Membrane lipid profile alterations are associated with

Membrane lipid profile alterations are associated with the metabolic adaptation of the Caco-2 cells to aglycemic nutritional condition Vera F. Monteiro-Cardoso, Amélia M. Silva, Maria M. Oliveira, Francisco Peixoto & Romeu A. Videira Journal of Bioenergetics and Biomembranes ISSN 0145-479X Volume 46 Number 1 J Bioenerg Biomembr (2014) 46:45-57 DOI 10.1007/s10863-013-9531-y

1 23

Your article is protected by copyright and all rights are held exclusively by Springer Science +Business Media New York. This e-offprint is for personal use only and shall not be selfarchived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23

Author's personal copy J Bioenerg Biomembr (2014) 46:45–57 DOI 10.1007/s10863-013-9531-y

Membrane lipid profile alterations are associated with the metabolic adaptation of the Caco-2 cells to aglycemic nutritional condition Vera F. Monteiro-Cardoso & Amélia M. Silva & Maria M. Oliveira & Francisco Peixoto & Romeu A. Videira

Received: 8 July 2013 / Accepted: 26 September 2013 / Published online: 12 October 2013 # Springer Science+Business Media New York 2013

Abstract Cancer cells can adapt their metabolic activity under nutritional hostile conditions in order to ensure both bioenergetics and biosynthetic requirements to survive. In this study, the effect of glucose deprivation on Caco-2 cells bioenergetics activity and putative relationship with membrane lipid changes were investigated. Glucose deprivation induces a metabolic remodeling characterized at mitochondrial level by an increase of oxygen consumption, arising from an improvement of complex II and complex IV activities and an inhibition of complex I activity. This effect is accompanied by changes in cellular membrane phospholipid profile. Caco-2

Electronic supplementary material The online version of this article (doi:10.1007/s10863-013-9531-y) contains supplementary material, which is available to authorized users. V. F. Monteiro-Cardoso : M. M. Oliveira : F. Peixoto : R. A. Videira (*) Chemistry Department, School of Life and Environmental Sciences, University of Trás-os-Montes and Alto Douro (UTAD), P.O. Box 1013, 5001-801 Vila Real, Portugal e-mail: [email protected] V. F. Monteiro-Cardoso : R. A. Videira Animal and Veterinary Research Centre (CECAV), UTAD, P.O. Box 1013, 5001-801 Vila Real, Portugal A. M. Silva Department of Biology and Environment, School of Life and Environmental Sciences, UTAD, P.O. Box 1013, 5001-801 Vila Real, Portugal A. M. Silva : F. Peixoto Centre for the Research and Technology of Agro-Environmental and Biological Sciences (CITAB), UTAD, P.O. Box 1013, 5001-801 Vila Real, Portugal M. M. Oliveira Chemistry Center – Vila Real (CQ-VR), UTAD, P.O. Box 1013, 5001-801 Vila Real, Portugal

cells grown under glucose deprivation show higher phosphatidylethanolamine content and decreased phosphatidic acid content. Considering fatty acid profile of all cell phospholipids, glucose deprivation induces a decrease of monounsaturated fatty acid (MUFA) and n-3 polyunsaturated fatty acids (PUFA) simultaneously with an increase of n-6 PUFA, with consequent drop of n-3/n-6 ratio. Additionally, glucose deprivation affects significantly the fatty acid profile of all individual phospholipid classes, reflected by an increase of peroxidability index in zwitterionic phospholipids and a decrease in all anionic phospholipids, including mitochondrial cardiolipin. These data indicate that Caco-2 cells metabolic remodeling induced by glucose deprivation actively involves membrane lipid changes associated with a specific bioenergetics profile which ensure cell survival.

