Phosphorylation of pyruvate kinase and glycolytic metabolism in three human glioma cell lines.

Tumor Biol 1991;12:339-352

© 1991 S. Karger AG. Basel 1010-4283/91 /0 126-0339S2.75/0

Phosphorylation of Pyruvate Kinase and Glycolytic Metabolism in Three Human Glioma Cell Lines Paschal A. Oude Weernink, Gert Rijksen, Gerard E.J. Staal Department o f Hematology, Laboratory o f Medical Enzymology, University Hospital. Utrecht, The Netherlands

Abstract. Three cell lines established from human gliomas were found to differ in the capacity to phosphorylate the glycolytic enzyme pyruvate kinase in vitro. Phosphorylation in the glioblastoma cell line U-138 was more pronounced than in the glioma cell line Hs 683 and in the glioblastoma cell line A-172. All 3 cell lines showed similar pyruvate kinase isozyme patterns and expressed about 90% K-type and 10% M-type subunits. So, differences in pyruvate kinase phosphorylation could not be explained by differences in the availability of the appropriate substrate, being pyruvate kinase type K. As in gliomas, phosphorylation could specifically and almost completely be inhibited by fructose-1,6-bisphosphate. In order to investigate a potential physiological significance of the phosphorylation of pyruvate kinase, we have characterized these cell lines for several glycolytic parameters. In U-138 cells, the production of lactate appeared to be 2 times higher as compared with A-172 and Hs 683 cells under normal growth conditions and even 4 times higher under low glucose culture regime. The efflux of lactate correlated with the pyruvate kinase phosphorylation pattern in the cell lines. In none of the cell lines could the lactate production be stimulated by glutamine as additional energy source under low glucose culture conditions. The higher glycolytic flux in U-138 cells was not accompanied by higher glycolytic enzyme activities. The isozyme patterns of hexokinase, pyruvate kinase, aldolase, enolase and lactate dehydrogenase in the cell lines were nearly identical and resembled the patterns previously described for solid gliomas. However, the isozyme composition of phosphofructokinase in the cell lines differed from the situation in gliomas. While in gliomas the expression of L-type phosphofructoki nase is favored, in the glioma cell lines, we found an increase in the expression of C-type subunits.

Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/19/2018 8:59:15 AM

Key Words. Human glioma cell lines • Pyruvate kinase • Phosphorylation • Glycolysis • Lactate production • Glycolytic enzymes

Oude Weernink/Rijksen/Staal

Introduction

In malignant tumors, an altered expres sion of the glycolytic enzyme pyruvate ki nase (PK) has been observed. In these tu mors, the expression of K-type-containing isozymes is favored above M- and L-typecontaining isozymes. This phenomenon has been demonstrated in a variety of tumors, including tumors of the breast [1, 2], liver [3], thyroid gland [4], soft tissues [5], in tumors of childhood [6], and in neuroecto dermal tumors [7, 8], Generally, the PK iso zyme shift was found to correlate with the grade of malignancy of the neoplasm [9], The functional meaning of this preference of tumors for the K-type enzyme of PK is only partly understood. It has been argued that the presence of the K isozyme is neces sary for the process of aerobic glycolysis [10], A high rate of glycolytic activity in the pres ence of oxygen is a common characteristic of malignant cells and tumors [11], A change in the expression of isozymes, which respond differently to metabolites, might explain the absence of the Pasteur effect in tumor cells. Indeed, as far as PK is concerned this hy pothesis might be correct, as the K-type PK exhibits different kinetic features as com pared with the M and L type. For instance, the K type has a lower affinity to phosphoenolpyruvate and is more inhibited by ala nine than the M type [12, 13], Furthermore, K-type but not M-type PK can be phosphorylated by a cAMP-independent protein ki nase in vivo and in vitro [ 14, 15], It has been reported that this phosphorylation of PK-K results in inactivation of the enzyme [16, 17]. Recently, we have demonstrated that in human brain tumors, the expression of Ktype PK was accompanied by phosphoryla

tion of the enzyme [18]. Phosphorylation appeared to be cAMP-independent and oc curred exclusively on serine residues. In the normal brain, where the M type is predomi nant, no phosphorylation of PK was de tected. In the same study, phosphorylation of PK was also observed in 2 cultured glioma cell lines; however, the degree of phosphate incorporation into PK differed between the cell lines despite equal amounts of K-type molecules present [18]. In this report, we further investigated the phosphorylation of PK in 3 cell lines estab lished from human gliomas: the glioma cell line Hs 683 [19] and the glioblastoma cell lines A-172 [20] and U-138 [21], We were also interested to find out whether differ ences in PK phosphorylating capacity could be correlated with the glycolytic activity of the cell lines. Therefore, we determined the glycolytic flux of the cell lines by measuring lactate productions. The formation of lactate was measured both under normal growth conditions and under low or high glucose regime. The effect of glutamine as additional energy source was also studied, as glutamine has been recognized as an important sub strate for the energy metabolism of tumor cells, especially under low glucose conditions [22, 23]. In addition, we determined glyco lytic enzyme activities and isozyme compo sitions. Materials and Methods Cell Culture The human glioma cell line Hs 683 and the glio blastoma cell lines U-138 and A-172 were purchased from the American Type Culture Collection (Rock ville, Md., USA). The cells were grown as monolayers in Ham’s medium (F-10 nutrient mixture; Gibco, Paisley, Scotland), supplemented with 10% heat-inac tivated fetal calf serum (Gibco), 15 mM Hepes (pH


