Cancer Research Cancer Health Disparities Conference 2009
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Cancer Research 67, 7638-7645, August 15, 2007. doi: 10.1158/0008-5472.CAN-06-3459
© 2007 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Provent, P.
Right arrow Articles by García-Martín, M. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Provent, P.
Right arrow Articles by García-Martín, M. L.

Cell, Tumor, and Stem Cell Biology

Serial In vivo Spectroscopic Nuclear Magnetic Resonance Imaging of Lactate and Extracellular pH in Rat Gliomas Shows Redistribution of Protons Away from Sites of Glycolysis

Peggy Provent1,2, Marina Benito3, Bassem Hiba1,2, Régine Farion1,2, Pilar López-Larrubia3, Paloma Ballesteros4, Chantal Rémy1,2, Christoph Segebarth1,2, Sebastián Cerdán3, Jonathan A. Coles1,2 and María Luisa García-Martín3

1 Institut National de la Santé et de la Recherche Médicale, U836; 2 Université Joseph Fourier, Grenoble Institut des Neurosciences, Grenoble, France and 3 Instituto de Investigaciones Biomédicas "Alberto Sols," Consejo Superior de Investigaciones Cientificas/Universidad Autónoma de Madrid; and 4 Laboratory of Organic Synthesis and Molecular MRI, Faculty of Sciences, UNED, Madrid, Spain

Requests for reprints: María Luisa García-Martín, Resonancia Magnetica "Ntra. Sra. del Rosario," Principe de Vergara, 53, 28002 Madrid, Spain. E-mail: mlgarcia{at}iib.uam.es.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results and Discussion
 References
 
The acidity of the tumor microenvironment aids tumor growth, and mechanisms causing it are targets for potential therapies. We have imaged extracellular pH (pHe) in C6 cell gliomas in rat brain using 1H magnetic resonance spectroscopy in vivo. We used a new probe molecule, ISUCA [(±)2-(imidazol-1-yl)succinic acid], and fast imaging techniques, with spiral acquisition in k-space. We obtained a map of metabolites [136 ms echo time (TE)] and then infused ISUCA in a femoral vein (25 mmol/kg body weight over 110 min) and obtained two consecutive images of pHe within the tumor (40 ms TE, each acquisition taking 25 min). pHe (where ISUCA was present) ranged from 6.5 to 7.5 in voxels of 0.75 µL and did not change detectably when [ISUCA] increased. Infusion of glucose (0.2 mmol/kg·min) decreased tumor pHe by, on average, 0.150 (SE, 0.007; P < 0.0001, 524 voxels in four rats) and increased the mean area of measurable lactate peaks by 54.4 ± 3.4% (P < 0.0001, 287 voxels). However, voxel-by-voxel analysis showed that, both before and during glucose infusion, the distributions of lactate and extracellular acidity were very different. In tumor voxels where both could be measured, the glucose-induced increase in lactate showed no spatial correlation with the decrease in pHe. We suggest that, although glycolysis is the main source of protons, distributed sites of proton influx and efflux cause pHe to be acidic at sites remote from lactate production. [Cancer Res 2007;67(16):7638–45]


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results and Discussion
 References
 
Compared with normal tissue, extracellular pH (pHe) is normally lower in tumors, and intracellular pH (pHi) is often higher (1, 2). These differences seem to favor tumor growth and metastasis (2–6) and can also be exploited to target therapies (7–11). Inhibition of pH-regulating mechanisms in tumor cells can reduce proliferation, invasiveness, and metastasis (2, 12–14).

Protons that are extruded from tumor cells can diffuse to blood vessels and be carried away from the tumor, so pHe can remain lower than blood pH only if there is a continuous source of intracellular protons. In tumors, glycolysis is up-regulated (4, 15–17). Glycolysis produces protons with the stoichiometry 1/2(glucose) = lactate– + H+. The proton and the lactate ion are normally exported from cells together, either on a monocarboxylic acid cotransporter (MCT; refs. 13, 18) present on C6 cells (19) or by diffusion of Hlactate through the lipid bilayer. Hence, glycolysis associated with lactate export might, in the absence of other processes, explain both a low pHe and a pHi that is not abnormally low. However, there is evidence that the Na+/H+ exchanger, NHE1 (which is highly expressed in gliomas), and other membrane H+ transporters play important roles in maintaining pHe low (20–23). This low pHe favors remodeling of the extracellular matrix in adjacent normal tissue and facilitates tumor progression (2, 3). Because the protons produced by glycolysis exit the cell in association with lactate on MCTs, where do the protons exported by these additional pumps come from?

As a step toward answering this question, we have used in vivo magnetic resonance spectroscopic imaging to map the distributions of lactate and pHe in C6 gliomas in rat brain. pH-sensitive magnetic resonance imaging has the advantages that it is noninvasive and that it can be used to image relatively inaccessible tumors, such as brain tumors (16). We have used an exogenous pHe probe molecule with a 1H resonance whose frequency shifts with pH (24, 25). The probe is injected in the blood and diffuses into the interstitium of C6 gliomas because their blood-brain barrier is leaky (26). This technique gives reliable absolute values of pHe and has the further advantage that the same magnetic resonance setup can be used to acquire images of the distributions of several metabolites in the same experiment.

