High glycolytic activity of tumor cells leads to underestimation of electron transport system capacity when mitochondrial ATP synthase is inhibited

This study sought to elucidate how oligomycin, an ATP synthase blocker, leads to underestimation of maximal oxygen consumption rate (maxOCR) and spare respiratory capacity (SRC) in tumor cells. T98G and U-87MG glioma cells were titrated with the protonophore CCCP to induce maxOCR. The presence of oligomycin (0.3–3.0 µg/mL) led to underestimation of maxOCR and a consequent decrease in SRC values of between 25% and 40% in medium containing 5.5 or 11 mM glucose. The inhibitory effect of oligomycin on CCCP-induced maxOCR did not occur when glutamine was the metabolic substrate or when the glycolytic inhibitor 2-deoxyglucose was present. ATP levels were reduced and ADP/ATP ratios increased in cells treated with CCCP, but these changes were minimized when oligomycin was used to inhibit reverse activity of ATP synthase. Exposing digitonin-permeabilized cells to exogenous ATP, but not ADP, resulted in partial inhibition of CCCP-induced maxOCR. We conclude that underestimation of maxOCR and SRC in tumor cells when ATP synthase is inhibited is associated with high glycolytic activity and that the glycolytic ATP yield may have an inhibitory effect on the metabolism of respiratory substrates and cytochrome c oxidase activity. Under CCCP-induced maxOCR, oligomycin preserves intracellular ATP by inhibiting ATP synthase reverse activity.

underestimation of SRC in tumor cell lines 14 . We therefore recently proposed that max OCR and SRC in tumor cells should preferably be estimated in the absence of oligomycin 14 .
The present study aimed to further characterize and identify the mechanisms that lead to the underestimation of max OCR and SRC in tumor cells when ATP synthase is inhibited. The results indicate that the inhibitory effect of ATP synthase blockers on max OCR induced by the protonophore CCCP in tumor cells is associated with high glycolytic activity and maintenance of intracellular ATP levels.

Results
Occurrence of oligomycin-induced underestimation of max OCR and SRC in T98G glioma cells under different experimental conditions. The concentration of oligomycin normally used in experimental protocols is 1 µg/mL, while the minimal concentration to inhibit ATP synthase completely in intact tumor cells is approximately 0.1 µg/mL 14,15 . A wide range of oligomycin concentrations (0.3, 1.0 and 3.0 µg/mL) was tested on OCR parameters in T98G cells. Similar underestimation of CCCP-induced max OCR and SRC was observed with the oligomycin concentrations tested (Fig. 1A-C). Figure 1D shows that the different oligomycin concentrations induced similar inhibitory effects on basal OCR, reflecting the fraction of oxygen consumption related to ATP synthesis and indicating that the oligomycin concentrations tested were equally efficient at inhibiting ATP synthase. Figure 1E depicts the effect of oligomycin on SRC when T98G glioma cells were incubated at a glucose concentration found under normoglycemic conditions (i.e., 5.5 mM), instead of 11 mM. Under this condition, the SRC value was 31.6 ± 4.2% lower when estimated in the presence of oligomycin than in the control (vehicle (DMSO) without oligomycin).
The influence of sodium bicarbonate (Fig. 2) and FBS (Fig. 3) on the oligomycin-induced underestimation of max OCR and SRC was also tested in T98G cells. Figure 2 shows the max OCR and SRC for intact T98G cells incubated in the presence or absence of sodium bicarbonate. Increased max OCR and SRC ( Fig. 2) was observed in the absence of sodium bicarbonate under control conditions (DMSO). Nevertheless, similar oligomycin-induced underestimation of max OCR and SRC was observed both in the presence and absence of sodium bicarbonate. The oligomycin-induced inhibition of SRC in the presence and absence of sodium bicarbonate was 38.0 ± 2.5% and 35.2 ± 1.5%, respectively. Next, the effect of FBS was evaluated on the oligomycin-induced inhibition of max OCR and SRC in T98G cells (Fig. 3). As expected, a lower concentration of CCCP (not shown) was required to achieve max OCR in the absence of FBS as this protonophore can bind non-specifically to FBS proteins. We observed a non-significant trend toward lower max OCR and SRC in the absence of FBS. Oligomycin-induced underestimation of max OCR was 19.9 ± 3.9% in the presence of FBS and 15.9 ± 2.2% in its absence, whereas underestimation of SRC was 33.7 ± 4.3% in the presence of FBS and 30.0 ± 2.4% in its absence.