340

7.4), 2 m M L-glutamine, 100 pg/ml streptomycin and 100 U/ml penicillin. For the U-138 cells, MEM nonessential amino acids (Gibco) were added to the me dium. Cultures were incubated at 37 °C in a humidi fied 5% COi atmosphere. Cells were harvested by trypsinization (0.05% trypsin and 0.02% EDTA in phosphate-buffered saline) before confluency was reached. Detached cells were centrifuged at 200 g for 10 min, washed 3 times with PBS and stored as a dry pellet at - 8 0 ° C or used immediately. The cultures were routinely checked for mycoplasma contamina tion. Special culture conditions during lactate mea surements are described below. Population doubling times o f the cell lines were calculated by counting cells in a hemocytometer after preparing single cell suspensions by trypsinization. In our culture system, doubling times were: A -172 = 71.5 h; Hs 683 = 19.7 h; U-138 = 33.1 h. Sample Preparation Frozen or freshly prepared cells (2 X 107 cells) were resuspended in 0.5 ml ice-cold extraction buffer containing 50 mM Tris-HCl (pH 7.5), 100 mM KC1, 100 mM sucrose, 10 mM MgCL, 10 mM glucose, 1 mAf EDTA, 1 mM dithiothreitol, 0.05% Nonidet P40, 1 mM phenylmethylsulfonyl fluoride, 1 mM paminobenzamidine, 250 pg/1 leupeptin and 0.1 mM p-tosyl-L-lysine chloromethylketone-HCl. Cell lysis was performed by sonication for 2 X 5 s on ice at an amplitude o f 16 pm with a 150-W ultrasonic desintegrator MK2 (MSE Scientific Instr., Crawley, UK). After cell disruption, the lysate was centrifuged at 48,000 g for 30 min at 4 °C to obtain a cytosolic frac tion. Normal brain tissue was obtained at necropsy from a patient who died o f non-neoplastic disease. The tissue was homogenized in a Potter homogenizer in the above-mentioned extraction buffer, and further processed as described for the cell pellets. Phosphorylation o f PK Cytosolic extracts, prepared as described above, were passed over Sephadex G-25 columns (P D -10, Pharmacia, Uppsala, Sweden), previously equili brated in 50 m M Tris-HCl (pH 7.5), 20 m M MgCh, 2 m M dithiothreitol and 1 m M EDTA. Phosphoryla tion o f PK was performed in the same buffer after addition o f 5 m M NaF and 50 \iM sodium vanadate. Reaction mixtures containing 3 units o f PK activity in a total volume o f 100 pi were incubated for 1 h at

341

4 °C in the presence o f 500 pA/ [y-32P]ATP with a specific activity o f 0.25 Ci/mmol (New England Nu clear, Boston, Mass., USA). Control experiments were performed in the presence of 0.1 mM fructose-1,6bisphosphate (FDP), which is known to inhibit the phosphorylation o f PK [18]. After phosphorylation, PK was immunoprecipitated from the mixtures using a purified rabbit polyclonal antibody against both human M- and K-type PK as described before [18]. The immunoprecipitates were analyzed by electro phoresis on 8% polyacrylamide slab gels in the pres ence o f sodium dodecyl sulfate (SDS). The gels were stained with Coomassie brilliant blue and dried be tween cellophane membranes. Autoradiographs were obtained by exposure o f the dried gels against X-ray films (Kodak X-omat) using intensifying screens (Cronex-Dupont) at - 80 °C. The autoradiographs were scanned at 540 nm in a Gilford Response spec trophotometer using a gelscan program. Enzyme Assays and Kinetics Activities o f glycolytic enzymes in the cytosolic fractions were determined according to Beutler [24] and expressed in units/milligram protein. One unit o f enzyme activity is defined as that amount o f enzyme which converts 1 pmol o f substrate per minute at 37 °C. Protein content was measured according to Lowry et al. [25], using bovine serum albumin as a standard. Determination o f soluble versus mitochon drial bound hexokinase (HK) was performed as de scribed before [26]. Substrates and auxiliary enzymes were obtained from Boehringer (Mannheim, FRG). Separation o f Isozymes Isozymes o f HK, PK, aldolase and enolase were separated by electrophoresis on cellulose acetate strips (Chemetron, Milan, Italy) and subsequently stained for enzyme activity as described [4, 27-29]. Isozyme composition was quantified by scanning the zymograms at 540 nm in a Beckman CDS-200 densi tometer. Isozymes o f lactate dehydrogenase (LDH) were analyzed by electrophoresis on agarose gels, us ing the Titan Gel LD Isoenzyme System (Helena Lab oratories, Gateshead, UK). Electrophoresis o f glucose-6-phosphate dehydrogenase (G6PD) on cellulose acetate at pH 7.5 was carried out according to the method o f Rattazzi et al. [30], Characterization o f the phosphofructokinase (PFK) subunit composition was performed essentially according to Heesbeen et al. [31]. PFK was partially


Glycolytic Metabolism in Glioma Cell Lines

342

Lactate Measurements Cells were seeded in 80 cm2 polystyrene tissue cul ture flasks and were allowed to grow under normal culture conditions until reaching about 80% confluency. At the beginning o f the experiment, the me dium was removed and the monolayers were washed and subsequently incubated with serum-free basal medium. Basal medium is Ham’s F-10 medium lack ing glucose, glutamine and glutamic acid. Glucose and glutamine could be added to the desired concen trations; the glucose concentration o f the medium was changed between 0.1 g/1, 1 g/1, and 10 g/1 (0.55, 5.5, and 55 mM , respectively) and the glutamine concen tration was 0 or 4 mM. The glucose concentration o f normal Ham’s F-10 medium is 5.5 mM. The time course o f lactate release into the culture medium was determined by removing aliquots (0.5 ml) from tripli cate flasks initially containing 20 ml o f medium. The samples were frozen immediately and processed as described below. Cell numbers were counted and pro tein content was determined prior to and at termina tion o f the experiment; protein was determined ac cording to Lowry et al. [25]. The collected samples were thawed and centrifuged at 800 g to remove cell debris. Lactate in the medium was measured enzy matically on a Kontron lactate analyzer 640. Total product formed includes amounts removed in prior aliquots. ATP Assay ATP concentrations were determined with a luciferin-luciferase technique in a Packard Picolite 6100 luminometer according to Akkerman et al. [35], Monolayers in the logarithmic phase o f growth were extracted with an ice-cold 6% (w/v) HCIO4 solution

Table 1. PK isozyme composition and PK phosphorylating capacity o f the glioma cell lines

Isozyme composition, % K. KjM KiM; Total K subunits, % PK phosphorylation relative

A -172

Hs 683

U-138

65.2 34.8

64.7 35.3

52.6 41.5 15.8 86.6

-

-

91.3

91.2

1

2

12

Quantification o f the PK isozyme composition and eventually the percentage o f K-type subunits was assessed after densitométrie analysis o f electrophoretograms. PK phosphorylating capacities of the cell lines are given as relative values based on scans o f the autoradiograph o f figure 1.

