Fetal calf serum, Dulbecco’s modified Eagle’s medium (DMEM), and tissue culture reagents were purchased from Invitrogen Corporation (Carlsbad, CA, USA). Carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP), 2-aminoethoxydiphenyl borate (2-APB) and cyclosporin A were from Sigma-Aldrich chemical Co. (St. Louis, MO, USA). JTV-519 and MCU-i4 were from Tocris BioScience (Bristol, UK). All other chemicals were of reagent grades and were from Sigma-Aldrich. BT474 cells were purchased from American Type Culture Collection (Manassas, VA, USA), and were cultured in RPMI-1640 medium supplemented with L-glutamine (2 mM), 10% fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 μg/mL) at 37°C (98.6°F) with 5% humidified CO₂.

BT474 cells were cultured in 96-well plates at a density of 1.5 × 10⁴ and were treated with different agents for 48 h. DMSO was added to the medium as solvent control (final concentration = 0.1%). 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenltetrazolium bromide (MTT; final concentration of 0.5 mg/mL) was added to each well and further incubated for 4 h. Culture medium was discarded and DMSO (100 μL) was added to each well for another 15 min with mild shaking to dissolve precipitates. Absorbance at 595 nm was measured by an ELISA reader; absorbance was used to indicate cell viability or metabolic activities. A reduction in MTT absorbance indicates cell death, reduced cell proliferation or metabolic activities. Number of viable cells was quantified by the trypan blue exclusion method: viable cells were unstained by trypan blue and were counted with a hemocytometer. An FITC annexin V apoptosis detection kit (BioLegend, San Diego, CA, USA) and flow cytometer (BD Biosciences, San Jose, CA, USA) were used to quantify apoptosis. Caspase-9 level was measured using an ELISA kit (cat. # E-EL-H0663; Elabscience, Houston, TX, USA) following the instructions in the manufacturer’s manual.

Cytosol Ca²⁺ concentration was measured using fura-2 as a fluorescent probe [22]. Cells were incubated with 5 μM fura-2 AM (Invitrogen) at 37°C (98.6°F) for 1 h; cells were then washed in bath solution (mM): 140 NaCl, 2 CaCl₂, 1 MgCl₂, 4 KCl, 10 HEPES (pH was adjusted to 7.4 by NaOH). Intracellular Ca²⁺ release was examined in Ca²⁺-free bath solution; the latter being the same as the bath solution, except that Ca²⁺ was removed and EGTA (100 μM) was added. Cells were excited by 340 and 380 nm alternately (switching frequency = 1 Hz) by an optical filter changer (Lambda 10-2, Sutter Instruments, Novato, CA, USA). Emission was collected at 500 nm. and data were captured by a CCD camera (CoolSnap HQ2, Photometrics, Tucson, AZ, USA), which was connected to a Nikon (Tokyo, Japan) TE2000-U microscope. Data were analyzed with an MAG Biosystems Software (Sante Fe, MN, USA). Experiments were conducted at 25°C (77°F). Changes in 340/380 ratio were analyzed at a region of interest of single cells; the same experimental procedures were repeated multiple times to obtain the mean.

Microfluorimetric quantification of Ca²⁺ concentration within the mitochondrial matrix was performed using a Ca²⁺-sensitive dye Rhod-2 AM [23]. The cells were incubated with 5 μM Rhod-2 AM (Invitrogen, Carlsbad, CA, USA) at 37°C (98.6°F) for 1 h. Cells were then permeabilized and washed with a digitonin (30 μM)-containing intracellular solution, which contained (mM): 140 KCl, 8 NaCl, 1 CaCl₂, 1 MgCl₂, 1.85 EGTA, 10 HEPES, and 8 MgATP (KOH was used to adjust pH to 7.25). Free [Ca²⁺] in this intracellular solution was calculated to be 114 nM. Excitation wavelength was 540 nm and emission wavelength was 605 nm. Images were captured by a CCD camera (CoolSnap HQ2, Photometrics, Tucson, AZ, USA) connected to an inverted microscope (Nikon TE 2000-U). An MAG Biosystems Software (Sante Fe, MN) was used for analysis. All experiments were performed at room temperature (25°C) (77°F).