In previous work, we obtained one in vivo pHe image and one set of metabolite images for each C6 glioma (25). We have now used improved methods to take this investigation considerably further. Instead of the probe molecule used previously, IEPA [(±)2-(imidazol-1-yl)3-ethoxycarbonylpropionic acid], we have synthesized and used a new one, ISUCA [(±)2-(imidazol-1-yl)succinic acid; refs. 24, 27]. In addition, we have implemented faster spectroscopic imaging techniques based on spiral acquisition in k-space. We first showed that the 1H spectrum and pharmacokinetics of ISUCA are superior to those of its predecessor, IEPA. We then imaged the changes in pHe and metabolites caused by hyperglycemia, which increases lactate concentration in gliomas (28, 29) and has been reported to decrease pHe in other tumor models (30–33). The results suggest that although glycolysis is indeed the main source of overall extracellular acidity in C6 gliomas, the net proton efflux is redistributed away from the sites of glycolysis.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results and Discussion
 References
 
Synthesis of ISUCA. ISUCA (Fig. 1A ) was prepared by Michael thermal addition of imidazole to diethyl fumarate (100°C for 24 h) to yield the diethyl-2-imidazol-1-ylsuccinate. The ester was purified by column chromatography and hydrolyzed with two equivalents of NaOH (1.2%; 100°C for 24 h; refs. 24, 27). Purity was confirmed by IR-attenuated total reflection spectroscopy and 1H and 13C magnetic resonance spectroscopy.


Figure 2
View larger version (12K):
[in this window]
[in a new window]

  		Figure 2. Buildup and decline of ISUCA during and after infusion. A, concentrations of ISUCA (full symbols) and IEPA (open symbols) in blood plasma. Bars, SE (n = 3 rats). B, concentrations in urine. For ISUCA (full symbols), n = 3 rats. For IEPA (open symbols), n = 3 (or n = 1 where no error bar is shown.) C, evolution of the area of the ISUCA H2 peak and the sum of the areas of the H4 and H5 peaks in a 4 x 4 x 4 mm3 monovoxel surrounding a C6 glioma during infusion of ISUCA with the protocol used for the spectroscopic imaging: 0.33 mmol/min·kg body weight for 20 min and then 0.20 mmol/min·kg body weight for 90 min. The experiments were done on eight rats, but in some experiments, spectra were not obtained at all time points.

 

Figure 1
View larger version (8K):
[in this window]
[in a new window]

  		Figure 1. A, structure of ISUCA. B, spectrum of ISUCA (100 mmol/L) in PBS at 37°C, 11.7 T, TE = 40 ms. Inset, enlarged section of the spectrum showing the change in chemical shift of the H2 peak as the pH changes. C, calibration curve of the H2 peak in plasma at 37°C. The chemical shifts of the H2 peaks were fitted with the Henderson-Hasselbalch equation: pH = pKa + log{({delta}1 – {delta}) / ({delta} – {delta}2)}, where {delta} is the chemical shift and {delta}1 > {delta}2 are the asymptotic values.

 
Calibration of the pH dependence of the chemical shift of the ISUCA H2 resonance. Solutions of ISUCA (20 mmol/L) in rat plasma were prepared at different pHs (measured at 37°C) by addition of HCl. 3-(Trimethylsilyl)-propionic acid d4, sodium salt (2 mmol/L; Aldrich) was used as a chemical shift reference. 1H spectra were obtained on a 9.4 T magnet (AM400 NB, Bruker) at 37°C with presaturation of the water signal followed by a 90° hard pulse for observation.

Measurement of T1 and T2 in vitro. T1 and T2 for the H2, H4, and H5 protons were measured at 37°C in a solution of ISUCA in fetal bovine serum (FBS; pH 7.2) using inversion recovery and Carr-Purcell-Meiboorn-Gill sequences. The values of T1 were as follows: H2, 1,056 ± 10 ms (SE); H4, 2,831 ± 259 ms; and H5, 1,268 ± 101 ms, n = 3. The values of T2 were as follows: H2, 80.0 ± 1.5 ms; H4, 86.58 ± 0.53 ms; and H5, 59.5 ± 1.4 ms, n = 5.

In vitro cytotoxicity of ISUCA. The amount of lactate dehydrogenase (LDH) released to the medium from C6 cells was measured. Cells were grown to confluence in DMEM containing 5% FBS and incubated for 1, 3, 6, and 24 h with concentrations of ISUCA from 5 to 100 mmol/L, in Krebs solution containing (in mmol/L) NaCl (119), NaHCO3 (24), KCl (4.7), MgSO4 (1.2), KH2PO4 (1.2), CaCl2 (1.3), and glucose (1). At the end of the experiment, cells were broken by freeze thawing thrice, and the total releasable LDH was determined spectrophotometrically in a microplate reader (SpectraMax, Molecular Devices) at 340 nm using a buffer mixture containing 50 mmol/L HEPES, 5 mmol/L sodium pyruvate, and 0.35 mmol/L NADH (pH 7.2).

Animal care. All procedures involving animals conformed to European Council Directive 86/609/EEC (French licenses 380321, A3851610004, and B385161000).

Pharmacokinetics. Male Wistar rats (200–250 g) were anesthetized with xylazine (8 mg/kg) and ketamine (80 mg/kg) i.p. A 2 mol/L solution of ISUCA was infused in the right jugular vein at 2 mL/h for 20 min and then 1.2 mL/h for 40 min. Samples of blood (0.2 mL every 10 min) were taken from the left jugular vein. Urine samples were collected every 5 min. ISUCA concentration was measured by high-performance liquid chromatography using 80% methanol as eluant under isochratic conditions.

Tumor implantation. C6 cells (34) from the American Type Culture Collection were grown in DMEM containing 25 mmol/L glucose and 2 mmol/L glutamine (product 31966-021 from Invitrogen) to which was added 10% FBS (Invitrogen) and antibiotics. Male Wistar rats (200–230 g) were anesthetized with chloral hydrate (400 mg/kg), and 105 C6 cells in DMEM were injected stereotaxically in the right caudate nucleus (35). Nuclear magnetic resonance (NMR) experiments were done ~25 days after tumor implantation, by which time the rats weighed 275 to 380 g.

