Requirement of ABC transporter inhibition and Hoechst 33342 dye deprivation for the assessment of side population-defined C6 glioma stem cell metabolism using fluorescent probes

C6 cells were maintained under 5 % CO₂ at 37 ̊C in Dulbecco’s modified Eagle’s medium (DMEM, Wako) supplemented with 10 % fetal bovine serum (FBS) and 50 U/ml penicillin and 50 μg/ml streptomycin (Gibco). Detection and isolation of SP cells were performed as previously described [5]. Briefly, C6 cells were harvested by trypsinization, resuspended in 2 % FBS-DMEM at a final density of 1 × 10⁶ cells/ml, and incubated in dark with 5 μg/ml Hoechst 33342 dye for 90 min at 37 ̊C in the presence or absence of 50 μM verapamil. One μg/ml propidium iodide was used to discriminate dead cells. SP and MP cells were separated by FACS Aria II (BD Biosciences) equipped with 355 nm UV laser. The fluorescence of Hoechst 33342 dye was monitored through 405/20 nm band pass filter and 670 nm long pass filter.

Three fluorescent probes were purchased from Molecular Probes, dissolved in dimethylsulfoxide (DMSO) (100 mM, 500 μM, and 1 mg/ml stock concentrations, respectively), and stored at −20 ̊C in dark. Experimental conditions were indicated as follows: cells were resuspended in 2 % FBS-DMEM at a density of 1 × 10⁵ cells/ml and incubated at 37 ̊C with 5 μM CellROX for 30 min, 100 nM MTG for 30 min, or 1 μg/ml JC-1 for 40 min. The fluorescent probes were excited with 488 nm (for MTG and JC-1) or 633 nm (for CellROX) laser, and monitored through 530/30 nm band pass (for MTG and JC-1 monomers), 575/25 nm band pass (for JC-1 aggregates) or 660/20 nm band pass (for CellROX) filter. The mean fluorescent intensity (MFI) was calculated using FlowJo software version 7.6.5 (TOMY Digital Biology).

All data are shown as the means ± standard deviation (SD) from three independent experiments. Significance between experimental groups was determined by Student’s t-test. A two-tailed value of less than 0.05 was considered significant.

To examine whether fluorescent probes are applicable to the assessment of CSC metabolism, we first sorted SP and MP cells (defined as CSCs and non-CSCs, respectively) from the C6 glioma cell line by FACS. The sorted cells were subsequently stained with three fluorescent probes, CellROX (for ROS levels), MTG (for mitochondrial amount) or JC-1 (for mitochondrial membrane potential). To see the effect of ABC transporters known to be present at higher levels in SP cells than MP cells and expel the Hoechst 33342 dye, the staining process was done either in the presence or absence of 50 μM verapamil, an inhibitor of ABC transporters. The exclusion of Hoechst 33342 dye from C6 cells was negligible at this concentration of verapamil (data not shown). The treatment of verapamil significantly enhanced the CellROX fluorescence in SP cells, whereas the fluorescence of CellROX in MP cells was not affected (Fig. 1a and b), indicating that CellROX was excluded from SP cells through ABC transporters and that oxidative stress as measured by ROS formation in SP and MP cells was comparable. Similar results were obtained from experiments using MTG, which indicates that the mitochondrial amount is comparable in SP and MP cells (Fig. 1c and d). The mean fluorescent intensity (MFI) of JC-1 monomers and aggregates were 6.9-fold and 5.0-fold respectively lower in SP cells than MP cells without verapamil (Fig. 1e, f, and g). The MFI of JC-1 monomers in SP cells was significantly increased by verapamil treatment and C6 SP and MP cells displayed similar levels of fluorescent intensities of JC-1 monomers in the presence of verapamil (Fig. 1f, the third and fourth columns), which is comparable to that in MP cells without verapamil (Fig. 1f, the second column). The MFI of JC-1 aggregates in SP cells was significantly increased (by 193 %) by verapamil, and the MFI of JC-1 aggregates of MP cells was also increased to a lesser extent (by 26 %) (Fig. 1g). The MFI of JC-1 aggregates in SP cells was 46 % of that in MP cells in the presence of verapamil, whereas that in SP cells, in the absence of verapamil, was only 20 % of that in MP cells (Fig. 1g). These data indicate that the three metabolic indicators CellROX, MTG and JC-1 are excluded from SP cells through ABC transporters, and thus the influence of ABC transporters must be considered for the fluorescent probe-based assessments of the metabolic status of SP-defined CSCs. Considering the lower level of JC-1 aggregates in SP cells in the presence of verapamil (Fig. 1g), C6 glioma CSCs may seem to have lower mitochondrial membrane potential than non-CSCs.

