This article is mentioned in:

Introduction

Cancer is a major cause of morbidity and mortality worldwide. Several types of malignant diseases have been characterized to date. Leukemia, which is a term used to describe a group of hematological malignancies, is one of the major causes of cancer-related mortality globally (1). Treatment of leukemia with chemotherapy is a common approach. Anthracyclines and their derivatives, including doxorubicin (Dox), are anticancer drugs commonly used as chemotherapy for hematological malignancies in the clinical setting (2–4). Pirarubicin (Pira) is a Dox derivative that exhibits higher toxicity compared with Dox against cancer cells, but has been associated with lower cardiotoxicity than Dox in hamsters (5–7). Pira is used for the treatment of several malignancies, such as head and neck cancer, ovarian cancer, lymphoma and leukemia (8). There is evidence indicating that anthracyclines induce reactive oxygen species (ROS) generation, resulting in DNA damage (2–4). Mizutani et al (8) investigated the mechanisms of action of Pira in the induction of apoptosis via ROS generation, and revealed that Pira induced hydrogen peroxide (H2O2) generation and apoptosis of both HL-60 and HL-100 cells. The authors suggested that Pira induced cell apoptosis through H2O2 generation.

However, chemotherapy with anthracycline drugs for leukemia is often limited by the development of multidrug resistance (MDR) in cancer cells (9,10). MDR is observed when cells exhibit simultaneous resistance to various drugs without similar chemical structure and with different cellular targets. The mechanism identified by these MDR studies is considered to involve P-glycoprotein-mediated efflux of anticancer drugs. P-glycoprotein is an ATP-dependent plasma membrane protein (11–13). Certain compounds have exhibited the ability to reverse MDR in cancer cells, resulting in intracellular accumulation of anticancer drugs and making them a good choice for MDR cancer treatment; however, these MDR-reversing agents, such as cyclosporin A, verapamil, reserpine, quinidine and tamoxifen, are associated with unwanted side effects (13,14). Therefore, finding a new approach to overcoming MDR in cancer cells that includes MDR-reversing agents without unwanted side effects poses a challenge.

Several studies have reported the potential anticancer effects of natural products on various cancer cells (15–19). Gallic acid (GA) is a 3,4,5-trihydroxybenzoic acid that is distributed in various natural products (20,21). GA has exhibited anticancer activity without damage to normal cells (22–26). Based on that evidence on the effects of GA on cancerous and normal cells, it appears that GA may hold promise for overcoming MDR of cancer cells. In addition, it was hypothesized that GA may help to enhance the anticancer activity of Pira in MDR cancer cells. However, GA has been revealed to have strong antioxidant activity (27–29), which may also affect the effectiveness of cancer treatment by Pira, as Pira-induced cell death occurs via ROS generation. Therefore, these findings raised the question of whether GA (which has antioxidant properties) could modify the cytotoxicity of Pira (which has pro-oxidant properties) in cancer cells. The present study was undertaken to investigate the effect of GA on the anticancer activity of Pira in K562 and K562/Dox MDR leukemia cells and elucidate the potential underlying mechanisms.

Materials and methods

Chemicals

Pira (product no. P8624; Fig. 1A), GA (product no. 398225; Fig. 1B), rhodamine B (product no. 83689), MTT (product no. M2003) and resazurin (product no. R7017) were purchased from Sigma-Aldrich; Merck KGaA. FBS (cat. no. FBS-11A) and penicillin/streptomycin (cat. no. PS-B) were purchased from Capricorn Scientific. RPMI-1640 medium (product no. RPP10) was purchased from Caisson Labs.

Figure 1. Chemical structure of (A) Pira and (B) GA. Effect of Pira on (C) K562 and (F) K562/Dox cancer cell viability. Effect of GA on (D) K562 and (G) K562/Dox cancer cell viability. Effect of GA/Pira combination on (E) K562 and (H) K562/Dox cancer cell viability. Treatment times: ▬48 and ▭72 h. aP<0.05 when compared with control and bP<0.05 when compared with 10 nM Pira. Pira, pirarubicin; GA, gallic acid; Dox, doxorubicin.