Animal preparation for NMR experiments. Anesthesia was induced with air containing 5% isoflurane (Baxter) and maintained with 0.6% to 1.2% isoflurane in air enriched with oxygen to 36%. A catheter was placed in the femoral vein for infusion of ISUCA and glucose and another in the femoral artery to measure arterial pressure and to take 0.1 mL samples for measurement of PaO2, PaCO2, pH, and hematocrit (ABL510, Radiometer). The rat was tracheostomized and ventilated. Rectal temperature was maintained at 37°C with a water heating coil under the abdomen.

In vivo NMR measurements. The console (Surrey Medical Imaging Systems) was interfaced to a 20-cm horizontal bore, 7 T, magnet (Magnex Scientific), with actively shielded gradients (200 mT·m–1, diameter of 12 cm) driven by series 7700 gradient amplifiers (Techron). A 25-mm-diameter homemade surface coil was used for radiofrequency transmission and reception (36).

T2-weighted, spin-echo images. To select tumors, scout T2-weighted images were obtained with repetition time (TR) = 3000 ms, echo time (TE) = 80 ms, slice thickness = 1 mm, seven to nine slices in both coronal and transverse orientations.

1H spectroscopic imaging of pHe and metabolites. A PRESS (37) voxel (15 x 15 x 3 mm3) in the transverse plane was selected with five-lobe sinc RF pulses (bandwidth, 6 kHz) and centered on the tumor. The FLATNESS automatic shimming method (38) was used to give full-width half-maximum line widths of typically 25 Hz for the water signal. For optimum signal-to-noise ratio of the ISUCA H2 peak, we used a short TE of 40 ms. This TE is unsuitable for imaging lactate because mobile lipids, present particularly in necrotic parts of C6 gliomas (39), have a prominent methylene group peak at 1.22 ppm, close to the complex lactate resonance near 1.33 ppm. The T2 for these lipids is shorter than that of lactate, so their relative contribution is reduced at longer TEs. We used TE = 136 ms, which, in our conditions in vivo, gives a lactate doublet fused into a single negative peak; this facilitates detection of any remaining lipid contamination (Fig. 5 in ref. 25). For acquisitions with either TE, the spectroscopic imaging sequence was preceded by a VAPOR water suppression module (40), composed of a set of three-lobe sinc RF pulses (bandwidth, 500 Hz). In-plane spatial encoding following PRESS volume selection was achieved with out-and-in spiral encoding (36). With TE = 136 ms, 2 spatial and 32 temporal interleaves were applied; the spectral bandwidth was 3,125 Hz. Each spatial interleave sampled k-space along eight out-and-in spirals, each of eight turns during the outward k-space trajectory and eight during the inward. The acquisition time of the metabolite images [number of averages (NA) = 4, TR = 3 s] was 12.8 min. In acquisitions with TE = 40 ms, 4 spatial and 32 temporal interleaves were applied; the bandwidth was 6,250 Hz. Here, each spatial interleave sampled k-space along eight out-and-in spirals, each of four turns during the outward trajectory and four during the inward. The acquisition time of the ISUCA images (NA = 4, TR = 3 s) was 25.6 min. Before each spectroscopic acquisition, a water reference image was acquired with NA = 1, TR = 3 s to give acquisition times of 3.2 min (with TE = 136 ms) and 6.4 min (with TE = 40 ms). To account for deformation by imperfections in the gradient channels and by eddy currents, the k-space trajectories were measured using the Fourier transform technique (36, 41) and the real trajectories were used for image construction.

Spectroscopic imaging data processing. Spectroscopic images were reconstructed following the method of Hiba et al. (36). Voxel spectra from the acquisitions with 40 and 136 ms TEs were phase corrected to align the water reference signals. Peak areas in the 136 ms TE spectra were obtained by fitting a function composed of Lorentzian line shapes. For the 40 ms TE spectroscopic images, an acquisition made before infusion of ISUCA allowed us to check the spectrometer settings and to verify that there were no endogenous resonances overlapping the ISUCA peaks. In the presence of ISUCA, Lorentzian line shapes were fitted to the H2, H4, and H5 peaks to obtain the peak areas and chemical shifts. pHe was determined in each voxel from the chemical shift of the H2 peak using the titration curve obtained in vitro (Fig. 1C). The fitting procedure did not accept a signal-to-noise ratio of <2, so for voxels not meeting this criterion for H2, no pHe was calculated. In 136 ms TE images, the fitting procedure was instructed to fit a Lorentzian to the positive peak corresponding to mobile lipids at 1.2 ppm. If it did so, the value of the (negative) lactate peak for that spectrum was excluded from analysis. The final voxel size was 0.5 x 0.5 x 3 mm3.

T1-weighted imaging. After the spectroscopic imaging acquisitions, regions of extravasation within the tumor were assessed by infusing a bolus of Gd-DOTA (Dotarem, Guerbet), 0.2 mmol/kg body weight, and acquiring, in the same slices as the previous T2-weighted images, a series of T1-weighted images at 40-s intervals with TR = 500 ms, TE = 30 ms.

Infusion protocol. A 1 mol/L solution of ISUCA was prepared in deionized water and the pH was adjusted to 7.2 with HCl. ISUCA was first infused at 0.33 mmol/min·kg body weight during 20 min and then at 0.20 mmol/min·kg body weight during 90 min (Fig. 3A). The infusion was then switched from ISUCA to 1 mol/L glucose at 0.20 mmol/min·kg body weight (Fig. 3A).


Figure 3
View larger version (25K):
[in this window]
[in a new window]

  		Figure 3. Imaging pHe and metabolites. A, experimental protocol showing the relation between the periods of data acquisition and the infusion of ISUCA or glucose. The rates of infusion were adjusted for the weight of the rat; the values shown are for a typical body weight of 300 g. WSI, water spectroscopic imaging. B, T2 spin-echo image. The tumor ROI (white square) has 8-mm sides. C, histologic image, H&E stain. D, spectra with TE = 40 ms from sample voxels in the tumor and in the contralateral brain as indicated. The ISUCA peaks H2, H4, and H5 are present only in the tumor.