Keywords Cancer cells lipidomics . Warburg effect . Metabolic remodeling . Mitochondrial bioenergetics . Phospholipids

Abbreviations CL GluL PA PC PE PI PS SM TLC TMPD

Cardiolipin Glycophospholipids Phosphatidic acid Phosphatidylcholine Phosphatidylethanolamine Phosphatidylinositol Phosphatidylserine Sphingomyelin Thin layer chromatography N,N,N ′,N′-tetramethyl-1,4-benzenediamine dihydrochloride

Author's personal copy 46

Introduction Colorectal carcinoma is a heterogeneous disease and the most common gastrointestinal malignancy in men and the third most common in woman (Kanthan et al. 2012). It develops through a multipathway sequence of events characterized, at cellular level, by a number of well described physiological alterations—i) self-sufficiency in growth signals; ii) insensitivity to growth-inhibitory signals; iii) programmed cell death evasion (apoptosis); iv) limitless replicative potential (immortalization); v) sustained angiogenesis and in advanced stages; vi) tissue invasion and metastasis (Hanahan and Weinberg 2000). Furthermore, cancer cells have long been known to exhibit profound alterations on their metabolic-bioenergetic profile (Warburg 1956). Reprogramming of cellular energetic metabolism, with an increase of glycolysis and reduced mitochondrial oxidative phosphorylation even in the presence of oxygen, first recognized by Otto Warburg in 1920, has also emerged as one of the most important hallmark of cancer (Warburg 1956; Weinberg and Chandel 2009; Smolkova et al. 2010). Tumor development and growth are dependent on the ability of cancer cells to adapt their metabolism in order to supply the biosynthetic and bioenergetic requirements for fast and uncontrolled growth and ultimately for tissue invasion (Munoz-Pinedo et al. 2012). In fact, cancer cells are continuously forced to rewire alternative metabolic routs for their energetic and anabolic needs due to tumor microenvironment subtract and oxygen availability/deprivation cycles. These metabolic switches are connected with both oncogenes and tumor suppressor (e.g. Ras, C-MYC, P53), which regulate the expression of key cellular proteins involved in the metabolic pathways that make cancer cells adapt to microenvironmental conditions and survive (Jose et al. 2013). Under glucose deprivation, cancer cells are able to use glutamine as main substrate to generate ATP via oxidative phosphorylation process and to ensure the biosynthetic precursors required for cell survival, suggesting the existence of alternative metabolic pathways to the classic forward direction of tricarboxylic acid cycle (TCA) (Reitzer et al. 1979; Filipp et al. 2012). An in vitro study using glioblastoma cells provides some evidences that glutamine conversion to α-ketoglutarate is the main pathway used by these cells to obtain malate, oxaloacetate and NADPH used for fatty acid biosynthesis (DeBerardinis et al. 2007). Additionally, it was shown that rout glutamine through TCA cycle is connected to oncogenic-induced tumor growth (Weinberg et al. 2010) and inhibition of mitochondrial glutamine metabolism suppress tumor metastasis and oncogenic transformation (Shelton et al. 2010; Wang et al. 2010). Despite glutamine being an effective carbon source able to supply metabolic and bioenergetics cancer cells requirements, when glucose is available in the tumor microenvironment, cancer cells metabolism sifts towards a glycolytic phenotype

J Bioenerg Biomembr (2014) 46:45–57

even in the presence of high oxygen concentration. Aerobic glycolysis, with high lactate production rate, provides metabolic advantages for proliferating cells which ensure both adequate levels of ATP and metabolic intermediates for biosynthesis of macromolecules (e.g. lipids) necessary to sustain all mitotic process (Warburg 1956; Greiner et al. 1994). Enhanced lipid biosynthesis is a key feature to support the rapid synthesis of new membrane and organelles for daughter cells (Menendez and Lupu 2007). The majority of fatty acids in cancer cells are derived from de novo fatty acid biosynthesis, supported by an increased enzymatic activity of acetylCoA carboxylase and fatty acids synthase (Mashima et al. 2009). Changes in membranes lipid composition, mainly in the phospholipid monounsaturated fatty acids and cholesterol contents, lead to alterations in their biophysical properties, cellular signal transduction and gene expression (Murai 2012; Pegorier et al. 2004). Additionally, reorganization of plasma membrane components leading to lipid rafts formation, enriched with protein caveolin-1 and phosphoinositides for increased signaling of cell growth receptors, is intimately associated with metastatic dissemination process (Yamaguchi and Oikawa 2010). In this view, the present study aimed to characterize how the membrane lipid profile is correlated with metabolic adaptations triggered in Caco-2 cells by glucose deprivation.