(1 ml for 2 X 106 cells). The solution was neutralized with a 0 .2 -3 /Tris maleate buffer (pH 7.4) centrifuged in an Eppendorf centrifuge, and the ATP content in the supernatant was measured.

Results

Phosphorylation o f PK Endogenous phosphorylation of the cyto solic fractions of the 3 glioma cell lines was carried out with [y-32P]ATP as phosphate donor in the presence of magnesium ions. The experiments were done in either the presence or absence of 0.1 mM FDP. PK was immunoprecipitated from the phosphorylated extracts using a rabbit polyclonal anti body and subsequently analyzed by SDS polyacrylamide gel electrophoresis and auto radiography (fig. 1). Equal amounts of en zyme activity were precipitated in all assays. PK subunits migrate with an apparent mo lecular mass of about 60 kD (fig. la). In pre-


purified from the cytosolic fractions by Cibacron Blue F3GA affinity chromatography (Pierce Chemical Co., Rockford, 111., USA). The eluted fraction containing PFK activity was concentrated with an Amicon C30 microconcentrator and subjected to SDS electropho resis in 6 % polyacrylamide gels according to Laemmli [32]. The sample contained 1 pg o f PFK protein and had a specific activity of about 200 U/mg protein. Normal human brain PFK, which was completely purified according to Dunaway and Kasten [33], was used as a standard. The gels were silver-stained as described by Wray et al. [34], and the PFK subunit bands were quantified using a Gilford Response spec trophotometer with gel-scanning accessory.

Oude Weemink/Rijksen/Staal


343

Fig. 1. SDS-polyacrylamide gel electrophoresis o f immunoprecipitates from phosphorylated extracts o f A172 (lanes 1 and 4), Hs 683 (lanes 2 and 5), and U-138 (lanes 3 and 6). Phosphorylation was performed in the absence (lanes 1-3) or presence (lanes 4 -6 ) o f 0.1 m.V/ FDP. a Coomassie staining pattern, b Correspond ing autoradiograph. Prestained low-molecular-weight markers were used as standards. The position of PK is indicated by the arrow.

reduced the radiolabeling of the 60-kD pro tein, indicating that it was PK itself that served as the substrate in the phosphoryla tion reaction (fig. lb, lanes 4-6). Absence o f Correlation o f PK Phosphorylation with PK Isozyme Composition The PK isozyme composition of the 3 glioma cell lines was determined to see whether this parameter might be correlated with the degree of PK phosphorylation in the cell extracts. The results are given in ta ble 1. The PK isozyme composition is almost the same for all 3 cell lines, the K4 homotetramer being the predominant form with a smaller amount of K3M. Only in U-138 was a small amount of the K2M2 hybrid detected. As a result, all cell lines express about 90% K-type subunits which is similar to the iso-


vious experiments, the identity of this protein to be PK was confirmed by its specific reac tivity with a PK-type-K specific antibody, raised against a synthetic peptide [18, 36]. The autoradiograph shows that PK was phosphorylated in the extracts of all 3 cell lines (fig. lb, lanes 1-3). The strongest phos phorylation could be observed in the glio blastoma cell line U-138. Evidence that the black spots on the autoradiograph were in deed caused by radiolabeled PK molecules and not by another coprecipitating and co migrating polypeptide was obtained by per forming the phosphorylation reaction in the presence of 0.1 mM FDP. Recently, we have demonstrated that the addition of FDP in this concentration completely inhibits the phosphorylation of PK in biopsies of human gliomas [18]. Indeed, the presence of FDP during the phosphorylation reaction in the glioma cell extracts greatly and specifically

344


zyme composition in biopsies of poorly dif ferentiated human gliomas and glioblasto mas [37], So, differences in PK phosphoryla tion observed between the cell lines cannot be attributed simply to the absence of the proper substrate, i.e., K-type PK, in 1 of the cell lines. Lactate Production The lactate efflux from the glioma cell lines A-172, Hs 683 and U-138 is shown in figure 2. The total amount of lactate present in the growth medium was determined after several time intervals and as a function of the glucose and glutamine concentrations in the medium. Several conclusions can be drawn from these data. In the first place, during the course of incubation, the production of lac tate appeared to be linear with time. At the

higher glucose concentrations in the medium (5.5 and 55 mM) lactate production was even linear up to 24 h, however, the produc tion of lactate levelled off after 7.5 h of incu bation in the presence of 0.55 mM glucose alone or 0.55 mM glucose together with 4 mM glutamine due to depletion of nutrients (results not shown). Secondly, the lactate production by U138 cells was significantly higher as com pared with A-172 and Hs 683 cells (fig ures 2, 3). At normal glucose concentrations (5.5 mM), the lactate production by U-138 cells was about 2 times higher than the efflux from the other 2 cell lines; at low glucose concentrations, the production by U-138 cells was even 4 times higher than by A-172 cells (fig. 3). Thirdly, the cell lines responded differ ently to changes in the glucose concentra-


Fig. 2. Release o f lactate into the culture medium o f the cell lines A -172, Hs 683, and U -138. Monolayers at about 80% confluency were grown in 80-cm2 culture flasks and exposed to fresh serum-free basal medium, containing 55 m.M (high; •), 5.5 m M (normal; o), 0.55 m M glucose (low; ■), or 0.55 rriM glucose and 4 mM glutamine (□) for 7.5 h. At the times indicated, aliquots were removed for analysis o f lactate. Each point represents the mean o f 3 culture flasks, each assayed in duplicate.


