Metformin and phenethyl isothiocyanate combined treatment in vitro is cytotoxic to ovarian cancer cultures

Metformin, PEITC, cisplatin, and most other reagents were purchased from Sigma Aldrich (St. Louis, MO). Tissue culture media, DMEM and RPMI were purchased from Thermo-Fisher Scientific (Waltham, MA). Fetal bovine serum (FBS) was purchased from Atlanta Biologicals (Atlanta, GA). Fluorescent dyes: MitoSOX, 2’,7’-dichlorodihydrofluorescein (DCF), dihydroethidine (DHE), dihydrorhodamine 123 (DHR), Sytox green, and MitoTracker Green were all purchased from Invitrogen (Grand Island, NY). Antibodies to detect AMPK, PARP, cleaved PARP were purchased from Cell Signaling Technology (Beverly, MA). Antibodies to detect MnSOD were purchased from BD Pharmingen (San Diego, CA).

Cell lines were incubated in a 37 C, 5% CO2 humidified chamber. OVCAR3, CAOV3, SKOV3, and PA-1 were obtained from ATCC. A2780 and A2780cis were received as a kind gift from Dr. Subhash Chauhan. OVCAR3, CAOV3, SKOV3, and PA-1 were maintained in DMEM supplemented with 10% fetal bovine serum, 100 units/mL penicillin, 100 μg/mL streptomycin, and 250 ng/mL amphotericin B. A2780 was maintained in RPMI with 10% FBS, 100 units/mL penicillin/streptomycin, and 250 ng/mL amphotericin B. A2780cis cells were maintained in complete RPMI media supplemented with 1 μM cisplatin every 3rd passage to maintain selection for cisplatin resistance.

Multiple methods were used in cell counting. For cells counted by glass hemacytometer with Trypan blue exclusion, adherent cells were harvested by trypsinization. For live and dead cell counts, media was also collected and added to the cell suspension. Cells were pelleted by centrifugation (200 x g) and each sample was resuspended in an equal volume. Samples were then mixed with an equal volume of 0.4% Trypan blue and counted using a glass hemacytometer. Cells excluding Trypan blue were counted as live, stained cells were counted as dead. Sytox green (Invitrogen) fluorescent dye was also used for dead cell analysis. In a 96-well plate format, cells were plated in phenol red free media 24 hours prior to treatment with drugs of interest. At desired time points, Sytox green dye was added to wells at 1 μM final concentration and incubated for 10 minutes to stain DNA of dead cells. Fluorescence was measured using a Spectramax M5 fluorescence plate reader at excitation 485, emission 530, cutoff 515. Total cell fluorescence was measured by adding Triton X-100 at a final concentration of 0.1% to all wells, incubating for 10 minutes, and remeasuring Sytox fluorescence.

Cells were plated at 300–500 cells per 100 mm dish 24 hours before beginning treatment conditions. Cells were allowed to grow for 7–10 days before fixation. Colonies were fixed using 70% ethanol, and stained using 1% crystal violet dissolved in 10% ethanol. Number of colonies was obtained using image analysis software (AlphaInnotech). Colonies were defined as a minimum of 50 cells in a group.

Samples were harvested in an SDS based buffer, sonicated and heated before running. Proteins were separated by SDS-PAGE and transferred to PVDF membrane by semi-dry transfer apparatus. Antibodies for detection of proteins of interest were diluted in blocking buffer, 5% BSA in TBST. Signal was detected using HRP-conjugated secondary antibodies and Amersham ECL prime detection reagent (GE healthcare). Exposures were captured either by photographic film or CCD camera imaging system (UVP).

After varying treatment periods, adherent cell cultures were detached with trypsin and transferred to culture tubes as a cell suspension. Cells were then incubated with fluorescent dyes to determine levels of oxidative stress products. MitoSOX (Invitrogen) was used to measure mitochondrial specific superoxide. Dihydroethidine (DHE) was used to measure total cellular superoxide. Absorbance was measured with the FL2 channel (Ex / Em) of an Accuri C6 flow cytometer. Data was collected and analyzed using Cflow Plus software (BD-Accuri).

Cells were cultured on a chambered coverglass (Thermo scientific) in phenol red free DMEM with 10% FBS. After completing desired drug treatment courses, cells were stained with dyes of interest in phenol red free DMEM with 5% FBS. MitoSOX was used at 5 μM final concentration and MitoTracker green at 100 nM final concentration. Cells were incubated with dye for 15 minutes at 37 °C, and then washed twice with PBS. Phenol free DMEM with 5% FBS was used during live cell imaging. An Olympus FV1000 imaging system on an Olympus IX81 inverted microscope with heated 5% CO₂ stage chamber was used for image capture. Image processing and analysis was completed using Olympus FluoView software.

Error bars displayed are standard deviations. Comparison of two groups was done using a two-tailed student’s t-test. Analysis of multiple groups was performed by ANOVA or Chi-squared test using R, the statistical computing program.

CalcuSyn (BioSoft) was used to calculate combination index values for drug combinations. Drug combinations were given at constant ratios and effect measured was either cell death or cell growth inhibition.

