The environmental carcinogen benzo[a]pyrene induces a Warburg-like metabolic reprogramming dependent on NHE1 and associated with cell survival

Cancer cells display alterations in many cellular processes. One core hallmark of cancer is the Warburg effect which is a glycolytic reprogramming that allows cells to survive and proliferate. Although the contributions of environmental contaminants to cancer development are widely accepted, the underlying mechanisms have to be clarified. Benzo[a]pyrene (B[a]P), the prototype of polycyclic aromatic hydrocarbons, exhibits genotoxic and carcinogenic effects, and it is a human carcinogen according to the International Agency for Research on Cancer. In addition to triggering apoptotic signals, B[a]P may induce survival signals, both of which are likely to be involved in cancer promotion. We previously suggested that B[a]P-induced mitochondrial dysfunctions, especially membrane hyperpolarization, might trigger cell survival signaling in rat hepatic epithelial F258 cells. Here, we further characterized these dysfunctions by focusing on energy metabolism. We found that B[a]P promoted a metabolic reprogramming. Cell respiration decreased and lactate production increased. These changes were associated with alterations in the tricarboxylic acid cycle which likely involve a dysfunction of the mitochondrial complex II. The glycolytic shift relied on activation of the Na+/H+ exchanger 1 (NHE1) and appeared to be a key feature in B[a]P-induced cell survival related to changes in cell phenotype (epithelial-to-mesenchymal transition and cell migration).


In silico tests for B[a]P-affected metabolic signatures
The single sample Gene Set Enrichment Analysis (ssGSEA) method was applied on the published microarray dataset of B[a]P-exposed HepG2 GSE40117 (Doktorova et al., 2013), to determine the B[a]P-affected metabolic signatures. Data were downloaded from the InSilico DB Genomic Datasets Hub (https://insilicodb.com/; Coletta et al., 2012), and analyzed by applying the ssGSEA method available in the GenePattern software. (http://genepattern.broadinstitute.org/). The ssGSEA method provides a representation of the gene expression data by assigning to each individual sample an Enrichment Score (ES) with respect to each gene set. ssGSEA. Heatmap visualization of the ssGSEA was then performed using GENE-E (http://www.broadinstitute.org/).

Glucose uptake assay
Glucose uptake was measured according to Kim et al (Kim et al., 2010) with some modifications. Briefly, after 48 hours of B[a]P treatment, cells were washed twice with serum-free, glucose-free William's medium supplemented with 0.1% BSA and pre-incubated with this medium for 3 hours at 37°C. After a starvation period, cells were washed twice with Krebs-Ringer-Bicarbonate buffer (KRB), and incubated further for 30 minutes at 37°C with 100 nM insulin (or not, for the negative control). To initiate glucose uptake, 2-deoxy-[1-3 H]-glucose (1μCi/mL), diluted in 0.1mM 2-deoxyglucose solution, was added to each well and further incubated for 10 minutes at 37°C. After incubation, cells were washed twice with ice-cold KRB and solubilized with 0.1N NaOH. Half of the content of each well was transferred into scintillation vials, and 10 mL of scintillation cocktail, Ultima Gold LLT, were added. The radioactivity incorporated into cells was measured using a liquid scintillation counter (Hewlett Packard, USA). The protein content was assayed for each point on the remaining half with the Pierce, bicinchoninic acid enzymatic kit (Pierce, France) after cell lysis in 0.1N NaOH. The results were expressed as the radioactivity incorporated related to the protein content.

Flow cytometry analysis of cell cycle
After a 48h-B[a]P treatment, cells were harvested and washed with phosphate-buffer saline (PBS). Cell nuclei were stained with propidium iodide using the Cycle Test™ PLUS DNA Reagent Kit (Becton Dickinson, San Jose, CA). DNA content of 20 000 cells/analysis was then monitored with a FACSCalibur flow cytometer (Becton Dickinson). Analysis of the cell cycle parameters was performed using the Modfit software (Becton Dickinson).

Analysis of cell phenotype using Transmission Electron Microscopy (TEM)
After drug exposure, cells were rinsed with 0.15 M Na cacodylate buffer and fixed by drop-wise addition of glutaraldehyde (2.5%) for 1 h. After fixation, the specimens were rinsed several times with 0.15 M Na cacodylate buffer and postfixed with 1.5% osmium tetroxide for 1 h. After further rinsing with cacodylate buffer, the samples were dehydrated with a series of ethanol solutions from 70 to 100%. The specimens were infiltrated in a mixture of acetone-Eponate (50/50) for 3 h and then in pure Eponate for 16 h. Finally, the specimens were embedded in DMP30-Eponate for 24 h at 60 °C. Sections (0.5 μm) were cut on a Leica UC7 microtome and stained with toluidine blue. Ultrathin sections (90 nm) were obtained, collected onto copper grids, and counterstained with 4% uranyl acetate and then with lead citrate. Examination was performed with a JEOL 1400 transmission electron microscope operated at 120 kV.