Cell lines and cell culture

The K562 human leukemia cell line and its Dox-resistant counterpart, K562/Dox (P-glycoprotein-overexpressing cells) were provided by author CU. The cells were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated FBS and 1% penicillin/streptomycin in a humidified atmosphere with 5% CO2 at 37°C. The initial cell density was 1×105 cells/ml and reached 8–10×105 cells/ml 72 h later. Cells were routinely sub-cultured at 72 h.

For the experiments, the initial cell density was 5×105 cells/ml and reached 8–10×105 cells/ml 24 h later. Cells were found to be at the exponential growth phase.

Cytotoxic assay

Cytotoxicity was assessed using the resazurin assay as previously described (30). Cells (5×104 cells/ml) were incubated with various concentrations of GA (0, 0.1, 1, 10, 100 and 200 µM) or Pira (0, 1, 10, 100, 300 and 500 nM) in complete RPMI-1640 medium and placed into a humidified atmosphere with 5% CO2 at 37°C for 48 and 72 h. Subsequently, 100 µl resazurin (blue and non-fluorescent dye, 0.1 mg/ml) was added to the system. After 4 h, red fluorescent dye resorufin was obtained. The fluorescence emission at 590 nm and excitation wavelength at 570 nm were assessed by a spectrofluorometer (LS55; Perkin Elmer, Inc.). Of note, there were several in vitro studies on anticancer activities of GA which used concentrations of GA ranging from 10–500 µM (31–37).

Co-treatment

Cells (5×104 cells/ml) were incubated with 10 or 100 µM GA and 10 nM Pira in complete RPMI-1640 medium in a humidified atmosphere with 5% CO2 at 37°C for 48 and 72 h. Next, 100 µl resazurin (0.1 mg/ml) was added to the system. After 4 h, red fluorescent dye resorufin was obtained. The fluorescence emission at 590 nm and excitation wavelength at 570 nm were assessed by a spectrofluorometer (LS55; Perkin Elmer, Inc.).

Synergism quotient (SQ) calculation

The SQ was calculated by subtracting baseline values from all treatments, then dividing the net effect of the combination drugs (GA/Pira) by the sum of individual effects of drugs (GA + Pira). SQ >1.0 indicated a synergistic effect.

Mitochondrial activity determination

The mitochondrial activity was determined by MTT-reduction assay. The solvent used to dissolve the purple formazan was dimethyl sulfoxide (DMSO). Mitochondrial reductase reduced MTT, a yellow tetrazole, to purple formazan in living cells. Cells (5×104 cells/ml) were incubated with GA (10 and 100 µM) or Pira (10 nM) or GA/Pira combination (10/10 and 100 µM/10 nM) in complete RPMI-1640 medium in a humidified atmosphere with 5% CO2 at 37°C for 3 h. Next, 200 µM MTT was added to the system. After 20 min, purple formazan was obtained. The absorption intensity at 560 nm was measured by the UV-visible spectrophotometer (Agilent Technologies, Inc.).

Mitochondrial membrane potential (ΔΨm) measurement

The ΔΨm was measured by using the non-invasive functional spectrofluorometric method previously described (38–40). Briefly, 2×106 cells/ml were incubated with GA (10 and 100 µM) or Pira (10 nM) or GA/Pira combination (10/10 and 100 µM/10 nM) for 3 h and were then suspended with 40 nM rhodamine B in 2 ml HEPES-Na+ buffer (pH 7.25) and vigorously stirred at 37°C. The fluorescence spectrum of rhodamine B (emission wavelength at 582 nm and excitation wavelength at 553 nm) was recorded as a function of time. Next, 200 µM MTT was added to the system after cells had been incubated with rhodamine B for 20 min, resulting in a progressive decrease in rhodamine B fluorescence. ΔΨm was calculated by the following equation: ΔΨ=−61.51 logVi-258.46 mV, where Vi=initial rate of decrease in rhodamine B fluorescence=(dF/dt) × (CT/F0), nM/sec. dF/dt=slope of the tangent to the curve F=f(t) after the addition of MTT. CT=rhodamine B concentration, F0=rhodamine B fluorescence intensity. In the present study, ΔΨm values are presented as an absolute |ΔΨm| value.

