Open Full Page

[CANCER RESEARCH 46, 3768-3774, August 1986] 31PNuclear Magnetic Resonance Study of a Human Colon Adenocarcinoma Cultured Cell Line1

Franck Desmoulin, Jean-Philippe Galons, Paul Cantoni, Jacques Marvaldi, and Patrick J. Cozzone2 Laboratoire de Biologie Physicochimique, Institut de Chimie Biologique, UniversitÃ©d'Aix-Marseille, Place Victor Hugo 13003 Marseille (France)

ABSTRACT Saccharomyces cerevisiae (14-16). Investigation of mammalian cell lines has received only scant appraisal because of several â€¢"Pnuclear magnetic resonance (NMR) spectroscopy has been used to technical difficulties. Work has been limited to a few cell types, monitor the energy metabolism in a human colon adenocarcinoma cell essentially HeLa cells (17, 18), Ehrlich ascites tumor cells (19), line (HT 29). NMR spectra were recorded at 80.9 MHz on approximately 2.5 x 10" cells continuously perfused with culture medium within a 20- and normal and transformed fibroblasts (20, 21). In addition, a detailed analysis of the metabolite content of radiation-induced iiini NMR sample tube. Typical NMR spectra display a series of well-resolved resonances fibrosarcoma cells by using multinuclear NMR has been re assigned to nucleoside triphosphates (mainly adenosine S'-triphosphate), cently published (22). uridine diphosphohexose derivatives (uridine 5'-diphosphate-A/-acetyl- A major difficulty to overcome in the study of perfused cells glucosamine, uridine 5/-diphosphate-Ar-acetylgalactosamine, uridine 5'- is the preservation of physiological conditions in the NMR diphosphate-glucose), intra- and extracellular inorganic phosphate, and sample tube where the accumulation of metabolite by-products phosphomonoesters (mainly phosphorylcholine and glucose 6-phos- and the required bubbling of gas may affect the viability of the phate). Measurement of phosphorylated metabolite concentrations from cells. Various perfusion techniques have been developed to the intensity of NMR signals is in good agreement with the results ensure the protection of cell metabolic integrity (23-26). These provided by conventional biochemical assays. techniques have been discussed in the recent review by Evan- 31P NMR allows to follow noninvasively the effect of anoxia on HT 29 cells. The results indicate that the cells are able to maintain about ochko et al. (27) on the application of NMR to the study of isolated tumor cells, excised tumors, and in vivo tumors. 60% of their initial nucleoside triphosphate level after 2 h of anaerobic In this report, we present results of the 3IP NMR study of a perfusion. Cells accumulate inorganic phosphate during anoxia and the intracellular-extracellular pH gradient increases from 0.5 in well-oxygen human colon adenocarcinoma cell line (HT 29 line) established ated cells to more than l pH unit under anoxic conditions. The value of in 1964 by J. Fogh (28). All NMR measurements were carried intracellular pH of well-oxygenated HT 29 cells is 7.1. out on cell preparations constantly perfused with the culture The effect of glucose starvation upon energy metabolism has also been medium while in the NMR sample tube. The perfusion medium examined in real time by NMR: a rapid decline of adenosine 5'-triphos- provided the cells with various substrates and oxygen while phate down to 10% of the initial value is observed over a period of 2 h. removing damaging metabolite by-products. The perfusion sys In contrast, the level in uridine diphosphohexoses reaches a new steady tem was kept at constant temperature in order to maintain the state value representing 60% of the initial one. Refeeding the cells with 25 HIMglucose leads to a dramatic drop of internal pH reflecting the cells under the steady state conditions required for NMR signal activation of the glycolytic pathway. averaging. Intracellular levels of phosphate metabolites and external and cytosolic pH values were simultaneously measured from the NMR spectra of well-oxygenated cells or hypoxic INTRODUCTION cells. The effect of glucose starvation which has been shown to alter the energetic metabolism of HT 29 (29, 30) has also been Over the past few years, development of instrumental tech nology in NMR3 has permitted the observation of phosphorus investigated in this NMR study. The results illustrate the pos sibility offered by NMR to analyze in a noninvasive manner containing metabolites in living systems. The technique has precise metabolic variations of cultured cell lines as a valuable been extensively applied to the study of either perfused organs contribution to the understanding of specific metabolic events or suspensions of whole cells. More recently, it has been ex related to cell growth and cell differentiation. tended to whole animals and humans because of the availability of wide-bore horizontal magnets and the development of meth ods which provide a spatial localization of the observed metab MATERIALS AND METHODS olites (1-5). 3IP NMR spectroscopy provides a unique way of Cell Culture and Growth Conditions. HT 29 cells were routinely simultaneously measuring extra- and intra-cellular pH on the grown as monolayers in ISO-cm2 sterile plastic flasks (Corning) in basis of the chemical shift, and determining stationary intracel Dulbecco's modified Eagle's medium containing 25 HIMglucose and lular concentrations of phosphate metabolites and their fluctua supplemented with 10% fetal calf serum. The culture medium was tions under a variety of metabolic events. So far, the majority replaced every 2 days. Cells were cultured for 2 weeks in an incubator of the studies on cells by 3IP NMR has been dedicated to at 37Â°Cin a humidified atmosphere containing 95% air and 5% COz prokaryotic cells including Escherichia coli (6-8), Staphylococ- until they reached confluency. They were harvested at confluency after cus aureus (9), Tetrahymena (10), Rhodopseudomonas sphae- a 5-min treatment at 20Â°Cwith 0.05% Trypsin-0.5 HIM EDTA in a Ca2+- and Mg2+-free 150 mM phosphate buffer, pH 7.2. For NMR roides (11), and yeast, Acanthamoeba castellani (12, 13), and experiments, trypsinized cells were pelleted by centrifugation. The Received 12/4/84; revised 1/21/86; accepted 4/17/86. pellet was washed twice with fresh culture medium and the cells were 1This work was supported by the Centre National de la Recherche Scientifique resuspended in 1 ml of the same medium to give a final count of 2.5 x (UÃ€202) and grants from the MinistÃ¨re de la Recherche et de la Technologie 10*cells confined into a 0.7- x 10-cm flat dialysis membrane (Spectra- (GBM 83-M-0802, 84-M-0804 and 85-M-0564), the CNAMTS (Contrat de Recherche Externe 1983-1986), thÃ©MinistÃ¨rede l'Education, and thÃ©Fondation por). Viability of cells was tested by their ability to exclude trypan blue. Typically, viability values of 90-95 and 70-90% were measured before pourâ€¢'Towhome la Recherche requests MÃ©dicale. for reprints should be addressed. and after the NMR experiment. Similar results on cell viability were 'The abbreviations used are: NMR, nuclear magnetic resonance; DPDE, obtained by the 51Crrelease technique. diphosphodiesters (mainly uridine diphosphohexose or diphosphohexosamine derivatives); G-6-P, glucose 6-phosphate; NTP. nucleoside triphosphates; PÂ¡â„¢, Cell Extracts. After removal of the medium from the flasks, the cell intracellular PÃ¢;P", extracellular PÂ¡;NDP, nucleoside diphosphates. layers were rinsed once with ice-cold 0.9% NaCl solution and then were 3768

