The CML lines K562 and FEPS (K562/DNR), were grown in RPMI-1640 medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS) at 37 °C and 5% CO₂. FEPS is a stable K562-derived cell line established upon exposure to increasing doses of DNR (26). FEPS cell line was cultivated in the presence of 500 nM of DNR to maintain its resistant profile up to 72 h prior to carrying out the experiments. Mycoplasma contamination was routinely tested by PCR using primers described elsewhere (28).

Assessment of oxygen consumption was performed on a high-resolution respirometry system (OROBOROS Instruments – Oxygraph2k – Innsbruck, Austria). 2×10⁶ intact cells were incubated in Dulbecco’s Modified Eagle’s Medium without glucose, L-glutamine, phenol red, sodium pyruvate and sodium bicarbonate supplemented with 11 mM glucose and 2 mM glutamine (equivalent concentrations to RPMI-1640 culture medium). After the respiration stabilization period, the following modulators of mitochondrial function were added: oligomycin (1 µg/mL), carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazine - FCCP (50 nM), rotenone (0.5 µM) and antimycin A (2 µg/mL). From the addition of these modulators, it is possible to evaluate the following respiratory parameters: (i) routine oxygen flow – in absence of modulators; (ii) oxygen flux uncoupled from ATP synthesis – through treatment with oligomycin, ATP synthase inhibitor; (iii) oxygen flux coupled to ATP synthesis – by subtracting the routine flux values ​​from those obtained after the addition of oligomycin; (iv) maximum capacity of the ETS – due to the acceleration of the electron transport necessary for the maintenance of the membrane potential resulting from FCCP treatment, which dissipates the proton gradient; (v) residual oxygen flow – through the inhibition of complexes I and III by treatment with rotenone and antimycin A, respectively, which makes it possible to measure the oxygen flow not related to mitochondrial functions; (vi) reserve capacity – by subtracting the routine flow values ​​from those obtained after adding FCCP (maximum capacity). To acquire the graphs and analyze the values, the DatLab version 5.1 program (OROBOROS Instruments – Oxygraph2k) was used.

For lactate release, 4×10⁵ cells per condition were incubated for 24 h at 37°C and 5% CO₂ in DMEM without phenol red and FBS supplemented as described before (11 mM glucose and 2 mM glutamine) in a reaction buffer pH 9.2 containing 0,6 M glycine, 0,35 M hydrazine and 15 mM β-NAD⁺. The lactate was quantified using the formation of nicotinamide adenine dinucleotide (NADH) and pyruvate in the presence of excess NAD⁺. The reaction was catalyzed by the addition of lactate dehydrogenase (LDH) (2 U/mL), and the concentration of lactate was calculated from the extinction coefficient of NADH (6,22 × mM^-1 × cm^-1) in the Spectramax spectrophotometer M4 Multi-Mode Microplate Reader at 340 nm.

ATP content was measured from 10⁵ cells following the protocol guidelines of the ATP Determination kit (Thermo Fisher). This assay measures the luminescence of the reaction, based on the luciferin-luciferase system, in a dark 96-well plate. Cells were placed in the presence of reaction medium (1 mM 1,4-dithiothreitol (DTT), 20× Reaction Buffer, 0.5 mM luciferin and 1.25 mg/ml luciferase) at the time of reading. The bioluminescence generated, directly proportional to the ATP content, was measured by spectrophotometry in the device’s default condition. The ATP was quantified from the correlation to the values obtained in the standard curve.

Total RNA was isolated from cells using TRIzol reagent (Invitrogen) following the manufacturer’s instructions. Total RNA was quantified spectrophotometrically. 1 µg RNA was treated with 1 U of RNAse-free DNAse for 30 minutes at 37 °C. Reactions were stopped by adding 1 µL of ethylenediamine tetraacetic acid (EDTA) 20 mM and heating for 10 minutes at 65 °C. cDNA synthesis was performed using the DNAse treated RNA using the High-Capacity cDNA Reverse Transcription Kit from Applied Biosystems in accordance with the manufacturer’s instructions.