Material and methods Chemicals Chloroform, methanol and ethanol were purchased from Panreac (Barcelona, Spain). TLC silica gel 60 plates were purchased from Merck (Darmstadt, Germany). Primuline Acetyl-CoA, DTNB as well as other chemicals used in bioenergetics studies were purchased from Sigma-Aldrich. All the compounds were the highest commercially available quality. Cell culture The human epithelial Caco-2 cells (Caco-2; ATCC, Rockville, MD) were maintained in Glucose Dulbecco’s Modified Eagle Media (DMEM), containing 25 mM glucose (glucose medium) supplemented with 10 % (v/v) FBS (Gibco, Life technologies), 2 mM L-glutamine (Gibco, Life technologies) and antibiotics (100 U/L penicillin and 100 μg/L streptomycin, Life technologies) at 37 °C in normal atmosphere of 5 % CO2 in air. Alternatively, Caco-2 cells were grown using the same culture conditions in glucose-deprived medium, in which glucose was isosmotically replaced by 25 mM galactose (glucose-deprived medium). For bioenergetics and lipidomic studies, cells were grown in T75 flasks during 7 days with medium replacement every

Author's personal copy J Bioenerg Biomembr (2014) 46:45–57

2 days. Cells were harvested by washing with HBSS followed by trypsin-EDTA treatment. In order to compare the morphology of cells exposed to glucose and glucose deprivation, during growth, cells were routinely observed, by optic microscopy, and for the longer period of growth, a set of cells were observed by electronic microscopy. For both growth conditions, no morphologic differences were observed. Cell growth and viability Caco-2 cells were seeded onto 96-well microplates at a cell density of 2×103 cells per well and placed at 37 °C in a CO2 incubator. For estimation of the cell survival rate, 10 % (v/v) AlamarBlue (Invitrogen Corporation) was added directly to each well, and absorbance was monitored at 570 and 620 nm, after 2 h culture, as described by the manufacturer’s protocol, and at several timescale points (daily intervals) as indicated in the Results section. The percentage of AlamarBlue reduction was calculated using the equations recommended by the manufacturer’s protocol. At the same time, cell proliferation studies were carried out by planting Caco-2 cells at a density of 2×104 cells per well on 12-well plates containing 1 mL of glucose or glucose-free medium. At daily intervals, cells were harvested by trypsinization and counted using a Neubauer chamber. Respiration measurements Oxygen consumption was measured on intact and permeabilized cells in a Clark-type oxygen electrode (Hansatech, Norfolk, UK), at 37 °C using 3 mg of cellular protein (4.8× 106 ±0.2×106 cells), as determined by the biuret method (Gornall et al. 1949). Routine respiration- Oxygen consumption of intact cells was evaluated in glucose or glucose-deprived growth medium without FBS, thus the main metabolic substrates are glucose and glutamine, respectively. The reaction was stopped by adding 2 mM potassium cyanide (KCN). Permeabilized cells respiration- Cells were resuspended in a MIR05 buffer (110 mM sucrose, 0.5 mM EGTA, 3 mM MgCl2, 80 mM KCl, 60 mM K-lactobionate, 10 mM KH2PO4, 20 mM taurine, 20 mM HEPES, 1 g/L BSA, pH 7.1) and immediately permeabilized with digitonin (0.005 %, w/v) (Stadlmann et al. 2006). Mitochondrial respiration was activated by 10 mM pyruvate plus 10 mM malate followed by addition 2.5 mM ADP in order to activate oxidative phosphorylation. Basal respiration was accessed after addition 3 mg/mL oligomycin. The reaction was stopped by adding 3 μM rotenone, 3 μM antimicine and 2 mM KCN, specific inhibitor of complexes I, II and IV, respectively. Parallel assays were performed in the presence of 150 μM cytochrome c to evaluate the integrity of the outer mitochondrial membranes of

