842

A Metabolite of an Antineoplastic Ether Phospholipid May Inhibit Transmembrane Signalling Via Protein Kinase C 1 Wim J. v a n Blitterswijk*.a, Rob L. v a n der Benda, i'Jsbrand M. Kramerb, Arthur J. V e r h o e v e n b , H e n k Hilkmanna a n d John d e Widta aDivision of Cell Biology, The NetherlandsCancer Institute(Antoni van Leeuwenhoek-Huis), Amsterdam, The Netherlands, and bCentral Laboratory of the Netherlands Red Cross Blood TransfusionService, Amsterdam, The Netherlands In our search for the mechanisms by which the drug 1-O~ alkyl-2-O-methylglycero-3-phosphocholine (AMG-PC) inhibits tumor growth and metastasis, we have detected a metabolite, 1-O-alkyl-2-O-methylglycerol (AMG), in membranes of MO4 mouse fibrosarcoma cells grown in the presence of the drug. Synthetic AMG inhibited the activation of highly purified human protein kinase C by diacylglycerol in the presence of phosphatidylserine. Furthermore, AMG also inhibited the receptor-specific binding of 3H-phorbol-12,13-dibutyrate to human HL-60 promyeloid leukemia cells in a dose~dependent fashion. AMCr PC was not effective or much less so in these assays. We suggest that interaction of the metabolite AMG with protein kinase C may inhibit stimulus-response coupling in tumor cells and may thus potentially contribute to the mechanism by which AMG-PC exerts its anticancer activities. Lipids 22, 842-846 (1987).

The drug 1-O-alkyl-2-O-methylglycero-3-phosphocholine (AMG-PC) has been shown to specifically inhibit the growth of tumor cells in vitro and in vivo, to inhibit tumor cell invasion and metastasis and to enhance the tumoricidal capacity of macrophages (reviewed in ref. 1). The mechanisms involved in these actions are unknown. Besides the possibility that accumulation of the drug as such in cellular membranes (2) may alter their structure and functioning, possible metabolic products of the drug should also be considered. Since AMG-PC is not a substrate for the alkyl cleavage enzyme (3), possible metabolites may only be formed via enzymatic cleavage of the phosphate bond. In the present report, we show that the metabolite 1-Oalkyl-2-O-methylglycerol (AMG) is detectable in membranes of tumor cells grown in the presence of AMG-PC. Given the structural analogy of this metabolite to 1,2-diacylglycerol, a natural activator of protein kinase C (PKC) (4), we have investigated the possibility Chat AMG interacts with this key enzyme in cell signal transduction. First, we demonstrate that AMG inhibits the diacylglycerol-stimulated activity of a highly purified preparation of human PKC. Second, we show that AMG 'Presented at the symposium on "Ether Lipids in Oncology," Gi)ttingen, Federal Republic of Germany, December 1986. *To whom correspondence should be addressed at the Division of Cell Biology, The Netherlands Cancer Institute (Antoni van Leeuwenhoek-Huis), 121 Plesmanlaan, 1066 CX Amsterdam, The Netherlands. Abbreviations: AMG-PC, 1-O-alkyl-2-O-methylglycero-3-phosphocholine; AMG, 1-O-alkyl-2-O-methylglycerol;PKC, protein kinase C; diCs, 1,2-dioctanoyl-sn-glycerol;PS, phosphatidylserine;PMSF, phenylmethylsulfonylfluoride; ATP, adenosine 5'-triphosphate; PDBu, 3H-phorbol-12,13-dibutyrate;EGTA, ethylene-glycol-bis(betaaminoethylether)-N,N,N',N'-tetraacetic acid; TLC, thin layer chromatography. LIPIDS,Vol. 22, No. 11 (1987)

inhibits the binding of a biologically active phorbol ester to its receptor (PKC) on human HL-60 cells. MATERIALS AND METHODS

