Abstract
The invasive transformation of A-459 lung epithelial carcinoma cells has been linked to the autocrine regulation of malignant phenotypic changes by transforming growth factor β (TGF-β). Here we demonstrate, using stable 13C glucose isotopes, that the transformed phenotype is characterized by decreased CO2 production via direct glucose oxidation but increased nucleic acid ribose synthesis through the nonoxidative reactions of the pentose cycle. Increased nucleic acid synthesis through the nonoxidative pentose cycle imparts the metabolic adaptation of nontransformed cells to the invasive phenotype that potentially explains the fundamental metabolic disturbance in tumor cells: highly increased nucleic acid synthesis despite hypoxia and decreased glucose oxidation.
Introduction
Members of the TGF 3 -β family decrease cell differentiation marked by a significant decrease in surfactant protein synthesis in two distinct pulmonary adenocarcinoma cell lines with bronchiolar (NCI-H441–4) and alveolar (NCI-H820) cell characteristics (1) . Hepatocyte growth factor and TGF increase cell number and tritiated thymidine incorporation into nucleic acid in H441 cell cultures. This proliferative phenotype shows a significant increase in the phosphorylation of the c-met proto-oncogene receptor and phosphorylation of the intracellular signaling molecules p42 and p44 mitogen-activated protein kinases (2) . It has recently been revealed that cyclosporine treatment increases TGF-β levels and promotes the transformation of several human tumor cells to obtain an invasive phenotype in vitro (3) . The regulation of TGF-β mRNA synthesis, tumor cell hypoxia, glucose utilization, and cell transformation are closely linked in liver tumor cells (4) ; therefore, it is likely that specific reactions of intracellular glucose metabolism impart the metabolic adaptation of human cells to malignant transformation in response to TGF-β2 treatment. Here we demonstrate that TGF-β primarily targets glucose carbon deposition into nucleic acid through the nonoxidative transketolase pathway while decreasing 13CO2 production from[ 1,2-13C2]glucose through the oxidative G6PD pathway in lung epithelial cells. Other glucose metabolic pathways, such as the TCA cycle anaplerotic flux, are also affected by TGF-β2 treatment.
Materials and Methods
Cell Line and Culture.
Human lung adenocarcinoma cells (A-549 and H441 cells; American Type Culture Collection, Rockville, MD) and normal human fibroblasts (CRL-1501; American Type Culture Collection) were grown in MEM in the presence of 10% fetal bovine serum at 37°C in 95% air-5% CO2. To compare glucose utilization rates, ribose synthesis, lactate production, and glutamine oxidation by H441, A549 lung carcinoma cells, and CRL-1501 normal fibroblasts (control), 75% confluent cultures of each cell line were incubated in[ 1,2-13C2]glucose-containing media (180 mg%, 48% isotope enrichment). Cultures for the study were selected with the same cell number (6 × 107) that was achieved using an inverted binocular microscope and standard cell counting techniques. Two control cultures were incubated with unlabeled glucose in the media (180 mg%, 0% enrichment). Treatments with TGF-β were carried out using 0.01, 1, and 10 ng/ml porcine TGF-β2 (R&D Systems) dissolved in the culture media of cells for 24 and 48 h. On days 2 and 3, 500 μl of the media were removed from the cultures, and glucose levels (Cobas Mira chemistry analyser; Roche) as well as 13C:12C ratios were measured in released CO2 by a Finnegan Delta-S ion ratio mass spectroscope (GC/C/IRMS). 13CO2 release was used to estimate glucose carbon utilization through oxidation by the three cell lines and expressed as atom percent excess, which is the percentage of 13C produced by the cultured cells above background in calibration standard samples (5) . Data obtained with the H441 lung epithelial carcinoma cell line was reported herein in comparison with CRL-1501 fibroblasts.