 
Comparison of different images. The peak areas for resonances of interest in each acquisition were normalized with respect to the water peak in the preceding water spectroscopic image (Fig. 3A).

The measurement of pHe depends not on peak area but on the chemical shift of the ISUCA H2 peak. In voxel spectra for TE = 40 ms, the SD of the chemical shifts of the total creatine (tCr) peak was 0.075 ppm in acquisition pH(glc) (five rats, 2,602 voxels). In the same acquisition, the SD of the chemical shifts of the pH-sensitive H2 peak was 0.119 ppm (994 voxels). With TE = 40 ms, the H2 peak was generally better defined than that of tCr, so certainly a major part, and probably most, of the variation of the measured chemical shift of H2 would have been due to local variations in pHe. To check that images from acquisitions with the two TEs were in register, we compared tCr acquired with TE = 40 ms with tCr acquired with TE = 136 ms. For the acquisitions during glucose infusion, the correlation coefficient was 0.769 (1,446 voxels; P < 0.0001). Considering that the tCr peak in the TE = 40 ms spectra was noisy, this correlation is satisfactory. Because the acquisition was spiral, images were not displaced as a function of chemical shift (although the fields of view were, and there was some blurring); hence, the H2 image (~8 ppm) was in register with that of lactate (1.33 ppm).

Single voxel spectroscopy. To follow the infiltration of the tumors by ISUCA, parallel experiments were done in which a single 4 x 4 x 4 mm3 voxel was placed around the tumor. Spectra were acquired with TE = 40 ms using a PRESS sequence. Shimming, water suppression by VAPOR, and RF pulses were the same as for the spectroscopic imaging. TR was 3,300 ms, the signal was accumulated 80 times, and spectra were acquired every 5 min.

Histology. At the end of each spectroscopic imaging experiment, the rat brain was rapidly removed and frozen in isopentane at –80°C. Cryosections, 10 µm thick, were cut and stained with H&E (39).


   Results and Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results and Discussion
References
 
ISUCA as a probe for pHe. The pH-sensitive H2 resonance of ISUCA, in the range 7.7 to 8.8 ppm, is well separated from endogenous peaks in either normal tissue (Fig. 3D) or tumoral tissue (data not shown). Other ISUCA peaks, from the CH2 group, are present, near 2.9 and 3.1 ppm (Fig. 1B), and, in in vivo spectra, could contaminate the peaks for tCr (3.0 ppm) and total choline (tCho; 3.2 ppm). However, there is no ISUCA resonance near that of lactate (1.33 ppm), the metabolite of most interest in the present work. The pKa of ISUCA, measured in blood plasma at 37°C, was 6.92 ± 0.05 so that ISUCA has close to maximum sensitivity over the pH range of 6.4 to 7.4 (Fig. 1C). In experiments on C6 cells in vitro, the pH measured with ISUCA corresponded to pHe, not pHi, indicating that ISUCA does not readily cross the membranes of these cells (data not shown).

During venous infusion of ISUCA, its concentration in the blood plasma reached an approximately steady level of ~40 mmol/L in 10 min (Fig. 2A ). In cytotoxicity tests, C6 cells in culture did not release detectable amounts of LDH after 6 h in concentrations of ISUCA up to at least 50 mmol/L (data not shown). This suggests that, in the in vivo experiments, the ISUCA did not greatly affect the tumor metabolism. ISUCA was excreted in the urine but at lower concentrations than IEPA (Fig. 2B). This might explain why, in the plasma, [ISUCA] was higher than [IEPA] for the same infusion protocol (Fig. 2A).

With an approximately constant concentration of ISUCA in the blood, the areas of the H2 and H4 + H5 peaks from a monovoxel, including the tumor, continued to increase over at least 2.5 h (Fig. 2C). Some of this increase may have been due to diffusion of ISUCA into less accessible parts of the tumor.

Protocol for in vivo spectroscopic imaging. T2-weighted spin-echo images showed that the approximate diameters of the five tumors studied ranged from 5.5 to 8.5 mm (mean, 6.7 mm). The first spectroscopic imaging acquisition with TE = 136 ms ["Metab(pre)" in Fig. 3A ] gave the baseline voxel spectra for the metabolites. After 59 min of ISUCA infusion, the first acquisition with TE = 40 ms in the presence of ISUCA gave an initial pHe image, "pH1," which was followed by a second, "pH2." We then infused glucose (see below) and obtained images "pH(glc)" and "Metab(glc)." For example, "tCho(pre)" refers to the image of tCho obtained during acquisition "Metab(pre)." The complete protocol of Fig. 3A was carried out for four rats. For an additional fifth rat, the experiment was ended before the second metabolite acquisition.

In vivo measurement of pHe. ISUCA was detected mainly within the tumors (Fig. 3D) presumably because blood vessels in C6 tumors are leaky (26). A region of interest (ROI) embracing the tumor was selected (Figs. 3B and 5A), and the area of the H2 peak in each voxel was measured to give an image of the distribution of ISUCA (Fig. 4A ). Voxels where the S/N of the H2 peak was <2 were left blank (black). From the chemical shift of H2, the pH was calculated using the calibration curve (Fig. 1C) and an image of the pHe distribution was constructed (Figs. 4B and 5D ). pHe was heterogeneous within the tumor, ranging from about 6.5 to 7.5, and was generally lower than the 7.4 of the blood: the mean value for acquisition pH1 for the five tumors taken together was 6.97 (SD, 0.127; n = 649 voxels). This falls between values previously reported in C6 gliomas of 7.08 (25) and 6.87 (42).