The possible interference of multiple CCCP additions with a consequent long exposure time to oligomycin and CCCP to assess max OCR and SRC in cell lines was investigated by performing a single addition of CCCP to T98G cells (Fig. 4). First, a suboptimal CCCP concentration (3 µM) was added, and this resulted in similar stimulation of OCR in the absence or presence of oligomycin (Fig. 4A). A 6 µM CCCP concentration was then tested, and a higher OCR was observed under the control condition (DMSO) but lower stimulation of OCR was observed when oligomycin was present (Fig. 4B). Finally, max OCR was achieved under DMSO conditions using 9 µM CCCP, although progressive inhibition of max OCR occurred immediately after addition of 9 µM CCCP under both the control and oligomycin conditions (Fig. 4C).
Taken together, these results indicate that the inhibitory effect of oligomycin on max OCR and, consequently, SRC is related neither to the characteristics of the medium (i.e., glucose concentration and the presence or absence of bicarbonate buffer and FBS) nor to excess concentrations of or long exposure to oligomycin or CCCP.
The high glycolytic activity of tumor cells leads to underestimation of max OCR and SRC in the presence of oligomycin. As we reported previously 14 , supplementing the medium with pyruvate only slightly decreased the oligomycin-induced underestimation of max OCR and SRC, suggesting that this effect is not associated with a limited supply of respiratory substrates to mitochondria. We hypothesized that this effect might be associated with the high glycolytic activity of tumor cells. To investigate the role of glycolysis in the underestimation of max OCR and SRC in the presence of oligomycin, the influence of this metabolic pathway was minimized in two different ways. First, cells were incubated in DMEM without glucose and pyruvate but containing 4 mM glutamine, a substrate that is metabolized to produce α-ketoglutarate, an intermediate of the citric acid cycle. For comparison, experiments were also conducted using DMEM containing all metabolic substrates (11 mM glucose, 1.25 mM pyruvate and 4 mM glutamine) (Figs 5 and 6). Second, cells were incubated in supplemented DMEM (sDMEM) containing the glycolytic inhibitor 2-deoxyglucose (2-DG; 40 mM), a glucose analog metabolized by hexokinase at the expense of ATP, generating the non-metabolizable molecule 2-deoxyglucose-6-phosphate and thus partially inhibiting glycolysis 16,17 (Figs 7 and 8). The use of these two approaches to investigate the role of glycolysis is important since 2-DG also has effects on signal transduction that are unrelated to glycolysis inhibition 18 . Figures 5 and 6 show that using only glutamine as the metabolic substrate, the underestimation caused by oligomycin on max OCR (Figs 5C and 6A) and SRC (Figs 5D and 6B) was not observed in either T98G or U-87MG cells. Measurements of OCR in T98G cells showed a higher basal OCR with DMEM containing only glutamine than with DMEM containing all the substrates (Fig. 5A,B), an effect likely related to the absence of the glycolytic metabolism-induced inhibition of oxidative phosphorylation (i.e., the Crabtree effect). Notably, SRC in the presence of glutamine alone was lower than in the presence of all the substrates under DMSO conditions. This decrease in SRC was expected because basal respiration increases with glutamine but max OCR does not. The data in Fig. 5E indicate that in medium containing only glutamine, oxidative metabolism in T98G cells cannot   ). Statistically significant difference in relation to the control (DMSO), **P < 0.01. Statistically significant difference in relation to the corresponding condition with sodium bicarbonate, # P < 0.05 and ## P < 0.01.  When ATP synthase was inhibited using oligomycin, no significant drop in ATP levels was observed in T98G and U-87MG cells incubated with sDMEM (Figs 7C and 8C), indicating that the ATP measured is produced mainly by glycolysis. When cells were incubated in sDMEM plus CCCP, ATP levels decreased by 26.6 ± 6.6% in T98G cells and 21.6 ± 6.1% in U-87MG cells in relation to the control (DMSO). Under this condition, mitochondrial ATP synthase is expected to hydrolyze ATP to restore the ∆Ψ m that the protonophore CCCP dissipated 20,21 . In fact, when CCCP was added in the presence of oligomycin, cellular ATP levels were maintained because ATP synthase was prevented from operating in the reverse mode. In addition, the results in Fig. 9 show that the decrease in ATP levels induced by CCCP was accompanied by increased ADP/ATP ratios in both T98G and U-87MG cells in a mechanism sensitive to oligomycin.
