Introduction

Erythrina poeppigiana is distributed in South America, Africa, and Asia, and is planted as an ornamental plant in Okinawa Prefecture, Japan. The plant contains various types of isoflavonoids and alkaloids [1,2,3,4], some of which exhibit anti-viral activity, including inhibition of human immunodeficiency virus-1 replication [5], as well as anti-microbial [6, 7], anti-tumor [8], and estrogenic [9, 10] properties. In the current study, during the isolation of bioactive constituents from the stem bark of E. poeppigiana (Family: Leguminosae), we isolated erypoegin K, an isoflavone, for the first time and reported its inhibitory activity against recombinant human glyoxalase I (hGLO I) [11].

GLO I is an enzyme involved in the detoxification pathway of the toxic compound methylglyoxal (MG) produced as a by-product of glycolysis [12]. It catalyzes the reaction by which MG is converted into hemimercaptal using the reduced form of glutathione (GSH). Hemimercaptal is subsequently converted by glyoxalase II (GLO II) into the nontoxic product D-lactic acid [12, 13]. Tumor cells typically depend on glycolysis, rather than on the Krebs cycle, for ATP generation under both aerobic and anaerobic conditions. This phenomenon is referred to as the Warburg effect [14]. Multiple human cancer types, including breast [15], colon [16], lung [17], and liver [18, 19] cancers, as well as melanoma [20], express elevated levels of GLO I. The increase in GLO I expression is thought to mediate the survival of cancer cells under the condition of high MG levels. GLO I is also associated with multidrug resistance in cancer cells [21, 22] and metastasis in certain types of carcinoma [23].

S-p-bromobenzylglutathione (BBG) is a known GLO I inhibitor, which is a bromobenzyl derivative of GSH. Because BBG is a water-soluble compound, a membrane-permeable BBG derivative, S-p-bromobenzylglutathione cyclopentyl diester (BBGC), is utilized for in vivo studies [17, 21, 24]. BBGC reportedly exhibits cytotoxicity and anti-tumor activity against several tumor cell lines and in tumor-bearing model animals [17, 24]. These findings suggest that selective GLO I inhibitors may function as effective tumor-specific anti-tumor agents [25]. In the present study, we investigated the effect of erypoegin K on the viability of cancerous HL-60 cells and compared it with that on human peripheral lymphocytes.

Results and discussion

Erypoegin K is an isoflavone compound (chemical structure illustrated in Fig. 1) containing a characteristic furan ring at the A-ring.

Fig. 1
figure 1

Chemical structure of erypoegin K. The numbering and the A-, B-, and C-ring characteristics of isoflavones are shown

Full size image

We examined the effects of erypoegin K on the viability of human acute promyelocytic leukemia HL-60 cells. The cells were treated with various concentrations of erypoegin K, and their viability was evaluated by MTS assay after incubation for 48 h. As depicted in Fig. 2, erypoegin K potently reduced the viability of HL-60 cells. The IC50 value was found to be 0.175 ± 0.004 μM, which was 20-fold more potent than that of a known GLO I inhibitor, BBGC (IC50 = 3.59 ± 0.10 μM).

Fig. 2
figure 2

Effects of erypoegin K on the viability of HL-60 cells. HL-60 cells (1.2 × 104 cells/well) were treated with various concentrations of erypoegin K for 48 h. The values are presented as mean ± SE of three independent experiments. Triplicate measurements were obtained in each experiment. Open circle, BBGC; closed circle, erypoegin K

Full size image

We used human peripheral lymphocytes as a model to examine the effect of erypoegin K on normal (noncancerous) cells. As illustrated in Fig. 3, the cytotoxic effect of erypoegin K on normal cells was weak, suggesting that the cytotoxic activity of erypoegin K might be specific to cancerous cell lines such as HL-60.