Downloaded from cancerres.aacrjournals.org on October 3, 2021. © 1986 American Association for Cancer Research. 31P NMR OF HUMAN COLON ADENOCARCINOMA CELLS frozen by flotation of the flask on liquid nitrogen. The frozen cell layer MDPA Pi was scraped with 2.4 ml of ice-cold 0.9 M HC1O4. The mixture was kept 10 min at 0Â°Cand then centrifuged for 5 min at 5000 x g and 4Â°C.The clear supernatant was neutralized with 5 N KOH and the NTPa NAD solution centrifuged for 10 min at 8000 x g and 4Â°C.Extracts were + + stored at -70Â°C. NDPa DPDE NMR Spectroscopy and Cell Perfusion. 31PNMR spectra of perfused HT 29 cells were recorded on a Nicolet NT200 wide-bore spectrometer operating at 80.9 MHz. Each spectrum corresponds to the Fourier NTP/3 transform of the sum of 600 free induction decays (10-min accumula tion). Flip angle of 60Â°with repetition time of 1.2 s was selected in order to optimize the signal-to-noise ratio and to minimize the satura tion of the signals arising from intracellular metabolites and I',". Under those conditions, a partial saturation (15%) of external PÂ¡wasobserved. All chemical shifts were expressed as ppm relative to 85% phosphoric acid, using as secondary reference the signal at 20.8 ppm of hexachlo- rocyclotriphosphazene contained in a capillary (31) for in vivo experi ments. In some cases, the glycerophosphorylcholine signal at 0.50 ppm was used as internal reference at pH 7.1 and 37Â°C. The dialysis membrane containing the cell suspension as described above was folded 4-5 times in order to fit into a 20-mm NMR sample r ' tube. The 1-ml solution bathing the cells was continuously renewed by 20 10 -10 -20 PPM Fig. 1. Proton-coupled 31PNMR spectrum (80.9 MHz) of the perchloric acid perfusing the system with 180 ml of culture medium. The volume of extract of HT 29 cells. The solution contains 1 mM methylenediphosphonic acid fresh medium inside the NMR tube was around 10 ml. The perfusate (MDPA) as internal reference. The spectrum corresponds to the average of 10 was maintained at pH 7.5 and oxygenated with a 5% CO2-95% air blocks of 624 scans each. The signal-to-noise ratio was improved by multiplying mixture using a membrane oxygenator (Scimed). the free induction decay with an exponential (5 Hz line broadening). The extract 31P NMR spectra of the cell perchloric acid extracts were recorded (8 ml total volume) was prepared from 400 x IO6 cells and corresponds to a protein concentration of 20 mg/ml. PC, phosphorylcholine; GPE, glycerophos- under fully relaxed conditions in order to accurately quantitate the content in phosphorylated metabolites. Typical conditions were 90Â° phorylethanolamine; PCr, phosphocreatine. GPC, glycerophosphorylcholine. pulses and 30-s delays between acquisitions. 31P resonances in the Table 1 Phosphorylated metabolite content of HT 29 cells as determined by" P spectrum of the extracts were referenced to internal mÃ©thylÃ¨nediphos- NMR and biochemical analysis phonic acid (triplet centered at 17.8 ppm). Concentration" Proton NMR spectra were recorded at 200 MHz in 5-mm tubes. protein)MetaboliteNTP (nmol/mg Proton chemical shifts were referenced to external sodium-2,2- dimethyl-2-silapentane-5 sulfonate. The residual HDO signal was re assay23.2 duced by homonuclear presaturation. A spectrum width of 2000 Hz, 90Â°-pulseangles, and 1-s delays were usually selected. I3C NMR spectra of the cell extracts were recorded at 50.3 MHz in ATP XTP* 7.63 20-mm tubes with 60Â°pulses, 2-s recycling time, and a spectrum width NDP of 11,000 Hz. Noise-modulated broadband proton decoupling was used ADP XDP*Biochemical with high-level power (5 W) during the aquisition time and low-level 0.8NMR305 power (1.5 W) during the delay. DPDEUDP-hexoseUDP-/V-acet) IhexosaminePhosphocreatineGlucose RESULTS NMR Spectra of Perchloric Acid Extracts of HT 29 Cells. 6-phosphatePhosphory'cholineGlycerophophorylcholineP,2.124.212.77.13313.78.219.410.535 The proton-coupled 31P NMR spectrum of a perchloric acid extract of HT 29 cells is shown in Fig. 1. Methylenediphosphonic acid in known concentration was " Concentrations based on biochemical analysis have been derived from the added to the extract in order (a) to simplify peak identification work by Wice et al. (32). They are compared with values calculated from 3IP NMR spectra of perchloric acid extracts. on the basis of chemical shift values and (/;) to quantitate the * Sum of GTP, UTP, and CTP. phosphate metabolite contents in the extract. For an unambig c Sum of GDP, UDP, and CDP. uous assignment of the resonances, the titration behavior of the signals from the extract has been followed as a function of pH metabolites as shown on Fig. 1 is directly proportional to their and compared with the titration profile of authentic samples concentration in the perchloric acid extract since the NMR under the same experimental conditions. The main resonances spectrum was acquired under fully relaxed conditions. The can be assigned to PÂ¡andNTPâ€ž,NTP^, and NTPT. The intense calculated concentrations, using the methylenediphosphonic signals at â€”12.5ppm and â€”11.1ppm correspond to diphospho- acid resonance as an internal standard, are given in Table 1. diesters which have been shown to accumulate in nondifferen- For the sake of comparison, the contents in the same metabo tiated HT 29 cells (30). These compounds have been identified lites measured in HT 29 by Wice et al. (32) using conventional by high performance liquid chromatography as UDP-glucose, biochemical assays is also given in Table 1. The agreement UDP-galactose, UDP-/V-acetylglucosamine, and UDP-A'-ace- between the two sets of data is very satisfactory. tylgalactosamine (32). In the phosphomonoester region (4-6 The proton NMR spectrum of the extract is given in Fig. 2. ppm), the peaks at 4.1 and 5.4 ppm arise from phosphorylcho- Based on literature values and on the resonances of pure line and glucose 6-phosphate, respectively. The weak signal at compounds added to the perchloric acid extracts, the following 4.9 ppm could correspond to nucleoside monophosphates or to resonances may be identified: in the aliphatic region (Fig. 2/4, phosphorylethanolamine. 0-5 ppm), the signal from the methyl group of the lactic acid The area of the NMR signals arising from phosphorylated at 1.28 ppm (doublet); and signals from amino acids (leucine, 3769

