Thiazolyl blue Tetrazolium bromide (MTT), verapamil, fetal bovine serum (FBS) and 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl-imidacarbocyanine iodide (JC-1) were purchased from Sigma-Aldrich (St. Louis, MO, USA). All cell culture flasks and dishes were obtained from Corning (Corning, NY, USA). Nigericin, salinomycin, valinomycin, cyclosporine A (CsA), bongkrekic acid (BKA), N-ethylmaleimide (NEM), 1,4-dithioerythritol (DTE) and spermine were purchased from Sigma-Aldrich. Adenosine diphosphate (ADP) was purchased from Boehringer (Mannheim, Germany). AGE provided by Wakunaga Pharmaceutical Co. Ltd. (Hiroshima, Japan) was manufactured as follows: Garlic cloves were sliced, immersed in a water-ethanol mixture solution and naturally extracted for >10 months at room temperature, as previously described (33). The AGE powder used in our experiments was prepared by lyophilization. It contained approximately 28.6% (w/v, 286 mg/ml) solid material, 0.63% (6.3 mg/ml) arginine and 0.1% SAC (calculated on a dry weight basis) as a marker compound for standardization (34). The AGE powder was freshly dissolved in Ham's F-12 or in RPMI-1640 medium prior to each experiment.

A human colon adenocarcinoma cell line (LoVo WT) isolated from a metastatic nodule, its MDR variant (LoVo DX), a gastric adenocarcinoma cell line (AGS), a melanoma cell line (M14 WT), isolated from an epidermal melanoma, the corresponding MDR variant M14 ADR2, a human cervical adenocarcinoma cell line (HeLa) and a human ovarian carcinoma cell line (A2780) were used in this study. The LoVo, M14 and HeLa cells were a kind gift from Professor E. Dolfini (University of Milan, Milan, Italy), Dr A. Molinari (National Institute of Health, Rome, Italy) and Professor M.T. Conconi (University of Padova, Padova, Italy), respectively. The A2780 cells were purchased from ECACC (Sigma-Aldrich) cat. no. 93112519. The MDR cell line, LoVo DX, was obtained by the prolonged culture of drug-sensitive parental LoVo WT cells in medium containing doxorubicin (Adriblastina, Pharmacia & Upjohn, Milan, Italy) as previously described by Grandi et al (35). The M14 ADR2 cell line was obtained by us (36) by culturing the M14 ADR or M14 DX cell line, previously selected for resistance to adriamycin by Molinari et al (37), in medium containing 10 µM doxorubicin constantly in each passage. Both resistant cell lines are also resistant to other chemotherapeutic agents, such as etoposide and vincristine (38,39). The AGS cell line (homo sapiens gastric adenocarcinoma) was obtained from the American Type Culture Collection (ATCC^® CRL-1739™; ATCC, Manassas, VA, USA). The LoVo and AGS cell lines were grown in Ham's F-12 medium containing glutamine supplemented with 10% FBS, 1% MEM vitamins, 1% MEM non-essential amino acid, penicillin (100 U/ml) and streptomycin (100 µg/ml). The M14 cells were grown in RPMI-1640 with glutamine, 10% FBS, 1% MEM non-essential amino acid, penicillin (100 U/ml) and streptomycin (100 µg/ml). The HeLa and A2780 cells were grown in Nutrient Mixture F-12 (Ham's) and RPMI-1640, respectively. In total, 1.5 g/l NaHCO₃ (Sigma-Aldrich), 10% heat-inactivated fetal calf serum (Biowest, Nuaillé, France), 100 U/ml penicillin, 100 µg/ml streptomycin and 0.25 µg/ml amphotericin B (all from Sigma-Aldrich) were added to both media.

All cell lines were incubated in a humidified atmosphere of 5% CO₂ in a water-jacketed incubator at 37°C. For each passage, exponentially growing M14 cells were harvested with 10 mM EDTA. LoVo and AGS cells were harvested by further addition of 0.25% trypsin solution. The trypsin activity was quenched by the addition of complete F-12 medium.