Figure 5
View larger version (38K):
[in this window]
[in a new window]

  		Figure 5. Glucose-induced changes in pHe and lactate. Anatomic (A) and Gd-DOTA (B) images of rat brain. The tumor ROI (white square) is 10 mm2. Lactate and pHe before (C and D) and during (E and F) glucose infusion and images of the changes (G and H). Note that acidification (red in H) does not match increases in lactate (green in G). Yellow background, no measurement was possible.

 

Figure 4
View larger version (52K):
[in this window]
[in a new window]

  		Figure 4. A, map of the peak areas of the ISUCA H2 resonances, acquisition pH1. B, map of pHe calculated from the chemical shifts of the H2 peaks. Yellow background, no measurement. C to F, maps of metabolite peak areas in the tumor ROI from acquisition Metab(pre) with TE = 136 ms. Pixels are 0.5 x 0.5 mm2.

 
To see if the presence of ISUCA in the interstitial fluid changed pHe, we compared pH1 and pH2. Values for the peak area and chemical shift of the pH-sensitive H2 resonance of ISUCA were obtained in both acquisitions for 448 voxels in the tumor ROIs in five rats. The mean change in area of the H2 peaks in pH2 compared with pH1 was an increase of 29.6% (Table 1A ) presumably due to continued diffusion of ISUCA into the tumor. Despite this increase in [ISUCA], we were unable to detect a significant change in mean pHe in these voxels (Table 1A). To estimate the upper limit for any possible effect of ISUCA on pHe, we calculated, by extrapolation of the pHe values to zero [ISUCA], an estimate of true pHe and obtained a value for the possible effect of ISUCA during acquisition pH1 of 0.025 pH units (=0.0075 x 100 / 29.6; see Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1. Changes in values

 
Distributions of metabolites: qualitative observations. The qualitative features of our metabolite images confirmed earlier imaging studies (25, 39). Briefly, N-acetyl aspartate (NAA) was essentially in normal tissue outside the tumor (Fig. 4F). Choline (tCho) was higher near the periphery of the tumor (Fig. 4C). Creatine (tCr) was low near the center of the tumor (Fig. 4E). Lactate was detected almost exclusively within the tumors (Fig. 4D).

Baseline distributions of pHe and lactate. Because glycolysis is the most obvious source of tumor protons, it is noteworthy that, in voxels where both pHe and lactate peak area could be quantified, the distributions of acidity and lactate seemed to be different (Fig. 4B and D). To make an objective test, we calculated the Spearman correlation coefficient for pHe (from acquisition pH1) versus lactate peak area [from acquisition Metab(pre)] for all the voxels of the five tumors for which these two values were available. The correlation coefficient r was small, +0.125, showing that, for any one voxel, pHe could not be predicted from lactate peak area, but strikingly, r was positive, and the relation between the two distributions was significantly different from random (P < 0.03); that is, pHe tended to be less acid where lactate was high. The pHe images also suggest that pHe was often more acid near the peripheries of the tumors (Figs. 4B and 5D) as also seen in tumors in skin-fold chambers (Fig. 3B of ref. 4).

Changes in pHe and metabolite peaks induced by glucose infusion. Hyperglycemia increases lactate concentration in brain gliomas (28, 29, 43) and is known to decrease pHe in tumors outside the brain (30–33). We therefore looked at the effects of glucose infusion both to check that ISUCA detected changes in pHe and to shed more light on the relation between pHe and lactate.

Hyperglycemia increased lactate peak areas in our C6 gliomas, the average increase in voxels where both Lac(pre) and Lac(glc) were available being 54% (Table 1D). There was also a general decrease in pHe (Table 1D). Strikingly, however, the decreases in pHe were not necessarily greater where lactate increased most (Fig. 5G and H). We calculated the change in lactate peak area and the change in pHe in all the voxels where the four necessary quantities were available: Lac(pre), Lac(glc), pHe2, and pHe(glc). Because [lactate] is linearly related to lactate peak area, we expressed H+ ion activity on a linear scale rather than the logarithmic pH scale: pHe = –log({gamma}H[H+]e), where {gamma}H is an (unknown) activity coefficient and [H+]e is in mol/L (44). On a plot of change in {gamma}H[H+]e versus change in lactate peak, the points were widely scattered, with many voxels where [lactate] increased with little change in {gamma}H[H+]e and others where {gamma}H[H+]e increased but [lactate] did not (Fig. 6Ac , see also Fig. 5G and H).


Figure 6
View larger version (16K):
[in this window]
[in a new window]

  		Figure 6. Spatial redistribution of protons. Aa, comparison of peak areas for tCr (normalized to the water peaks) in voxel spectra from acquisitions Metab(pre) and Metab(glc). The dots correspond to all the voxels in the tumor ROIs of four rats for which the tCr peak could be quantified (in some voxels there may have been slight contamination from an ISUCA CH2 peak). The scales are in arbitrary units. Solid line, no change in peak area. This figure illustrates the case of a peak area that changed little during glucose infusion. The Pearson correlation coefficient for the two acquisitions was 0.8547 (1,387 voxels; P < 0.0001). Ab, [H+]e and lactate had different spatial distributions. The [H+]e scale is absolute (in nmol/L); the lactate scale is the same as for (Aa). Ac, the glucose-induced changes in [H+]e and in lactate also had different spatial distributions. Ad, the glucose-induced increase in [lactate] tended to be greater where [lactate] was initially high. Solid line, no change. B, a scheme that combines the present results with the previously known importance of glycolysis, membrane proton pumps, and the diffusion of H+ into normal tissue surrounding the tumor.