ATP measurement

Cellular ATP was extracted by the single-step ATP extraction method using boiling water as previously described (41). Briefly, 2×106 cells/ml were incubated with GA or Pira or GA/Pira combination in 2 ml HEPES-Na+ buffer (pH 7.25) at 37°C for 3 h. The cells were centrifuged at 2,000 × g for 1 min and then cell pellets were collected. Cellular ATP was extracted by adding 1 ml boiling water (60°C) to the cell pellets, followed by vortexing and centrifugation at 2,000 × g for 1 min, room temperature. The supernatant was used for ATP measurement. The supernatant was purified by thin-layer chromatography (TLC). The stationary phase was silica-coated aluminum (TLC silica gel 60 F254; Merck KGaA). The mobile phase consisted of butanol:acetic acid:water at a ratio of 4:1:5 (v/v). ATP standard was used as a reference. The retardation factor was 0 for the ATP standard when applied to the stationary and mobile phases. Purified ATP samples were measured by spectrofluorometry. The fluorescence spectrum of purified ATP samples was recorded at a fluorescence emission range of 390–500 nm, with an emission peak at 410 nm (excitation at 360 nm).

Determination of P-glycoprotein function in living K562/Dox cancer cells

The P-glycoprotein function in living K562/Dox cancer cells was assessed using a non-invasive functional spectrofluorometric method (LS55; Perkin Elmer, Inc.) as previously described (40,42). Briefly, 2×106 cells/ml were suspended in 2 ml HEPES-Na+ buffer (pH 7.25) at 37°C and were continually stirred using a 1-cm quartz cuvette (Hellma Analytics). Pira (1 µM) was added to the system. The fluorescence intensity of Pira at 590 nm (excitation at 480 nm) was recorded as a function of time. At the steady state, GA at various concentrations (0, 10, 50, 100 and 200 µM) was added. Subsequently, 5 µl of 4% Triton X-100 solution was added, resulting in the system entering the equilibrium state.

In the present study, the P-glycoprotein function is presented as P-glycoprotein-mediated active efflux coefficient (ka). The ability of GA to inhibit the function of P-glycoprotein was determined using the following ratio: kai/ka° where kai=P-glycoprotein-mediated active efflux of Pira in the presence of GA; and ka°=P-glycoprotein-mediated active efflux of Pira in the absence of GA. When kai/ka°=1, GA cannot inhibit active efflux. When kai/ka°=0, GA can completely inhibit active efflux.

The overall concentration of Pira accumulation in cells (Cn) and concentration of Pira accumulation in the cytoplasm (Ci) were also determined.

Resistance factor (RF) calculation

RF in drug-resistant cells with P-glycoprotein expression contributed to the kinetics of passive and active efflux of Pira, which was previously described (43). RF could theoretically be calculated as follows: RF=1 + ka/k+ where k+=passive influx coefficient; and ka=P-glycoprotein-mediated active efflux coefficient.

Intracellular pH (pHi) measurements

In order to analyze the kinetics of Pira transport into the cells, the free Pira concentration in the cytoplasm may be calculated as previously described (44): Ci=Ce (1+10pKa-pHi)/(1+10pKa-pHe), which may be re-written as follows: pHi=pKa-log [Ci (1+10pKa-pHe)]/Ce where Ci=intracellular free Pira concentration in the steady state, Ce=extracellular free Pira concentration in the steady state, pKa=7.7 for Pira, pHe=7.25 at 37°C

Statistical analysis

Statistical analysis was performed using Microsoft Excel 2010 (Microsoft Corporation) and OriginPro 2015 (OriginLab Corporation). The number of repeats was 5. The data in the present study are expressed as the mean ± standard error of the mean (SEM). To evaluate statistical differences in the mean values between each treated group and the non-treated control group, one-way ANOVA was used to assess the significance of GA concentration. The post hoc test (Tukey's test) was used to evaluate statistical differences in the mean values between each group. P<0.05 was considered to indicate a statistically significant difference.