The LoVo WT, LoVo DX, M14 WT, M14 ADR2 and AGS cells were seeded in a 96-well plate and incubated for 24 h to allow for the complete reattachment of the cells to the plates. After changing with fresh medium, the cells were incubated in the presence of 0, 1, 5, 7 and 10 mg/ml AGE for 24, 48 and 72 h. The anti-proliferative effects of AGE on human tumor cells were examined by MTT assay. Briefly, MTT was added to each well. Following 3 h of incubation at 37°C, dimethyl sulfoxide was added to dissolve the crystals. The absorbance was determined at 577 and 660 nm using a spectrophotometer multi-mode plate reader [Synergy HT BioTek, serial no. 270204; BioTek, Bernareggio (MB), Italy].

The M14 WT cells were incubated in the presence of 0, 1, 3, 6 and 10 mg/ml AGE at 37 and 42°C for 1 h. After washing the cells with PBS with 1% bovine serum albumin (BSA) twice, the cells were seeded into 96-well plate and incubated in RPMI-1640 complete medium containing 0, 1, 3, 6 and 10 mg/ml of AGE at 37°C for 48 h followed by MTT assay as described above.

The changes in Δψm in whole cells were assayed using the lipophilic cationic probe, JC-1 dye. The LoVo WT, LoVo DX and AGS cells were seeded in a 12-well plate and incubated for 24 h to allow the cells to adhere to the plates. After changing with fresh medium, the cells were incubated with 0, 1, 5 and 10 mg/ml of AGE for 24 h at 37°C. Subsequently, the cells were stained with 2.5 µg/ml of JC-1 for 20 min at 37°C. In the LoVo cells, 100 µM verapamil was further added to the solution in order to inhibit the P-glycoprotein-mediated efflux of JC-1 in the LoVo DX cells. Exposure to verapamil significantly increased the fluorescence intensity of JC-1 in the LoVo DX cells, while it did not affect fluorescence in LoVo WT cells. The detached cells were washed with PBS and then resuspended in PBS. The samples were then analyzed using a BD Accuri C6 flow cytometer (BD Biosciences, San Jose, CA, USA). JC-1 was excited using an argon laser at a wavelength of 488 nm (using a BD Accuri C6 flow cytometer). The emitted green (JC-1 monomer) and red (JC-1 aggregate) fluorescence were detected at the FL-1 channel (533/30 nm) and FL-2 channel (585/40 nm), respectively. At least 10,000 events/sample were acquired in log mode. The ratio of red (FL2)/green (FL1) fluorescence intensity was used to represent the Δψm.

The Δψm of the HeLa and A2780 cells was evaluated using the BD™ MitoScreen kit (BD Pharmigen, San Diego, CA, USA) containing JC-1. The cells (1.5×10⁵) were seeded in a 24-well cell culture plate. following incubation for 24 h, AGE was added to the complete medium at 5 and 10 mg/ml, and the cells were incubated for a further 72 h. Following treatment, the cells were centrifuged at 1,000 × g at room temperature, resuspended in JC-1 working solution and incubated for 30 min at 37°C in a CO₂ incubator. Following incubation, the cells were washed twice, resuspended in assay buffer and analyzed using a FACSCanto II flow cytometer (Becton-Dickinson, Mountain View, CA, USA). The results are presented as dot plots and as the percentages of cells with an energized or depolarized mitochondrial membrane.

A total of 50 male Wistar rats, 2 months old, weighing approximately 150 g, were used in our experiments. The rats, housed in the animal facility of the Department of Biomedical Sciences, University of Padova (Padova, Italy), were maintained under controlled conditions (temperature 20-22°C, relative humidity 48-50%, water with antibacterial control and a 12:12 h light/dark cycle) and provided with water and a standard diet (4RF25) purchased by Mucedola s.r.l., Settimo Milanese (MI), Italy.

The experimental procedures were approved by the local Ethics Committee for Animal Experimentation (CEASA) (protocol no. 3619, 15.1.2014) and performed in agreement with the international guidelines as well as European Communities Council Directive and National Regulations (CEE Council 86/609 and DL 116/92).