Mitochondrial membrane potential was measured by a mitochondrial membrane potential assay kit (#12664; Cell Signaling, Danvers, MA, USA) as described in a previous report [24]. The cells were plated at a density of 5 × 10⁴ cells per well to settle overnight. The cells were treated with DMSO or other agents for 24 h. JC-1 (2 μM) was added to each well for 30 min. Fluorescent emission was measured by a Varioskan LUX multimode microplate reader (Thermo Fisher Scientific, Waltham, MA, USA). Excitation wavelength was at 485 nm and dual emission wavelengths were at 520 and 590 nm. Mitochondrial membrane potential was calculated as the ratio RFU of red emission (590 nm)/RFU of green emission (520 nm).

ROS was measured as described in a previous report [25]. The cells were treated with various agents for 4 h, and were then incubated in serum-free DMEM supplemented with 2,7-dichlorodihydrofluorescein diacetate (DCFH2-DA, 20 μM, Sigma, St. Louis, MO, USA) at 37°C (98.6°F) for 30 min in darkness. Cells were then washed, trypsinized at 37°C for 3 min, and washed again three more times in phosphate-buffered saline using centrifugation. The cells were then dispersed in phosphate-buffered saline, put in polystyrene tubes for FACS (fluorescence-activated cell sorting). The samples (1 × 10⁵ cells/sample) were analyzed by an FACS Canto flow cytometer system (BD Biosciences, San Jose, CA, USA). Data analysis was performed with the aid of a BD FACSDIVA™ software (BD Biosciences).

Cellular ATP content was quantified as described in a previous report [26]. The cells were seeded in 10-cm dishes at a density of 6 × 10⁶ cells per dish and treated with DMSO or 30 μM MCU-i4 for 24 h. The cells were subsequently trypsinized and cell viability was quantified by the trypan blue exclusion method. The cells were sonicated (33 Hz, 90 s) in ice bath, centrifuged (10,000 g, 10 min) and 30 μL supernatant was taken for ATP quantification using an ATP assay kit (catalogue # E-BC-K57-M; Elabscience, Houston, TX, USA). Samples were analyzed by a BioTek Epoch 2 microplate spectrophotometer (Winooski, VT, USA).

Quantification of lactate was used as an indicator of glycolysis [27]. The cells were seeded in 96-well plates at a density of 1 × 10⁴ cells/per well overnight. The cells were then treated with DMSO or 30 μM MCU-i4 for 3 h, and 5 μL supernatant was taken for lactate quantification using a glycolysis assay kit (catalogue # ECGL-100; BioAssay Systems, Hayward, CA, USA), and samples were analyzed by a BioTek Epoch 2 microplate spectrophotometer (Winooski, VT, USA).

Statistical analysis and graphing were performed using Origin8.5 (OriginLab, MA, USA). Results are presented as means ± standard error of mean (S.E.M.). Paired or unpaired Student t-test was employed where appropriate to compare two groups. When multiple groups were analyzed, ANOVA and the Tukey’s HSD post-hoc test were used. Statistical significance was considered to be reached if the p-value is less than 0.05.

Treatment of BT474 cells with MCU-i4 (3–30 μM) for 2 days resulted in a concentration-dependent decrease in cell viability (Fig. 1A). Since the MTT assay shown in Fig. 1A could not discriminate between cell death and reduced cell proliferation, we used the trypan blue exclusion method to quantitate the number of viable cells (Fig. 1B). Cell proliferation was suppressed by 3 μM MCU-i4, while higher concentrations (10–30 μM) concentration-dependently caused cell death. We investigated whether cell death was necrotic or apoptotic. There was a 10-fold increase in annexin-positive/propidium iodide-negative cells (Q4), suggesting early apoptosis had taken place (Fig. 1C). Caspase-9 has been known as a marker of apoptosis [28]. Consistently, MCU-i4 treatment also resulted in a moderate increase in the level of caspase-9 (Fig. 1D).