Results

Effect of GA and Pira on K562 and K562/Dox cancer cell viability

Fig. 1C, D, F and G revealed the effects of Pira and GA on K562 (Fig. 1C and D) and K562/Dox (Fig. 1F and G) cancer cell viability at 48 and 72 h. These results demonstrated that Pira and GA decreased K562 cancer cell viability in a dose- and time-dependent manner. The viability of K562/Dox cancer cells in the presence of GA (0.1, 1 and 10 µM) and Pira (1, 10, 100 and 300 nM) did not significantly change at 48 and 72 h compared with the control group. High concentrations of GA (100 and 200 µM) were associated with a change in cell viability to 79 and 31% at 48 h, and to 72 and 25% at 72 h, respectively. Treatment with 500 µM Pira was associated with a change in cell viability to 73% at 48 h and 64% at 72 h.

The results indicated that GA and Pira decreased the viability of K562 and K562/Dox cancer cells in a concentration- and treatment time-dependent manner.

Effect of GA/Pira combination on K562 and K562/Dox cancer cell viability

The data demonstrated that 10 nM Pira significantly decreased the viability of K562 cancer cells at 48 and 72 h to 65 and 60% (Fig. 1C), respectively, and decreased the viability of K562/Dox cancer cells at 72 h to 95% (Fig. 1F), compared with the control group. The viability of K562/Dox cancer cells treated with 10 nM Pira did not change at 48 h. Fig. 1E and H revealed the percentage (%) of cell viability in treated (with a combination of Pira and GA) K562 and K562/Dox cancer cells and non-treated cells at 48 and 72 h.

In the GA/Pira combination group, the viability of K562 cancer cells decreased following treatment with a combination of 10 nM Pira with 10 or 100 µM GA, specifically to 36 or 28% at 48 h and to 27 or 22% at 72 h, respectively, compared with the control group (Fig. 1E). The viability of K562/Dox cancer cells decreased following treatment with a combination of 10 nM Pira and 10 or 100 µM GA, specifically to 80 or 69% at 48 h and to 78 or 60% at 72 h, respectively, compared with the control group (Fig. 1H).

In addition, as revealed in Table I, the SQ calculation of cell viability (48 and 72 h) indicated a synergistic effect of the GA/Pira combination. These results indicated that GA enhanced the cytotoxic effect of Pira on both K562 and K562/Dox cancer cells in a concentration- and treatment time-dependent manner.

Table I. SQ calculation.

Table I.

SQ calculation.

	SQ values
	K562 cancer cells		K562/Dox cancer cells
Combinations	48 h	72 h	48 h	72 h
Pira 10 nM/GA10 µM	1.13	1.10	2.42	1.54
Pira 10 nM/GA 100 µM	1.10	1.01	1.47	1.23

[i] SQ, synergism quotient; Pira, pirarubicin; GA, gallic acid; Dox, doxorubicin.

The microscopic examination revealed rough, shrinking, irregularly shaped cells, and decreases in cell density as shown in Fig. 2. These findings were linked to cell death. Of note, the present study used higher concentrations of Pira and GA in Fig. 2 because there were clearly observable cell morphological changes.

Figure 2. Microscopic images (magnification, ×100) of treated K562 and K562/Dox cancer cells and non-treated cells at 72 h. Pira, pirarubicin; GA, gallic acid; Dox, doxorubicin.

Effect of GA/Pira combination on mitochondrial activity in K562 and K562/Dox cancer cells

The mitochondrial activity was determined by the reduction of MTT to insoluble purple formazan. The reaction was followed by the alteration of absorbance intensity at 560 nm. Therefore, increments of absorbance intensity at 560 nm were considered as an indication of increased mitochondrial activity. Fig. 3A and B revealed the alterations in absorbance intensity at 560 nm of treated K562 and K562/Dox cancer cells and non-treated cells. The absorbance intensity at 560 nm was decreased in K562 and K562/Dox cancer cells treated with 10 nM Pira and 10 or 100 µM GA compared with the control group.