Fig. 3
figure 3

Effect of erypoegin K on the viability of human peripheral lymphocytes. Human peripheral lymphocytes (1.0–1.5 × 105 cells/well) and HL-60 cells (1.2 × 104 cells/well) were treated with various concentrations of erypoegin K for 48 h. The values are presented as mean ± SE of three independent experiments. Triplicate measurements were obtained in each experiment. The significance of differences was evaluated using Student’s t test. *p < 0.01 vs. HL-60 cells. Open circle, human peripheral lymphocytes; closed circle, HL-60 cells

Full size image

To determine whether erypoegin K-induced cell death occurred by apoptosis, we evaluated the enzymatic activity of caspase-3, a marker of apoptotic cell death. As depicted in Fig. 4, caspase-3 activity was strongly enhanced in HL-60 cells treated with 15 μM erypoegin K, and with 15 μM BBGC used as a positive control of apoptosis-inducing agent in HL-60 cells [24]. To confirm this indication, we examined the effect of a caspase-3 inhibitor, Z-DEVD-FMK, on the viability of HL-60 cells treated with erypoegin K. As depicted in Fig. 5, the pretreatment of the cells for 2 h with the Z-DEVD-FMK rescued the viability in cells treated with erypoegin K for 6 h or 24 h, indicating the involvement of caspase-3 in the signaling pathway of cell death by erypoegin K. We further examined nuclear condensation and genome DNA fragmentation, both of which are characteristic features of apoptotic cells. The nuclear condensation was observed in cells treated with 10 μM erypoegin K for 6 h, whereas no condensation was observed in cells treated with DMSO (Fig. 6). A typical DNA fragmentation with a ladder-like pattern on agarose gel electrophoresis was visible in cells treated with 10 μM erypoegin K for 10 h (Fig. 7). These data clearly demonstrated that erypoegin K-induced cell death is mediated by apoptosis.

Fig. 4
figure 4

Induction of caspase-3 activity in HL-60 cells treated with erypoegin K or BBGC. HL-60 cells (5.0 × 105 cells/well) were treated with 15 µM erypoegin K, 15 µM BBGC, or DMSO for the indicated time period. Caspase-3 activity was measured using a Caspase-3 Fluorometric Assay Kit, and the activity was expressed as an arbitrary unit of fluorescence intensity in proportion to the reaction product. The values are presented as mean ± SE of three independent experiments. Triplicate measurements were obtained in each experiment. The significance of differences was evaluated using Student’s t-test. *p < 0.01 vs. control. Open circle, erypoegin K; closed circle, BBGC; open square, DMSO (control)

Full size image
Fig. 5
figure 5

Effect of the caspase-3 inhibitor on the viability of HL-60 cells treated with erypoegin K. HL-60 cells were treated with various concentrations of erypoegin K in the presence or absence of the caspase-3 inhibitor Z-DEVD-FMK for 6 h (a) or 24 h (b). The inhibitor was added 2 h prior to treatment with erypoegin K. Duplicate measurements were obtained in each experiment. Open circle, without caspase-3 inhibitor; closed circle, with caspase-3 inhibitor (100 µM)

Full size image
Fig. 6
figure 6

Detection of DNA condensation. HL-60 cells (8.0 × 105 cells/well) were treated with 10 µM erypoegin K or DMSO for 6 h. The nuclei were stained with Hoechst 33342 and fixed with 5% paraformaldehyde. a erypoegin K; b DMSO (control)

Full size image
Fig. 7
figure 7

Detection of apoptotic genome DNA ladder. HL-60 cells (2.0 × 105 cells/dish) were treated with 10 µM erypoegin K or DMSO for 10 h. Genomic DNA was extracted from the cells using Apoptosis Ladder Detection Kit Wako and separated by 1.5% agarose gel electrophoresis. Lane 1, 100-bp size marker; lane 2, erypoegin K; lane 3, DMSO (control)

Full size image

MG is a reactive dicarbonyl compound that is mainly formed as a metabolic by-product of glycolysis [24]. Reportedly, extracellular MG triggers the intracellular production of free radicals, which in turn easily interact with proteins, lipids, and DNA [26, 27]. These reactions may be closely associated with its cytotoxicity [28]. In fact, exogenously added MG was found to induce apoptosis in human leukemia HL-60 [29] and U937 [30] cells as well as in human hepatocyte HepG2 cells [31] and human osteoblasts [32]. Furthermore, BBGC has been shown to cause MG generation by HL-60 cells and to induce apoptosis [24]. We have previously demonstrated that erypoegin K can inhibit human recombinant GLO I enzyme activity [11]. This observation prompted us to measure MG levels in HL-60 cells treated with erypoegin K to determine whether erypoegin K can actually inhibit intracellular GLO I. Consequently, MG was not detected in the lysates of HL-60 cells treated with 10 μM erypoegin K (data not shown); however, high levels of MG were detected in the culture medium after 24 h of the treatment; the levels were much higher than those in the medium obtained from HL-60 cells treated with 10 μM BBGC (Fig. 8). These data strongly suggested that erypoegin K inhibits intracellular GLO I activity, which leads to MG production and accumulation in the culture medium.