345

Fig. 3. Production o f lactate by the cell lines A-172, Hs 683 and U-138 after 7.5 h of incubation as a function o f glu cose concentration in the medium. Cul ture conditions are described in Materi als and Methods and in the legend to fig ure 2. □ = 0.55 m.W. ■ = 5.5 m M ■ = 5.5 m.V/ glucose.

by elevating the glucose concentration. This suggests that for U-138 cells, 0.55 mM glu cose is not rate-limiting for a sufficient sup ply of energy and metabolites. However, for A-172 and Hs 683 cells, the glucose concen tration seemed to be rate-limiting and the production of lactate could be stimulated by increasing the glucose concentration from 0.55 to 5.5 mM, but this same effect could not be achieved by the addition of 4 mM glutamine. ATP Content ATP was measured in HCIO4 extracts of cell monolayers. The cell lines oentained about equal amounts of ATP, expressed as nanomols/106 cells: A-172 = 16.3 ± 1.2, Hs 683 = 23.8 ± 1.8, and U-138 = 21.1 ± 2.4 nmol/106 cells. Glycolytic Enzyme Activities The specific activities of various glyco lytic enzymes, expressed as U/mg protein, are summarized in table 2. The results show a tendency for higher specific enzyme activi ties in U-138. However, these higher activi ties can be explained by the lower protein


tions in the medium (fig. 2, 3). The U-138 cell line already produced lactate at a maxi mum rate in the presence of 0.55 mM glu cose; elevation of the glucose concentrations to 5.5 or 55 mM could not increase the pro duction of lactate further. Lactate produc tion in Hs 683 cells could be stimulated by increasing the glucose concentration from 0.55 to 5.5 mM, but an increase to 55 mM had no further effect. In A-172 cells, the pro duction of lactate reached a (apparent) max imum only at the highest glucose concentra tion of 55 mM. These differences might re present differences in the uptake capacity of glucose between the cell lines. The importance of glutamine as an addi tional substrate for the tumor cells was in vestigated by incubating the cells at the low glucose concentration (0.55 mM) in the pres ence or absence of 4 mM glutamine. In all 3 cell lines, the addition of 4 mM glutamine to the incubation medium had no effect on the production of lactate (fig. 2). In U-138 cells, this impotence of glutamine to increase the efflux of lactate could be expected, as lactate production is already maximal at 0.55 mM glucose and could equally not be enhanced

Table 2. Glycolytic enzyme activities and protein content o f the glioma cell lines A -172

Hs 683

U-138

HK PFK PK

0.17 + 0.04 0.1 6 ± 0 .0 3 4.7 9 ± 0 .2 4

0.14 ±0.06 0.18 ±0.04 4 .79± 0.50

0.26 ± 0.09 0.19 ±0.01 7.84 ±0.43

G6PD 6PGD PGI

0.28 ±0.08 0.08 ± 0.01 1.13 ± 0.12

0.38 ± 0.16 0.08 ± 0.01 1.83 ± 0.26

0.13 ± 0.06 0.06 ±0.03 2.57 ±0.32

GDH PGK PGM

3.97 ± 0.56 3.12 ±0.78 0.45 ±0.05

3.76± 0.34 2.61 ±0.39 0.85 ±0.07

6 .78± 0.85 4.27 ± 0.31 1.52 ± 0.22

Aldolase Enolase LDH

0.17 ±0.05 0.43 ±0.07 3.20 ±0.66

0.18 ±0.03 0.90 ± 0.30 6.85 ± 1.03

0.26 ±0.03 0.98 ±0.07 9.09 ± 1.48

Total protein

267 ± 2 7

187 ±31

104 ± 13

Enzyme activities are expressed in U/mg protein, protein content in pg/106 cells. Values are mean ± SD (n = 4). 6PGD = 6-phosphogluconate dehydro genase; PGI = glucose phosphate isomerase; GDH = glyceraldehyde-3-phosphate dehydrogenase; PGK = phosphoglycerate kinase; PGM = phosphoglycerate mutase.

content of U -138 cells. With this in mind, the rather low G6PD activity in U-138 cells is remarkable. It is of interest to note that PK activity is very high as compared with the activities of the other regulatory enzymes PFK and HK in all 3 cell lines. Isozyme Composition The isozyme compositions of HK, PFK. enolase, aldolase, and LDH in the 3 glioma cell lines are listed in table 3. The appear ance of HK type II is a finding common to a number of tumors and has also been re ported for the neoplastic human brain [38, 39]. While in A-172 only 1 isozyme could be detected (type I), both Hs 683 as well as U-

138 exhibited 2 HK bands. The second band was identified as HK type II. Only a minor amount of HK activity was measured in the particulate fraction (i.e., bound to the mito chondrial membrane). No differences be tween the cell lines were observed in this respect. In human gliomas, there is a relative in crease in the expression of the liver type sub unit of PFK accompanied by a decrease in the epxression of the M subunit, whereas the percentage of C subunit remains about the same (table 3). In all 3 glioma cell lines, we also found a decrease in the expression of PFK type M; however, in A-172 and Hs 683, this decrease was exclusively in favor of the expression of the C type, and not the L type. Only in U-138, a slight elevation in the expression of the L subunit was observed, but also in this cell line, the expression of the C subunit was predominant (table 3). In the normal brain, the eux, ay, and yy dimers of enolase are present [29], In all 3 cell lines, act enolase was the predominant form, whereas about one third of the enzyme was present as the ay hybrid isozyme. The neuron-specific yy isozyme was not detected (table 3). In gliomas and metastatic brain tumors, a shift can be observed towards the expression of A subunits of aldolase [40, 41]. Also, in the glioma cell lines, the presence of A sub units was by far predominant, resulting in primarily A4 homotetramers (table 3). Be sides A4, small amounts of the A3C and A2C2 hybrids were present. Identical patterns were observed in all 3 cell lines. In most mammalian tissues, LDH exists in 5 tetrameric isozymic forms, produced by the random association of 2 different poly peptides, A and B. The 5 possible tetramers are therefore designated as B4, BjA, B2A2,