Chemicals. The AMG-PC analogs racemic 1-O-octadecyland 1-O-hexadecyl-2-O-methylglycero-3-phosphocholine as well as 1-O-hexadecyl-2-O-methyl-rac-glycerol (AMG) were purchased from Bachem AG (Bubendorf, Switzerland). 1,2-Dioctanoyl-sn-glycerol (diCJ was obtained from Avanti Polar Lipids (Birmingham, AL). Phosphatidylserine (PS; from bovine brain), 1,2-dioleoyl-rac-glycerol (diolein), leupeptin, phenylmethylsulphonyl fluoride (PMSF), soybean trypsin inhibitor, aprotinin, histone IIIS {from calf thymus) and adenosine 5'-triphosphate (ATP) were obtained from Sigma Chemical (St. Louis, MO). 32P-ATP (3 Ci/~mol) and 3H-phorbol-12,13-dibutyrate (PDBu, 12.5 Ci/mmol) were from New England Nuclear (Boston, MA), and cholesterol and ethylene-glycolbis(beta-aminoethylether)-N,N,N',N'-tetraacetic acid (EGTA) from BDH (Poole, U.K.). All chemicals were of analytical grade. Cells. MO4 cells are virally transformed C~H mouse fibroblastic cells that are invasive in vitro (5) and produce invasive and metastasizing fibrosarcomas in syngeneic mice (6). They were cultured in minimum Eagle's medium (Rega I; Gibco Europe, Paisley, U.K.) supplemented with 10% fetal calf serum and 0.05% glutamine. Cells were also grown in the presence of AMG-PC (octadecyl-type, 7 ~g/ml culture medium), added 48 hr prior to membrane lipid analysis. The concentration of the drug used permitted growth to about 80% of that of control cells (5). Human HL-60 promyeloid leukemia cells (7) were maintained as a suspension culture in RPMI 1640 medium, supplemented with 20% fetal calf serum. AMG detection in membranes. A crude membrane fraction from MO4 cells was prepared by centrifugation of the low-speed supernatant of a cell homogenate at 105,000 • g for 1.5 hr, as described before (8). Lipids were extracted with chloroform/methanol (2:1, v/v) followed by partition according to Folch et al. (9) and were separated by thin layer chromatography (TLC) on precoated silica gel plates (Merck, Darmstadt, FRG) using hexane/ ether/acetic acid (60:40:1.6, v/v). The spot corresponding to AMG was scratched off, extracted and trimethylsilylated with a mixture of pyridine-hexamethyldisilazane-trimethylchlorosilane ~12:5:2, v/v/v; obtained from Pierce Chemical Co., Rockford, IL) for 2 hr at room temperature and analyzed by gas liquid chromatography according to Myher and Kuksis (10), using a 25 m CP-Sil 58 column. AMG was also enzymatically prepared from AMG-PC (octadecyl-type) using phospholipase C (20 U/ ml) from Clostridium perfringens (obtained from Sigma) in 10 mM CaCl~, 0.1 mM ZnCl~, 20 mM Tris-HC1 (pH 7.4) at 37 C for 15 hr. The AMG formed (in low yield) was separated from AMG-PC by TLC and trimethylsilylated

