL-10, and TNF-α increased remarkably (red bars). However, the expression levels of IL-6, IL-1α, IL-10, and TNF-α were continuously decreased with the concentration of TG-FeCORM rising from 20 to 40 μmol/L. The result showed that the treatment of HeLa cells with TG-FeCORM significantly decreased the expression levels of IL-6, IL-1α, IL-10, and TNF-α within the LPS-stimulated HeLa cells in a concentration-dependent manner. This downregulation of inflammatory cytokines indicated that TG-FeCORM exhibited the anti-inflammatory characteristics related to the intracellular CO release in cancer cells, and thus the anti-inflammation effect of TG-FeCORM should be attributed to the released CO in cells. In addition, the treatment with TG-FeCORM reduced the nuclear shift of p50 and p65 in LPS-stimulated HeLa cells (Figure S11, Supporting Information). By contrast, the total levels of ß-actin and proliferating cell nuclear antigen were unaffected after being treated with TG-FeCORM (Figure S11, Supporting Information). It was speculated that CO release from TG-FeCORM effect on inflammatory cytokines in LPS-stimulated HeLa cells might be associated with the NF-ΚB signaling pathway.[9,23] These results are consistent with the CO release from TG-FeCORM, triggered by endogenous GSH in cancer cells.

Figure 4

Effect of TG-FeCORM on the expression of TNF-α, IL-1α, IL-6, and IL-10 in LPS-stimulated HeLa cells. LPS (10 μg/mL) was employed to stimulate to produce inflammatory cytokines. Untreated cells were employed as controls. The incubation time of TG-FeCORM was 24 h. Differences were defined as significant at *p < 0.05.

Conclusions

In conclusion, our experimental results demonstrated for the first time that the water-soluble TG-FeCORM could act as an endogenous GSH-responsive CORM. We observed that CO release from TG-FeCORM was clearly dependent on the concentration of GSH, both in tubes and in cells. As indicated by our results, significant differences of endogenous GSH between normal cells and cancer cells could lead to the site-specific CO release from TG-FeCORM in cancer cells. Moreover, the anti-inflammatory properties and the longer releasing time contributed to the potential therapeutic role of TG-FeCORM for medical applications.

Experimental Section

Materials and Apparatus

Myoglobin and 1-thioglycerol were bought from Aladdin and employed as received. HeLa, HL-7702, and BEL-7402 cell lines were bought from the Chinese Academy of Sciences.

Images were obtained by a Zeiss LSM 710 confocal laser point-scanning microscope. The ELISA test was measured using a finite M200 microplate absorbance reader, Tecan. Myoglobin kinetics assay was measured with a UV spectrophotometer (UV-2600). All reagents were purchased from commercial sources and utilized without further disposal. All solvents were freshly distilled prior to use.

Myoglobin Kinetics Assay[26,27]

For the myoglobin assay, all solutions were prepared in the phosphate buffer at pH = 7.4. A myoglobin solution (2 mg/mL) was degassed through pouring into nitrogen for more than 15 min. Then, a freshly prepared solution of sodium dithionite (24 mg/mL) with 1:10 dithionite/dexy-Mb (v/v) added to the above degassed solution, which gave a 108 μM/mL dexy-Mb. An appropriate amount of TG-FeCORM and GSH was added to the dexy-Mb solution (Table S1). The solution was moved quickly into a room. UV–vis spectra were taken at 37 °C at predetermined time points by a UV–vis spectrophotometer and measured with a wavelength from 600 to 500 nm by an interval of 2 nm. Quantification of CO release was calculated from the obtained spectra according to the equation below (eqs –3).

Equation was used for counting the total myoglobin quantity of saturated Mb-CO solution. ε represents the extinction coefficient while Mb-CO is 15.4 mM–1 cm–1 and OD540 is the absorbance of Mb-CO solution while the wavelength is 540 nm.

Intermediate quantities of Mb-CO are counted by the OD540. A novel extinction coefficient (ε2) must be calculated to consider the altered absorbance at 540 nm (ΔOD540). To improve the calculation accuracy, another wavelength serves as a constant reference point. There are four isosbestic (ODiso) points (510, 550, 570, and 585 nm) in Mb-CO and deoxy-Mb spectra. The data at 510 nm (ODiso510) were utilized in this set of experiments. ε2 was calculated by eq .