346

347


BA3 and A4, or alternatively as LDH 1-5, respectively. Neoplastic diseases have been characterized by a shift in LDH isozyme pat terns, with a marked increase in the A4 (LDH 5) isozyme in neoplasms as compared to both normal tissues and benign tumors [42, 43], Both in Hs 683 and in U-138, LDH 5 is the most predominant isozyme (table 3). The other isozyme forms are also present, but in much lesser amounts. In U-138, where more than 60% of LDH activity is present as LDH 5, the LDH 1 isozyme is even absent. In A-172, the shift towards the expression of A-type subunits is less pronounced, resulting in LDH 4 to be the prominent isozyme. Electrophoresis of G6PD on cellulose acetate was performed to inspect a possible altered mobility of the enzyme in 1 of the cell lines. After staining for enzyme activity, only 1 band with normal mobility was present in all 3 cell lines (results not shown).

Table 3. Isozyme composition o f glycolytic en zymes in 3 human glioma cell lines Glioma

A -172

Hs 683 U-138

100 12.9

64.3 35.7 13.3

64.2 35.8 16.9

22.8 50.0 27.2

36.2 38.3 25.5

42.4 32.7 24.9

44.5 21.1 34.4

32.5 26.0 41.5

63 37 -

71 29 -

65 35 -

12.8 11.5 15.0 21.2 39.4

88.4 18.0 13.6

90.2 17.1 12.7

91.6 16.1 12.2

Normal brain

HK

HK I HK II Bound HK, %

-

PFK

C M L

17.3 34.1 48.5

Enolase a -a a -y y-y

Aldolase A4 A3C A?C~> AC3

c4

-

-

-

-

-

-

12.3 18.2 28.2 40.8 20.7

13.7 14.2 10.8 29.2 52.2

13.2 18.5 27.2 61.1

LDH

To gain more insight into the glycolytic metabolism in gliomas and especially in the role of the PK phosphorylation in this pro cess, we have chosen the model of estab lished glioma cell lines for further study. Cells grown in tissue culture offer several advantages in studies of this type, primarily because a single cell type, i.e., the tumor cell, can be used. For instance in phosphorylation reactions, no protein kinases or protein phosphatases originating from other cell types can mask the results, and isozyme pat terns cannot be influenced by variance in the cell types present in a solid tumor. In our study, we used the human glioma cell line Hs 683 and the glioblastoma cell lines U-138 and A-172. Indeed, as pre-

LDH LDH LDH LDH LDH

1 2 3 4 5

-

The data o f PFK in gliomas are derived from Staal et al. [49]

viously observed in the glioma biopsies, in the cultured glioma cell lines, PK became phosphorylated as well after incubation of cytosolic extracts with radiolabeled ATP (fig. 1). Although PK phosphorylation was found in all 3 cell lines, differences in the degree of phosphate incorporation were no ticed. The strongest phosphorylation was ob served in extracts of U-138 cells. These dif ferences in PK phosphorylating capacity of


Discussion

Fig. 4. Relation between the capacity to phosphorylate PK (relative values) and the lactate production at 0.55 m M glucose after 7.5 h of incubation (gm ol/106 cells). PK phosphorylation in the A -172 cell line was arbitrarily set at a value o f 1.

the cell lines could not be explained by the availability of the appropriate substrate (PK type K), as the cell lines had about the same PK isozyme composition (± 90% K sub units) and equal amounts of PK enzyme activity were used in the assays. The strong PK phosphorylation in U-138 cells was accompanied by a higher glycolytic flux in these cells as compared with A-172 and Hs 683 cells. This was concluded from measurements of lactate production. Lactate production by the glioma cells was deter mined by measuring the lactate content in the culture medium. As the efflux of lactate from cultured cells occurs down a concentra tion gradient, it is not necessary to measure intracellular lactate concentrations [44]. U138 cells were characterized by a high lactate production even at low glucose concentra tions in the medium (fig. 2, 3). This high pro duction of lactate in U-138, about 2 times as high as in A-172 and Hs 683 at normal glu


cose and even 4 times as high at low glucose culture conditions, was not accompanied by an altered ATP content of the cells. A-172 cells responded best to an increase in the medium glucose concentration by a higher lactate production (fig. 2, 3). Interestingly, the capacity to phosphorylate PK appears to correlate with the production of lactate, es pecially under low glucose growth conditions (fig- 4). Glutamine has been recognized as a ma jor substrate for the energy metabolism of rapidly growing tumor cells [22, 23]. It has been proposed that under conditions of car bohydrate limitation, all lactate produced would be derived from glutamine [10, 45]. PK is believed to play an essential role in the switch from glycolysis to glutaminolysis at reduced glucose concentrations. At low glu cose concentrations, the level of FDP in the cell will decrease, thereby allowing PK type K to become phosphorylated. This phosphorylation of PK would result in inac tivation of the enzyme and a blockade in gly colysis [10, 17]. Cells which exhibit a strong PK phosphorylation might better be able to use glutamine as an energy source. However, we found no increase in lactate production in any cell line after addition of glutamine to a low glucose culture medium (fig. 2). In A172 and Hs 683 cells, the lactate production could be stimulated by increasing the glucose concentration; in U-138 cells, this was not possible, indicating that the availability of glucose was not rate-limiting for this cell line. We have to practise some caution in interpreting the effects of glutamine as an additional energy source, as we have only measured the production of lactate. Besides lactate, other metabolites as well have been described as end products of glutaminolysis, for instance glutamate and aspartate [46].