Gene expression analysis was performed using 7500 Real-Time PCR (Applied Biosystems) and power SYBR-GREEN PCR master MIX (Applied Biosystems). For this test primer pairs were synthesized based on GenBank sequences of mRNA. The comparative Ct method was used to measure changes in gene expression levels (29). Actin was employed as an endogenous control.

10⁶ cells were lysed during 3 minutes in a Potter-Elvehjem homogenizer in ice, in a buffer containing 10 mM Tris-HCl pH 7.4, 0.25 M sucrose, 20 mM sodium fluoride (NaF) and 5 mM EDTA (cell lysis buffer). The homogenates were centrifuged for 5 minutes at 4°C at 1000×g. The supernatant was collected and centrifuged for 15 minutes at 4°C at 10000×g. The supernatant (cytosolic fraction) was set aside, and the pellet (mitochondrial fraction) suspended with the cell lysis buffer. Protein concentration was assayed using the Bradford method.

Cell pellets were lysed as described in mitochondrial preparations. 40 µg of protein extracts were fractionated by standard 10% SDS-PAGE and transferred to nitrocellulose membranes by electro blotting in a buffer consisting of 39 mM glycine, 48 mM Tris-base, sodium dodecyl sulfate (SDS) 0.037% (w/v) and methanol 20% (v/v). Proteins were detected using primary antibodies diluted in Tris-buffered saline (TBS), Tween 20 0.1% (v/v) and bovine serum albumin (BSA) 5% (w/v). Rabbit anti-human HKII and anti-G6PDH polyclonal primary antibodies were obtained from Cell Signaling Technology^®. Rabbit anti-human GR primary antibody was obtained from Abcam^®. The secondary antibody for all incubations was the IRDye 680RD goat anti-rabbit immunoglobulin. β-actin was used as loading control because its expression not being altered between K562 and FEPS (25). Bands were visualized in a Li-Cor Odyssey Western blot imaging.

Clotrimazole (Sigma-Aldrich), an imidazole derivative, was dissolved in dimethylsulfoxide (DMSO) to a stock concentration of 2.9 mM which was used for cell treatments at the final concentrations of 15 µM, 35 µM and 50 µM for 1, 3, 6 and 24 h. The previous incubation of cells with this compound was employed to induce detachment of HKII from the outer mitochondrial membrane, since this effect is already well described in the literature (30–32). The cells were treated with 11 mM 2-deoxy-D-glucose (2DG) – equimolar to the concentration of glucose present in the RPMI-1640 medium – for 72 h in combination with 500 nM DNR or 60 nM VCR (VCR). Alternatively, cells were previously incubated for 2 h with 1 mM NAC before the assays.

Cell viability was assayed by crystal violet and cell count with trypan blue. Crystal violet interacts with viable cell DNA while trypan blue permeates and stains only dead cells. For the crystal violet assay, 2×10⁴ cells were fixed with 4% paraformaldehyde after treatment under the tested conditions and stained with 0.05% crystal violet for 30 minutes under agitation and protected from light. Then, the wells were washed with deionized water to remove excess dye and the crystals were solubilized with 70% ethanol. Intensity of staining, directly proportional to the percentage of viable cells, was quantified by spectrophotometry at 570 nm. Trypan blue was employed at 0.08% in a Neubauer chamber for counting under an optical microscope after treatment.

The mitochondrial superoxide ( O 2 − ) was measured from 10⁵ cells treated as described before, following the protocol of the reagent MitoSOX™ Red (Thermo Fisher). MitoSOX™ Red is a fluorogenic dye specifically targeted to mitochondria in live cells. Oxidation of MitoSOX™ Red reagent by superoxide produces fluorescence, selectively targeting mitochondria of live cells. It is rapidly oxidized by superoxide generating a fluorescent product but not by other ROS and reactive nitrogen species. The oxidized product is highly fluorescent upon binding to nucleic acid. Maximum absorption/emission: ∼510/580 nm.