Figure 3. Effect of Pira, GA and Pira/GA on mitochondrial activity, ΔΨm and ATP levels in K562 and K562/Dox cancer cells. Abs at 560 nm of (A) treated K562 and (B) treated K562/Dox cancer cells and non-treated cells. The abs at 560 nm decreased, reflecting a decrease in the mitochondrial activity. The abs at 560 nm was decreased in K562 cancer cells treated with combination of 10 nM Pira and 10 or 100 µM GA, as it changed to 81 or 72% compared with the control group, respectively. The abs at 560 nm was significantly decreased in K562/Dox cancer cells treated with a combination of 10 nM Pira and 10 or 100 µM GA, as it changed to 72 or 70% compared with the control group, and to 82 or 80% compared with the 10 nM Pira alone group, respectively. |ΔΨm| in (C) treated K562 and (D) treated K562/Dox cancer cells and non-treated cells. The |ΔΨm| decreased, reflecting a decrease in the ΔΨm. The |ΔΨm| of K562 cancer cells significantly decreased with the combination of 10 nM Pira and 10 or 100 µM GA, as the |ΔΨm| changed to 83 or 82% compared with the control group, and changed to 87 or 86% compared with the 10 nM Pira alone group, respectively. The |ΔΨm| of K562/Dox cancer cells significantly decreased with the combination of 10 nM Pira and 10 or 100 µM GA, as the |ΔΨm| changed to 87 or 86% compared with the control group, and to 92 or 90% compared with the 10 nM Pira alone group, respectively. Percentage (%) of change of ATP levels in (E) treated K562 and (F) treated K562/Dox cancer cells and non-treated cells. The minus % change of ATP levels indicates the decreasing ATP. aP<0.05 compared with control and bP<0.05 compared with 10 nM Pira. Abs, absorbance intensity; Pira, pirarubicin; GA, gallic acid; ΔΨm, mitochondrial membrane potential; Dox, doxorubicin.

In the GA/Pira combination group, absorbance intensity at 560 nm was decreased in K562 cancer cells treated with a combination of 10 nM Pira and 10 or 100 µM GA, as it changed to 81 or 72%, respectively, compared with the control group (Fig. 3A). The absorbance intensity at 560 nm was decreased in K562/Dox cancer cells treated with a combination of 10 nM Pira and 10 or 100 µM GA, specifically changing to 72 or 70% compared with the control group (P<0.05), and to 82 or 80% compared with the 10-nM Pira alone group, respectively (Fig. 3B).

These results indicated that GA significantly enhanced the Pira-induced mitochondrial activity that was decreased in K562/Dox cancer cells. Moreover, GA enhanced the Pira-induced mitochondrial activity that was decreased in K562 cancer cells, but the difference was not statistically significant.

Effect of GA/Pira combination on the |ΔΨm| of K562 and K562/Dox cancer cells

Fig. 3C and D revealed the effect of GA/Pira combination on |ΔΨm| of K562 and K562/Dox cancer cells. This data demonstrated that 10 nM Pira decreased the |ΔΨm| of K562 and K562/Dox cells by 5% compared with the control group, whereas 10 or 100 µM GA decreased the |ΔΨm| by 10 or 14% and by 4 or 8% for K562 and K562/Dox cancer cells respectively, compared with the control group.

In the GA/Pira combination group, the |ΔΨm| of K562 cancer cells significantly decreased with the combination of 10 nM Pira with 10 or 100 µM GA, specifically changing to 83 or 82% compared with the control group, and to 87 or 86% compared with the 10 nM Pira alone group, respectively (Fig. 3C). The |ΔΨm| of K562/Dox cancer cells decreased following treatment with the combination of 10 nM Pira with 10 or 100 µM GA, specifically changing to 87 or 86% compared with the control group (P<0.05) and to 92 or 90% compared with the 10 nM Pira alone group, respectively (Fig. 3D).

These results indicated that GA enhanced the Pira-induced ΔΨm disruption in both K562 and K562/Dox cancer cells in a concentration-dependent manner.

Effect of GA/Pira on ATP levels in K562 and K562/Dox cancer cells

The ATP level was assessed by determining the fluorescence emission intensity at 410 nm. Subsequently, the changes in fluorescence emission intensity in treated cell groups were compared with the control group. The change in intensity was considered to reflect ATP level changes and the minus symbol indicates decrement. Fig. 3E and F revealed the percentage (%) of change of ATP levels in treated K562 and treated K562/Dox cancer cells and non-treated cells. The percentage (%) of decrease in ATP levels was observed in K562 and K562/Dox cancer cells treated with 10 nM Pira and 10 or 100 µM GA compared with the control group.