 
Lactic acid, being a weak acid, tends to distribute across cell membranes according to the pH gradient. For tumor cells, pHe < pHi so [lactate]i > [lactate]e (7). The facts that intracellular space exceeds extracellular space, and that lactate is produced within cells and exported, will increase the proportion that is intracellular. In contrast, the pH measured by ISUCA is extracellular. These different locations would give a spatial mismatch of measured {gamma}H[H+] and of [lactate] over distances on the order of microns, but this is much less than the observed mismatch of hundreds of microns.

Glucose infusion also induced small increases in the apparent peak areas for tCho, tCr, and NAA in the tumor ROIs, and the NAA peak increased outside the tumor ROI (Table 1B and D). However, the peaks tCho(glc) and tCr(glc) within the tumor may have been contaminated by the CH2 peaks of ISUCA (Fig. 1B) and that of tCho(glc) additionally by glucose.

The cause of the glucose-induced changes in [lactate] and pHe. Lactate and H+ might accumulate in a tumor if the blood flow were reduced. However, venous infusion of glucose (in contrast to i.p. injection) seems to have little effect on tumor blood flow (32, 45, 46), and the small effects reported were for quantities of glucose greater than we used. C6 cells in vitro have a high Km of 7.7 mmol/L for glucose uptake and metabolism (47), so their glycolytic flux increases when extracellular [glucose] increases above normal in vivo values. We suggest that the glucose-induced increases in lactate we observed were mainly due to increased lactate production rather than reduced lactate clearance. These increases tended to be greater in voxels where lactate was already high (Fig. 6Ad), an observation which corroborates previous conclusions that lactate is associated with well-perfused, metabolically active, regions of tumors (29).

Terpstra et al. (29) have shown that, on a time scale of tens of minutes, the quantity of extra lactate that constitutes the glucose-induced increase in [lactate] in C6 gliomas is small compared with the quantity that is cleared to the blood. Changes in the total lactate concentration are the main determinant of changes in the flux of lactate out of the cells (pH changes also playing a role) and through the extracellular clefts to the blood. For small changes, the change in flux will be approximately proportional to the change in the concentration. We calculated the fractional increase in the total amount of lactate as [{Sigma}Lac(glc) – {Sigma}Lac(pre)] / {Sigma}Lac(pre), where the sums are over the 117 voxels for which Lac(pre) and Lac(glc) were available. This gave a fractional increase of 73%.

Glycolysis produces equal numbers of H+ ions and lactate– ions, and an increase in glycolytic flux is expected to produce equal increases in the rates of production of the two. We therefore asked whether, despite the local mismatch, the increase in overall lactate production in the whole tumor was approximately equal to the increase in H+ production. Because H+ ions in the extracellular clefts can diffuse readily to the blood (48), a change in pHe mainly reflects a change in flux. On arrival in the intercellular clefts, nearly all the H+ ions that are exported from the cells will combine with mobile buffer molecules (49), the diffusion of these buffer molecules to the blood greatly facilitating the clearance of H+ (48). Changes in this flux are determined mainly by changes in the gradient of [H+] from just outside the cells to the blood where, for pH {approx} 7.4, {gamma}H[H+]b {approx} 39.8 nmol/L. We calculated

Formula

for the 524 voxels for which both pH2 and pH(glc) were available. The mean increase in "[H+] gradient" was 85.2%.

These numbers (73% for lactate and 85% for H+) are only very approximate estimates of the changes in the fluxes, particularly because, in some rats, lactate was measured in only a few voxels. They do seem to be compatible with the glucose-induced increases in the production of lactate and H+ being equal.

A hypothesis to explain the different distributions of lactate and [H+]e. Decades of research support the idea that glycolysis is the main source of excess protons in tumors (4, 17) and the present results seem not to contradict this. It is also known that lactate normally leaves cells in association with protons mainly on MCT cotransporters (3, 18). How can we reconcile these ideas with our observation that lactate and [H+]e, considered voxel by voxel, have different distributions and also with the well-documented evidence for active extrusion of protons by other proton pumps, notably the Na+/H+ exchanger NHE1 (2) pump, which require an intracellular source of protons? Our suggestion is that many of the protons that leave tumor cells with lactate reenter the same cell or cells in the vicinity. This reentry could occur simply through a proton "leak," for which the electrochemical driving force is greater than in normal cells or, perhaps, indirectly by HCO3–/Cl– exchange (22). The cells of C6 gliomas are weakly coupled by gap junctions (50) and we hypothesize that these junctions might allow some flow of intracellular protons from one cell to another. In this scheme, pHe would not necessarily change where lactate was extruded, and intracellular protons would be available for the NHE1 proton pumps (Fig. 6B); protons could be extruded into the extracellular space where the tumor is invading normal tissue. Indeed, a minimum in pHe has been observed near the tumor periphery (our Figs. 4B, 5D and F, and 3B of ref. 4).

Conclusion. ISUCA is a satisfactory probe molecule for magnetic resonance spectroscopic imaging of pHe in gliomas and probably in other tumors. The results suggest that, within tumors, protons produced by glycolysis are redistributed toward sites of net proton efflux.


   Acknowledgments
 
Grant support: Institut National de la Santé et de la Recherche Médicale-Consejo Superior de Investigaciones Cientificas Collaborative Grant; Association pour la Recherche sur le Cancer; Cancéropôle Région Rhône-Alpes; Région Rhône-Alpes; Institute of Health Carlos III, Spain, PI051845 and PI051530 (M.L. García-Martín and S. Cerdán); Ministry of Education and Science, Spain SAF 2004-03197 and NAN2004-09125-C07-03 (S. Cerdán); and Institut National de Cancer (C. Rémy).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Laura Barrios and Anne Ziegler for technical advice.