843

1-ALKYL-2-METHOXYGLYCEROL INHIBITS PROTEIN KINASE C as described above. The AMG thus prepared and that isolated from MO4 cell membranes were identical, as unambiguously shown in Figure 1. AMG was quantitated by the detector response, using a known amount of heptadecanoic acid methyl ester as a reference. Phorbol ester receptor assay. Binding of 3H-PDBu to HL-60 cells in the presence or absence of inhibitors, supplied in dimethyl sulfoxide (0.5 % final content), was performed by the method of Goodwin and Weinberg (11) in polypropylene tubes, using 5-10 • l0 s cells and 20 nM 3H-PDBu per assay (0.2 ml), unless otherwise stated. AMG and 3H-PDBu were added simultaneously. Following incubation at room temperature for 30 min, excess cold assay buffer was added, and free and receptor-bound ~H-PDBu were immediately separated by filtration through glass-fiber filters (Whatman GS/A). The filters were washed rapidly with excess buffer and prepared for liquid scintillation counting. The results are presented as specific binding, that is, the difference between 3H-PDBu bound in the presence and absence of 20 ~M unlabeled PDBu. Purification of protein kinase C. Human PKC was isolated from a platelet/lymphocyte fraction of 50 buffy coats of donor blood. The purification procedure follows essentially the methods of Kikkawa et al. (12) and Girard et aL (13) and will be described in detail elsewhere. In brief, buffy coats were centrifuged (1300 • g, 8 re_in, 20 C) to remove granulocytes and erythrocytes. The rest of the procedure was performed at 0-4 C. Platelets and lymphocytes were spun down at 1700 • g (45 rain) and taken up in 5 vol homogenization buffer containing 0.25 M sucrose, 10 mM EGTA, 2 mM EDTA and protease inhibitors (leupeptin, trypsin inhibitor, PMSF and aprotinin). Cells were broken in a Omni homogenizer (Sorvall) and the 100,000 X g supernatant of the cell homogenate underwent three chromatographic steps: DEAE cellulose, Phenyl sepharose CL-4B and affinity column chromatography using phosphatidylserine and cholesterol in the column matrix (13). PKC was eluted from the affinity column with 10 ml of 5 mM Tris-HC1 (pH 7.5), 0.2 M NaCl, 5 mM 2-mercaptoethanol, 2 mM EGTA in a yield of about 50 ~g protein. The PKC preparation thus obtained was 7,500-fold purified relative to the cytosolic fraction and showed one single band of 81 kD (see Fig. 2) upon SDSpolyacrylamide gel electrophoresis (14). The enzyme was stored in 40% sucrose at - 7 0 C until use. Enzyme assays. The activity of PKC in the presence or absence of stimulatory/inhibitory lipids was determined essentially according to Kikkawa et al. (15). The assays were performed in a 2 mM Ca2*-EGTA buffer (16) yielding 10 ~M free Ca 2*or in 2 mM EGTA. The reaction mixtures (60 ~ final volume) furthermore contained 10 mM Mg '§ 20 mM Tris-HC1 (pH 7.5), 200 ~g/ml histone IIIS, 20 ~M leupeptin, 10 ~M 32P-ATP (0.5-1 ~Ci per assay) and PKC (10 ~l preparation). Lipids were present in amounts indicated in the footnote of Table 1 and the legend of Figure 2. They were added as liposomes (10-~l samples), prepared by sonication for 1 min in a Branson Sonifier equipped with a standard probe (50 W). Incubations were started by addition of ATP and performed for 5 min at 30 C. The reaction was stopped by precipitation on Whatman paper filters (3 Chr) in cold 10% trichloroacetic acid plus 10 mM K~HPO~. The filters were extensively washed and prepared for liquid scintillation counting.

RESULTS A N D DISCUSSION

In a previous study (5), it was demonstrated that the anticancer drug AMG-PC inhibits the invasion of MO4 fibrosarcoma cells into normal embryonic chick heart tissue, under conditions where growth of the tumor cells in tissue culture was only minimally affected. In the companion article (2), we have shown that under these conditions AMG-PC is accumulated in the tumor cell membranes, predominantly in the plasma membrane: MO4 cells grown in the presence of 5-10 t~g/ml AMG-PC for 48 hr incorporated the drug to about 10% of the total phospholipids (measured in purified plasma membranes). Figure I shows

B

C

D

E

!

1() retention time (min)

12

FIG. 1. Detection of 1-O~ctadecyl-2-(~methylglycerol (AMGI in membranes from MO4 cells grown in the presence of 1-O-octadecyl-2-Omethyiglycero-3-phosphocholine (AMG-PC). Gas liquid chromatography of trimethylsilyl derivatives. The first peak, at a retention of 8.7 min, was not identified. (A) Control cells; (Bt cells grown with AMG-PC; (C) A M G prepared from AMG-PC (octadecyl-typel using phospholipase C from Cl. perfringens; (D) C added to A; IE) C added to B.