In the GA/Pira combination group, the percentage (%) of decrease in ATP levels was significant in K562 cancer cells treated with a combination of 10 nM Pira and 10 or 100 µM GA compared with the control group (Fig. 3E). The percentage (%) of decrease in ATP levels was significant in K562/Dox cancer cells treated with a combination of 10 nM Pira and 10 or 100 µM GA compared with the control group (Fig. 3F).

These results indicated that GA enhanced the Pira-induced decrease in ATP levels in both K562 and K562/Dox cancer cells in a concentration-dependent manner.

Effect of GA on P-glycoprotein function in K562/Dox cancer cells

The kinetic parameters of Pira uptake into GA-treated K562 and GA-treated K562/Dox cancer cells and non-treated cells are revealed in Table II. The V+ and k+ values were significantly decreased in K562 cancer cells following treatment with 50, 100 and 200 µM GA, whereas these two values did not significantly change in GA-treated K562/Dox cancer cells compared with the control group. However, Va and ka values were decreased in K562/Dox cancer cells.

Table II. Kinetics of Pira uptake into K562 and K562/Dox cancer cells after treatment with various GA concentrations.

Table II.

Kinetics of Pira uptake into K562 and K562/Dox cancer cells after treatment with various GA concentrations.

	K562 cancer cells		K562/Dox cancer cells
GA, µM	V+, nM/sec	k+ ×10−12 l/cell.sec	V+, nM/sec	k+ ×10−12 l/cell.sec	Va, nM/sec	ka ×10−12 l/cell.sec
0	2.2±0.2	2.24±0.21	1.4±0.1	1.41±0.14	1.5±0.2	4.06±0.98
10	1.7±0.1	1.86±0.09	1.2±0.2	1.24±0.22	1.3±0.3	4.17±0.93
50	1.5±0.2a	1.47±0.20a	1.3±0.2	1.26±0.18	0.9±0.1	2.65±0.48
100	1.5±0.1a	1.48±0.06a	1.3±0.2	1.26±0.24	0.9±0.2	2.48±0.58
200	0.9±0.2a	0.94±0.02a	1.1±0.2	1.06±0.19	0.6±0.1	1.34±0.31

a P<0.05 compared with control. V₊, rate of Pira uptake into the cells [(dF/dt) × (C_T/F₀), where C_T is the total drug concentration added to the cells and F₀ the Pira fluorescence at time=0]; V_a, rate of P-glycoprotein-mediated active efflux of Pira (referred to P-glycoprotein function); k₊, passive influx coefficient; k_a, P-glycoprotein-mediated active efflux coefficient (referred to P-glycoprotein function); Pira, pirarubicin; GA, gallic acid; Dox, doxorubicin.

The Cn and Ci values were increased in K562/Dox cancer cells treated with 50, 100 and 200 µM GA, whereas these values did not change in GA-treated K562 cancer cells at all GA concentrations compared with the control group (Fig. 4A and B). Fig. 4C revealed the ratio of kai/ka° values in GA-treated K562/Dox cancer cells. The data revealed decreases in the kai/ka° ratio values in GA-treated K562/Dox cancer cells at 50, 100 and 200 µM GA.

Figure 4. Pira accumulation and pHi in K562 and K562/Dox cancer cells. (A) Cn, (B) Ci and (C) kai/ka° ratio values in GA-treated K562 cells, GA-treated K562/Dox cancer cells and non-treated-cells. (D) RF values in GA-treated K562/Dox cancer cells. (E) Percentage (%) of change of pHi values in GA-treated K562 and GA-treated-K562/Dox cancer cells and non-treated control group. aP<0.05 when compared with control. Cn, overall concentration of Pira accumulation in cells; Ci, concentration of Pira accumulation in cytoplasm; kai/ka°, inhibition characterization of the P-glycoprotein-mediated efflux of Pira by GA; RF, resistance factor; Pira, pirarubicin; GA, gallic acid; Dox, doxorubicin; pHi, intracellular pH.

These data indicated that GA inhibited P-glycoprotein from pumping Pira out of the cells, resulting in increased intracellular Pira accumulation.

Effect of GA on RF values

The RF values were revealed to be decreased in K562/Dox cancer cells treated with 50, 100 and 200 µM GA compared with the non-treated control group (Fig. 4D). This result indicated that GA may reverse drug resistance in K562/Dox cancer cells.