47

digitonin-permeabilized cells. Oxygen consumption was expressed as nmol O2/min./mg of cell protein. Mitochondria isolation Caco-2 mitochondria were isolated as previously described (Frezza et al. 2007), with some modifications. Cells (from 5× 107 to 7×107) were homogenized in 210 mM manitol, 70 mM sucrose, 5 mM HEPES, 1 mM EGTA, 0.5 % (w/v) BSA, pH 7.4 buffer using a glass/teflon Potter Elvehjem at 1,600 rpm, for 5 min at 4 °C. Homogenate was centrifuged 5 min at 1,000g (Sigma 2-16K; rotor 12139). The supernatant was collected and centrifuged 15 min at 14,000g. Then, the pellet was resuspended in the washing buffer (210 mM manitol, 70 mM sucrosose, 5 mM HEPES, pH 7.2) and centrifuged 15 min at 14,000g. The final concentration of mitochondrial protein was determined by the biuret method (Gornall et al. 1949), using BSA as standard. The functionally of the mitochondrial preparation was evaluated determining the respiratory control ratio (RCR), using a similar procedure described by permeabilized cells. Only those mitochondrial preparations with RCR ≥ 2 were used to assess mitochondrial enzymes activities. Determination of mitochondrial enzymes activities Before assay enzymatic activities, mitochondria were submitted to four cycles of freezing/thawing to give substrate access to the mitochondrial matrix. Mitochondrial enzymatic assays were conducted at 30 °C using 0.3 mg of mitochondrial protein. Complexes I, II and IV and FoF1-ATPase activities were normalized with citrate synthase activity. Citrate synthase (CS) activity was determined as described previously (Shepherd and Garland 1969) with slight modifications. The reactions were conducted in 200 mM Tris–HCl buffer (pH 8.0) supplemented with 0.02 % (v/v) Triton X-100, 10 μM DTNB and 1 mM oxaloacetic acid. DTNB reduction was monitored at 412 nm using a Varian Cary 50 Bio spectrophotometer (Agilent Technologies, Santa Clara, CA, USA) after adding 0.37 mM acetyl-CoA . The specific activity of citrate synthase was calculated by DTNB (DO 13,600 M−1 cm−1) reduction rate, determined in the linear range of the plot and expressed as nmol/min./mg protein. Complex I (NADH-ubiquinone oxidoreductase) activity was evaluated spectrofluorometrically as described previously (Melo et al. 2012). NADH (50 μM) fluorescence intensity was evaluated before and after 162.5 μM dodecylubiquinone were added to the reaction medium (25 mM KH2PO4, 10 mM MgCl2; pH 7.4) supplemented with 1 mM KCN using a Varian Eclipse fluorescence spectrophotometer (Agilent Technologies, Santa Clara, CA, USA) at 450 nm, setting excitation at 366 nm. In the final phase of each assay, a specific complex I inhibitor (rotenone 3 μM) was added. Specific complex I activity was expressed in arbitrary units/min./mg.


Complex II (Succinate-Coenzyme Q reductase) activity was evaluated in a 25 mM potassium phosphate buffer pH 7.4, supplemented with mitochondrial suspension, 2 mM KCN, 6.5 μM rotenone, 6.5 μM antimycin A, 0.05 mM DCIP (2,6-dichloroindophenol) and 0.1 mM decylubiquinone. DCIP reduction was followed at 600 nm after adding 20 mM succinate (Barrientos et al. 2009). The reaction was stopped by adding 0.5 mM oxaloacete (complex II inhibitor). The specific complex II activity was calculated in the linear range of the plot corresponding to DCIP reduction and expressed as nmol/min./mg protein. Complex IV enzyme activity was evaluated following oxygen consumption (Melo et al. 2012), in an Clark-type oxygen electrode (Hansatech, Norfolk, UK), associate to cytochrome c oxidation promoted by 5 mM ascorbate/2.5 mM N,N,N′,N′tetramethyl-1,4-benzenediamine dihydrochloride (TMPD). The reaction occurred in standard buffer (130 mM sucrose, 50 mM KCl, 5 mM MgCl2, 5 mM KH2PO4 and 5 mM HEPES; pH 7.2) supplemented with 3 μM rotenone (specific inhibitor of complex I), 0.1 μM antimycin A (inhibitor of complexe III) and 15 μM cytochrome c. In order to achieve specific complex IV activity 2 mM KCN (inhibitor of complex IV) were added. Complex IV activity was expressed as nmol O2/min./mg protein. FoF1-ATPase activity was determined by an electrometric technique (pH electrode) (Madeira et al. 1974) using Crison pH evaluation system (Barcelona, Spain) connected with Kipp and Zonen recorder (Omni Instruments, Dundee, UK). Assay buffer (130 mM sucrose, 60 mM KCl, 0.5 mM HEPES and 2.5 mM MgCl2; pH 7.0) was supplemented with 3 μM rotenone reactions and 2 mM ATP-Mg initiated protons production. The reaction was completely abolished after adding oligomycin (1 μg). At the end of each assay, pH titration was performed to system calibration, using a standard HCl solution. The specific activity of FoF1-ATPase was expressed in nmol H+/min./mg of protein. Extraction of lipid crude Total lipids from Caco-2 cells were extracted with the Bligh and Dyer method (Bligh and Dyer 1959), with some modifications. Total lipids extraction was performed using a solvent combination of methanol/chloroform/water (2:1:0.8, v/v/v). A volume of cells suspension corresponding to 5 mg of protein was mixed with 2 mL methanol and 1 mL Chloroform and then centrifuged at 2,000g (Sigma 2-16K, rotor 11192) for 5 min at 4 °C. The pellet was discharged and an additional volume of 1 mL chloroform and 2 mL H2O were added to supernatant following vigorous mixing. Finally, samples were centrifuged at 2,000g for 5 min at 4 °C, in order to obtain a two-phase system: an aqueous top phase and an organic bottom phase from which lipids were collected. The extracts were dried in a nitrogen stream and stored at −20 °C.