Lac LEU

Lac

0 PPM

80 60 40 20 0 PPM Fig. 3. Region between 0 and 80 ppm of the "C NMR spectrum of the perchloric acid extract of HT 29 cells. The spectrum corresponds to 13,200 scans accumulated overnight (5 Hz line broadening). IK. lactate; I'KO. proline; GLU, glutamic acid; O7..V.glutamine; TAU, taurine; dl. Y.glycine; LEU, leucine; PCR, phosphocreatine; CR, creatine; ASN, asparagine; X, non-identified resonances.

NAD DPDE

Fig. 2. Proton NMR spectrum (200 MHz) of the perchloric acid extract of HT 29 cells. A, high-field region of the spectrum (240 scans; 0.5 Hz line broadening). II. low-field region of the spectrum (240 scans; 1 Hz line broadening). asparficVertical acid;expansion Asn, asparagine; is 20 times Clio, higher choline; than that Cr, ofcreatine; spectrum Glu, A. glutamic Ala, alanine; acid;Asp, ('Â¿In. glutamine; G/y, glycine; Lac, lÃ¡clate;Leu, leucine; PCr, phosphocreatine; Pro, proline; Tau, taurine;. l. adenine; C, cytosine; G, guanine; I . uracyl; Glc, glucose; Tyr, tyrosine; Phe, phenylalanine.

proline, glycine, glutamic acid, glutamine), taurine, phospho creatine, and creatine. In the aromatic region (Fig. 2Ã„,5-9 ppm), the spectrum bears resonances arising from glucose and glucose 6-phosphate (5.2 and 5.35 ppm, respectively), from the H'-l ofribose in nucleotidesand nucleosides(6-6.3 ppm),from aromatic protons of tyrosine and phenylalanine amino acids, and from purine and pyrimidine bases. The proton-decoupled I3C NMR spectrum of the perchloric acid extract of HT 29 cells confirms the presence of high levels of lactic acid in these cancer cells (Fig. 3). Signals from amino acids, taurine, phosphocreatine, plus creatine appear also in the 50-5 -10 -15 -20 -25 PPM 0-80 ppm region of the spectrum. Two resonances labeled X at Fig. 4. "!' NMR spectra of perfused human colon adenocarcinoma cells. 45.2 and 65.1 ppm remain to be assigned; they might reflect Spectra were recorded at 80.9 MHz without proton decoupling. They correspond to the signals from 2.5 x 10*cells perfused with the culture medium in a 20-mm the presence of hydroxybutyrate. diameter NMR sample tube at 37'C. Each spectrum represents the average of 31P NMR Spectra of Perfused HT 29 Cells. The 80.9-MHz 1200 free induction decays, (2 blocks: 20 min) accumulated in 4000 time domain 3IP NMR spectrum of HT 29 cells obtained from 2.5 x IO8 addresses.. I, spectrum of well-oxygenated cells in presence of 25 HIMglucose; li. spectrum recorded after l h of perfusion under anoxic conditions (the air t X).. cells placed within a dialysis membrane and perfused as de mixture was replaced by the N2-CO2,95-5% mixture). PME, phosphomonoesters; scribed in "Materials and Methods" is shown in Fig. 4A. The GPC, glycerol-3-phosphorylcholine; PCr, phosphocreatine. 3770