RNA ribose was isolated by acid hydrolysis of cellular RNA after Trizol purification of cell extracts. Ribose was purified using a tandem set of Dowex1/Dowex50 ion exchange column (Sigma). Ribose was derivatized to its aldonitrile acetate form using hydroxil amine in pyridine and acetic anhydrate. We monitored the ion cluster around the m/z 256 (carbons 1–5 of ribose; chemical ionization) and m/z 217 (carbons 3–5 of ribose) and m/z 242 (carbons 1–4 of ribose; electron impact ionization) to find molar enrichment and positional distribution of 13C labels in ribose (6 , 7) .
Stable[ 1,2-13C2]d-glucose isotope was purchased with >99% purity and 99% isotope enrichment for each position (Isotec, Inc., Miamisburg, Ohio). For isotope incubation and drug treatment studies, fibroblasts and tumor cells were seeded in T175 tissue culture flasks after adjusting the number of cells to the values reported above. During the study, the cultures were supplied with unlabeled (natural) or[ 1,2-13C2]glucose dissolved in otherwise glucose- and sodium pyruvate-free DMEM with 2.5% fetal bovine serum. The final glucose concentration was adjusted to 180 mg/100 ml with ∼50% isotope enrichment in each culture. Glucose mass isotope analysis of the medium before cell incubations showed that the actual labeled glucose enrichment was 48% in the culture media, and this number was used for further calculations to determine maximum labeled glucose enrichment in the molecules that we isolated. The carbon fragments in biomolecules and their isotopomers under three experimental conditions were analyzed: 1, no label–no drug treatment; 2,[ 1,2-13C2]glucose label–no TGF-β2 treatment; and 3,[ 1,2-13C2]glucose label with increasing doses of TGF-β. Under condition of high enrichment (48%), singly labeled ribose molecules recovered from RNA (m1) on the first carbon position were used to measure the ribose molar fraction, which is produced by direct oxidation of glucose through the G6PD pathway, after subtracting the fraction of the singly labeled product which came from the TK pathway calculated by the m1 = m2(m3/m4) formula (7) . Doubly labeled ribose molecules (m2) on the first two carbon positions were used to measure the molar fraction produced by transketolase. Doubly labeled ribose molecules (m2) on the fourth and fifth carbon positions were used to measure molar fraction produced by triose phosphate isomerase and TK. Isotopomers with three labels were used to estimate ribose production by combining recycled products of the G6PD reaction through the TK and transaldolase reactions. Isotopomers with four labels (m4) were used to estimate synthesis through TK and triose phosphate isomerase, as described before (7) .
Lactate from the cell culture media (0.2 ml) was extracted by ethylene chloride after acidification with HCl. Lactate was derivatized to its propylamine-HFB form, and the m/z 328 (carbons 1–3 of lactate; chemical ionization) was monitored for the detection of m1 (recycled lactate through the PC) and m2 (lactate produced by the Embden-Meyerhof-Parnas pathway) for the estimation of PC activity (6) . Fragmental lactate analysis was not necessary because the[ 1,2-13C2]glucose tracer labels lactate on the third (m1) or the second and third (m2) carbon positions, which are clearly distinguished in the molecular ion.
Glutamate from the cell culture media (0.2 ml) was isolated by 6% perchloric acid treatment, centrifugation, and neutralization. Dowex50 (H+) column was used to bind amino acids, which were eluted by 60 ml of 2 N-ammonium hydroxide. Purified amino acids from the media were converted to their TAB using the method of Leimer (8) . The molecular ion for TAB-glutamate was detected at m/z 356. Under electron impact ionization conditions, ionization of TAB-glutamate gives rise to two fragments, m/z 198 and m/z 152, corresponding to C2—C5 and C2—C4 of glutamate.
Data Analysis and Statistical Methods.