Cholesterol quantification Cholesterol of total lipid extracts was quantified according to Liebermann-Burchard method with some modifications (Huang et al. 1961). Shortly, lipid sample (50–100 μL) were dried with nitrogen stream and dissolved in 0.1 mL acetic acid. Then, 5 mL Liebermenn-Burchard reagent (120 mL anhydride acetic acid, 20 mL H2SO4, 60 mL CH3COOH, 4 g Na2SO4) were added and the reaction occurred for 30 min at room temperature. Standards from 0 to 250 μg cholesterol underwent the same sample treatment. The absorbance of standards and samples were measured at 630 nm in a Varian Cary 50 Bio spectrophotometer. Cholesterol content in total lipid crude was expressed in nmol/mg of cell protein. Phospholipids carbohydrate quantification The carbohydrate content of glycolipids from Caco-2 lipid extracts was determined spectrofotometrically by phenolsulfuric acid assay (Mazurek et al. 2012). Lipid sample (50– 100 μL) were dried with nitrogen stream and 0.1 mL of destilled water were added and then, mixed with 0.5 mL of 5 % (v/v) phenol solution in distilled water and 2.5 mL concentrated H2SO4, with continuous stirring. Standards from 0 to 100 μg glucose were prepared underwent the same sample treatment. The reaction occurred during 30 min at room temperature and the absorbance was measured at 490 nm in a Varian Cary 50 Bio spectrophotometer. Carbohydrate content in total lipid crude was expressed in nmol glucose equivalents/mg of cell protein. Phospholipids quantification In order to quantify membrane phospholipid (PL) content, phosphorus assay was performed according to the method outlined by Bartlett and Lewis (Bartlett 1959). Briefly, 0.65 mL of concentrate perchloric acid (70 %, w/v) were added to lipid samples (50–100 μL) previously dried with nitrogen stream and then incubated for 2 h at 180 °C. About 3.3 mL of H2O, 0.5 mL of 1 % (w/v) ammonium molybdate and 4 % (w/v) ascorbic acid were added to all samples, followed by incubation for 10 min at 100 °C in a water bath. Standards from 0 to 250 nmol phosphate underwent the same sample treatment. Finally, the absorbance of standards and samples were measured at 800 nm in a Varian Cary 50 Bio spectrophotometer. Separation of phospholipids classes by thin-layer chromatography (TLC) Phospholipid classes from the total lipid extract were separated by TLC using silica gel plates. Prior to separation, plates