Downloaded from cancerres.aacrjournals.org on October 3, 2021. © 1986 American Association for Cancer Research. "P NMR OF HUMAN COLON ADENOCARCINOMA CELLS resonance at -2.45 ppm is assigned to phosphocreatine based grown in the presence of 25 mM glucose then harvested by upon its characteristic chemical shift relative to 85% H3PO4. It trypsinization using the procedure described in "Materials and can be taken as an internal reference together with the signal Methods." They were subsequently perfused in the NMR sam at 0.5 ppm from glycerophosphorylcholine. The major reso ple tube with a growth medium deprived of glucose. Fig. 5 nances of the in vivo spectrum can be easily assigned to nucleo- displays a sequence of 3IP NMR spectra illustrating the effect side triphosphates (NTP,,, NTPâ€žplusNDPâ€ž,NTPTplus NDP,). of glucose starvation on the intracellular concentration of phos- The nucleoside phosphate resonances include contributions phorylated compounds. Each spectrum corresponds to the sum from both purine and pyrimidine nucleotides which are not of NMR signals acquired over 30-min periods. The NTP con resolved in the spectrum of intact cells. Other resonances arise tent of the cells decreased rapidly (Fig. 5/1); within 1-2 h, the from diphosphodiesters, NAD, PÂ¡,and various phosphomon- NTP level fell to 10% of the reference level. The reference level, oesters. Sharp 3IP resonances were observed in the spectrum of to which an arbitrary value of 100 is assigned, corresponds to perchloric acid extracts after chromatography of the solution the first spectra recorded in the sequence, i.e., l h after the through a Chelex-100 column (Fig. 1). They allow a better onset of glucose starvation. Observation of the time course assignment of 3IP resonances from in vivo spectra, in particular reveals that the cells began to lose high-energy phosphates as in the low-field region. Resonance-labeled phosphomonoesters soon as the incubation without glucose is started (Fig. 6). It is (3.48 ppm) corresponds to an overlap of signals from G-6-P, also noteworthy that the DPDE level which is found to be very phosphorylethanolamine and/or AMP, and phosphorylcholine. high in this cell line when grown in glucose-supplemented Resonances at 2.90 and 2.44 ppm correspond to external and medium, decreases sharply within 2 h of glucose starvation to internal inorganic phosphate (Pi" and Piin) respectively. The reach a new steady state value (Fig. 6). Another pronounced DPDE region is less easy to interpret. The spectrum of the change caused by glucose starvation pertains to the level of extract indicates that these resonances correspond to a mixture internal PÂ¡(seePÂ¡insignalin Chart 5, B and Q. An increase of of uridine diphosphoglucose and uridine diphosphohexose de the PÂ¡inresonance is observed as a function of time. These rivatives such as UDP-A'-acetylglucosamine, UDP-jY-acetylga- changes have been quantified based on the integrations of peak lactosamine, and UDP-glucuronate (32). areas and are shown in Fig. 6, bottom. A very slight increase in In the spectrum shown in Fig. 4.-Ã•.asingle major I', peak internal pH value is additionally observed (Fig. 6, top). appears at 2.90 ppm with a shoulder at 2.44 ppm. In order to In order to document the role of glucose as energetic fuel on determine the contribution of intra- and extracellular PÂ¡tothese the metabolism of phosphorylated metabolites in HT 29, the resonances, N2 was substituted to O2 in the gas mixture. After l h of anoxia, profound changes were apparent in the PÂ¡spectral region (Fig. 4B). The intensity of the PÂ¡inresonance increased dramatically. In addition, this resonance was shifted upfield from 2.44-1.39 ppm thus reflecting the expected acidification of the intracellular medium under anaerobic conditions. In contrast, the insensitivity of the resonance at 2.90 ppm confirms its assignment to external phosphate (PÂ¡ex). The comparison of the intensity of the signal corresponding to the phosphate of NTP in the spectra of normoxic and anoxic cells (Fig. 4) indicates that HT 29 cells are able to maintain 60 Â±20% (SD) of their NTP stores after l h of anaerobic perfusion. The large deviation is due to the low signal to noise obtained in 20 min of accumulation. Intracellular pH Measurements. The intracellular and exter nal pH values can be obtained by using the known dependence of the Pj chemical shift upon pH. The corresponding reference curve was constructed under conditions which best approximate the intracellular environment, since the pKâ€žvalueof PÂ¡isknown to be affected by ionic strength and magnesium ions. PÂ¡was titrated to construct the reference curve in presence of free Mg2+ at a concentration of 150 fiM, i.e., the concentration estimated in HT 29 from the chemical shift difference between the a and ÃŸpeaksof ATP on the 3IP spectra of the intact cells. More than a 2-fold change in magnesium concentration was shown to be necessary to affect significantly the pH determi nation. One can then calculate intracellular pH values of 7.10 and 6.30 for well-oxygenated and anoxic HT 29 cells, respec tively. The extracellular pH (perfusion medium) remains stable at a value of 7.50. The statistical error in chemical shift deter OmM Glc mination corresponds to uncertainties of Â±0.08pH unit at pH 7.5 and Â±0.05 pH unit between pH 6 and 7.1. Taking into 10 10 20 PPM account that intracellular free Mg2+ concentration and ionic Fig. 5. Effect of glucose starvation on the "P NMR spectrum of HT 29 cells. Cells were harvested by trypsinization and perfused with a growth medium strength are relatively stable, we estimate the systematic error deprived of glucose (2x10* cells). Blocks of spectra corresponding to IO min in our pH measurements to be 0. l pH unit. accumulation were sequentially recorded. Spectra (3 blocks) were recorded after 30 (A), 90 (B), and 150 min (C) of glucose starvation. After 6.5 h of starvation, Effect of Glucose Starvation and Refeeding on HT 29 Energy cells were perfused with a medium containing 25 mM glucose and spectra were Metabolism. In this experiment, HT 29 cells were initially recorded at 7.5 (D), 8 (Â£),and 8.5 h (F). Glc, glucose. 3771