Equation was used to count the unknown Mb-CO extinction coefficient. ΔODiso510 is the altered absorbance at the isosbestic point; ΔOD540 is the altered absorbance at 540 nm; and Mb-COmax is the maximum concentration of myoglobin.

From ε2 and the altered absorbance of 510 and 540 nm, the unknown myoglobin concentration will be obtained as given in eq , which was used to count the Mb-CO concentration.

t1/2 values are given as the time when the concentration of Mb-CO is equal to half the start concentration of TG-FeCORM.

Assessing the Cytotoxicity of TG-FeCORM Using MTT Assay

HeLa cells (180 μL, 1 × 105 cells mL–1) were seeded into 96 well microtitre plates and treated with 24 h. Then, the medium in wells was abandoned and added a medium including a mixture of TG-FeCORM. In addition, the concentrations of TG-FeCORM were 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100 μmol·L–1. The cells were then treated with another 24 h before the treatment medium took the place of the complete medium. Cells in a well without the addition of TG-FeCORM were used as a control (100% in cell viability). Each concentration was detected in five wells of the same plate. It was repeated three times to test the reproducibility of the assessment.

Probe COP-1 Fluorescence Response by Confocal Microscopy Imaging

Probe COP-1 was synthesized according to the literature.[9,22] Images were obtained by a Zeiss LSM 710 confocal laser point-scanning microscope with a 40× oil objective lens, and a numerical aperture of 1.3. COP-1 was excited by a 488 nm laser, and Hoescht 33342 was excited using a diode laser 405 nm, and they were read at green (λem = 500–550 nm) and blue (λem = 420–470 nm), respectively. Cells were imaged at 37 °C and 5% CO2 throughout the course of the experiment. HeLa cells (5 × 105 cells mL–1) were seeded in laser confocal culture dish 2 days before the experiment. Culture conditions were the same, as used for the routine cell passage using RPMI-1640 medium. HeLa cells were preincubated with nothing (top row), 20 μM CORM-3 (middle row), and 20 μM TG-FeCORM (bottom row) for 30 min, respectively, and then all of them were incubated with 1 μM COP-1 for 30 min. Images were taken every 2 min using representative images of the experiment. Mean of total fluorescence intensity of treated TG-FeCORM or CORM-3 versus untreated cells was compared using representative images of the independent experiment. Data are presented in the graphs (Figure S7) as a mean of total fluorescent intensity.

Measurement of GSH Concentration by Commercial Assay Kit

GSH concentration of HL-7702, BEL-7402, and HeLa cells were measured by an 5,5-dithio-bis(2-nitrobenzoic) acid method used in the commercial GSH assay kit (Beyotime Institute of Biotechnology, Jiangsu, China). The absorbance at 412 nm was measured using a microplate reader.[22]

Cytokine Modulation by CO Release[9]

Growth medium levels of the tested chemokines were quantified using Human IL-1α, IL-6, IL-10, and TNF-α Mini ELISA Development kits (PeproTech, sensitivity range of 0.063 ng/mL to 4 μg/mL) and revealed using 3,3′,5,5′-tetramethylbenzidine substrate reagent set (BD Biosciences) on the basis of the manufacturer’s protocol. The absorbance in each well was detected at 450 nm with a microplate reader (Infinite M200 microplate absorbance reader, Tecan). Cells were plated with the 2.5 × 105 cell/well in 6 well plates. Five groups were tested: cells were incubated with LPS or without LPS and then treated with TG-FeCORM (20–40 μM), 48 h after seeding. Supernatants were collected at 24 h post-treatment.

Western Blot Analysis[28]

Total cell lysates were collected by extracting with the NEPER kit. Concentrations of collected proteins were detected through a protein assay kit. Equal amounts of cellular total proteins were separated on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then moved to polyvinylidenedifluoride membranes and sealed with 5% nonfat milk in TBST buffer (150 mM NaCl, 0.05% Tween 20, 20 mM Tris, pH = 8.0) for 1 h. Then, the membranes were treated with the primary antibodies (p50 and p65) at 4 °C for 24 h. After a subsequent washing step, the membrane was treated with the appropriate secondary antibodies conjugated with horseradish peroxidase for 2 h at 25 °C and washed 3 times with TBST. The immune reactivity was determined by Amersham ECL Plus western blotting detection reagents.