Effect of GA on pHi in K562 and K562/Dox cancer cells

Fig. 4E revealed the percentage (%) of change in pHi values in GA-treated K562 and K562/Dox cancer cells and in the non-treated control group. The changes in the pHi values in GA-treated K562/Dox cells were illustrated but no significant changes were observed when compared with the non-treated control group for all concentrations of GA.

Discussion

GA is a natural phenolic compound found in several fruits and medicinal plants. It has been reported to have several health-promoting effects such as an anti-inflammatory activity that involved MAPK and NF-κB signaling pathways (45). GA and its derivatives have exhibited potential anticancer activity in vitro and in vivo (46,47). The anticancer activity has been revealed in several types of cancer cells, including ovarian cancer cells and HCT-15 colon cancer cells (48–52). There is evidence suggesting that GA is able to induce apoptosis and cell death (50,51). In the present study, it was observed that GA inhibited the proliferation of K562 and K562/Dox cancer cells in a concentration and time-dependent manner. Moreover, the morphology of treated K562 and K562/Dox cancer cells displayed apoptotic characteristics, such as rough cells and shrinking cells. These findings were consistent with the findings of Subramanian et al (50). That study investigated the anticancer activity of GA on HCT-15 cancer cells. The authors determined that GA induced HCT-15 cancer cell death via apoptotic pathways and further revealed evidence of apoptosis based on images of cell morphology. The images revealed typical signs of apoptosis, such as membrane blebbing and shrinking cells (50). In addition, the present study found that GA disrupted the cellular energy status by decreasing mitochondrial activity, ΔΨm and ATP levels in both K562 and K562/Dox cancer cells. Hence, it was inferred that GA was able to induce apoptotic cell death in K562 and K562/Dox cancer cells.

In the present study, the question of whether GA (an antioxidant compound) can modify the cytotoxicity of Pira (a pro-oxidant compound) in cancer cells was raised. It was herein demonstrated that the GA/Pira combination exhibited higher antiproliferative activity compared with Pira alone in K562 and K562/Dox cancer cells. Hence, the mechanism of action of the GA/Pira combination was investigated in K562 and K562/Dox cancer cells. The results demonstrated that GA/Pira decreased mitochondrial activity, ΔΨm and ATP levels in both K562 and K562/Dox cancer cells. These biological effects of the combination were higher compared with those of Pira alone in both types of cancer cells. The findings of the present study indicated that the synergistic actions of the combination treatment could aggravate the disturbance of mitochondrial activity, ΔΨm and ATP levels. Therefore, it was inferred that GA was able to enhance Pira-induced K562 and K562/Dox cancer cell death through cellular energy status disruption.

GA and its derivative compounds did not only enhance the anticancer activity of Pira, but also enhanced the effects of other anticancer drugs on various cancer cells. Aborehab and Osama (31) studied the effect of GA on the anticancer effects of paclitaxel on HeLa cervical cancer cells. The authors found that the combination of GA and paclitaxel exerted cytotoxic effects and induced apoptosis. The combination treatment also increased the levels of P53 and caspase 3. The authors suggested that GA augmented the effects of paclitaxel on cervical cancer cells. Wang et al (53) studied the effect of GA on the anticancer activity of cisplatin in the H446 human small cell lung cancer cell line. The authors revealed that the combination of GA and cisplatin exerted cytotoxic effects and induced ROS generation and apoptosis. The combination treatment increased the levels of apoptosis regulators, such as Bax, Apaf-1 and P53, in H446 cells. The authors suggested that GA enhanced the anticancer activity of cisplatin in H446 cells. The authors proposed that the mechanism of GA enhancement was through a ROS-dependent mitochondrial apoptotic pathway. In addition, Rajagopalan et al (54) studied the effect of GA-derived compounds on the anticancer activity of 5-fuorouracil in the A431 human squamous carcinoma cell line. The authors found that the combination of GA derivatives and 5-fuorouracil exhibited higher cytotoxicity compared with 5-fuorouracil alone in A431 cancer cells. Therefore, it was deduced that GA could enhance the efficacy of anticancer drugs, including paclitaxel, cisplatin, 5-fuorouracil and Pira, to induce cancer cell death.