Across many human epithelial ovarian cancer cell lines, metformin was found to significantly lower cell proliferation (Figure 1A). This effect was greater over longer treatment periods (Compare 48 hour treatment Figure 1A and 24 hour treatment 5A below). Live cell proliferation was measured using a glass hemacytometer and Trypan blue exclusion. With drugs of interest that target cellular metabolism, direct live cell counting using a glass hemacytometer is preferred over assays such as the MTT which rely on cellular respiration to indirectly measure number of cells. In addition to inhibiting cell growth, anchorage dependant colony formation was strongly inhibited by metformin (Figure 1C and 1D). In the cisplatin resistant cell line SKOV3, metformin was still effective at preventing colony formation (Figure 1C).

When the same cell lines were treated with PEITC, their growth was also significantly inhibited (Figure 1B). While metformin is highly effective at reducing cell proliferation, cell death is not induced. On the contrary, PEITC was cytotoxic to many ovarian cancer cell lines (Figure 2). Cell death induced by PEITC was dose dependent and significantly higher than controls in the low micromolar range. Examination of PARP showed a shift from intact full length PARP to the cleaved form when cells were treated with PEITC (Figure 3). This shift suggests apoptosis as the primary cause of cell death for PEITC. PARP cleavage was not observed in cells treated with metformin over the same time period. This is consistent with the results showing a lack of metformin-induced cell death in the cell lines examined.

Because previous research has found metformin to inhibit complex I and PEITC to inhibit complex III of the mitochondrial electron transport chain, ROS levels were examined as a potential mechanism for each drug's anti-cancer effects. Inhibition of complex I increases the aberrant flow of electrons to oxygen and creates superoxide within the mitochondrial matrix. Inhibition of complex III also increases superoxide production, but can generate these ROS species in either the cytosol or mitochondria because of its position in the mitochondrial membrane [18]. When live cells were treated with metformin or PEITC, they showed increased ROS as measured by MitoSOX (Figure 4A). MitoSOX is an ethidine based mitochondrial targeted probe that preferentially reacts with superoxide to produce a fluorescent product. Dihydroethidine (DHE) is a non-targeted ROS detector and can assess total cellular oxidative stress. In cells treated with metformin, fluorescence of DHE as detected by flow cytometery was also increased (Figure 4B). These results suggest that previously identified mechanisms of ETC complex inhibition by both metformin and PEITC results in an increase of ROS within the cancer cell.

In order to determine the role of ROS in each drug's effects on cell proliferation or apoptosis, ROS scavenging compounds were used. When the ROS scavenger N-acetyl-L-cysteine was used to pretreat cancer cells before addition PEITC, cell death induction by the drug was completely reversed (Figure 5). This finding suggests PEITC is heavily dependent on ROS production and oxidative stress to induce apoptosis.

While PEITC-induced cell death is reversed by addition of a ROS scavenger, the effects of metformin on cell proliferation are not strongly affected. This may be due to metformin's general action of growth inhibition versus PEITC's induction of apoptosis. Metformin's complex mechanism, working through several pathways may also play a role. In situations where complex I inhibition does not produce high amounts of ROS to damage a cancer cell's mitochondria, metformin's action through AMPK and mTOR may still allow growth inhibition to be achieved. However, treatment of cells with methyl succinate partially reversed the ability of metformin to inhibit growth of PA-1 cells (Figure 6). Methyl succinate can bypass electron transport through complex I because it donates electrons directly to complex II of the mitochondrial electron transport chain. These results suggest inhibition of electron flow may be responsible for some of metformin's growth inhibition effects. As metformin concentration increases, the effect of methyl succinate is overcome, which is likely a result of metformin-induced activities such as AMPK activation.

The results above indicate that metformin and PEITC function through different mechanisms and have different consequences on the growth and survival of ovarian cancer cells. Therefore we hypothesized that the two drugs would act synergistically when used in combination. In the cell line CAOV3, combination therapy significantly increased cell death over either drug individually (Figure 7A). Attaining the high levels of cell death observed for the combined drugs requires much higher doses with the individual drugs. Furthermore, this combination continues to be effective in the cisplatin resistant cell line A2780cis. Even high doses of cisplatin were not able to match the cell death obtained through metformin and PEITC combination in the A2780cis cell line (Figure 7B). A Combination index (CI) is a value used to quantify the degree of synergy or antagonism in drug combinations[19]. As a general guideline, CI < 0.7 is synergistic, 0.7 > CI <1.2 is additive, and CI > 1.2 is antagonistic. When determined using CalcuSyn software (BioSoft), the metformin and PEITC combination was found to be synergistic at several effective doses in the cisplatin resistant line A2780cis (Figure 7C). Calculated synergy is higher the closer the CI is to 0. With a CI of 0.00015 at ED50 metformin and PEITC doses, the combination is highly synergistic and increases with higher ED combinations. In this cell line, cisplatin remains ineffective at inducing cell death at doses as high as 33 μM or 10 μg/mL.