Author's personal copy J Bioenerg Biomembr (2014) 46:45–57

were washed with chloroform/methanol (1:1, v/v) and treated with 2.3 % (w/v) boric acid in ethanol. The plates with spots containing between 20 and 30 μg of phosphorous were developed in solvent mixture chloroform/ethanol/water/ triethylamine (35:30:7:35, v/v/v/v). Lipid spots on TLC plates were detected with a UV lamp (λ=254 nm) after primuline (50 μg/100 mL in acetone/water, 80:20; v/v) exposure and identified by PL standards comparison. The spots were scraped off the plates to lipid extraction and quantification. The percentage of each PL class was calculated, relating the amount of phosphorus in each spot to the total amount of phosphorus in the sample, thus giving the relative abundance of each PL class. Preparation of esters derivate of fatty acids Fatty acid methyl ester (FAME) was obtained by acidcatalyzed transmethylation (Peixoto et al. 2004). The lipid extract (400–900 nmol) was dissolved in 5 mL 5 % (v/v) of HCL in methanol (freshly prepared) and 10 mM C17:0 was used as internal standard. Mixture was vigorously vortexed for 1 min and incubated at 70 °C for 120 min. After cooling down at room temperature, 5 mL n-hexane were added and mixed followed by centrifugation at 2,000g for 5 min. The organic phase (hexane containing FAME) was collected, dried with Na2SO4 anhydrous, filtered into a new tube and finally evaporated under a stream of nitrogen. The dry residue was solubilized in a small volume of n-hexane. Fatty acid composition analysis The hexane FAME solution coming from derivatization step was analyzed by Gas Chromatography (GC), using a ThermoFinningan:Trace Gas Chromatograph connected to the mass spectrometer Polaris Q MSn equipped with an Ion Trap analyzer. The column used was a Supelcowax 10 M (30 m, 0.32 mm I.D., 0.50 mm film thickness, 0.45 mm O.D.) from Supelco, Bellefonte, Pennsylvania, USA. Injections were made in splitless mode (1 min) with an injection volume of 1 mL. The column oven temperature was settled at an initial value of 140 °C (1 min hold), increased to 240 °C with a ramp of 2.0 °C/min. (9 min hold), and finally increased to 280 °C with a ramp of 4.0 °C/min. Total runtime was 70 min using Helium as the carrier gas, at a constant flow of 1 mL/min. The injector temperature was set at 250 °C. Data acquisition and treatment of results were carried out with an Xcalibur data system (V2.0, ThermoFinnigan, San Jose, CA, USA). Fatty acids were identified by comparison with retention time and fragmentation profile of reference standard mixtures FAME 37 (Supelco 37 Component FAME Mix) and quantified using peak area of the internal standard. The peroxidability index (PI) and unsaturation index (UI) were calculated, as previously described (Peixoto et al. 2004).

49

Statistical analysis All the data presented in this study correspond to the mean value of n experiments ± SEM, with n ≥3. Comparison of two data sets was performed with the Student’s t test, using KaleidaGraph. Data sets were considered statistically different when p PS> > PI>> > PE > PC (Kagan et al. 2009). Therefore, our results suggest that membrane lipids protection against peroxidation is one important feature behind the metabolic remodeling process to aglycemic nutritional condition allowing cells survival.

Discussion Tumor proliferation is a complex and integrated process that requires changes in the basic mechanisms of coupling energyproduction with energy-utilization (metabolic network) to endow the cells with properties which favor the mitotic process and detract apoptosis. Cancer cells exhibit metabolic remodeling characterized by high rate of aerobic glycolysis which ensure both energetic and biosynthetic subtracts for growth and proliferation. Even though, mitochondrial oxidative phosphorylation is reduced it is known that most cancer cells retain functional mitochondria and thus the capacity to adapt their metabolism to environmental changes, including availability of subtracts (e.g. glucose). Additionally, functional mitochondria are essential to citrate production, through tricarboxylic acid cycle, required for de novo fatty acids synthesis and consequently membrane lipids (Weinberg and Chandel 2009; Menendez and Lupu 2007; Reitzer et al. 1979). In this study, we investigate the bioenergetics adaptations of

