H4IIE rat hepatocellular carcinoma cells were obtained from the Korean Cell Line Bank (Seoul, Korea) and maintained in Dulbecco's minimal essential medium (DMEM, 1 g/L glucose) with 10% fetal bovine serum (FBS). At the beginning of the experiments, H4IIE cells were incubated in serum-free DMEM overnight. Cells were washed twice with Dulbecco's phosphate-buffered saline (D-PBS) and again incubated in serum- and glucose-free DMEM (GFM) supplemented with 2 mM pyruvate and 20 mM lactate for 30 minutes before treatment with reagents.

DMEM, Hank's-balanced salt solution (HBSS), D-PBS, trypsin-ethylenediaminetetraacetic acid solution, metformin, compound C, 3-(4,5-dimethylthiazol-2-Yl)-2,5-diphenyltetrazolium bromide (MTT), 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR), SB202190, SP600125, wortmannin, rapamycin, H33342, and acridine orange (AO) were purchased from Sigma Chemical Corp. (St. Louis, MO, USA). FBS was obtained from Life Technologies, Inc. (Rockville, MD, USA). Polyclonal antibodies against poly ADP ribose polymerase (PARP), cleaved caspase-3, autophagy-related (3, 5, 7), β-actin, and monoclonal antibodies against beclin-1 and microtubule-associated protein 1 light chain 3B (LC3B) were purchased from Cell Signaling Technology (Danvers, MA, USA). Polyclonal antibodies against phospho-AMPK were obtained from Millipore (Billerica, MA, USA). Polyclonal anti-PARP and anti-rabbit goat immunoglobulin G-horseradish peroxidase (HRP) secondary antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Electrophoresis reagents (including Bis-Tris gels, running buffer, and polyvinylidene difluoride [PVDF] membranes) were obtained from Invitrogen (Carlsbad, CA, USA).

Cell viability was analyzed using the MTT assay as previously described [16]. Briefly, H4IIE cells in 24-well plates were pre-incubated in serum-free DMEM overnight, washed twice with D-PBS and then further incubated for 24 hours after pretreatment and treatment with reagents in GFM. After the treatments, the same amount of MTT reagent (1 g/L in D-PBS) was added to each well. After a 30 minute incubation at 37℃, cells were washed with D-PBS, and the blue-colored formazan product was subsequently solubilized in 0.5 mL of 2-propanol for 20 minutes. The absorbance of the converted dye was measured at a wavelength of 570 nm.

H4IIE cells in 6-well plates were pre-incubated in serum-free DMEM overnight, washed twice with D-PBS, and then further incubated in GFM-containing reagents. After treatment, the cells were lysed in an ice-cold lysis buffer (50 mM Tris-HCl, 1% nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM sodium orthovanadate, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 mM aprotinin, 1 mM leupeptin, and 1 mM pepstatin A). Equal amounts (10 to 20 µg) of protein were separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis on 4% to 12% polyacrylamide gels and transferred onto PVDF membranes. The membranes were incubated in blocking buffer (5% nonfat dry milk in Tris-buffered saline [TBS]-0.1% Tween-20 [TBS-T]) for 1 hour at room temperature, after which the membranes were probed with different primary antibodies (at dilutions of 1:1,000 to 1:2,000). After a series of washes, the membranes were further incubated with the respective HRP-conjugated secondary antibodies at a dilution of 1:10,000. The signal was detected using the enhanced chemiluminescence detection system (Intron, Seongnam, Korea).

Acidic intracellular vesicles, a marker of the onset of autophagy were visualized using AO staining. After treatment, the cells were washed twice with D-PBS and stained in 2% acetone in D-PBS containing AO (1 µg/mL) for 15 minutes at room temperature. To observe the degree of nuclear condensation (a marker of apoptosis), H33342 (1 µg/mL) was added alone or in combination with AO. After removal of the AO/H33342 staining solution, the cells were immediately observed under a fluorescent microscope (IX70; Olympus, Tokyo, Japan) and imaged using a digital camera (DP-70; Olympus).

The experimental results are presented as the mean±standard error. The significance of the differences among groups was determined using Student t-test. P<0.05 was considered statistically significant.

Our preliminary results showed that the viability of H4IIE cells was not significantly affected within the tested ranges of the glucose concentration (0 to 5.5 mM) after 24 hours (data not shown). Thus, we regarded the cells incubated in GFM as the control group in the experiments. The culture conditions mimic glucose-deprived environments for proliferating cancer cells with respect to the induction of apoptosis and/or autophagy.