Fig. 8
figure 8

MG concentration in the culture medium of HL-60 cells treated with erypoegin K or BBGC. HL-60 cells (1.5 × 106 cells/dish) were treated with 10 µM erypoegin K, 10 µM BBGC, or DMSO (control) for 24 h. MG concentration in the culture medium was measured using HPLC/UV. The values are presented as mean ± SE of three independent experiments. Duplicate measurements were obtained in each experiment. The significance of differences was evaluated using Student’s t test. *p < 0.05 vs. control

Full size image

The cytotoxic effect of exogenously added MG was investigated at a sub-mM to mM level [29,30,31,32]. In our preliminary study, we found that the level of MG added in culture medium containing 10% fetal bovine serum (FBS, without cells) decreased rapidly, possibly due to covalent bond formation with proteins and/or amino acids present in the medium (data not shown). Therefore, there is a possibility that much higher levels of MG are released from the cells treated with erypoegin K than those shown in Fig. 8.

Erypoegin K that naturally occurs in Erythrina plants is a racemic mixture containing a chiral carbon at the C-2′′ position (Fig. 1). Thus, our next aim is to elucidate whether there are any differences in the biological activities of the two optical isomers of erypoegin K, especially in their apoptosis-inducing activity against cancerous cell lines. These points are now under investigation.

In the present study, we demonstrated that erypoegin K, a compound isolated from E. poeppigiana, potently reduced the viability of and induced apoptosis in HL-60 cells through GLO I activity inhibition. The cytotoxic effect of erypoegin K was specific to cancerous HL-60 cells and not to human peripheral lymphocytes. These findings suggest that erypoegin K might provide the basis for a novel and selective cancer therapy.

Experimental

Preparation of erypoegin K

The methods used to prepare erypoegin K, including methanol extraction, silica gel chromatography, and reversed-phase HPLC, were conducted as previously described [11]. The purity of the compounds was evaluated using reversed-phase HPLC [11].

HL-60 cell cultures

Human acute promyelocytic leukemia HL-60 cells were obtained from RIKEN cell bank (Riken, Tokyo, Japan) and cultured in RPMI1640 medium supplemented with 10% FBS, 100 units/ml penicillin, and 100 μg/ml streptomycin at 37 °C in a humidified atmosphere of 5% CO2 and 95% air.

Human peripheral lymphocyte cultures

Human peripheral lymphocytes from healthy volunteers were isolated using the Percoll–Hypaque density gradient method [33] with a Percoll isolation kit (Lymphoprep, Axis-Shield, Oslo, Norway), according to the manufacturer’s instructions. The isolated lymphocytes were precultured for 24 h in RPMI1640 medium supplemented with 10% FBS, 100 units/ml penicillin, and 100 μg/ml streptomycin at 37 °C in a humidified atmosphere of 5% CO2 and 95% air.

Cell viability assay

HL-60 cells in the log phase (approximately 1 × 106 cells/ml) were plated in a 96-well plate (Corning, NY, USA) at a density of 1.2 × 104 cells/well. Human peripheral lymphocytes were plated at a density of 1.0–1.5 × 105 cells/well. After 24 h of preincubation, the cells were treated with the test compound and cultured for an additional 48 h. Erypoegin K and BBGC were dissolved in DMSO (25 mM), and the final DMSO concentration in the cell culture medium was less than 0.4%. Cell viability was assessed using an MTS assay (CellTiter 96™ AQueous One Solution Cell Proliferation Assay Kit, Promega, Madison, WI, USA) based on the production of water-soluble tetrazolium by mitochondrial dehydrogenase. In cells treated with the caspase-3 inhibitor Z-DEVD-FMK (R&D Systems Inc., Minneapolis, MN, USA), the inhibitor was added 2 h prior to the addition of the test compound.