In vitro experiments were carried out in duplicates using four cultures each time for each treatment regimen. Mass spectroscopic analyses were carried out by three independent automatic injections of 1-μl samples by the automatic sampler and accepted only if the standard sample deviation was <1% of the normalized peak intensity. Statistical analysis was performed using the parametric unpaired, two-tailed independent sample t test with 99% confidence intervals (μ ± 2.58ς), and P < 0.05 was considered to indicate significant difference in glucose carbon metabolism.
Results
For the present study, we measured glucose consumption, ribose synthesis, 13CO2 release, lactate production and TCA cycle anaplerotic flux to characterize the overall metabolic effect of TGF-β on tumor cell glucose metabolism. We compared glucose utilization rates and metabolic changes observed in H441 and A-549 lung adenocarcinoma cells to that of CRL-1501 normal fibroblasts in the presence of[ 1,2-13C2]glucose and TGF-β. Treatments with TGF-β were carried out using 0.01, 1, and 10 ng/ml porcine TGF-β2 for 24 and 48 h. These doses of TGF-β were selected for the study based on reports that normal plasma TGF-β levels in experimental animals are in the range of 1–6 ng/ml, and >600 ng/ml TGF-β levels were detected after treatment with lipopolysaccharide in mice (9) . After cyclosporine treatment, the measured TGF-β levels were in the 0.50–1.00 ng/ml range in the culture media of A-549 cells (3) .
Fibroblast cultures used minimal amounts of glucose from the media throughout the experiment because their media demonstrated steady glucose levels (∼180 mg%) close to the initial values. Glucose consumption by H441 and A-459 cells was ∼10-fold higher, and TGF-β-treated cultures maintained a high rate of glucose consumption. Accordingly, 13CO2 production by CRL-1501 cells was undetectable by ion ratio chromatography, whereas tumor cells showed high levels of 13C:12C ratio in released CO2, indicating high glucose oxidation rates. Labeled CO2 production from[ 1,2-13C2]glucose by the oxidation of the first carbon of glucose was decreased dose dependently by TGF-β in both H441 and A-549 cells, indicating that glucose metabolism in tumor cells follows a nonoxidative path in response to this TGF (Fig. 1) ⇓ .
Comparison of 13CO2 production by CRL-1501 fibroblasts and H441 lung epithelial carcinoma cells in the presence of [1,2-13C2]glucose in the culture media as measured by 13C:12C ratios using ion ratio mass spectroscopy. CRL-1501 cells did not release measurable amounts of labeled CO2 from labeled glucose because the 13C:12C ratios were at the reference level in both untreated and TGF-β2-treated cultures. Lung carcinoma cell cultures, on the other hand, produced significant amounts of 13C that increased 13C:12C ratios dramatically in the culture media. TGF-β2 treatment decreased 13CO2 production by H441 cells, which indicates that the utilization of glucose follows a nonoxidative path after treatment with this growth factor responsible for cell transformation. The decrease represents significant values after all doses after 24 and 48 h of TGF-β2 treatment (P < 0.01). n =4; bars, SD.
In the next step, RNA was extracted from tumor cells and ribose analyzed using established MS methods (6 , 7) . We observed high label incorporation from glucose to RNA ribose in H441 cells. Chemical ionization of RNA ribose by monitoring the m/z: 255–261 ion cluster revealed that 43.6% of ribose molecules carried 13C labels from[ 1,2-13C2]glucose in the form of m1, m2, m3, and m4 of ribose. The m0 unlabeled species was 56.4%. Because labeled glucose enrichment was 48% in the media, the theoretical maximum label accumulation (0.48) into RNA ribose was almost achieved by the measured values of heavy carbon ribose labeling in tumor cells. This high portion of glucose label recovery from ribose indicates that >90% of ribose molecules in H441 cell RNA was derived from glucose. On the basis of these findings, glucose molecules play a significant role in H441 cell proliferation as the primary source of ribose carbons for de novo nucleic acid synthesis.