Caco-2 cancer cells to microenvironmental conditions of glucose deprivation and its relationship with membrane lipid profile. Our results indicate that although glucose deprivation limits cell replicating ability, bioenergetics metabolic remodeling coupled with characteristic membrane phospholipid profile ensures the functionality and survival of cancer cells in nutritional hostile environment. Cell viability and proliferation studies performed in the presence and absence of glucose throughout 144 h shown that glucose-free medium has less ability to sustain Caco-2 cells proliferation, as detected by cell number, but enhances the reducing power of living cells, as evaluated by AlamarBlue assay (Fig. 1). Cell health in glucose free medium was also confirmed by 3-fold increase of its endogenous oxygen consumption rate, in relation with glucose rich medium (Fig. 2) and by the significant improvement of the mitochondrial coupled respiration as detected in digitonin-permeabilized cells assays (Table 1). These data indicate functional changes in mitochondria, as detected by an inhibition of complex I and by an enhancement of complexes II and IV activities. Low complex I activity associated with high complex II activity, detected in Caco-2 cells grown in glucose-free medium, suggests that this cells use a truncated citric acid cycle to accumulate fumarate or malate, which may be transported to the cytosol increasing the reducing power of cancer cells cytoplasm, supporting the high rate of biosynthetic pathways needed to cell proliferation (Filipp et al. 2012; DeBerardinis et al. 2007). Furthermore, the mitochondrial ability to synthesize ATP by oxidative phosphorylation is preserved since the FoF1-ATPase activity is similar in both nutritional conditions (Fig. 3). Mitochondrial responses of cancer cells to changes of carbon source, namely glucose to glutamine switch, have also been reported in other cancer cell lines. Rossignol and coworkers showed that HeLa cells, specialized to produce ATP via glycolic pathway, can drive their energy through oxidative phosphorylation when glucose is not available by remodeling mitochondrial activity, detected by an increase of respiratory rate and cytochrome c oxidase (complex IV) activity (Rossignol et al. 2004). HTB-126 cells, derived from ductal carcinoma of human breast, can also turn on their oxidative phosphorylation capacity when glucose is replaced by galactose. It was shown that glucose deprivation leads to an increase of HTB-126 routine respiration which is accompanied by a stimulation in the expression of oxidative phosphorylation proteins (e.g . complex IV), indicating a general improvement of oxidative phosphorylation process (Smolkova et al. 2010). Likewise, glucose withdrawal from HepG2 hepatoma cell culture medium stimulates oxidative phosphorylation system with 2-fold increase of complex IV activity and elevate the mitochondrial DNA levels, including mitochondrial DNA encoded mRNA and proteins (Weber et al. 2002).



Fig. 4 a Relative abundance (mol%) of each phospholipid classes in total lipid extract obtained from Caco-2 cells grown in glucose rich medium (dark bars) and glucose deprived medium (grey bars); PC phosphatidylcholine, SM sphingomyelin, PI phosphatidylinositol, PS phosphatidylserine, PE phosphatidylethanolamine, PA phosphatidic acid, CL cardiolipin, GluL1 glycophospholipid (TLC spot 7), GluL2 glycophospholipid (TLC spot 8). b Fatty acid profile of all phospholipid obtained from Caco-2 cells grown in glucose (dark bars) and glucosedeprived (grey bars) culture medium; C12:0 - lauric acid, C14:0 - myristic acid, C16:0 - palmitic acid, C18:0 - stearic acid, C20:0 - arachidic acid, C22:0

- behenic acid, C24:0 - lignoceric acid, C14:1 - myristoleic acid, C16:1 palmitoleic acid, C18:1 - oleic acid, C20:1 - eicosenoic acid, C22:1 - erucic acid, C24:1 - nervonic acid, C18:2n-6 - linoleic acid, C18:3n-6 - γ-linolenic acid, C20:2n-6 - eicosadienoic acid , C20:4n-6 - arachidonic acid, C22:2n-6 decosadienoic acid, C22:5n-6 - docosapentaenoic acid n-6, C18:3n-3 - αlinolenic acid, C20:3n-3 - eicosatrienoic acid, C20:5n-3 - eicosapentaenoic acid (EPA), C 22:5n-3 - docosapentaenoic acid n-3, C 22:6n-3 docosahexanoic acid (DHA). Values are means ± SEM with n ≥3. *Significant differences between glucose-deprived and glucose nutritional conditions, with p