As shown in Figs. 1 and 4, high concentrations of phosphorylcholine, glucose 6-phosphate, and various diphosphodiesters (predominantly UDP-hexose derivatives) have been found in these cells, by contrast to what was observed in several other mammalian cells. This result confirms the observation initially made by several authors on the basis of traditional biochemical analyses (29, 32). Paris et al. (29) have linked the high level of sugar phosphates to the activity of glycogen-metabolizing en zymes. HT 29 cells are known to store high levels of glycogen in the stationary phase as compared to the normal homologous cells (33, 34). It is then possible that the control of glycogen metabolism is not operative in these cancer cells where glycogen synthesis would be achieved by a glycogen synthetase b allo- sterically activated by the high concentration of glucose 6- phosphate (29). The presence of very high levels of DPDE (and then uridine diphosphoglucose), is also indicative of an elevated rate of the flux through the glycogen biosynthesis break-down Fig. 6. Time course of changes in phosphorylated metabolite concentrations cycle. The concentration of G-6-P in HT 29 cells as determined and intracellular pH in glucose-deprived HT 29 cells. Intracellular pH (top) has by NMR spectroscopy is around 1.4 nmol/106 cells, corre been determined from the chemical shift of the PÂ¡resonanceby using a calibration curve. Concentration of metabolites (bottom) is expressed in relative intensity sponding to 8 nmol/mg protein. This value is in good agreement with respect to an arbitrary' initial value of 100 for each metabolite. Vertical line, with the value recently published for this cell line (29); it is onset of cell refeeding (25 mM glucose in the perfusate). (A), NTP; (â€¢),Pe;(â€¢), almost 10 times higher than in the corresponding normal DPDE. epithelium. It is also higher in comparison with the level found in liver cells which are well known to metabolize glucose at an cells were refed with glucose-containing (25 HIM)perfusate after elevated rate. In addition to the fast flux through the glycogen 6.5 h of starvation. Changes in the level of metabolites were cycle, the accumulation of G-6-P can be due to a large increase monitored by NMR as illustrated in Fig. 5, spectra D-F. The in glucose utilization through the glycolytic pathway in tumor prominent effect of glucose refeeding is a decrease in intracel cells as compared to their normal counterparts. It can reason lular PÂ¡concentration correlated with a dramatic drop of intra ably be questioned whether or not the glycolysis is to be consid cellular pH (Fig. 6, top). Starting from pH 7.1 in starving cells, ered as the main source of energy in malignant cells. We have the internal pH evolves to a very acidic value of about 6.3 at a rate of 0.1 pH unit/15 min. The sharp decline of pHin is no direct evidence suggesting a preponderance of glycolysis to provide cell energy; however, taking into account that glucose followed by a slower recovery back to a less acidic value (pH is present at high concentrations (10-25 mM) in regular culture 6.45), but still lower than the value found for glucose-deprived media, this carbohydrate has been generally assumed to be the cells. The concurrent fall of internal PÂ¡content starts at the major source of energy for cultured cells. In our study, the onset of glucose infusion; however, a lag time of l h is observed observation that HT 29 cells are able to maintain more than for ATP recovery which reaches 130% of its reference value 60% of their NTP level after l h of perfusion under anaerobic after 4 h of glucose infusion (Fig. 6, bottom). conditions (Fig. 4) supports this hypothesis. Another interesting feature among the changes occurring in Previous reports on 3IP NMR measurements have shown phosphorylated metabolites upon glucose refeeding is the vari that the internal I', resonance constitutes a reliable probe to ation of the intracellular level of DPDE. Surprisingly, as illus study noninvasively transmembrane gradients (35). The data trated in the spectrum of Fig. 5D, the DPDE content starts to reported here indicate that in HT 29 cells perfused with the decrease as the glucose is added to the perfusion medium and standard culture medium at 37Â°C,the pH'" differs from the drops down to a value that represents 30% of the reference pHcx. Typically, the | pHin-pHex | gradient corresponds to 0.2- level. Fig. 6 shows that the process of DPDE accumulation is 0.5 pH units, in well-oxygenated cells, depending on the values triggered when the ATP level reaches 100% of the reference of external pH. In all cases, pH1" is clearly less than pH", as value. A test of viability after 12 h of experiment indicated that confirmed by the chemical shift position of the glucose 6- less than 15% of the cells had been damaged. phosphate resonance on the NMR spectrum. This observation contrasts with several others studies of mammalian cells in suspension (18, 21, 36). It must be pointed out that HT 29 cells DISCUSSION have been collected prior to the NMR experiment by trypsini- The present results illustrate the usefulness of 3IP NMR zation and the physiological consequence of this harvesting spectroscopy in understanding the metabolism of cancer cells. method is actually not well documented. Levels of the main phosphorylated metabolites can be nonin- It has been recently reported that intracellular pH regulation vasively determined and their variations can be observed as a in response to acid loads is dependent on the Na+/H+ exchange and NaMinked C1~/HCCV exchange in acid extrusion (37); function of time and under a variety of metabolic events and growth conditions. In addition, the resonance from PÂ¡displays accordingly, the interpretation of the mechanism of intracellu a chemical shift dependence upon pH of the environment which lar acidification in harvested HT 29 cell could involve a proteo- allows the accurate determination of the value of intracellular lytic inactivation of membrane proteins modifying transmem and external pH while respecting cell integrity. Under appro brane ion transport. priate conditions, narrow lines can be obtained for metabolites Since 3IP NMR is particularly adapted to being monitored in intact HT 29 cells and the integration of the "P NMR signal in real time changes in concentrations of phosphorylated intra areas provides a direct determination of the intracellular con cellular metabolites, we have documented the effect of glucose centrations of the corresponding phosphorylated metabolites. starvation upon energy metabolism in HT 29 cells. This cancer 3772