Incubating the cells in sDMEM containing 2-DG resulted in a significant drop in ATP levels of 75.6 ± 4.0% in T98G cells and 83.5 ± 1.9% in U-87MG cells (Figs 7C and 8C), further suggesting that most cellular ATP in glioma cell lines is produced by glycolysis 19 . Induction of the Crabtree effect by 2-DG 22 may also contribute to ATP depletion under this condition. When CCCP, oligomycin or oligomycin plus CCCP were present with 2-DG, the ATP levels in T98G cells dropped by 91.0 ± 1.2%, 86.0 ± 1.7% and 86.8 ± 1.8%, respectively (Fig. 7C). The same pattern was observed when U-87MG cells were tested with 2-DG (Fig. 8C).
To investigate the importance of the inhibitory effect of oligomycin on the reverse activity of ATP synthase, we conducted experiments with citreoviridin. Low concentrations of citreoviridin inhibit the forward reaction of ATP synthase, with a minor effect on the reverse activity of this enzyme. However, higher concentrations of citreoviridin can inhibit forward and reverse activities 23,24 . Two different concentrations of citreoviridin were tested: 5 µM and 20 µM. Figure 10(A,B) shows that 5 µM citreoviridin did not affect max OCR and SRC. However, as shown previously 14 , in the presence of 20 µM citreoviridin, max OCR and SRC were underestimated by 18.9 ± 4.5% and 26.7 ± 5.9%, respectively, an effect similar to that observed in the presence of oligomycin.
The inhibitory effect of citreoviridin on the forward activity of ATP synthase was assessed based on its effect on basal OCR and compared with the effect of 1 µg/mL oligomycin 14 . The inhibitory effect on the reverse activity of ATP synthase was estimated by measuring the dissipation of ∆Ψ m under conditions in which the respiratory chain was inhibited by antimycin A (Fig. 10C). In our previous study, the forward activity of ATP synthase, which was measured indirectly as the fraction of OCR due to ATP turnover, was inhibited by 84.8 ± 1.7% in the presence of 5 µM citreoviridin and was completely inhibited by 20 µM citreoviridin (see Fig. 5A,B in 14 ). However, the membrane potential maintained by the reversed activity of ATP synthase was only 30.0 ± 3.0% and 58.5 ± 4.7% sensitive to 5 µM and 20 µM citreoviridin, respectively (Fig. 10C,D). These results indicate that a low concentration (5 µM) of citreoviridin can almost completely inhibit the forward reaction of ATP synthase and has a small effect on its reverse activity. Interestingly, this low concentration of citreoviridin does not reflect the underestimation of max OCR and SRC produced by oligomycin.
Next, the effect of CCCP on ATP levels was measured in the presence of 5 and 20 µM citreoviridin (Fig. 10E). Compared with the control condition (DMSO), CCCP reduced ATP by 27.9 ± 6.9%. The lower concentration of citreoviridin (5 µM) did not significantly prevent the ATP drop induced by CCCP (ATP drop of 17.6 ± 8.6%). However, CCCP did not induce a significant drop in ATP levels in the presence of 20 µM citreoviridin. This latter result is in accordance with the important inhibitory effect of citreoviridin at a high concentration (20 µM) on the reverse activity of ATP synthase. ATP, but not ADP, inhibits max OCR in digitonin-permeabilized T98G and U-87MG glioma cells. To investigate the direct influence of ATP on max OCR, digitonin-permeabilized cells were incubated in the presence of ADP or ATP (i.e., ADP plus phosphocreatine and creatine phosphokinase, an ATP regeneration system) (Fig. 11). Permeabilized cells were used to allow easy adjustment of extramitochondrial ADP and ATP levels and to assess their effects on CCCP-induced max OCR. The presence of only ADP did not inhibit max OCR in either cell line; however, the presence of ATP inhibited max OCR by 26.0 ± 2.7% in T98G cells and 19.3 ± 1.8% in U-87MG cells.