348

Comparing our 3 cell lines, the highest glycolytic flux in U-l 38 cells was not associ ated with the shortest population doubling time (see Materials and Methods). However, in cultures of U-138 cells, a relatively large amount of cell death was observed, resulting in an apparent slower growth rate. We further investigated whether the high rate of glycolytic metabolism in U-138 cells was accompanied by changes in the activities of several enzymes of glycolysis and the hexose monophosphate shunt, and whether changes in isozyme patterns had occurred. No striking differences were observed be tween the enzyme activities of the cell lines (table 2). The apparent higher specific en zyme activities in U-138 seem to be due to the lower protein content of these cells. The higher activities in U-138 were found for all enzymes studied, except for 6-phosphogluconate dehydrogenase and especially for G6PD whose activity was markedly de creased in U-138 cells. This lower G6PD activity was not accompanied by an aberrant electrophoretic mobility of the enzyme. The enzyme activities in the glioma cell lines are in good agreement with the values previously reported for human gliomas [41]. Both gliomas and our glioma cell lines ex hibit a PK hyperactivity as compared with the activities of the other regulatory enzymes HK and PFK. When we arbitrarily assign the activity of HK as unity, the activity of PK in the cell lines reaches the value of 30. This is significantly higher than the activity in the normal brain, where PK activity is only 11 times higher than HK activity [47]. How ever, the situation in gliomas is not excep tional, for in skeletal muscle, the relative activity of PK is no less than 258 [47], This hyperactivity of PK with respect to the other regulatory enzymes HK and PFK questions

349

the significance of PK activity per se in the regulation of glycolysis. The isozyme patterns of PK, PFK, HK. LDH, aldolase and enolase in the cell lines were quite similar. Only the isozyme compo sition in A-172 was somewhat different. A172 cells expressed LDH type A to a some what lesser extent, resulting in LDH 4 to be the predominant isozyme instead of LDH 5, as in U-138 and Hs 683 cells. Furthermore, HK type II was not expressed in A-l 72 cells. However, the presence of HK II in cultured cells is reported to be highly dependent on the growth conditions [27]. The relative low proportion (about 15%) of HK activity asso ciated with mitochondria is remarkable, as in tissues which are critically dependent on glucose metabolism, like brain and neoplas tic tissues, HK is predominantly found in the mitochondrial-bound form [39, 48], Identical isozyme patterns as for the glioma cell lines have been reported for glio mas and other tumors: preferential expres sion of PK type K [7], a-enolase [29], aldo lase type A [41 ] and LDH type A [42] were observed. The increase of L-type PFK in human gliomas was not seen in the glioma cell lines [49], Instead, the cell lines exhib ited a shift towards the expression of C-type subunits as compared with the normal brain (table 3). PFK type C has been recognized to be the most insensitive type of subunit with regard to the allosteric effectors citrate, ATP, and fructose-2,6-bisphosphate [50], The dis crepancy between gliomas and glioma cell lines in this respect deserves further investi gation. In conclusion, comparing 3 human glioma cell lines, we have shown that U-138 cells have the highest capacity to phosphorylate PK. This phenomenon appeared to be associated with a high glycolytic flux. No



350

gross differences were observed between the cell lines in glycolytic enzyme activities and isozyme patterns; only the activity of G6PD was markedly decreased in U-138 cells. Fur ther, we have demonstrated that the glyco lytic isozyme patterns in the 3 glioma cell lines are similar to the patterns in solid glio mas, except for the PFK pattern.

7

8

9

Acknowledgments The authors are indebted to G. Gorter for the ATP measurements and the Department o f Clinical Chemistry for the LDH isozyme determination. G. Spierenburg is thanked for the mycoplasma screen ing. This work was supported by grant UUKC 87-2 from the Netherlands Cancer Foundation.

10

11 12

References 1 Ibsen KH, Orlando RA, Garratt KN, Hernandez AM, Giorlando S, Nungaray G: Expression o f multimolecular forms o f pyruvate kinase in nor mal, benign and malignant human breast tissue. Cancer Res 1982;42:888-892. 2 Hennipman A, van Oirschot BA, Smits J, Rijksen G, Staal GEJ: Heterogeneity of glycolytic enzyme activity and isozyme composition o f pyruvate ki nase in breast cancer. Tumor Biol 1988;9:178— 189. 3 Hammond K.D, Balinsky D: Isozyme studies of several enzymes o f carbohydrate metabolism and cell cultures. Cancer Res 1978;38:1323-1328. 4 Verhangen JN, van der Heijden MCM, De Jongvan Dijken J. Rijksen G, Der Kinderen PJ, Van Unnik JAM, Staal GEJ: Pyruvate kinase in nor mal thyroid tissue and thyroid neoplasms. Cancer 1985;55:142-148. 5 Staal GEJ, Rijksen G, van Oirschot BA; Roholl PJM: Characterization o f pyruvate kinase from human rhabdomyosarcoma in relation to immu nohistochemical and morphological criteria. Can cer 1989;63:479-483. 6 Cottreau B, Rousseau-Merck MF, N ezelof C, Kahn A: Pyruvate kinase and phosphofructoki-