To measure the contribution of the oxidative G6PD and the nonoxidative TK reactions to ribose synthesis in H441 cells, the m1:m2 and m3:m4 ratios of ribose isotopomers were found. These ratios describe the contribution of the oxidative and nonoxidative pathways of the PC based on the fact that G6PD and TK reactions generate different distribution of labels recognized as mass isotopomers of ribose from glucose as a single tracer method (6 , 7) . During metabolism through the oxidative G6PD pathway, the first carbon of glucose is lost as labeled 13CO2, whereas the nonoxidative steps of the PC do not separate the first two labeled carbons of glucose. Therefore, the m1:m2 and m3:m4 ratios of ribose isotopomers sensitively characterize the nonoxidative synthesis of ribose over the oxidative synthesis from glucose. The m1-m1′ (where m1′ = m2×[m3/m4]) formula was used to measure the molar fraction synthesized through the G6PD pathway, as described previously (6 , 7) . This method has been used to estimate the contribution of the G6PD and TK reactions to hepatoma (HepG2) and MIA pancreatic adenocarcinoma cell ribose synthesis, which use the nonoxidative PC predominantly for ribose synthesis (>70%) over the G6PD pathway (<30%). In H441 cells, the m3:m4 ratios were relatively low (0.89) as compared with the published value (1.6) for MIA pancreatic adenocarcinoma cells. The m1-m1′ formula yielded a value of 0.02 (or 2%) in H441 cells, which indicates that 98% of 13C-labeled ribose molecules in H441 cells was synthesized by the nonoxidative TK reaction, whereas only 2% of ribose molecules was derived directly through the oxidative pathway. TGF-β2 significantly increased 13C-label accumulation from[ 1,2-13C2]glucose into RNA ribose after treatment with 1 ng/ml TGF-β2 (Fig. 2) ⇓ , with a P of 0.0024.
Molar enrichment (ME; ME = Σmixn) of 13C in ribose (Y axis) synthesized from 13C labeled glucose after 0.01 ng/ml and 1 ng/ml TGF-β2 treatment (X axis) in H441 lung epithelial carcinoma cells after 48 h of treatment. These calculations allow the determination of how intensively glucose carbons are deposited into RNA ribose. One ng/ml TGF-β2 treatment increased significantly glucose carbon deposition into RNA ribose, with P approaching 0.002. n = 4; bars, SD.
PC activity was measured by the m1:m2 ratios in lactate using the published formula for PC calculations: PC = (m1/m2)(3+ m1/m2) (Refs. 6 and 10 ). As expected by the decrease of labeled 13CO2 release of treated tumor cells, the m1:m2 ratio significantly decreased after TGF-β2 treatment. The decrease in lactate m1 production (the recycled species through the PC) is shown in Fig. 3 ⇓ . Decreased m1 in lactate is a strong indicator of decreased G6PD activity and recycling of glucose through the nonoxidative steps of the PC. In a recent study, poor prognosis was found in nasopharyngeal cancer patients with low G6PD activity (11) . Because glucose carbon deposition into RNA ribose is increased after TGF-β treatment, the decrease in lactate m1:m2 ratio and the decrease in labeled CO2 production indicate that glucose carbon metabolism follows the nonoxidative path within the PC after treatment with this growth factor responsible for cell transformation.
PC activity measured by m1:m2 ratio in lactate produced by H441 lung epithelial carcinoma cells in the presence of [1,2-13C2]glucose in the culture media using GC/MS as the percentage of glycolysis. TGF-β2 treatment decreased m1:m2 ratios in lactate, which indicates that utilization of glucose decreased through the G6PD oxidative pathway in the PC. The decrease in m1:m2 lactate ratio represents significant values after 48 h of TGF-β2 treatment (P < 0.05). n = 4; bars, SD.