Effects of oligomycin and CCCP on mitochondrial membrane potential (∆Ψm) and OCR in human glioma cell lines. To further assess the underestimation of max OCR and SRC that occurs in the presence of oligomycin, OCR and mitochondrial membrane potential (∆Ψ m ) were determined in parallel in intact T98G and U-87MG cells (Figs 12 and 13). A stable fluorescence signal of TMRM (500 nM) in sDMEM was obtained after approximately 8 min (not shown), after which cells were added. Oligomycin (or its vehicle DMSO) was then added to inhibit the activity of ATP synthase, and titration of the protonophore CCCP was carried out to progressively decrease ∆Ψ m and achieve max OCR. In DMSO without oligomycin, max OCR was reached with 12 µM and 9 µM CCCP for T98G and U-87MG cells, respectively (Figs 12A and 13A), whereas in the presence of oligomycin, lower CCCP concentrations (6 µM and 3 µM) were required (Figs 12B and 13B). However, higher protonophore concentrations were required to completely dissipate ∆Ψ m in both cell lines: 15 µM in the absence of oligomycin and 9 µM in its presence. These results show that protonophore-induced max OCR occurred at a low ∆Ψ m , but complete dissipation of ∆Ψ m led to inhibition of OCR (Figs 12C and 13C). This finding is in accordance with the inhibition of mitochondrial transport of substrates that can occur when ∆Ψ m is dissipated 10,[25][26][27] . Figures 12D and 13D show that max OCR was significantly underestimated in both cell lines when oligomycin was present and that max OCR was observed under higher ∆Ψ m in the presence of oligomycin than in its absence (DMSO condition).

Discussion
The results presented here indicate that the underestimation of CCCP-induced max OCR in tumor cell lines treated with oligomycin is caused by the high glycolytic activity of these cells. This conclusion is supported by the elimination of the inhibitory effect of oligomycin on max OCR when the glycolytic pathway was minimized using either glutamine as the only respiratory oxidative substrate (Figs 5 and 6) or the glycolytic inhibitor 2-DG (Figs 7 and 8). Furthermore, parallel determinations of OCR and ∆Ψ m showed that max OCR was observed in the presence of a low ∆Ψ m (Figs 12 and 13), which was higher when the estimation was performed in the presence of oligomycin (Figs 12D and 13D). This result indicates that max OCR cannot be reached in the presence of oligomycin because of limiting factors, as will be discussed later. (Figs 7-9). When glycolysis was limited, either by using glutamine as the only metabolic substrate or by the presence of 2-DG, cellular ATP levels were not sustained. Under these conditions, a similar max OCR was obtained in both the presence and absence of oligomycin (Figs 5-8). In the absence of this ATP synthase inhibitor, the decrease in glycolytic ATP in the presence of CCCP was mostly due to ATP consumption by the reverse activity of mitochondrial ATP synthase (Figs 5E, 7C and 8C). Even though slow reverse operation of ATP synthase is expected in respiration-impaired mitochondria with an intact inner membrane 28 , higher reverse activity occurs under conditions of strong uncoupling, e.g., in CCCP-induced max OCR or when the integrity of the inner mitochondrial membrane is disrupted.

Measurements of cellular ATP levels and ADP/ATP ratios showed an association between the maintenance of intracellular ATP levels by glycolysis and the inhibitory effect of oligomycin on CCCP-induced max OCR
The results of the experiments conducted with the ATP synthase inhibitor citreoviridin, a compound that affects the forward and reverse activity of ATP synthase differently 23,24 (Fig. 10C,D), further support the proposition that cellular ATP levels are reduced by the reverse activity of ATP synthase under uncoupling conditions. A low concentration of citreoviridin (5 µM), which almost completely inhibited the forward activity of ATP synthase 14 but had only a minor effect on the reverse activity (Fig. 10D), neither prevented a decrease in ATP levels by CCCP nor inhibited CCCP-induced max OCR. However, at a higher concentration (20 µM), citreoviridin inhibited the reverse activity of ATP synthase to a greater degree, preventing the drop in ATP caused by CCCP and resulting in the inhibition of CCCP-induced max OCR.