Received 9/18/06. Revised 5/25/07. Accepted 6/12/07.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results and Discussion
 References
 

  1. Gillies RJ, Raghunand N, Karczmar GS, Bhujwalla ZM. MRI of the tumor microenvironment. J Magn Reson Imaging 2002;16:430–50.[CrossRef][Medline]
  2. Cardone RA, Casavola V, Reshkin SJ. The role of disturbed pH dynamics and the Na+/H+ exchanger in metastasis. Nat Rev Cancer 2005;5:786–95.[CrossRef][Medline]
  3. Gatenby RA, Gillies RJ. Why do cancers have high aerobic glycolysis? Nat Rev Cancer 2004;4:891–9.[CrossRef][Medline]
  4. Gatenby RA, Gawlinski ET, Gmitro AF, Kaylor B, Gillies RJ. Acid-mediated tumor invasion: a multidisciplinary study. Cancer Res 2006;66:5216–23.[Abstract/Free Full Text]
  5. Rofstad EK, Mathiesen B, Kindem K, Galappathi K. Acidic extracellular pH promotes experimental metastasis of human melanoma cells in athymic nude mice. Cancer Res 2006;66:6699–707.[Abstract/Free Full Text]
  6. Stock C, Gassner B, Hauck CR, et al. Migration of human melanoma cells depends on extracellular pH and Na+/H+ exchange. J Physiol 2005;567:225–38.[Abstract/Free Full Text]
  7. Gerweck LE, Seetharaman K. Cellular pH gradient in tumor vs normal tissue: potential exploitation for the treatment of cancer. Cancer Res 1996;56:1194–8.[Abstract/Free Full Text]
  8. Raghunand N, He X, van Sluis R, et al. Enhancement of chemotherapy by manipulation of tumour pH. Br J Cancer 1999;80:1005–11.[CrossRef][Medline]
  9. Gerweck LE, Vijayappa S, Kozin SV. Tumor pH controls the in vivo efficacy of weak acid and base chemotherapeutics. Mol Cancer Ther 2006;5:1275–9.[Abstract/Free Full Text]
  10. Ojugo ASE, McSheehy PMJ, Stubbs M, et al. Influence of pH on the uptake of 5-fluorouracil into isolated tumour cells. Br J Cancer 1998;77:873–9.[Medline]
  11. Andreev OA, Dupuy AD, Segala M, et al. Mechanism and uses of a membrane peptide that targets tumors and other acidic tissues in vivo. Proc Natl Acad Sci U S A 2007;104:7893–8.[Abstract/Free Full Text]
  12. Lu X, Qin W, Li J, et al. The growth and metastasis of human hepatocellular carcinoma xenografts are inhibited by small interfering RNA targeting to the subunit ATP6L of proton pump. Cancer Res 2005;65:6843–9.[Abstract/Free Full Text]
  13. Mathupala SP, Parajuli P, Sloan AF. Silencing of monocarboxylate transporters via small interfering ribonucleic acid inhibits glycolysis and induces cell death in malignant glioma: an in vitro study. Neurosurgery 2004;55:1410–9.[CrossRef][Medline]
  14. Petrangolini G, Supino R, Pratesi G, et al. Effect of a novel vacuolar-H+-ATPase inhibitor on cell and tumor response to camptothecins. J Pharmacol Exp Ther 2006;318:939–46.[Abstract/Free Full Text]
  15. Warburg O, Wind F, Negelin E. The metabolism of tumors in the body. J Gen Physiol 1927;8:519–30.[Free Full Text]
  16. Gillies RJ, Raghunand N, Garcia-Martin ML, Gatenby RA. pH imaging. A review of pH measurement methods and applications in cancers. IEEE Eng Med Biol Mag 2004;23:57–64.[Medline]
  17. Matoba S, Kang JG, Patino WD, et al. p53 regulates mitochondrial respiration. Science 2006;312:1650–3.[Abstract/Free Full Text]
  18. Halestrap AP, Price NT. The proton-linked monocarboxylate transporter (MCT) family: structure, function and regulation. Biochem J 1999;343 Pt 2:281–99.[CrossRef][Medline]
  19. Volk C, Kempski B, Kempski OS. Inhibition of lactate export by quercetin acidifies rat glial cells in vitro. Neurosci Lett 1997;223:121–4.[CrossRef][Medline]
  20. Shrode LD, Putnam RW. Intracellular pH regulation in primary rat astrocytes and C6 glioma cells. Glia 1994;12:196–210.[CrossRef][Medline]
  21. Volk C, Tobias A, Kempski OS. A proton-translocating H+-ATPase is involved in C6 glial pH regulation. Biochim Biophys Acta 1998;1372:28–36.[Medline]
  22. McLean LA, Roscoe J, Jorgensen NK, Gorin FA, Cala PM. Malignant gliomas display altered pH regulation by NHE1 compared with nontransformed astrocytes. Am J Physiol Cell Physiol 2000;278:C676–88.[Abstract/Free Full Text]
  23. Philippe JM, Dubois JM, Rouzaire-Dubois B, et al. Functional expression of V-ATPases in the plasma membrane of glial cells. Glia 2002;37:365–73.[CrossRef][Medline]
  24. Gil MS, Zaderenzo P, Cruz F, Cerdan S, Ballesteros P. Imidazol-1-ylalkanoate acids as extrinsic 1H NMR probes for the determination of intracellular pH, extracellular pH and cell volume. Bioorg Med Chem 1994;2:305–14.[CrossRef][Medline]
  25. García-Martín ML, Hérigault G, Rémy C, et al. Mapping extracellular pH in rat brain gliomas in vivo by 1H magnetic resonance spectroscopic imaging: comparison with maps of metabolites. Cancer Res 2001;61:6524–31.[Abstract/Free Full Text]
  26. Fonchy E, Lahrech H, François-Joubert A, et al. A new gadolinium-based contrast agent for magnetic resonance imaging of brain tumors: kinetic study on a C6 rat glioma model. J Magn Reson Imaging 2001;14:97–105.[CrossRef][Medline]