Anaplerotic substrate flux in the TCA cycle is a good indicator of cell intermediate metabolism for anabolic reactions through measuring carbon 13C distribution in glutamine (12) . TCA cycle activity in H441 cells was characterized through glutamine oxidation using the glutamine-glutamate equilibrium and their 2–4 and 2–5 carbon fragments by GC/MS in the selective ion monitoring mode for the present study (13) . Anaplerotic flux activity calculated by the accumulation of glucose labels into glutamate showed a significant increase after TGF-β treatment, which indicates increased amino acid synthesis from glucose in the TCA cycle as shown in Fig. 4 ⇓ .
TCA cycle anaplerotic flux as measured by m1:m2 ratio in glutamate produced by H441 lung epithelial carcinoma cells in the presence of[ 1,2-13C2]glucose in the culture media using GC/MS. The increase in anaplerotic flux in the TCA cycle indicates that glucose carbons are increasingly needed in transforming tumor cells for intracellular anabolic processes, including the synthesis of glutamate. P < 0.05 after 10 ng/ml TGF-β, 48-h studies. n = 4; bars, SD.
Discussion
Data presented here provide direct evidence regarding the crucial role of glucose carbons in tumor cell nucleic acid and amino acid synthesis and further demonstrate that nonoxidative ribose synthesis pathways predominate and further increase in response to TGF-β in tumor cells. Cell transformation induced by TGF-β2 is accompanied by increased glucose carbon deposition into nucleic acid and amino acids through nonoxidative metabolic reactions. This increase in the nonoxidative metabolism of glucose in the PC may provide the explanation at the molecular level for the principal metabolic disturbance observed in human tumors: increased glucose utilization for nucleic acid synthesis despite decreased glucose oxidation and hypoxia.
The increasing role of glucose carbons in cell metabolism as the main source of intermediate metabolites of nucleic acid and amino acids in response to TGF treatment also explains how tumor cells are capable of maintaining a continuous proliferation rate in a weakening host. Glucose is a very reliable carbon source for anabolic intracellular reactions because it is regenerated and replenished if necessary at constant rates by the host organism until complete exhaustion. This is a significant distinction between normal and tumor cells, which is demonstrated in our study by comparing normal, differentiated fibroblasts, which do not depend on glucose carbons for their growth in the same environment as tumor cells do in culture.
Tumor cells show exceptional dependence on glucose carbons, and their level of transformation and malignancy correlates with increased metabolism of glucose; yet glucose tracers are used only in the diagnosis of cancer by positron emission tomography. This study demonstrates that the fate of glucose carbons in tumor cells is an increased use for intracellular synthetic reactions, therefore, macromolecule synthesis from glucose and glucogenic precursors should be considered critical for tumor cells. The identification of key metabolic enzymes specific for glucose using anabolic processes in tumor cells has the potential to identify new target sites for the treatment of cancer. One specific target site where such approach attacks tumor cells is where ribose for de novo synthesized nucleosides is produced from glucose through the nonoxidative reactions of the PC, which tumor cells use predominantly for nucleic acid production in response to a strong transforming factor, as reported herein.
Footnotes
-
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.
-
↵1 Supported by Grant PHS M01-RR00425 from the General Clinical Research Unit and Grant P01-CA42710 from the UCLA Clinical Nutrition Research Unit Stable Isotope Core.
-
↵2 To whom requests for reprints should be addressed, at the UCLA School of Medicine, Harbor-UCLA Research and Educational Institute, 1124 West Carson Street, RB-1 Room 125, Torrance, CA 90502. Phone: (310) 222-1883; Fax: (310) 533-0627; E-mail: boros{at}gcrc.humc.edu
-
↵3 The abbreviations used are: TGF, transforming growth factor; G6PD, glucose-6-phosphate-dehydrogenase; TCA, tricarboxylic acid; TK, transketolase; TAB, trifluoroacetamide butyl ester; PC, pentose cycle; GC/MS, gas chromatography/mass spectrometry.
- Received September 7, 1999.
- Accepted January 18, 2000.
- ©2000 American Association for Cancer Research.
Volume 60, Issue 5