Navon, G., Ogawa, S., Shulman, R. G., and Yamane, T. High resolution "P cell line has been subjected to extensive studies including mor nuclear magnetic resonance studies of metabolism in aerobic Eschericha coli phological and metabolic investigations in connection with cells. Proc. Nati. Acad. Sci. USA, 74: 888-891, 1977. hormonal effects (38) or utilization of different energetic sub Ugurbil, K., Rottenberg, H., Glynn, P., and Shulman, R. G. 31P nuclear strates or drugs (29, 39-41). The most striking result concern magnetic resonance studies of bioenergetics and glycolysis in anaerobic Escherichia coli cells. Proc. Nati. Acad. Sci. USA, 75: 2244-2248, 1978. ing substrate utilization is the occurrence of a brush border Ugurbil, K., Rottenberg. H., Glynn, P., and Shulman, R. G. Phosphorus-31 nuclear magnÃ©tiqueresonance studies of bioenergetics in wild-type and formation induced by a replacement of glucose by galactose in adenosine triphosphatase (1â€”)Escherichia coli cells. Biochemistry, 21:1068- the culture medium (40). The appearance of cells exhibiting 1075, 1982. morphological changes is correlated with modification of their Ezra, F. S., Lucas, D. S., Mustacich, R. V., and RÃ¼ssel,A.F. Phosphorus- carbohydrate metabolism (25) and it is likely that the enterocyte 31 and carbon-13 nuclear magnetic resonance studies of anaerobic glucose metabolism and lactate transport in Staphylococcus aureus cells. Biochemis differentiation of colon cancer cells in vitro is induced by the try, 22: 3841-3849, 1983. lack of glucose in culture medium rather than the presence of 10. Findly, R. C, Gillies, R. J., and Shulman, R. G. In vivo phosphorus-31 galactose. In our NMR study of glucose-deprived HT 29 cells nuclear magnetic resonance reveals lowered ATP during heat shock of Tetrahymena. Science (Wash. DC), 219: 1223-1225, 1983. illustrated in Fig. 5, a rapid decline of NTP (ATP) levels is 11. Nicolay, K., Kaptein, R., Hellingwerf, K. J., and Konings, N. N. 31Pnuclear observed over a period of 2 h. This observation is consistent magnetic resonance studies of energy transduction in Rhodopseudomonas sphaeroides. Eur. J. Biochem. 116: 191-197, 1980. with the work of Demetrakopoulos et al. (42) showing that in 12. Deslauriers, R., Jarrell, H. C., Byrd, R. A., and Smith, I. 31P NMR studies contrast to normal cells, several malignant cell types suffer a of metabolism in Ancanthamoeba castellani polyphosphate release from dramatic lowering in their ATP levels within the first hour of encysted cells. Biochem. Biophys. Res. Commun., 95: 1211-1217, 1980. 13. Deslauriers, R., Byrd, R. A., Jarrell, H. C., and Smith. 1.31P NMR studies glucose starvation. This finding is at variance with a recent of vegetative and encysted cells of Ancanthamoeba castellani. Eur. J. study carried out on HT 29 cells by Paris et al. (29), showing Biochem., ///: 369-375, 1980. 14. Salhany, J. M., Yamane, T., Shulman, R. G., and Ogawa, S. High resolution that ATP content after 24 h of starvation is still one-half the 31Pnuclear magnetic resonance studies of intact yeast cells. Proc. Nati. Acad. initial value. A possible explanation accounting for this differ Sci. USA, 72:4966-4970, 1975. ence could be that harvested cells in suspension have been used 15. Ogino, T., Den Hollander, J. A., and Shulman. R. G. 3'K, "Na, and 31P NMR studies of ion transpon in Saccharomyces cerevisiae. Proc. Nati. Acad. in our study instead of anchored cells in stationary phase. Sci. USA, 74:4909-4913, 1977. Although stores of ATP are very low within the first hour of 16. Den Hollander, J. A., Ugurbil. K., Brown, T. R., and Shulman, R. G. glucose deprivation and remain at an undetectable level by Phosphorus-31 nuclear magnetic resonance studies of the effect of oxygen upon glycolysis in yeast. Biochemistry. 20: 5871-5880, 1981. NMR spectroscopy during more than 5 h of perfusion, the loss 17. Evans, F. E., and Kaplan, N. O. 31P nuclear magnetic resonance studies of of viability of the cells is no more than 10-15% as indicated by HeLa cells. Proc. Nati. Acad. Sci. USA, 74:4909-4913. 1977. 18. Navon, G., Navon, R., Shulman, R. G., and Yamane, T. Phosphate metab the trypan blue method and by the ability of the cells to olites in lymphoids friend erythroleukemia, and HeLa cells observed by high- metabolize glucose after refeeding; indeed, the refeeding exper resolution 3IP nuclear magnetic resonance. Proc. Nati. Acad. Sci. USA, 75: iment reported here is consistent with a high metabolic activity 891-895, 1978. 19. Navon, G., Ogawa, S., Shulman, R. G., and Yamane, T. 31Pnuclear magnetic of the cells in presence of 25 mM glucose; in particular, the resonance studies of Ehrlich ascites tumor cells. Proc. Nati. Acad. Sci. USA, dramatic drop of pH upon refeeding (Fig. 6) is indicative of the 74:87-91, 1977. 