To test whether metformin affects the viability of hepatocellular carcinoma cells, H4IIE cells were treated with metformin in GFM. To remove the residual proliferative effects of serum, cells were pre-incubated in serum-free DMEM for 24 hours and then treated with metformin and AICAR, typical activators of AMPK, in fresh GFM. Stimulation of AMPK was detected using Western blotting analyses for phospho-AMPK after 4 hours of treatment (Fig. 1A). While lower doses of metformin (0.2 mM) or AICAR (0.1 mM) did not change the levels of phospho-AMPK, 1 mM AICAR weakly increased the levels of p-AMPK (approximately 1.3-fold over the non-treated control) and 2 mM metformin strongly increased the levels of p-AMPK (2.2-fold over the control). Pretreatment with compound C (10 µM), an inhibitor of AMPK, blocked the stimulation of AMPK by metformin or AICAR. Cell viability was measured using the MTT assay after 24 hours of treatment with metformin and AICAR. AICAR (1 mM) decreased MTT reactivity, but was significantly restored by pretreatment with compound C (50.44%±3.53% vs. 72.56%±0.88% of the non-treated control, respectively) (Fig. 1B). Metformin (2 mM) markedly decreased MTT reactivity, and such a decrease was also significantly restored by pretreatment with compound C (10.79%±1.13% vs. 57.8%±1.73% of the non-treated control, respectively) (Fig. 1C). These results indicated that pharmacological stimulation of AMPK by AICAR or metformin exerts cytotoxicity in H4IIE hepatocellular carcinoma cells. After determining that stimulation of AMPK and reduction of MTT reactivity were stronger with metformin (2 mM) than AICAR (1 mM) treatment, we next examined whether the decrease in MTT reactivity caused by metformin reflected an induction of apoptosis via the stimulation of AMPK. Induction of apoptosis was detected using fluorescent staining of condensed nuclear chromatin (apoptotic bodies) and Western blotting analyses for cleaved caspase-3 and PARP. In the control group (cells pre-incubated in serum-free DMEM for 24 hours and further incubated in GFM for an additional 24 hours), little change was observed in the degree of apoptotic body formation and the cleavage of caspase-3 and PARP (Fig. 1D and E). The addition of 2 mM metformin increased the number of apoptotic bodies (Fig. 1D) and the amount of cleaved caspase-3 and PARP (Fig. 1E). Compound C also suppressed the effects of metformin.

Recent studies have suggested that metformin induces autophagy in various cells [171819]. However, it is unknown if metformin (or AMPK) modulates autophagy with respect to its anti-proliferative activity in hepatocellular carcinoma. Because autophagy is initiated by the depletion of energy sources, including glucose [2021] or amino acids [22], we tested whether deprivation of glucose or amino acids induced autophagy in H4IIE cells. In addition, the effect of metformin on the induction of autophagy was examined. As shown in Fig. 2A, deprivation of glucose did not enhance the formation of acidic intracellular vesicles or apoptotic bodies per se. However, compound C stimulated the formation of acid vesicles (yellow fluorescence within cells) in most of the treated cells. The number of apoptotic cells increased with metformin, while pretreatment with compound C completely reversed the effect of metformin. In an enlarged fluorescent image, acidic vesicles were observed only in healthy, non-apoptotic cells, but were not observed in apoptotic cells (arrows in Fig. 2B) when co-treated with metformin and compound C in GFM. Compared to the results observed with glucose treatment, deprivation of amino acids (incubation in amino acid-free HBSS containing 5.5 mM glucose) markedly increased the number of apoptotic cells (Fig. 2C). In amino acid-deprived cells, metformin did not affect the viability of cells compared to the control, but AICAR accelerated cell death (Fig. 2D). However, in amino acid-fed cells (together with 5.5 mM glucose), metformin decreased cell viability and the cells that were sensitive to compound C (Fig. 2D). Although compound C enhanced the formation of acidic vesicles within cells, it failed to rescue cells from apoptosis in all of the treatments evaluated (Fig. 2C and D). These results indicated that metformin induced apoptotic cell death and inhibited autophagy in glucose-deprived environments, but not in amino acid-deprived conditions in H4IIE cells. Moreover, compound C suppressed the pro-apoptotic activity of metformin, indicating a role for AMPK in the control of apoptosis/autophagy in H4IIE cells.

Although compound C stimulated the formation of acidic intracellular vesicles, it is unknown whether metformin inhibits autophagy in an AMPK-dependent manner. Thus, the effect of metformin on the expression of various autophagy-related proteins (Atg3, Atg5, Atg7, Atg12, LC3B, and beclin-1) was examined. Serum-starved (for 24 hours) H4IIE cells were treated with different concentrations of metformin (0.016 to 2 mM) for 24 hours in GFM. Western blotting analyses revealed that metformin decreased the expression levels of all six proteins tested in a dose-dependent manner (Fig. 3). Pretreatment with compound C blocked the metformin-induced decrease of LC3B and beclin-1 proteins. Because these two proteins are essential components in the formation of the autophagosome compared to the other four proteins, these results suggested that metformin inhibits autophagy in an AMPK-dependent manner.

Our results showed that metformin induces apoptosis and inhibits autophagy in an AMPK-dependent manner in H4IIE cells under glucose-deprived conditions. However, little evidence is available regarding the involvement of signaling pathways other than AMPK for the pro-apoptotic and anti-autophagic activities of metformin. We tested the involvement of two major signaling pathways with regard to metformin activity. Members of the MAPK and phosphatidylinositol-3-kinase (PI3K)/mammalian target of rapamycin (mTOR) families play diverse roles in controlling the viability of normal and cancer cells. First, inhibitors against MAPKs were pre-incubated for 30 minutes prior to metformin treatment for 24 hours. Only the inhibition of p38MAPK by 50 µM SB202190 significantly suppressed metformin's pro-apoptotic activity (Fig. 4A, cleavage of caspase-3; Fig. 4B, MTT reactivity), whereas inhibition of ERKs (extracellular signal-regulated kinases) by 10 µM U0126 or inhibition of JNKs (c-Jun N-terminal kinases) did not suppress the pro-apoptotic activity of metformin. SB202190 also restored the reduced expression of beclin-1 by metformin (Fig. 4C). However, the reduced expression of LC3B by metformin was not restored by SB202190, whereas it was restored by compound C (Fig. 3). Taken together, these results suggest that a member of the MAPK family, p38MAPK, combined with AMPK, also mediate metformin's pro-apoptotic and anti-autophagic activity in H4IIE cells. In addition, inhibition of PI3K and mTOR by wortmannin (100 nM) and rapamycin (100 nM) failed to change metformin's pro-apoptotic activity (Fig. 4D and E). On the basis of these results, p38MAPK combined with AMPK is likely a key player contributing to metformin's pro-apoptotic and anti-autophagic activities.