Free radicals react with membrane lipids and cause peroxidation with the consequent formation of malondialdehyde (MDA). Following heating in the presence of thiobarbituric acid (TBA), the reaction forms a fluorescent compound that is quantified in a fluorimeter (excitation: 515 nm; emission: 553 nm) (33). For this assay, 1×10⁶ cells were macerated in phosphate buffer for protein extraction. Proteins were quantified for later normalization of results and part of the obtained sample was incubated with 20% trichloroacetic acid (TCA) and 0.67% TBA (in sodium sulfate) in a boiling bath for 2 h. Then, n-butanol was added and the samples were centrifuged for 3 minutes at 5000×g and the fluorescent product, present in the upper fraction, was quantified by spectrophotometry under excitation of 515 nm and emission of 553 nm. A standard curve with 10× diluted tetramethoxypropane (TMP) was used as reference for the quantification.

Mitochondrial reducing capacity was measured by colorimetric assay with MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazol bromide) after a 1 h incubation with 15 µM, 35 µM and 50 µM CTZ. The MTT solution was prepared at a concentration of 5 mg/mL diluted in 1× phosphate buffered saline (PBS). MTT reacts with mitochondrial dehydrogenases reducing tetrazolium salts, resulting in blue-purple formazan crystals. Those were diluted in DMSO and the absorbance, directly proportional to the number of formed crystals, was measured by spectrophotometry at 492 nm.

Glutathione levels were measured in accordance with the protocol described by Biswas et al. (34). Glutathione is a key antioxidant tripeptide (L-γ-glutamyl-L-cysteinyl-glycine) that represents the major non-protein thiol in many organisms. Within cells, glutathione exists in two different forms: the reduced form GSH and GSSG, the oxidized form. Oxidative stress has a profound effect on the cellular thiol balance and can lead to a decreased GSH/GSSG ratio. ROS - superoxide anion is neutralized by GSH through a cascade of detoxification mechanisms involving enzymes such as GR. The colorimetric reaction generates a product directly proportional to levels of glutathione, which was quantified by spectrophotometry at 415 nm.

Apoptotic cells were quantified using the Click-iT^® TUNEL Alexa Fluor^® imaging assay kit from 2×10⁵ cells, following the protocol guidelines. TUNEL assay is based on the incorporation of modified uridine triphosphate (dUTPs) by the enzyme terminal deoxynucleotidyl transferase (TdT) at the 3’-OH ends of fragmented DNA, a hallmark as well as the ultimate determinate of apoptosis. The Click-iT^® TUNEL Alexa Fluor^® imaging assays utilize a dUTP modified with an alkyne, a small, bio-orthogonal functional group that enables the nucleotide to be more readily incorporated by TdT than other modified nucleotides. Detection is based on a click reaction, 3-6 a copper-catalyzed reaction between an azide and alkyne. The fluorescence generated is directly proportional to the number of apoptotic cells and was quantified by spectrophotometry under excitation of 495 nm and emission of 519 nm.

The ABCB1 and ABCC1 transport assays were performed, respectively, with the use of the Rho 123 and 5(6)-carboxyfluorescein diacetate (CFDA) dyes (both from Sigma-Aldrich) as previously described (27). Rho 123 and CFDA passively distribute into the cell, and while the first is fluorescent and is actively extruded by ABCB1, the latter undergoes hydrolysis by nonspecific esterases in the cytosol, originates the fluorescent substrate carboxyfluorescein (CF) that only then is transported out by ABCC subfamily members, notably ABCC1 (35). Briefly, assays were performed in two 30-minute steps, sufficient for the accumulation and efflux of dyes. Both steps were carried on at 37 °C in 5% CO₂ in a light-protected environment. First, 2×10⁴ cells/mL were treated for 1 h with 35 mM CTZ as described prior to the assays. Then, 2×10⁵ FEPS cells were incubated in 96-well plates with 250 nM Rho 123 or 500 nM CFDA in fresh RPMI medium to allow accumulation of the dyes within cells (accumulation phase). Following, cells were centrifuged at 200×g for 7 minutes and then resuspended in fresh RPMI to allow efflux of the dyes (efflux phase). Inhibition of ABCB1 efflux was performed in the presence of the calcium channel blocker verapamil (VP) (36), and the quinoline derivative MK-571 was employed to inhibit ABCC efflux (37) (inhibited efflux). As negative control, cells were exposed to medium only (free efflux). Then, cells were again centrifuged, resuspended in cold PBS and maintained on ice until acquisition by flow cytometry.