Western blotting
After treatment, cells were harvested and lysed for 20 min on ice in RIPA buffer supplemented with 1 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol, 1 mM orthovanadate, and a cocktail of protein inhibitors (Roche). Cells were then centrifuged at 13,000g for 15 min at 4 °C.
Thirty μg of whole-cell lysates were heated for 5 min at 100 °C, loaded in a 4% stacking gel, and then separated by 10% sodium dodecyl sulfate-polymerase gel electrophoresis (SDS-PAGE). Gels were electroblotted overnight onto nitrocellulose membranes (Millipore). After membrane blocking with a Tris-buffered saline (TBS) solution supplemented with 5% bovine serum albumin, membranes were hybridized with primary antibodies against E-cadherin or vimentin overnight at 4 °C, and next incubated with appropriate horseradish peroxidaseconjugated secondary antibodies for 1 h. For protein loading evaluation, primary antibodies against HSC70 or β-actin were used. Immunolabeled proteins were then visualized by chemiluminescence using the LAS-3000 analyzer (Fujifilm). Image processing was performed using Multi Gauge software (Fujifilm).

Evaluation of mitochondrial pH
The pH m was monitored using two mitochondrial targeted plasmids (Aequotech, Ferrara, Italy), both coding for cytochrome c oxydase subunit 8A mRNA: the mt-HA-eGFP pH-sensitive Green fluorescent protein (λex.=488 nm, λem.=509 nm), and the mt-dsRed pH-insensitive red fluorescent protein (λex.=530 nm, λem.=583) used as a transfection rate control. Production of these two plasmids was performed using the PureLink HiPure Plasmid Filter Maxiprep Kit (Life Technologies). F258 cells were co-transfected for 24h with both plasmids using X-tremeGene HP DNA Transfection Reagent (Roche, Meylan, France). After 24 h, cells were treated with B[a]P for 24 or 48h. Cells were then collected, re-suspended in Cell Suspension Buffer (see Huc et al., 2004, for CSB composition), and analysed by flow cytometry. Mean Fluorescence Intensity (MFI) was determined on 40 000 cells using a FACSCalibur (BD Bioscience). Fluorescence intensity was analyzed using the standard laser 488 nm laser filter configuration with FL1-H and FL3-H channels for monitoring mt-HA-eGFP and for DsRed respectively. A standard curve was generated in situ on control cells exposed to calibration buffers containing ionophores nigericin and monensin, allowing the conversion of the MFI ratio 530/640 into pH value. FCCP, a mitochondrial protonophore, was used as control for pHm acidification.

Evaluation of complex I activity
Complex I activity was measured on frozen cells as described (Bénit et al., 2008). In brief, cells were mixed with 1 ml extraction buffer containing 0.25 M sucrose 20 mM Tris-HCl, 40 mM KCl, 2 mM EGTA, 1 mg/ml BSA, pH 7.2 (medium A). Digitonin (0.01% final) and Percoll (10%) were added for 5 min on ice. Cells were subsequently spun down at 2,500 g for 5 min and washed two times with 1 ml of medium A. Permeabilized cell pellet was subsequently used for enzyme measurement.

Cell survival Glycolytic reprogramming
Schematic diagram illustrating the impact of B[a]P on cell metabolism in F258 rat hepatic epithelial cell line. Upon B[a]P exposure, a shift from OXPHOS to glycolysis is induced. Alterations in OXPHOS rely upon a mitochondrial complex II dysfunction that might result from dissociation triggered by acidification of matrix pH (pHm). Such a dissociation is associated with a decrease of fumarate and a parallel increase in succinate concentration. Stimulation of glycolysis which results in lactate production and extracellular pH (pHe) acidification, is dependent on NHE1 activation as reflected by an increased intracellular pH (pHi). A role for AhR has also been detected with regard to glycolysis that might go through the NHE1 activation resulting from AhR and H 2 O 2 -dependent membrane remodeling (Tekpli et al., 2010(Tekpli et al., , 2012. Finally, a nitric oxide (NO) production also triggers mitochondrial dysfunction leading to membrane hyperpolarization (Hardonnière et al., 2015). In total, the B[a]P-induced glycolytic shift is involved in cell survival.