LIPIDS, Vol. 22, No. 11 (t987)

844

W.J. VAN BLITTERSWIJK ET AL. t h a t the incorporated drug was partly metabolized to AMG, presumably by a phospholipase C-like enzyme t h a t is able to split off or exchange the phosphocholine moiety. The gas chromatogram (Fig. 1B), pertaining to cells grown with AMG-PC, shows two main peaks {retention times 8.7 and 9.3 rain), the first of which is an unidentified compound t h a t was also found in the control cells (Fig. 1A), while the second peak was identified as AMG. The amount of AMG detected in the membrane preparation corresponded roughly to 1 ~g (or 3 nmol) per l0 s cells. It should be noted t h a t this is a steady-state value. Nothing is known about any possible further metabolism of AMG, such as conversion to a phosphatidic acid analogue or degradation by an alkyl cleavage enzyme (3). Many cellular activities, including proliferation, differentiation and secretion in endocrine and exocrine systems, depend on transmembrane stimulus-response coupling that may be mediated by PKC (4). Although not yet established, it is not unlikely t h a t this key enzyme in signal transduction is also involved in the mechanism of {tumor) cell adhesion and invasion, possibly through specific phosphorylation of cytoskeleton-associated proteins (4). PKC is t h o u g h t to be physiologically activated by diacylglycerol derived from r e c e p t o r - m e d i a t e d phosphatidylinositol hydrolysis and is the receptor for biologically active phorbol esters, such as P D B u (4). In the following experiments, we have studied the interaction of AMG-PC and AMG with PKC, in both an enzymatic assay and a PDBu-receptor assay. We have purified human PKC to apparent homogeneit y {Fig. 2, right) and used this preparation to s t u d y the lipid dependency of the enzyme in the presence of 10 ~M Ca 2§ and in submicromolar Ca 2§ {excess EGTA) {Table 1). In agreement with literature data (4,15,17), PS was found to be necessary to achieve substantial activation. At submicromolar free Ca 2§ like under physiological conditions in the resting cell, only diacylglycerol {diolein) is able to activate PKC in the presence of PS {six- to eightfold stimulation, relative to the basal level with PS only). Diolein alone had no effect (not shown here). At 10 ~M

Ca 2", the PS-dependent basal activity is much higher, and diolein is not able to increase this activity any further {exp. 2) or more than 30% (exp. 1). Neither AMG-PC nor AMG activate PKC by themselves or in the presence of PS. Rather, they inhibit the basal activities in the presence of PS some 10-20%. The largest inhibitory effect of AMG was found specifically on the diacylglycerol stimulation of the enzyme, precisely the activity that is physiologically relevant and t h a t is generated in the cell upon receptor stimulation. This inhibitory effect of AMG, shown in Figure 2, was dose-dependent and was present in excess E G T A as well as in 10 ~M Ca 2§ (the latter being relevant only in experiment 1, where diolein was indeed stimulating the PS-dependent activity of the enzyme; see Table 1). We have no clear explanation for the variability generally found in the degree of diolein stimulation of the PS-dependent enzyme activity among individual experiments. In contrast to AMG, AMG-PC does not have a significant effect on the diolein-stimulated PKC activity (Fig. 2}. As noted above, the drug only inhibits the basal PS-dependent activities somewhat, confirming to a certain extent results of Helfman et al. (18). These authors, however, used only a very crude PKC preparation, unphysiologically high Ca 2. concentrations and no diacylglycerol.

TABLE 1 Lipids Dependency of the Activity of Purified Human Protein Kinase C (PKC) toward Histone as a Substrate

PKC activity (pmol 3~p incorporated per min per 10 ~l PKC preparation)a 2 mM EGTA

10 pM Ca~*

Lipids in assayb

Exp. 1

Exp. 2

Exp. 1

Exp. 2

PS PS + diolein PS + AMG PS + AMG-PC AMG AMG-PC No lipid

0.4 3.1

0.5 2.9 0.4 0.4 0.4 0.4 0.3

5.0 6.4

6.2 6.4 5.4 5.0 0.3 0.4 0.2

0.3

0.3

PS, phosphatidylserine; AMG, 1-O-alkyl-2-O-methylglycerol;AMGPC, 1-O-alkyl-2-O-methylglycero-3-phosphocholine. aData are means of triplicates, with SD < 12%. bps, diolein, AMG and AMG-PC: 100, 5, 13.3 and 20 ~g/ml assay mixture, respectively. LIPIDS, Vol. 22, No. 11 (1987)