Assays were performed in similar conditions to the ABC-mediated efflux assays. FEPS cells were treated with CTZ as described, and then incubated in 96-well plates with 2.5 or 5.0 μM DNR for 1 h. In parallel, cells were incubated with the ABCB1 inhibitor VP, the ABCC inhibitor MK-571 or with reduced glutathione. As negative control, cells were exposed to medium only, and treated as described before acquisition by flow cytometry.

The median fluorescence intensities (MFI) from 10⁴ viable cells, gated in accordance with forward and side scatter parameters representative of cell size and granularity, were acquired using the FL1-H filter (for Rho 123 and CF) or the FL3-H filter (for DNR) on a BD FACS Calibur flow cytometer (BD Biosciences, San Jose, CA, USA) and represented on histograms. All post-analyses were performed on Summit version 4.3 software (Dako Colorado, Inc., Fort Collins, CO, USA).

Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software, inc.). Results are expressed as mean + SEM values for an independent experiment. Comparisons between groups were done by one-way ANOVA and a posteriori Dunnett’s test. When appropriate, unpaired Student’s t-tests or Mann-Whitney’s test were employed. Differences of p < 0.05 were considered significant.

Metabolism comprises a series of sequential regulated reactions to obtain energy or transform it, that can be produced through oxidative phosphorylation, coupled with oxygen consumption or as occurs in glycolysis (19, 21). Some tumors preferentially use one of these pathways to supply their energy demands depending on their metabolic profiles (16). One way to measure the mitochondrial contribution to these processes is through the quantification of oxygen flow. As shown in Figure 1, FEPS exhibited significantly lower basal oxygen flow (Figure 1A), coupled respiratory capacity (Figure 1B), maximal and spare respiratory capacities (Figures 1C, D) than K562. Basal oxygen flow represents the routine oxygen demand without any interference. Coupled respiratory capacity refers to mitochondrial oxygen consumption that is coupled to ATP synthesis. The maximum capacity is experimentally induced by titration with uncouplers, such as FCCP, to collapse the proton gradient across the inner mitochondrial membrane to measure the capacity of the ETS. Spare respiratory capacity reflects the extent of the mitochondria response to stressful situations or increased energy demand, essential for cell survival. These results suggest that adaptation to chemotherapeutic stress on FEPS cells reduced the oxidative metabolism compared to the parental K562.

Poorly oxidative cells preferentially resort to the glycolytic pathway to obtain energy. In this context, pyruvate is converted to lactate instead of feeding oxidative phosphorylation through the Krebs cycle (19). Results on Figure 2 show that FEPS released significantly more lactate than K562 (Figure 2A), supporting the hypothesis that MDR cells preferentially engage in glycolysis in detriment of the oxidative pathway. Nevertheless, cells need sufficient energy input to supply basic cellular processes, and the primary way to store it is through ATP production. Figure 2B showed that the ATP content of FEPS was approximately twice as that measured in K562. Considering the lower mitochondrial oxygen flux (Figure 1) associated with higher lactate release (Figure 2A), data suggest that MDR cells preferentially utilize the glycolytic pathway to increase ATP input.