RLM were isolated using the following method: The rats were starved overnight and sacrificed by cervical dislocation. The livers were rapidly explanted, immersed in ice-cold isolation medium containing 250 mM sucrose, 5 mM HEPES (pH 7.4), 0.5 mM EGTA and washed 4/5 times with the same medium. The livers were minced into small sections and washed with ice-cold fresh medium without EGTA. The suspension was transferred to a glass potter and homogenized using a Teflon pestle operating at 1,600 rpm, by 3-4 time strokes.

The homogenate was centrifuged at 700 × g for 5 min at 4°C and the obtained supernatant was centrifuged at 10,800 × g for 10 min at 4°C. The pellet was washed with isolation medium, resuspended and centrifuged at 15,900 × g for 10 min at 4°C. Finally, the obtained pellet, containing the mitochondria, was suspended in the standard medium (see the incubation procedure) (40). Mitochondrial proteins were measured by the biuret method with BSA, as a standard (41).

The mitochondria (1 mg protein/ml) were incubated in a water-jacketed cell at 25°C under continuous stirring. The standard medium contained 250 mM sucrose, 10 mM HEPES (pH 7.4), 5 mM Na-succinate, 1.25 µM rotenone and 1 mM Na-phosphate. Variations and/or other additions are provided with each experiment.

Δψm was calculated on the basis of the distribution of the lipid-soluble cation tetraphenylphosphonium (TPP⁺) (Sigma-Aldrich) measured across the inner membrane using a TPP⁺-selective electrode (42). Mitochondrial swelling was determined by measuring the apparent absorbance change of mitochondrial suspensions at 540 nm on a Kontron Uvikon model 922 spectrophotometer equipped with thermostatic control. The protein sulfydryl oxidation assay was performed as previously described by Santos et al (43). K⁺ was estimated in the supernatant by atomic spectroscopy as previously described by Crompton and Costi (44).

Nigericin, salinomycin, valinomycin are antibiotics that behave as ionophores for K⁺ in the mitochondria. They were all used at a 1 µM concentration, for approximately 10 min, to compare their effects with that of AGE. Mg²⁺ was used at a 1 mM concentration for approximately 10 min of incubation, as an inhibitor of K⁺/H⁺ exchanger. CsA is an immunosuppressant, while BKA is an inhibitor of AdNT, and both were used at a 1 µM concentration. ADP is a nucleotide, a substrate of AdNT, and was used at a 1 mM concentration. All these compounds were used, for approximately 30 min of incubation, as inhibitors of MPT. NEM, used at 10 mM concentration, is a thiol reagent, DTE, used at 1 mM concentration, is a reducing agent, while spermine, at 50 µM, is a ROS scavenger. These compounds were incubated for approximately 40 min, and used as MPT inhibitors.

The data were presented as the means ± SEM or means ± SD. Statistical analysis was performed using one-way ANOVA with the Tukey-Kramer post hoc test was used. A P-value <0.05 was considered to indicate a statistically significant difference.

The results of MTT assays revealed that treatment with AGE exerted anti-proliferative effects on all cell lines examined in a dose-dependent manner. Treatment with 10 mg/ml of AGE for 72 h decreased the viability of both the LoVo WT and LoVo DX cells to 73 and 65%, respectively, when compared with the untreated control cells (Fig. 1A and B). The proliferation of the M14 WT and M14 ADR2 cells was markedly decreased to approximately 43% of the control level following treatment with 10 mg/ml AGE for only 24 h (Fig. 1C and D). A further extension of the incubation time did not affect the viability of the cells. Long-term treatment with >5 mg/ml of AGE induced a marked inhibitory effect on growth of the AGS cells (Fig. 1E). An amount of AGE <5 mg/ml exerted a modest effect on all cell lines.