FIG. 2. Effect of 1-O-hexadecyl-2-O~methyl-rac-glycero-3-phosph~ choline (AMG-PCI and 1-GLhexadecyl-2-(~methyl-rac-glycerol (AMG) on the diacylglycerol-stimulated, phosphatidylserine IPS)-dependent activity of purified human protein kinase C (PKCI (basal PSdependent activity, given in Table 1, subtracted~ in the presence of 2 mM E G T A or 10 ~M Ca 2§ as indicated. Solid columns, controls I100%), containing only PS and diolein, 100 and 5 ~g/ml assay mixture, respectively; hatched columns, AMG-PC added in a ratio Iw/w) AMG-PC/diolein --- 2; open columns, A M G added in the ratios (w/w) indicated. Data are means of triplicate values, with SD as indicated. Experiments I and 2 correspond to those indicated in Table 1. Right panel: polyacrylamide gel showing the purity of the PKC preparation Ifrom human lymphocytes/platelets} used in the present experiments. The silver-stained gel shows one band at 81 kD. Positions of authentic marker proteins (M) are indicated.

845 1-ALKYL-2-METHOXYGLYCEROL INHIBITS PROTEIN KINASE C

! ~

1oo

200

100

60

50

20

1'2

18

2'4

0

50

total PDBu ,nM

1(~0

150

2()0

inhibitor (iuM)

FIG. 3. Binding of 3H-phorbol-12,13-dibutyrate (3H-PDBu) to intact HL-60 cells in the presence of various concentrations of 1-(g hexadecyl-2-O~methyl-rac-glycerol (AMG) as a function of the total 3H-PDBu concentration in the medium. 9 Control cells, without AMG; e , 25 pM; A, 100 ~M; A, 200 ~M AMG. A representative experiment is shown where data points are means of triplicates (SD

FIG. 4. Inhibition of 3H-phorbol-12,13-dibutyrate (3H-PDBu) binding to intact HL-60 cells by various concentrations of 1-(ghexadecyl-2-Omethyl.rac-glycero-3-phosphocholine (AMG-PC) ( e ) , 1-Ohexadecyl-2-(gmethyl-rac-glycerol (AMG} (&), 1,2-dioctanoyl-snglycerol (diCs) (A), in the presence of 20 nM 3H-PDBu. Data points are means of triplicates, with SD as indicated.

A s u b s t a n t i a l inhibition of the diacylglycerolstimulated PKC activity, as we have found with AMG, should also be demonstrable in more complicated biological systems, such as PKC-mediated differentiation of HL-60 cells. As a first approach, we have studied in this cell system the effect of AMG on the specific binding of a biologically active phorbol ester, 3H-PDBu, to its receptor (PKC). As shown in Figure 3, this binding is indeed inhibited by AMG in a dose-dependent fashion. At 20 nM PDBu, about 70% inhibition was reached with 200 ~M AMG. Under these conditions, we also determined the inhibitory capacity of AMG-PC and diCs, a membranepermeable synthetic diacylglycerol, known to be a potent activator of PKC (19,20). Figure 4 shows t h a t 3H-PDBu binding is only minimally inhibited by AMG-PC, whereas diC8 inhibits this binding twice as much as does AMG. However, if one takes into account t h a t generally only the 1,2-sn-enantiomers of diacylglycerol (analogs) are biologically active (17,20) and t h a t AMG is racemic, the inhibitory capacities of AMG and diCs may be the same. The concentration in the medium at which these compounds inhibit P D B u binding may seem high, b u t is, of course, dependent on their e x t e n t of incorporation (partitioning) into the apolar region of the cell membrane. This degree of uptake in the membrane is as yet unknown, but may likely be very low. As noted, we have found a steadystate content of about 1 ~g (3 nmol) AMG per 108 M O , cells cultured in the presence of a dose of AMG-PC t h a t only minimally inhibited cell growth. This cellular content of AMG is of the same order of magnitude as the amount of diacylglycerol generated in other cell types upon receptor stimulation (21). This notion together with the data on AMG inhibition of the diacylglycerol-

stimulated PKC activity (Fig. 2) could indicate t h a t the formation of AMG from AMG-PC in the cell membrane is physiologically relevant. In conclusion, the anticancer drug AMG-PC accumulating in tumor cell membranes has a relatively small inhibitory effect on the enzymatic and phorbol ester receptor activities of PKC. However, its metabolite, AMG, detectable in the membranes of tumor cells grown in the presence of the drug, has much larger effects. It inhibits dose-dependently both the binding of P D B u to its receptor and the diacylglycerol-stimulated activity of PKC. The latter effect of AMG on this key enzyme in transmembrane signalling may potentially contribute to the mechanism by which AMG-PC exerts its anticancer activities. Whether this is indeed of physiological relevance remains to be further investigated.