Of note, the present study demonstrated that GA was able to inhibit cell proliferation and induce cell energy status disruption in K562/Dox cancer cells. K562/Dox is an MDR leukemia cancer cell line, which is associated with the overexpression of P-glycoprotein in the cell membrane. P-glycoprotein is an ATP-dependent transmembrane transporter that transports drugs out of the cells, resulting in a decrease in intracellular drug concentration. A high P-glycoprotein function is commonly found in MDR cells (55,56). MDR poses a major challenge to leukemia chemotherapy. The present study found that GA was able to reverse drug resistance in K562/Dox cancer cells (low RF values). In addition, GA inhibited P-glycoprotein-mediated efflux of Pira in K562/Dox cancer cells (low kai/ka° values). This inhibitory effect led to an increase in the intracellular Pira concentration (high Cn and Ci values) in K562/Dox cancer cells. Hence, it was proposed that GA may reverse drug resistance in K562/Dox cancer cells by inhibiting the function of P-glycoprotein, resulting in increasing intracellular Pira accumulation that leads to K562/Dox cancer cell death. More importantly, it was demonstrated that the disruption of the cell energy status contributed to the reversal mechanisms of GA in K562/Dox cancer cells. The effect of verapamil on P-glycoprotein-mediated efflux of the Pira in K562/Dox was reported in a previous study by Mankhetkorn and Garnier-Suillerot (57). The authors reported that the verapamil concentrations required to obtain 50% inhibition of the P-glycoprotein-mediated efflux of the Pira (kai/ka° value=0.5) was 0.5±0.2 µM (57). To compare with GA in the present study, the effect of 50 and 100 µM GA was less than verapamil, whereas 200 µM GA was higher than verapamil. Of note, the results indicated that 10 µM GA did not affect kai/ka° whereas it still had an additive effect on cell survival. It was hypothesized that 10 µM GA may induce cell death via another mechanism, not via inhibition of P-glycoprotein function. In addition, GA-derived compounds were reported to have the ability to modify P-glycoprotein function in KB-C2 cells (58).

The strength of the present study was that it investigated cancer cells that were alive and in an active state. Therefore, these results may be considered as highly representative of the behavior of cancer cells in cancerous diseases. In addition, the results revealed the mechanisms through which GA enhanced the anticancer effect of Pira on living MDR cancer cells.

The biological effects of the GA/Pira combination were more prominent compared with those of Pira alone in K562 and K562/Dox cancer cells. The efficacy of this combination depended on the concentration of GA. In addition, these findings indicated that GA was able to reverse drug resistance in living K562/Dox cancer cells by inhibiting the function of P-glycoprotein, resulting in increased intracellular Pira accumulation. More importantly, it was demonstrated that the cellular energy status disruption contributed to the reversal mechanisms of GA in K562/Dox cancer cells. Finally, it was determined that GA enhanced the Pira-induced anticancer effects on living K562 and K562/Dox cancer cells through cellular energy status impairment and by inhibiting the function of P-glycoprotein. However, the anticancer properties of the GA/Pira combination require further evaluation in in vivo studies.

Acknowledgements

KTNA would like to thank the Ph.D. degree program in Biomedical Sciences, Faculty of Associated Medical Sciences, Chiang Mai University, under the CMU Presidential Scholarship (Chiang Mai, Thailand). The authors would like to thank the Department of Radiologic Technology, Faculty of Associated Medical Sciences, Chiang Mai University for their support (Chiang Mai, Thailand).

Funding

The present study was supported by the Ph.D. degree program in Biomedical Sciences, Faculty of Associated Medical Sciences, Chiang Mai University, under the CMU Presidential Scholarship. The present study was also partially supported by Chiang Mai University and Faculty of Associated Medical Sciences, Chiang Mai University.

Availability of data and materials

The datasets used and/or analyzed during the present study are available from the corresponding author on reasonable request.

Authors' contributions

KTA performed the acquisition, analysis and interpretation of data and drafted the manuscript. SW and BS performed acquisition and analysis of data. SK, CU, ST and JP performed interpretation of data and drafted the manuscript, revising it critically for important intellectual content. MT conceived and designed the study, as well as performed acquisition, analysis and interpretation of data, drafted the manuscript, revising it critically for important intellectual content, and provided final approval of the version to be published. All authors read and approved the final manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References