Exposing digitonin-permeabilized cells to exogenous ATP but not ADP resulted in partial inhibition of CCCP-induced max OCR (Fig. 11). Maintenance of a higher intracellular ATP/ADP ratio may limit CCCP-induced max OCR by inhibiting enzymes involved in the reduction of NAD + to NADH in the mitochondrial matrix, thereby restricting electron transfer from carbon substrates to ETS. According to previous studies, a higher ATP/ADP ratio in the mitochondrial matrix is associated with lower activity of pyruvate dehydrogenase (PDH), isocitrate dehydrogenase-3 (IDH-3) and glutamate dehydrogenase (GDH). ADP stimulates PDH activity by inhibiting pyruvate dehydrogenase kinase, which phosphorylates and inactivates PDH 29,30 ; a decrease in ATP/ADP ratio results in lower IDH-3 K m values for its substrates 31,32 , and GDH is subject to positive allosteric regulation by ADP 33,34 . Interestingly, a recent study with astrocytes revealed that ADP-stimulated GDH plays an important role under conditions of increased mitochondrial oxidative metabolism demand 35 . This pathway may play an important role in the supply of mitochondrial NADH to glutamine-addicted highly glycolytic tumor cells 19,36 . In addition, a high intramitochondrial ATP/ADP ratio may also limit max OCR by promoting partial inhibition of NADH oxidation. Kadenbach's group 37 demonstrated that the binding of ADP and ATP to cytochrome c oxidase (respiratory complex IV) regulates the activity of this enzyme. Allosteric inhibition of cytochrome c oxidase by ATP occurs in the presence of a high intramitochondrial ATP/ADP ratio and may inhibit max OCR even in the presence of sufficient NADH 37,38 .
In the present study, ATP contents and ADP/ATP ratios were determined in total cell extracts. Because of the negative-inside mitochondrial membrane potential, the ADP/ATP ratio is significantly different between the cytosol and the mitochondrial matrix 39 . However, our main inferences are based on experimental conditions in which membrane potential is mostly dissipated (i.e., in the presence of CCCP or CCCP plus oligomycin). Under such conditions an equilibrium between the cytosol and mitochondrial matrix ADP/ATP ratios is expected 39 . Significant inhibition of protonophore-induced max OCR was also observed by Kim et al. 40 in an oligomycintreated INS-1E pancreatic beta cell line. This effect was associated with time-dependent exposure to oligomycin that was likely causing a progressive loss of cell function in the absence of oxidative phosphorylation 40 . However, our results with glioma cell lines seem to differ in nature from those observed in INS-1E cells because the limitation on max OCR was observed very soon after addition of oligomycin to cells sustaining high intracellular ATP levels. Given that many studies have evaluated mitochondrial oxidative metabolism in highly proliferative tumor cell lines (i.e., highly glycolytic cells), the mechanism of oligomycin-mediated underestimation of max OCR reported here is likely to be more widespread in experimental protocols.
We conclude that high glycolytic activity leads to the underestimation of CCCP-induced max OCR and SRC in tumor cells treated with oligomycin. Under these conditions, oligomycin maintains cellular ATP levels by preventing the reverse activity of ATP synthase. Minimizing glycolytic activity may allow more accurate assessment of max OCR in the presence of ATP synthase inhibitors. Cell Lines and Cell Culture. The human glioblastoma T98G and U-87MG cell lines were purchased from the American Type Culture Collection (Manassas, VA, USA) and cultured as previously described 14 . On the day of the experiment, the cells were trypsinized and resuspended (16-32 × 10 6 cells/mL; >95% viability) in DMEM supplemented with 10% FBS, 100 U/mL penicillin and 100 µg/mL streptomycin (sDMEM) containing 20 mM HEPES. For experiments testing the components of the medium (glucose 5.5 mM or 11 mM, glutamine, sodium bicarbonate and FBS), cells were resuspended in the corresponding experimental medium as described in the figure legends. Cell suspensions were maintained at room temperature (ca. 23 °C) and used within 2.5 h.