< 10%).

ACKNOWLEDGMENTS

We thank Dr. Hidde L. Ploegh for stimulating discussions and for reading this manuscript. Liesbeth Pastoors and Gea Meijerink are thanked for technical and secretarial assistance, respectively.

REFERENCES

1. Berdel, W.E., Andreesen, R., and Munder, P.G. (1985) in Phospholipids and CeUularRegulation (Kuo, J.F., ed.) Vol. II, pp. 41-73, CRC Press, Boca Raton, FL. 2. Van Blitterswijk, W.J., Hilkmann, H., and Storme, G.A. (1987) Lipids 22, 820-823. 3. Hoffman, D.R., Hoffman, L.H., and Snyder, F. (1986) Cancer Res. 46, 5803-5809. 4. Kikkawa, U., and Nishizuka, Y. (1986) Ann. Rev. Cell Biol. 2, 149-178.

LIPIDS,Vol. 22, No. 11(1987)

846

W.J. VAN BLITTERSWIJK ET AL. 5. Storme, G.A., Berdel, W.E., Van Blitterswijk, W.J., Bruyneel, E.A., de Bruyne, G.K., and Mareel, M.M. (1985} Cancer Res. 45, 351-357. 6. Meyvisch, C., and Mareel, M. (1982} Invasion Metastasis 2, 51-60. 7. Collins, S.J., GaUo, R.C., and Gallagher, R.E. {1977)Nature 270, 347-349. 8. Van Blitterswijk, W.J., Emmelot, P., Hilkmann, H.A.M., Hilgers, J., and Feltkamp, C.A. (1979} Int. J. Cancer23, 62-70. 9. Folch, J., Lees, M., and Sloane-Stanley, G.H. {1957} J. Biol. Chem. 226, 497-509. 10. Myher, J.J., and Kuksis, A. {1982} Can. J. Biochem. 60, 638-650. 11. Goodwin, B.J., and Weinberg, J.B. {1982} J. Clin. Invest. 70, 699-706. 12. Kikkawa, U., Takai, Y., Minakuchi, R., Inohara, S., and Nishizuka, Y. {1982)J. BioL Chem. 257, 13341-13348. 13. Girard, P.R., Mazzei, G.J., Wood, J.G., and Kuo, J.F. (1985}Proc. Natl. Aca& Sci. USA 82, 3030-3034.

LIPIDS, Vol. 22, No, 11 (1987)

14. Laemmli, U.K. {1970} Nature 227, 680-685. 15. Kikkawa, U., Minakuchi, R., Takai, Y., and Nishizuka, Y., {1983} Methods Enzymol. 99, 288-298. 16. Fabiato, A., and Fabiato, F. (1979}J. PhysioL, Paris, 75, 463-505. 17. Boni, L.T., and Rando, R.R. {1985} J. BioL Chem. 260, 10819-10825. 18. Helfman, D.M., Barnes, K.C., Kinkade Jr., J.M., Vogler, W.R., Shoji, M., and Kuo, J.F. {1983} Cancer Res. 43, 2955-2961. 19. Ebeling, J.G., Vandenbark, G.R., Kuhn, L.J., Ganong, B.R., Bell, R.M., and Niedel, J.E. {1985} Proc. Natl. Acad. Sci. USA 82, 815-819. 20. Nomura, H., Ase, K., Sekiguchi, K., Kikkawa, U., and Nishizuka, Y. {1986} Biochem. Biophys. Res. Commun. 140, 1143-1151. 21. Preiss, J., Loomis, C.R., Bishop, W.R., Stein, R., Niedel, J.E., and Bell, R.M. {1986} J. Biol. Chem. 261, 8597-8600. [ R e c e i v e d J a n u a r y 14, 1987]