Hexokinase is the first enzyme of the glycolytic pathway; it is present in four isoforms (HKI-IV) and its main function is to convert glucose into G6P. Pyruvate generated at the end of glycolysis enters the Krebs cycle and fuels oxidative phosphorylation or is converted to lactate by LDH (16). Among its isoforms, only HKI and HKII have the ability to bind to the mitochondrial membrane (21, 38). Despite sharing this feature, they may be differentially expressed in a tissue-specific manner (39, 40). Figure 3A shows that although there are no significant differences in HKI and LDHA gene expression between FEPS and K562, HKII expression is approximately two-fold higher in MDR cells. An increase of HKII at gene level in FEPS was accompanied by increased expression of this enzyme in the mitochondrial fraction, where levels three times higher than those observed in K562 were detected (Figure 3B). Accordingly, the enzymatic activity of HKII on the mitochondrial fraction was four times higher in FEPS as compared to K562 (Figure 3C). Michaelis-Menten and Lineweaver-Burk plots (Figures 3D, E) confirm that the hexokinase isoform II is the most active. K562 and FEPS affinities for glucose were 0.1297 mM and 0.316 mM, similar to the Km values reported for HKII of 0.13-0.3 mM (39, 41, 42). The presence of HKII mostly in the mitochondrial fraction (Figure 3B) might relate to the high ATP content of FEPS (Figure 2B), as it can directly use and optimize its production by increasing the efficiency of ETS.

HK stimulate the PPP when associated with the outer mitochondrial membrane by providing G6P to the first enzyme of this pathway, G6PDH. This coenzyme, in turn, maintains adequate levels of NADPH and glutathione to protect cells from the oxidative damage caused by ROS (16). As depicted in Figure 4A, FEPS presented significantly increased G6PDH activity relative to K562. High activity of this enzyme may contribute to a reduction of O 2 − (Figure 4B) and MDA (Figure 4C) in MDR cells.

Lipoperoxidation occurs when ROS, namely the superoxide anion, react with phospholipids of the inner mitochondrial membrane, often resulting in cell death due to the accumulation of highly reactive compounds such as the unsaturated aldehyde MDA (33). Impairment of the PPP upon treatment with 2DG further reduced the viability of MDR cells compared to the sensitive ones (Figure 4D). Together with the observed increase in HKII activity (Figure 3C), this suggests that the PPP might play an important role for the viability of drug-resistant cells.

CTZ is a synthetic antimycotic agent of the azole class which induces detachment of HK from the outer mitochondrial membrane (30–32). To discriminate the activity of HKII coupled to mitochondria from the one in the cytosolic fraction, FEPS was incubated for 1 h with CTZ, and a sub-toxic concentration of 35 μM was selected for the following experiments (Figure 5A). Remarkably, CTZ showed no cytotoxic effect towards the parental cell line K562, possibly reflecting the lower expression of mitochondrial HKII (Supplementary Figure 1). FEPS showed a significant dose-dependent increase in cytosolic HKII activity after CTZ treatment, while the activity on the mitochondrial fraction was correspondingly reduced (Figure 5B), reversing the profile previously observed (Figure 3C). This profile was not observed in K562 sensitive cells (Supplementary Figure 2). The increase in cytosolic HKII activity was accompanied by an increase in the cytosolic expression as well. Likewise, the reduction in mitochondrial activity was followed by a reduction in mitochondrial HKII expression (Figure 5C). This shift corroborates earlier observations indicating CTZ as an interesting approach to detach HKII from mitochondria.

The factors that regulate the proportions of HK that must be coupled to the outer mitochondrial membrane or free in the cytosol are still controversial in the literature. However, it is plausible that interferences in this balance would reflect directly on cellular metabolism. Figure 6A shows that the reduced mitochondrial expression of HKII leads to increased ATP synthesis-uncoupled oxygen flow (leak). This parallels with a decrease in the ATP synthesis-coupled oxygen flow. Though CTZ does not affect the maximum capacity, the spare respiratory capacity (Figure 6B) and mitochondrial function (Figure 6C) were significantly reduced upon incubation with CTZ. Interestingly, despite the increased cytosolic HKII activity (Figure 5B), both the lactate release (Figure 6D) and ATP content (Figure 6E) were considerably reduced after a 6 h incubation with CTZ. These results show that cytosolic translocation of HKII markedly disturbs the metabolic profile of chemoresistant CML cells.