Chemicals
The data reported here are from experiments conducted over 18 months; as the cells were expanded from different frozen aliquots, and the components of the medium (e.g., FBS) were from different batches, absolute mean values of some variables can be expected to vary (e.g., max OCR can oscillate up to 25%) more than the standard error of the mean when experiments performed a couple of months apart are compared. Nonetheless, we would emphasize that each experimental protocol was conducted within 2 to 4 weeks and that these variations were not observed; moreover, the effects of treatments were consistent throughout the whole study.

Measurement of OCR in suspended tumor cells.
The OCR in intact and permeabilized suspended cells was determined using a respirometer (OROBOROS Oxygraph-2k, Innsbruck, Austria), as previously described 14,19 . In intact tumor cells this was carried out by incubating an aliquot of the cell suspension (2-4 × 10 6 cells) at 37 °C in a 2 mL chamber containing the reaction medium, as described in the figure legends. OCR was measured in permeabilized cells by incubating 3-4 × 10 6 cells at 37 °C in 2 mL of "permeabilization medium" containing 125 mM sucrose, 65 mM KCl, 2 mM K 2 HPO 4 , 1 mM MgCl 2 , 1 mM EGTA, 1 µg/mL oligomycin, 10 mM HEPES-K + pH 7.2 and respiratory substrates (1 mM glutamate, 0.5 mM malate and 1 mM pyruvate), as well as digitonin (30 µM) for plasma membrane permeabilization. The concentrations of the respiratory substrates in the "permeabilization medium" were chosen to better resemble those found in situ (submillimolar levels), enabling the inhibitory effect of adenine nucleotides on the oxidative metabolism of these substrates to be studied. for DMEM or 10 µM for sDMEM) was then added to some of the samples and incubated for an additional 10-12 min. Cell suspensions were immediately centrifuged (6,000 g, 5 min), the supernatants were discarded, and the pellets were homogenized in 1 mL of lysis buffer (25 mM TRIS-phosphate, pH 7.8, 2 mM dithiothreitol, 2 mM EDTA, 10% glycerol and 1% Triton X-100) compatible with the luciferase assay reagents and maintained for 5 min in an ice bath. Control experiments showed that the centrifugation step did not significantly impair cell viability (results not shown). Samples were then analyzed for ATP content using a luminescence assay (CellTiter-Glo Luminescent Cell Viability Assay, Promega, G7570) in a plate reader (SpectraMax M3 -Molecular Devices, Sunnyvale, CA, USA). Luminescence was read at 560 nm in endpoint mode with an integration time of 1 sec. SoftMax Pro 6.4 software was used for data acquisition. To estimate ADP content, 100 µM dCTP and 10 U/mL NDK were added and luminescence was read again after 10 min 41 . Figure 11. Effect of ADP and ATP on CCCP-induced max OCR in permeabilized T98G and U-87MG cells. T98G (1.5 × 10 6 /mL) and U-87MG (2 × 10 6 /mL) cells were incubated in "permeabilization medium" containing 30 µM digitonin, and max OCR was estimated by sequential additions of CCCP (0.05 µM each). The effects of ADP and ATP on max OCR were assessed by incubating the cells in the presence of 1 mM ADP or an ATP regeneration system (1 mM ADP, 10 mM phosphocreatine plus 50 µg/mL creatine phosphokinase). (A,B) Representative traces of OCR in permeabilized T98G and U87-MG cells. (C,D) Effects of ADP and ATP on CCCP-induced max OCR in permeabilized T98G and U87-MG cells. Statistically significant difference versus the corresponding control, **P < 0.01. Statistically significant difference in relation to " + ADP", ## P < 0.01.  Simultaneous measurements of OCR were taken in the chamber of the OROBOROS respirometer under identical experimental conditions. Membrane potential, expressed as ∆F/F, was calculated, where F is the fluorescence intensity after the last addition of CCCP (i.e., completely dissipated ∆Ψ m ) and ∆F is F minus any given fluorescence intensity 43 . Statistical analysis. The results are shown as representative traces and/or the mean ± standard error of the mean (SEM). Experiments were performed with cells from at least four different passages. Paired Student's t-test was applied to analyze differences between two groups. Multiple comparisons were performed by repeated-measures one-way analysis of variance (ANOVA) and the post hoc Bonferroni test 14 .

Data Availability
The datasets generated during the current study are available from the corresponding author on reasonable request.