Because the amounts of Hoechst 33342 dye in sorted SP and MP cells are different, with the latter being very high, the influence of Hoechst 33342 dye on the fluorescent intensity of JC-1 aggregates was further examined. As shown in Fig. 2a (upper panels), C6 cells stained with Hoechst 33342 dye displayed significantly higher levels of JC-1 aggregates. The difference was approximately 3.2 fold (Fig. 2b, left two columns). To explain this undesirable effect of Hoechst 33342 dye, we tested three hypotheses; 1) 355 nm UV laser for Hoechst 33342 might unexpectedly excite JC-1 aggregates, 2) fluorescence emitted from the UV-excited Hoechst 33342 might excite JC-1 aggregates, and 3) 488 nm argon laser for JC-1 excitation might unexpectedly excite Hoechst 33342 dye. However, even when UV laser was shut off, the fluorescent intensity of JC-1 aggregates in cells never decreased regardless of the staining without or with Hoechst 33342 and the MFI in the latter was always high (Fig. 2a, left panels and b). It is unknown why MFI of JC-1 aggregates is somewhat enhanced (by 11 %) following the shut off of the UV laser (Fig. 2b, second and fourth columns), but at least it is obvious that 355 nm UV laser does not excite JC-1 aggregates and UV laser-excited Hoechst 33342 does not excite JC-1 aggregates. When C6 cells were stained with Hoechst 33342 dye alone and excited by the argon laser, the emitted fluorescence from the cells is negligible through the filter for JC-1 aggregates (Fig. 2a, right panels). Finally, a possible influence of Hoechst 33342 staining on cell size and granularity, which might affect JC-1 aggregation, was examined. As shown in Fig. 2c, staining with Hoechst 33342 dye did not affect the size and granularity of cells (Fig. 2c). Therefore, we cannot yet figure out how Hoechst 33342 dye influences the fluorescence of JC-1 aggregates, but at least it is obvious that Hoechst 33342 staining interferes with the assessment of mitochondrial membrane potential using JC-1.

To achieve the accurate assessment of the mitochondrial membrane potential using the JC-1 dye, sorted SP and MP cells were recultured for 6 days to deprive intracellular Hoechst 33342 dye and subjected to staining with JC-1. These cells were confirmed to possess no or negligible Hoechst 33342 dye (Fig. 3a). The recultured SP and MP cells were then stained with Hoechst 33342 again to know SP/MP profiles of these cultured cells. As shown in Fig. 3b, 68.9 ± 1.7 % of SP-derived cells retained the SP phenotype, and MP-derived cells displaying SP phenotype were negligible (0.58 ± 0.48 %). As shown in Fig. 3c (upper panels) and d, the recultured MP cells deprived of initial Hoechst 33342 dye displayed higher levels of fluorescence whose wave length corresponds to JC-1 aggregates than SP cells in the absence of verapamil, as expected. However, SP- and MP-derived recultured cells exhibited similar levels of fluorescence corresponding to JC-1 aggregates in the presence of verapamil (Fig. 3c lower panels and d), suggesting that mitochondrial membrane potential of C6 CSCs are comparable to that of C6 non-CSCs. Taken together, our data indicate that the activity of ABC transporters and the presence of Hoechst 33342 dye in cells definitely interfere with the estimation of the metabolic status when fluorescent probes are applied to SP-defined CSCs.