  27. Mroczkiewicz A. Imidazolylalkanecarboxylic acids [in Polish]. Acta Pol Pharm 1984;41:435–40.
  28. Loginov VM, Vorob'ev IV, Khalanskii AS, Kachkov IA, Promyslov MS. Study of lactate content in glial brain tumors and brain tissue of experimental animals in short-term artificial hyperglycemia [in Russian]. Zh Vopr Neirokhir Im NN Burdenko 1985;31–5.
  29. Terpstra M, Gruetter R, High WB, et al. Lactate turnover in rat glioma measured by in vivo nuclear magnetic resonance spectroscopy. Cancer Res 1998;58:5083–8.[Abstract/Free Full Text]
  30. Thistlethwaite AJ, Alexander GA, Moylan DJ III, Leeper DB. Modification of human tumor pH by elevation of blood glucose. Int J Radiat Oncol Biol Phys 1987;13:603–10.[Medline]
  31. Sevick EM, Jain RK. Blood flow and venous pH of tissue-isolated Walker 256 carcinoma during hyperglycemia. Cancer Res 1988;48:1201–7.[Abstract/Free Full Text]
  32. Reinhold HS, van den Berg-Blok AE, van den Berg AP. Dose-effect relationships for glucose-induced tumour acidification and its erythrocyte flux. Eur J Cancer 1991;27:1151–4.[Medline]
  33. Helmlinger G, Sckell A, Dellian M, Forbes NS, Jain RK. Acid production in glycolysis-impaired tumors provides new insights into tumor metabolism. Clin Cancer Res 2002;8:1284–91.[Abstract/Free Full Text]
  34. Benda P, Someda K, Messer J, Sweet WH. Morphological and immunochemical studies of rat glial tumors and clonal strains propagated in culture. J Neurosurg 1971;34:310–23.[Medline]
  35. Julien C, Payen JF, Tropres I, et al. Assessment of vascular reactivity in rat brain glioma by measuring regional blood volume during graded hypoxic hypoxia. Br J Cancer 2004;91:374–80.[Medline]
  36. Hiba B, Faure B, Lamalle L, Decorps M, Ziegler A. Out-and-in spiral spectroscopic imaging in rat brain at 7 T. Magn Reson Med 2003;50:1127–33.[CrossRef][Medline]
  37. Bottomley PA. Spatial localization in NMR spectroscopy in vivo. Ann N Y Acad Sci 1987;508:333–48.[Abstract]
  38. Shen J, Rothman DL, Hetherington HP, Pan JW. Linear projection method for automatic slice shimming. Magn Reson Med 1999;42:1082–8.[CrossRef][Medline]
  39. Zoula S, Hérigault G, Ziegler A, et al. Correlation between the occurrence of 1H-MRS lipid signal, necrosis and lipid droplets during C6 rat glioma development. NMR Biomed 2003;16:199–212.[CrossRef][Medline]
  40. Tkac I, Starcuk Z, Choi IY, Gruetter R. In vivo 1H NMR spectroscopy of rat brain at 1 ms echo time. Magn Reson Med 1999;41:649–56.[CrossRef][Medline]
  41. Alley MT, Glover GH, Pelc NJ. Gradient characterization using a Fourier-transform technique. Magn Reson Med 1998;39:581–7.[CrossRef][Medline]
  42. García-Martín ML, Martinez GV, Raghunand N, et al. High resolution pHe imaging of rat glioma using pH-dependent relaxivity. Magn Reson Med 2006;55:309–15.[CrossRef][Medline]
  43. Terpstra M, High WB, Luo Y, et al. Relationships among lactate concentration, blood flow and histopathologic profiles in rat C6 glioma. NMR Biomed 1996;9:185–94.[CrossRef][Medline]
  44. Covington AK, Bates RG, Durst RA. Definition of pH scales, standard reference values, measurement of pH and related terminology. Pure Appl Chem 1985;57:531–42.
  45. Kalmus J, Okunieff P, Vaupel P. Effect of intraperitoneal versus intravenous glucose administration on laser Doppler flow in murine FSaII tumors and normal skin. Cancer Res 1989;49:6313–7.[Abstract/Free Full Text]
  46. Ward KA, DiPette DJ, Held TN, Jain RK. Effect of intravenous versus intraperitoneal glucose injection on systemic hemodynamics and blood flow rate in normal and tumor tissues in rats. Cancer Res 1991;51:3612–6.[Abstract/Free Full Text]
  47. Lust WD, Schwartz JP, Passonneau JV. Glycolytic metabolism in cultured cells of the nervous system. I. Glucose transport and metabolism in the C-6 glioma cell line. Mol Cell Biochem 1975;8:169–76.[CrossRef][Medline]
  48. Griffiths JR, McIntyre DJO, Howe FA, Stubbs M. Why are cancers acidic? A carrier-mediated diffusion model for H+ transport in the interstitial fluid. Novartis Found Symp 2001;240:46–62.[Medline]
  49. Engasser JM, Horvath C. Buffer-facilitated proton transport pH profile of bound enzymes. Biochim Biophys Acta 1974;358:178–92.[Medline]
  50. Sanchez-Alvarez R, Paino T, Herrero-Gonzalez S, Medina JM, Tabernero A. Tolbutamide reduces glioma cell proliferation by increasing Connexin 43, which promotes the up-regulation of p21 and p27 and subsequent changes in retinoblastoma phosphorylation. Glia 2006;54:125–34.[CrossRef][Medline]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Provent, P.
Right arrow Articles by García-Martín, M. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Provent, P.
Right arrow Articles by García-Martín, M. L.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online