The sensitivity of numerous cancer cell lines to an increase in the temperature >37°C is the basis of clinical hyperthermia. The potential of AGE to enhance the cytotoxic effects of hyperthermia was investigated in the M14 cells as shown in Fig. 2. Incubation at 42°C for 1 h suppressed the viability of the M14 WT cells by approximately 29% (data not shown). The M14 WT cells treated with 10 mg/ml of AGE at 42°C for 1 h exhibited a decrease in viability to 25% when compared to the untreated cells at the same temperature, while 10 mg/ml AGE at 37°C decreased cell viability only to 49% of the untreated cells.

To obtain information on mitochondrial functionality, a flow cytometric analysis of the control and treated cells loaded with the mitochondrial probe, JC-1, was performed on several cancer cell lines. Flow cytometric analysis revealed a tendency that AGE exerted toxic effects on the mitochondria in a dose-response manner in the LoVo WT, LoVo DX and AGS cells. Treatment with AGE induced an evident mitochondrial membrane depolarization that was higher in the AGS cells than in the LoVo WT and LoVo DX cells, as indicated by the ratio of red (FL2 emission)/green (FL1 signal) fluorescence intensity (Fig. 3).

A similar effect was also obtained when incubating the ovarian carcinoma A2780 (Fig. 4A) and cervical adenocarcinoma HeLa (Fig. 4B) cells in the presence of various concentrations of AGE (5 and 10 mg/ml) for 72 h. The decrease in red fluorescence in the dot plots of the treated cells indicated mitochondrial membrane depolarization. Nevertheless, the extent of the effect appeared to differ markedly between the two cell lines taken into consideration. In particular, AGE induced a detectable effect from the concentration of 5 mg/ml on the A2780 cells, with a percentage of cells exhibiting depolarized mitochondria that increased from 0.6% (untreated cells) to 4.0% (Fig. 4A). The ability of AGE to induce mitochondrial membrane depolarization was clearly confirmed at the maximum concentration considered, and indeed, at 10 mg/ml, the above-mentioned percentage increased to 85.4%. In the HeLa cells, the ability of AGE to provoke mitochondrial depolarization was markedly less pronounced. Nevertheless, a dose-dependent effect was also observed in this cell line, with the percentage of cells exhibiting mitochondrial membrane depolarization increasing from 6.4% (untreated) to 8.6 and 9.1% in the presence of AGE at 5 and 10 mg/ml, respectively.

RLM incubated in standard medium at the concentration of 1 mg protein/ml, as described in the Materials and methods, and energized by the oxidation of succinate in the presence of rotenone, exhibited a Δψm value of approximately 180 mV (Fig. 5, curve a). The addition of AGE at 0.5 mg/ml induced a hyperpolarization of approximately 15 mV that was maintained for several minutes (Fig. 5, curve b). This effect was similar to that observed with the ionophore, nigericin (45) (Fig. 5, curve c) and with the anticancer drug salinomycin (Fig. 5, curve d) (46). The presence of AGE also induced the release of K⁺ into the incubation medium (Fig. 5, insert). Higher AGE concentrations did not significantly increase Δψm (Fig. 8, below). The addition of another ionophore, valinomycin, after AGE, decreased the Δψm to the same value obtained prior to the addition of AGE (Fig. 6A, curve a). Indeed, the addition of K⁺ induced a further decrease in Δψm to a greater extent (Fig. 6A, curve b) (47). These results indicated that the addition of valinomycin provoked the electrophoretic uptake of K⁺ previously released by AGE (Fig. 5, insert) and are in close agreement with the mechanisms of the effect of valinomycin (Fig. 6A, curve c) that induces an electrophoretic uptake of exogenous K⁺ driven by Δψm (Fig. 7). Most probably, valinomycin was more efficient in inducing a decrease in Δψm with respect to the increased hyperpolarization induced by AGE. The observed effect of AGE may be explained by considering the mechanisms of action of the above-mentioned antibiotics (nigericin and salinomycin) on the mitochondrial membrane that activates an exchange between endogenous K⁺ and exogenous H⁺, as schematically depicted in Fig. 7. This exchange takes place to a large extent, due to the mitochondrial matrix that contains high concentrations of K⁺ (approximately 150 mM). Thus, the electrochemical gradient (Δμ_H⁺), that is the sum of the electrical and chemical gradient (Δμ_H⁺ = Δψ + ΔpH), is completely transformed in a K⁺ gradient, Δμ_K⁺ (Δμ_H ⁺ = Δμ_K⁺). Under this consideration, the large uptake of H⁺ completely collapses ΔpH, with a strong acidification of the matrix, and K⁺ gradient Δμ_K⁺ (that is Δμ_H⁺) becomes equal to Δψm. This explains the observed increase in Δψm following the addition of nigericin, salinomycin, and likely also of AGE. Most probably, the effect of AGE was attributable to an activation of the mitochondrial K⁺/H⁺ exchanger, while nigericin and salinomycin act as ionophores, forming a channel on the membrane for K⁺/H⁺ exchange. The effects of AGE on the K⁺/H⁺ exchanger were strongly supported by the observations that after its addition to the RLM, a rapid release of K⁺ was induced (Fig. 5, insert).