20. Ugurbil, K., Guernsey, D. L., Brown, T. R., Glynn, P., Tobkes, N., and accumulation of acid products from glycolysis. Levels of DPDE Edelman, I. S. 31P NMR studies of intact anchorage-dependent mouse and NTP after 6 h of refeeding amount to 50 and 130%, embryo fibroblast. Proc. Nati. Acad. Sci. USA, 78:4843-4847, 1981. respectively, of the values found in harvested cells grown in 21. Meiner, M. H., Sawyer, S. T., Evanochko, W. T.. Thian, C. N., Glickson, J. glucose-supplemented medium. In addition, a pH difference of D., and Puett, D. Phosphorus-31 nuclear magnetic resonance analysis of epidermal growth factor action in A-431 human epidermoidcarcinoma cells 1 unit has been detected between cytoplasmic and external and SV-40 virus transformed mouse fibroblasts. Biochemistry, 22: 2039- medium in reenergized cells whereas the internal pH remained 2042, 1983. 22. Evanochko, W. T., Sakai, T. T., Thian, C. NG., Krishna, N. R., Kim, H. D., relatively constant around pH 6.45. Zeidler, R. B., Chanta, V. T., Brockman, R. W., Schiffer, L. M., Braun In this study, we have used 3IP NMR spectroscopy to follow schweiger, P. G., and Glickson, J. D. NMR study of in vivo RIF-1 tumors. noninvasively metabolic processes as they take place within HT Analysis of perchloric acid extracts and identification of H-l, P-31 and C-13 resonances. Biochim. Biophys. Acta, Â«05:104-116, 1984. 29 cultured cells. Fluxes of phosphorylated metabolites and 23. Gonzalez-Mendez, R., Wemmer, D., Hahn, G., Wade-Jardetzky, N., and intracellular pH variations have been documented under con Jardetzky, O. Continuous-flow NMR culture system for mammalian cells. ditions of anoxia, glucose starvation, and refeeding. In most Biochim. Biophys. Acta, 720: 274-280, 1982. 24. Foxall, D., and Cohen, J. NMR studies of perfused cells. J. Magn. Reson., cases, the results reported extend further previous works con 52:346-349, 1983. ducted on the same cell line with more traditional techniques 25. Karczmar, G. S.. Koretsky, A. P., Bissell, M. J., Klein, M. P., and Weiner, M. W. A device for maintaining viable cells at high densities for NMR and provide new insights into the involvement of specific me studies. J. Magn. Reson., 53: 123-128, 1983. tabolite pathways in the process of malignant cell growth. The 26. Knop, R. H., Chen, C. W., Mitchell, J. B., Russo, A., McPherson, S., and NMR method will be of particular interest in the study of the Cohen, J. S. Metabolic studies on mammalian cells by 3IP-NMR using a continuous perfusion technique. Biochim. Biophys. Acta, 804: 275-284, relationship between UDP-hexose derivatives and the differ 1984. entiation process. 27. Evanochko, W. T., Thian, C. NG, and Glickson, J. D. Application of in vivo NMR spectroscopy to Cancer. Magn. Reson. Med., /: 508-534, 1984. 28. Fogh, J., and Trempe, C. New human tumor cell lines. In: J. Fogh (ed.), Human Tumor Cells in Vitro, pp. 115-141. New York: Plenum Publishing REFERENCES Corp., 1975. 29. Paris, H., Terrain, B., \ Â¡aliarti.V., Roussel, M., Zweibaum, A., and MurÃ¢t, 1. Ugurbil, K., Shulman, R. G.. and Brown, T. R. High resolution 31P and 13C J. C. Activity of glycogen metabolizing enzymes in glucose deprived HT 29 nuclear magnetic resonance studies of Escherichia coli cells in vivo. In: R. G. adenocarcinoma cell-line. Biochem. Biophys. Res. Commun., 7/0:371-377, Shulman (ed.), Biological Applications of Magnetic Nuclear Resonance, pp. 1983. 537-589. New York: Academic Press, Inc., 1979. 30. Zweibaum, A., Pinto, M., Chevalier, G., Dussaulx, E. Triadou, N., Lacroix, 2. Radda, G. K., and Seeley, P. J. Recent studies on cellular metabolism by B., Haffen, K., Brun, J. L., and Roussel. M. Emerocytic differentiation of a nuclear magnetic resonance. Annu. Rev. Physiol., 41: 749-769, 1979. subpopulation of the human colon tumor cell line HT 29 selected for growth 3. Roberts, J., and Jardetzky, O. Monitoring of cellular metabolism by NMR. in sugar-free medium and its inhibition by glucose. J. Cell. Physiol., 122: Â»iiii.-hiin.Biophys.Acta, 639: 53-76. 1981. 21-29, 1985. 4. Gordon, R. E., Hanley, P. I... and Shaw, D. Topical magnetic resonance. 31. Gard, J. K., and Ackerman, J. A. 3IP external reference for intact biological Prog. NucÃ.Magn. Reson. Spectr., 15: 1-47, 1982. systems. J. Magn. Reson.. 51: 124-127, 1983. 5. Bernard, M.. Canioni, P., and Cozzone, P. J. Etude du mÃ©tabolismecellulaire 32. Wice, B. M., Trugnan, G., Pinto, M., Rousset, M., Chevalier, G., Dussaulx, in vivo par rÃ©sonancemagnÃ©tiquenuclÃ©aireduphosphore-31. Biochimie, 65: E., Lacroix, B., and Zweibaum, A. The intracellular accumulation of IDI' 449-470, 1983. jV-acetylhexosamines is concomitant with the inability of human colon cancer 3773