13

14

15

16

17

18

nase isozymes in childhood cancer. Pediatr Res 1982;16:199-202. Bennett MJ. Timperlcy WR. Taylor CB, Hill AS: Fetal forms o f pyruvate kinase isozymes in tumors of the human nervous system. Neuropathol Appl Neurobiol 1975;1:347-356. Van Vcelen CWM, Vcrbicst H, Vlug AMC, Rijksen G, Staal GEJ: Isozymes o f pyruvate ki nase from human brain, meningiomas and malig nant gliomas. Cancer Res 1978;38:4681-4687. Staal GEJ, Rijksen G: Pyruvate kinase in selected human tumors; in Pretlow TG, Pretlow TP (eds): Biochemical and Molecular Aspects o f Selected Tumors. Orlando, Academic Press, 1991, in press. Eigenbrodt E, Fister P, Reinacher M: New per spectives on carbohydrate metabolism in tumor cells; in Beitner R (ed): Regulation o f Carbohy drate Metabolism. Boca Raton, CRC Press, 1985, vol 2, pp 141-179. Racker E: Why do tumor cells have a high aerobic glycolysis? J Cell Physiol 1976:89:697-700. Staal GEJ, Rijksen G: Regulation o f pyruvate kinase in normal and pathological conditions: in Beitner R (ed): Regulation o f Carbohydrate Me tabolism. Boca Raton, CRC Press, 1985, vol 1, pp 143-159. Imamura K, Noguchi T, Tanaka T: Regulation of isozyme patterns o f pyruvate kinase in normal and neoplastic tissues; in Staal GEJ, van Veelen CWM (eds): Markers o f Human Neuroectodermal Tumors. Boca Raton, CRC Press, 1986, pp 191— 222. Fister P, Eigenbrodt E, Presek P, Reinacher M, Schoner W: Pyruvate kinase type M2 is phosphorylated in the intact chicken liver cell. Biochem Biophys Res Commun 1983; 115:409-414. Rijksen G, van der Heijden MCM, Oskam R, Staal GEJ: Subunit-specific phosphorylation of pyruvate kinase in medullary thyroid carcinomas o f the rat. FEBS Lett 1988;233:69-73. Glossmann H, Presek P, Eigenbrodt E: Associa tion of the sre-gene product o f Rous sarcoma virus with a pyruvate kinase inactivating factor. Mol Cell Endocrinol 1981;23:49-63. Presek P, Reinacher M, Eigenbrodt E: Pyruvate kinase type M2 is phosphorylated at tyrosine resi dues in cells transformed by Rous sarcoma virus. FEBS Lett 1988;242:194-198. Oude Weemink PA, Rijksen G, van der Heijden


Oudc Weernink/Rijksen/Staal

19

20

21

22

23

24

25

26

27

28

29

30

MCM, Staal GEJ: Phosphorylation o f pyruvate kinase type K in human gliomas by a cyclic aden osine S'-monophosphatc-independent protein ki nase. Cancer Res 1990;50:4604-4610. Gershwin ME, Ikeda RM, Kawakami TG, Owens RB: Immunobiology of heterotransplanted hu man tumors in nude mice. J Natl Cancer Inst 1977;58:1455-1461. Giard DJ, Aaronson SA, Todaro GJ, Amstein P, Kersey JH, Dosik H, Parks WP: In vitro cultiva tion o f human tumors: Establishment o f cell lines derived from a series o f solid tumors. J Natl Can cer Inst 1973;51:1417-1423. Pontén J, Macintyre EH: Long-term culture of normal and neoplastic human glia. Acta Pathol Microbiol Scand 1968;74:465-486. Kallinowski F, Runkel S, Fortmeyer HP, Forster H, Vaupcl P: L-Glutamine: A major substrate for tumor cells in vivo? J Cancer Res Clin Oncol 1987;113:209-215. Medina MA, Sánchez-Jiménez F, Márquez FJ, Pérez-Rodriguez AR, Quesada AR, Nunez de Cas tro I: Glutamine and glucose as energy substrates for Ehrlich ascites tumour cells. Biochem Int 1988;16:339-347. Beutler E: Red cell metabolism: A manual o f bio chemical methods. New York, Gruñe & Stratton, 1975, pp 38-70. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ: Protein measurement with the Folin phenol re agent. J Biol Chem 1951;193:265-275. Sprengers ED, Koenderman AHL, Staal GEJ: Mi tochondrial and cytosolic hexokinase from rat brain: One and the same enzyme? Biochim Biophys Acta 1983;755:112-118. Kraaijenhagen RJ, Rijksen G, Staal GEJ: Hexoki nase isozyme distribution and regulatory proper ties in lymphoid cells. Biochim Biophys Acta 1980;631:402-411. Beemer FA, Vlug AMC, Rijksen G, Hamburg A, Staal GEJ: Characterization o f some glycolytic enzymes from human retina and retinoblastoma. Cancer Res 1982;42:4228-4232. Van den Doel EHM, Rijksen G, Roholl PJM, van Veelen CWM, Staal GEJ: Enolase isoenzymes in human gliomas. J Neurosurg 1986;65:345-353. Rattazzi MC, Bernini LF, Fiorelli G, Mannucci PM: Electrophoresis o f glucose-6-phosphate dehy drogenase: A new technique. Nature 1967;213: 79-80.