The presence of an antioxidant system able to prevent cell damage generated by ROS is essential for adaptation to chemotherapeutic stress. CTZ did not affect G6PDH and GR quantification (Figures 7A, B), two critical enzymes for the antioxidative machinery. Conversely, the activity of those enzymes were drastically reduced (Figures 7C, D). However, this reduction was not caused by direct effect of CTZ, given that the activity of both were similar when the fractions were treated with this drug (Supplementary Figure 3). Equilibrium of glutathione between the reduced (GSH) or oxidized (GSSG) forms vary in accordance with the oxidizing stimuli which cells are exposed to. GR is responsible for maintaining the redox status of GSH by recycling the oxidized form, in a way that a decrease in its activity result in low levels of GSH. Figure 7E shows that FEPS presented low GSH content, which relates to the superoxide levels (Figure 7F) and mitochondrial damage (Figure 7G). This effect is not observed in K562, since pre-incubation with CTZ did not reduce G6PDH activity and did not increase lipid peroxidation levels in this cell line (Supplementary Figure 4).

As demonstrated before, 1 h treatment with CTZ at 35 μM induced no evident cytotoxicity (Figure 5A). On the other hand, CTZ reduced FEPS proliferation by ~50% after 72 h, as shown in Figure 8A. Exposure to CTZ followed by treatment with the chemotherapeutic drugs VCR or DNR for 72 h further reduced cell viability (Figures 8A, B). As seen in Figure 8C, pre-incubation with CTZ caused collateral sensitization on FEPS cells, resulting in apoptotic death, as it increases the percentage of TUNEL-positive cells. These data corroborate reports that point to protection against apoptosis as one of HKII moonlight functions when attached to the mitochondria (16, 17).

Among the mechanisms enhanced during adaptation to stress, the roles of ABC transporters are the most discussed. ABCB1 and ABCC1 promote the MDR phenotype on a variety of cancer models, including the FEPS cell line (2, 26, 27). In this context, the impact of CTZ on the functional activity of both transporters was evaluated. Corroborating the results in Figure 8, CTZ promoted an increase in DNR uptake, most noticeably when FEPS was incubated with 5.0 μM of this drug. This profile manifested in the presence of the ABCB1 and ABCC1 inhibitors, VP and MK-571, respectively, which further increased DNR uptake in co-treatment with CTZ (Figure 9A). A distinct scenario occurred when the functional assays were performed by use of specific substrates for each ABC protein. Figure 9B showed that the fluorescence of Rho 123 was not altered after CTZ treatment, whereas the one for CF increased in absence (free efflux) and in presence of MK-571 (inhibited efflux) (Figure 9C). Thus, detachment of HKII from the outer mitochondrial membrane impaired ABCC1 activity with no effect on ABCB1, allowing for increased DNR retention in FEPS.

A crucial feature of the ABCC-mediated transport is the interaction with GSH. Transport of the majority of ABCC1 substrates requires the presence of GSH, either in conjugation or in cotransport (35). Based on this, we evaluated whether GSH could impact the inhibitory effect of CTZ towards DNR efflux in FEPS (Figures 10A, B). Of note, co-treatment of CTZ and GSH reduced the accumulation of this drug to similar levels to the untreated controls. Considering that the cytosolic localization of HKII reduced the levels of this antioxidant peptide (Figure 7E) it is plausible to assume that GSH would antagonize the metabolic consequences of CTZ treatment. The negative modulation of ABCC1 activity (Figure 9C) could be a consequence of this effect since GSH is a prime substrate for this transporter.