Caspase assay

The enzymatic activity of caspase-3 in HL-60 cells was assessed using a Caspase-3 Fluorometric Assay Kit (R&D Systems Inc., Minneapolis, MN, USA), according to the manufacturer’s instructions. HL-60 cells were plated in a 24-well plate (Corning, NY, USA) at a density of 5.0 × 105 cells/well. After 24 h, the cells were treated with the test compound for the indicated time period (h). Subsequently, the cells were washed thrice with phosphate buffered saline (PBS (−)) and were lysed using the lysis buffer supplied in the kit. Caspase-3 activity in the cell lysates was then measured using the Wallac 1420 ARVOsx microplate counter (excitation at 390 nm, emission at 510 nm; PerkinElmer, Waltham, MA, USA).

Detection of nuclear condensation

HL-60 cells were plated in a 24-well plate at a density of 8.0 × 105 cells/0.3 ml/well and treated with 10 μM erypoegin K for 6 h. The cells were collected into a 1.5-ml microtube by a brief centrifugation, with Hoechst 33342 reagent (Dojindo, Kumamoto, Japan) added (final concentration: 14.3 μg/ml), and stained for 30 min in a CO2 incubator at 37 °C. Subsequently, the cells were fixed with the same volume of 10% paraformaldehyde in PBS (−) at room temperature for 30 min. Then, the cells were collected by a brief centrifugation and washed twice with PBS (−). Finally, cells were mounted with the VECTASHIELD™ medium (Vector laboratories, Burlingame, CA, USA), transferred to a slide glass, and observed under a fluorescence microscope (excitation at 352 nm, emission at 461 nm; Zeiss Axiophoto 2, Carl Zeiss, Jena, Germany).

DNA fragmentation assay

HL-60 cells were plated in a 3.5-cm dish (Corning, NY, USA) at a density of 2.0 × 105 cells/ml/dish and treated with 10 μM erypoegin K for 10 h. The cells were collected into 1.5-ml microtubes by a brief centrifugation. Genome DNA was extracted from the cell pellet using an Apoptosis Ladder Detection Kit Wako (Wako Pure Chemical Industries, Osaka, Japan), according to the manufacturer’s instructions. The DNA fragments were separated by 1.5% agarose gel electrophoresis, stained with SYBR™ Green I, and observed with an UV transilluminator.

MG assay

MG levels in the culture medium and the cell lysate were determined as previously described [34] with some modifications. HL-60 cells were plated in a 6-cm dish (Corning, NY, USA) at a density of 1.5 × 106 cells/dish for 24 h and cultured in the presence of the test compound for the indicated time period (h). Subsequently, the cells were centrifuged (153 × g for 6 min), and the culture medium was collected and deproteinized using perchloric acid (final concentration: 0.5 M). The cells were washed thrice with PBS (−), homogenized by sonication (ultrasonic homogenizer VP-050, TAITEC, Saitama, Japan), and treated with perchloric acid (final concentration: 0.5 M). The samples were purified using a Sep-Pak Light tC18 cartridge (Waters, Milford, MA, USA). The flow-through fraction was collected, and MG was derivatized into 2-methylquinoxaline (2-MQ) by o-phenylenediamine for 4 h at room temperature. Following the addition of 5-methylquinoxaline (5-MQ) as the internal standard, the reaction solution was applied to the Sep-Pak Light tC18 cartridge. Bound 2-MQ and 5-MQ were eluted from the cartridge using acetonitrile, and the eluent was concentrated to approximately 100 µl using a SpeedVac SPD1010 system (Thermo Fisher Scientific, Waltham, MA, USA). 2-MQ and 5-MQ levels were measured using an HPLC platform (Jasco, Hachioji, Japan) equipped with an UV/VIS detector (UV-1570, wavelength: 315 nm; Jasco). The reversed-phase Develosil C30-UG-5 column (4.6 × 150 mm, 5 µm; Nomura Chemical, Aichi, Japan) was used, and the mobile phase comprised 10 mM KH2PO4 (pH 2.5; 80%) and acetonitrile (20%).

Statistical analysis

The significance of differences was estimated using Student’s t test with Microsoft Excel (Excel 2013). p < 0.05 was considered to be significant.