351

31 Heesbeen EC, Rijksen G, Batenburg JJ, van Golde LMG, Staal GEJ: Phosphofructokinase in alveolar type II cells isolated from fetal and adult rat lung. Biochim Biophys Acta 1989; 1002:388— 394. 32 Laemmli UK: Cleavage o f structural proteins dur ing the assembly o f the head o f bacteriophage T4. Nature 1970;227:680-685. 33 Dunaway GA, Kasten TP: Nature o f the rat brain 6-phosphofructo-1-kinase isozymes. J Biol Chem 1985;260:4180-4185. 34 Wray W, Boulikas T, Wray VP, Hancock R: Silver staining o f proteins in polyacrylamide gels. Anal Biochem 1981;118:197-203. 35 Akkerman JWN, Nieuwenhuis HK, Mommersteeg ME, Gorter G, Sixma JJ: ATP-ADP compartmentation in storage pool deficient platelets: Correlation between granule-bound ADP and the bleeding time. Br J Haematol 1983;55:135-143. 36 Oude Weernink PA, Rijksen G, Staal GEJ: Pro duction o f a specific antibody against pyruvate kinase type M2 using a synthetic peptide. FEBS Lett 1988;236:391-395. 37 Van Veelen CWM, Staal GEJ: Pyruvate kinase and human brain tumors; in Staal GEJ, van Veelen CWM (eds): Markers o f Human Neuroec todermal Tumors. Boca Raton, CRC Press, 1986, pp 63-83. 38 Bennett MJ, Timperley WR, Taylor CB, Hills AS: Isozymes o f hexokinase in the developing, normal and neoplastic human brain. Eur J Cancer 1978; 14:189-193. 39 Rijksen G, Staal GEJ: Hexokinase in health and disease; in Beitner R (ed): Regulation o f Carbohy drate Metabolism. Boca Raton, CRC Press, 1985, pp 87-103. 40 Staal GEJ, Schipper-Kester G, Willemen JN, Rijksen G, van Veelen CWM: Isozyme distribu tion of aldolase, enolase, and pyruvate kinase in metastatic human brain tumors; in Staal GEJ, van Veelen CWM (eds): Markers o f Human Neuroec todermal Tumors. Boca Raton, CRC Press, 1986, pp 173-180. 41 Beemer FA, Vlug AMC, Rousseau-Merck MF, van Veelen CWM, Rijksen G, Staal GEJ: Glyco lytic enzymes from human neuroectodermal tu mors of childhood. Eur J Cancer Clin Oncol 1984; 20:253-259. 42 Schapira F: Resurgence o f fetal isozymes in can cer: Study o f aldolase, pyruvate kinase, lactic de-



43

44

45

46

47 48

hydrogcnasc and hexosaminidase; in Rattazzi MC, Scandalios JG, Whitt GS (eds): Isozymes, Current Topics in Biological and Medical Re search. New York, Liss, 1981, vol 5, pp 27-75. Li SSL: Cancer-associated LDH is a tyrosylphosphorylatcd form o f human LDH-A, skeletal mus cle isoenzyme. Cancer Invest 1988;6:93-101. Schwartz JP, Lust WD, Lauderdale VR, Passonneau JV: Glycolytic metabolism in cultured cells o f the nervous system. II. Regulation o f pyruvate and lactate metabolism in the C-6 glioma cell line. Mol Cell Biochem 1975;9:67-72. Kovacevic Z, McGivan JD: Mitochondrial metab olism o f glutamine and glutamate and its physio logical significance. Physiol Rev 1983;63:547— 605. Lanks KW: End products o f glucose and gluta mine metabolism by L929 cells. J Biol Chem 1987;262:10093-10097. Newsholme EA, Start C: Regulation in Metabo lism. London, Wiley & Sons, 1973, pp 88-145. Wilson JE: Brain hexokinase: The prototype ambiquitous enzyme; in Horecker BL, Stadtman ER (eds): Current Topics in Cellular Regulation. New York, Academic Press, 1980, vol 16, pp 1-54.

Oude Wecrnink/Rijksen/Siaal

49 Staal GEJ, Kalff A, Heesbeen EC, van Veelen CWM, Rijksen G: Subunit composition, regula tory properties, and phosphorylation o f phosphofructokinase from human gliomas. Cancer Res 1987;47:5047-5051. 50 Foe LG, Kemp RG: Isolation and characteriza tion of phosphofructokinase C from rabbit brain. J Biol Chem 1985;260:726-730.

Received: May 13, 1991 Accepted: August 19, 1991 Dr. P.A. Oude Weernink Department o f Hematology Laboratory o f Medical Enzymology University Hospital Utrecht PO Box 85500 NL-3508 GA Utrecht (The Netherlands)


352

Glycolytic metabolism in cultured cells of the nervous system. II. Regulation of pyruvate and lactate metabolism in the C-6 glioma cell line.

Radiosensitivity, recovery and dose-rate effect in three human glioma cell lines.

Tenascin expression in human glioma cell lines and normal tissues.

Raf kinase inhibitor protein (RKIP) inhibits the cell migration and invasion in human glioma cell lines in vitro.

Endothelial cell chemotactic factor derived from human glioma cell lines.

Glycolytic metabolism in cultured cells of the nervous system. III. The effects of thiamine deficiency and pyrithiamine on the C-6 glioma and C-1300 neuroblastoma cell lines.

Expression of P-glycoprotein in human glioma cell lines and surgical glioma specimens.

Human erythrocyte pyruvate kinase.

Simultaneous phosphorylation of three human calpactins by kinase C.

Phosphorylation and inhibition of human erythrocyte pyruvate kinase by erythrocyte membranes.

Regulation of pyruvate metabolism and human disease.

Iododeoxyuridine incorporation and radiosensitization in three human tumor cell lines.

Response of malignant glioma cell lines to activation and inhibition of protein kinase C-mediated pathways.

A monomeric form of pyruvate kinase in human pyruvate kinase deficiency.

Normalisation of red blood cell pyruvate kinase in pyruvate kinase deficiency by riboflavin treatment.

Pyruvate Dehydrogenase Kinase-mediated Glycolytic Metabolic Shift in the Dorsal Root Ganglion Drives Painful Diabetic Neuropathy.

Differences in arachidonic acid metabolism by human myelomonocytic cell lines.

Acquired pyruvate kinase deficiency. The effect of maleic acid upon human erythrocyte pyruvate kinase.

Interspecies comparison of c-myc gene in human and rat glioma cell lines.

Amplification and expression of a multidrug resistance gene in human glioma cell lines.

Glycolytic metabolism in cultured cells of the nervous system. I. Glucose transport and metabolism in the C-6 glioma cell line.

Glycolytic Metabolism Plays a Functional Role in Regulating Human Pluripotent Stem Cell State.

Effect of human T lymphotropic retrovirus-I exposure on cultured human glioma cell lines.

Prediction of radiosensitivity using phosphorylation of histone H2AX and apoptosis in human tumor cell lines.