We investigated whether MCU-i4 affected cytosolic Ca²⁺ levels. MCU-i4 did not cause an immediate elevation in cytosolic Ca²⁺ concentration [Ca²⁺]_i (Fig. 2A). We then investigated whether a prolonged MCU-i4 pre-treatment (25 min) would affect the Ca²⁺ level. The cells in Ca²⁺-containing bath solution were treated with DMSO or MCU-i4 for 25 min (kept in the dark to avoid photobleaching) prior to microfluorimetric measurement. An elevated Ca²⁺ baseline was observed in the MCU-i4-treated cells (Fig. 2B). Treatment with MCU-i4 for 24 h (followed by fura 2 loading and microfluorimetric measurements) also resulted in an elevated Ca²⁺ baseline, suggesting prolonged MCU-i4 treatment raised Ca²⁺ concentration in the cytosol (Fig. 2C). We repeated the above protocol (25 min) in Ca²⁺-free bath solution (Fig. 2D). An elevated Ca²⁺ baseline was again observed in the MCU-i4-treated cells, suggesting the raised Ca²⁺ concentration in the cytosol was in part due to Ca²⁺ release. Inhibition of inositol 1,4,5-trisphosphate receptors (IP3R) by 2-APB strongly suppressed the elevation of Ca²⁺ baseline, while inhibition of ryanodine receptors (RYR) by JTV-519 only mildly alleviated it, suggesting the Ca²⁺ leak was mainly via IP3R and in part via RYR.

Ca²⁺ release from intracellular Ca²⁺ store caused elevation in cytosolic Ca²⁺ level under inhibition of mitochondrial Ca²⁺ uptake by MCU-i4 (Fig. 2). To show that inhibition of mitochondrial Ca²⁺ uptake by MCU-i4 led to a decrease in mitochondrial Ca²⁺ concentration, we used Rhod-2, a fluorescent probe for Ca²⁺ concentration in the mitochondrial matrix (Fig. 3). In the control where DMSO was added, there was a slow decrease in fluorescence which was due to inevitable photobleaching of the fluorescent dye; addition of MCU-i4 caused an immediate and persistent decrease in fluorescence when compared to the control, indicating a decrease in mitochondrial matrix Ca²⁺ concentration. This result, together with the data in Fig. 2, suggest that upon MCU-i4 inhibition of mitochondrial Ca²⁺ uptake, Ca²⁺ released from Ca²⁺ store failed to enter mitochondria and thus “spilled over” in the cytosol.

We next examined whether MCU-i4-induced lowered mitochondrial matrix Ca²⁺ level would affect ATP production. MCU-i4 treatment for 24 h resulted in a 51.9 ± 5.8% decrease in viable cell count (trypan blue exclusion test) but only moderately reduced ATP production by 23.9 ± 9.8%. When ATP production was normalized by the number of viable cells, MCU-i4 treatment significantly enhanced ATP production (Fig. 4A). Since lowered mitochondrial matrix Ca²⁺ level was not compatible with increased mitochondrial ATP production, we examined whether increased ATP production resulted from increased glycolytic activities. As shown in Fig. 4B, MCU-i4 caused a 1.6-fold elevation in secreted lactate concentration, which indicated an increase in glycolysis. Whether MCU-i4 elicited ROS formation was next examined. As shown in Fig. 5, treatment of cells with MCU-i4 for 4 h resulted in large production of ROS.

Using JC-1 as a fluorescent probe of mitochondrial membrane potential, we examined if MCU-i4 would depolarize mitochondrial potential. MCU-i4 at 10 μM caused marked depolarization whilst at 30 μM it caused collapse of mitochondrial membrane potential to an extent comparable to that caused by FCCP (Fig. 6A). Since mitochondrial membrane potential collapse could be a result of mitochondria permeability transition pore (MPTP) opening and that cyclophilin D is an integral part of the MPTP, we examined if cyclosporin A (an MPTP inhibitor by interacting with cyclophilin D) could reduce MCU-i4 cytotoxicity. However, MCU-i4-inflicted cell death was not prevented by cyclosporin A (Fig. 6B).