cells to differentiate. J. Biol. Chem., 260: 139-146, 1985. H* exchange by epidermal growth factor elevates intracellular pH in A-431 33. Rousset, M., Chevalier, G., Roussel, J. P., Robine-Leon, S., Dussaulx, E., cells. J. Biol. Chem., 258: 12644-12653, 1983. and Zweibaum, A. Kinetics of glycogen levels in asynchronous and synchro- 38- Laburthe, M., Rousset, M., Boissard, C, Chevalier, G., Zweibaum, A., and nous cultures of a human colon carcinoma cell line, HT 29. Front. Gastroin- R,Â°!^elm'19-Vasoactive intestinal peptide, a potent stimulator of adenosine- Â»p=Â« A. 11 70 1070 3 'â€¢*cychc-monophosphate accumulation in gut carcinoma cell lines in test. Kes.,4. /J-/y, iv/y. ,..-.., r A-, culture. Proc. Nati. Acad. Sci. USA, 75:2772-2775, 1978. 34. Rousset, M., Robine-Leon, S., Dussaulx, E., Chevalier, G., and Zweibaum, 39 pimo M Appay M D simon.Assman, p., chevalier, G., Dracopoli, N., A. Glycogen storage in foetal and malignant epithelial cells of the human Fogh, J., and Zweibaum, A. Enterocytic differentiation of cultured human colon. Front. Gastrointest. Res., 4:80-85, 1979. colon cancer cells by replacement of glucose by galactose in the medium. 35. Gillies, R. J., Alger, J. R., Den Hollander, J. A., and Shulman, R. G. Biol. Cell, 44:193-196, 1982. Intracellular pH measured by NMR: methods and results. In: R. Nuccitelli 40. Krug, E., Zweibaum, A., Schulz-Holstege, C., and Kepler, D. D-Glucosamine and D. W. Deamer (eds.), Intracellular pH: Its Measurement, Regulation, induced changes in metabolism and growth of colon carcinoma cells in and Utilization in Cellular Functions, pp. 79-103. New York: A. R. Liss, culture. Biochem J., 217: 701-708, 1984. lnc ,o81 41. Rousset, M., Paris, H., Chevalier, G., Terrain, B., Munii. J. C., and Zwei- 36.in /"-li-Gillies, DR. J., i r\Ogmo, â€¢ TT., Shulman,Kh i uR. r-G., ,and ,H u/,,,1Ward, r>D. rC. SID3'P â„¢Â¡nuclear cu|turedbÃ¤um, A. human Growth tumorcells related enzymatic Cancer Res control 44. 154.^ of glycogen f984. metabolism in magnetic resonance evidence for the regulation of intracellular pH by Ehrlich 42. Demetrakopoulos, G. E., Linn, B., and Amos, H. Rapid loss of ATP by ascites tumor cells. J. Cell Biol., 95: 24-28, 1982. tumor cells deprived of glucose: contrast to normal cells. Biochem. Biophys. 37. Rothenberg, P., Glaser, L., Schlesinger, P., and Cassel, D. Activation of Na+/ Res. Commun., 82: 787-794, 1978.

3774

Franck Desmoulin, Jean-Philippe Galons, Paul Canioni, et al.

Cancer Res 1986;46:3768-3774.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/46/8/3768

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/46/8/3768. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.