These results demonstrate that AGE induces the K⁺/H⁺ exchange, similar to nigericin or salinomycin, but through a different mechanism. A further confirmation of this effect is represented by the results reported in Fig. 6B. In fact, a previous addition of 1 mM Mg²⁺ to RLM completely prevented the increase in Δψm caused by AGE. Mg²⁺ has been reported to be a strong inhibitor of the K⁺/H⁺ exchanger (48). The authors proposed Mg²⁺ as a 'carrier brake' of the exchanger. Indeed the inhibition on K⁺/H⁺ exchanger by Mg²⁺ confirmed that AGE acts on the K⁺/H⁺ exchanger.

As shown in Fig. 8, the prolonged incubation of RLM with various concentrations of AGE, until 40 min, a decrease in Δψm to a varying extent. This observation appeared to be in contrast to the results reported in Fig. 6A, showing that AGE stimulated an increase in Δψm. In fact, the decrease in Δψm began after several minutes of incubation and was dependent on the AGE concentration. The decrease in Δψm could be ascribable to the strong acidification of mitochondrial matrix following H⁺ uptake in exchange with K⁺ that damages RLM. However, the collapse of Δψm may be due also to another event. In this regard, it was also found that the incubation of RLM with AGE, for approximately 15 min, caused mitochondrial swelling to a large extent (Fig. 9). This swelling was inhibited by CsA, BKA and ADP (Fig. 9), that are typical inhibitors of the mitochondrial permeability transition (MPT), a phenomenon closely related to cell death by intrinsic apoptosis. Generally, this phenomenon is induced in the presence of supraphysiological Ca²⁺ concentrations (primary inducer) and an oxidizing agent (secondary inducer or amplifier) (for reviews see refs. 49,50). However, Ca²⁺ was not added to the incubation medium; nevertheless, another experiment revealed that ruthenium red (RR), an inhibitor of Ca²⁺ transport, completely prevented mitochondrial swelling (data not shown). This indicates that in the presence of AGE, RR acts on MPT by an unknown mechanism that, however, does not involve exogenous Ca²⁺. A first conclusion of these results is that AGE, following prolonged incubation, induced MPT of large amplitude, as confirmed by the complete prevention due to the MPT inhibitors.

The experiment shown in Fig. 10 was performed in order to evaluate whether MPT was also related to oxidative stress. As shown in the figure, AGE induced a decrease in the mitochondrial sulfydryl groups with a dose response-dependent mechanism. Thus, this observation may demonstrate that AGE acts as an oxidant with the generation of disulfide bridge (see Discussion).

To support this result, another experiment on mitochondrial swelling induced by AGE, in the presence of antioxidant agents was performed. The results reported in Fig. 11 indicated that the alkylating reagent, NEM, completely prevented the swelling, while the reducing agent, DTE, or the scavenger of ROS, spermine, exerted negligible or only partial inhibitory effects, respectively. Very surprisingly, AGE was able to induce oxidative stress; however, MPT induction was not closely related to this event. Most probably, the oxidation of the SH groups by AGE was only partially involved in the induction of MPT, as demonstrated by the ineffectiveness of DTE and spermine. Of note, the inhibitory effects induced by NEM warrant further investigation.