Synthesis of enantiomerically pure glyceryl esters and ethers. I. Methods employing the precursor 1,2-isopropylidene-sn-glycerol

Structural variation of cationic lipids: minimum requirement for improved oligonucleotide delivery into cells

In vivo transfection efficiency (TE) using cationic liposome/oligonucleotide (ODN) complexes is often hampered by interactions with serum components. Novel cationic lipids with different hydroxyethyl or dihydroxypropyl ammonium backbones, esterified hydrocarbon chains and hydroxy substituents have been synthesized and applied in cationic liposome formulations with and without the helper lipid DOPE (1:1, m/m). Their properties for cellular ODN delivery were determined using fluorescently labeled ODNs (F-ODNs). Cationic lipids with hydrocarbon chains esterified to non-glycerol backbones in non-vicinal configuration were completely ineffective in nuclear ODN-delivery. Instead, an increased cytoplasmic localization of F-ODNs was observed. Cationic lipids equipped with only one hydrocarbon were completely incompetent for cellular ODN delivery. In the absence of serum, all cationic lipids tested with hydrocarbon chains in vicinal configuration esterified to a glycerol backbone (the respective N-(1,2-diacyl-dihydroxypropyl)-N,N,N-trimethyl-ammoniumchlorides or N-(1,2-diacyl-dihydroxypropyl)-N(hydroxyethyl)-N,N-dimethyl-ammoniumchlorides as well as N-(1,2-diacyl-dihydroxypropyl)-N(1,2-dihydroxypropyl)-N,N-dimethyl-ammoniumchlorides with lauroyl, myristoyl, palmitoyl, stearoyl and erucoyl chains) were able to transfect cells when combined with DOPE (20-80% nuclear fluorescence). Remarkably, only the analog esterified with two myristoyl chains was equally effective even in the absence of DOPE. By adding hydroxy groups to the N-alkyl residue, TE under serum conditions was improved yielding transfection rates of 55%, 75% and 90% for 0, 1 or 2 substituted hydroxy groups, respectively. For plasmid DNA, different requirements were identified. Again, the analog with two myristoyl chains was most effective but only in the presence of DOPE. However, the addition of hydroxy groups had no influence on the TE in the presence of serum.

Glyceryl-ether monooxygenase [EC 1.14.16.5]. A microsomal enzyme of ether lipid metabolism

The history, biological, and medical aspects of glyceryl ethers, as well as their chemical syntheses, biosynthesis, and their chemical and physical properties are briefly reviewed as background information for appreciating the importance of the enzyme glyceryl-ether monooxygenase, and for embarking on new studies of this enzyme. The occurrence, isolation and general properties of the microsomal, membrane-bound, glyceryl-ether monooxygenase from rat liver are described. Radiometric, nonradiometric, and coupled and direct spectrophotometric assays for this enzyme are detailed. The effects of detergents on the kinetics of this enzyme are described together with the stoichiometry and the effects of inhibitors. The structure-activity relationships of pterin cofactors and of ether lipid substrates, including their stereospecificities, have been summarized from enzyme kinetic data which are also tabulated. The mechanism of enzymic hydroxylation of glyceryl ethers and a model for the active site of glyceryl-ether monooxygenase are proposed from these apparent kinetic data. Notes on useful future studies of this monooxygenase have been made.

On the mechanism of lysophospholipase activity of secretory phospholipase A2 (EC 3.1.1.4): deacylation of monoacylphosphoglycerides by intrinsic sn-1 specificity and pH-dependent acyl migration in combination with sn-2 specificity

We show for the first time that secreted low-molecular weight phospholipase A2 (EC 3.1.1.4) catalyzes the deacylation of monoacylphosphoglycerides directly from the sn-1 position, although at a very low rate: purified phospholipase A2 enzymes from bee venom, crotalus atrox venom, and porcine pancreas hydrolyze the sn-1 ester bond in 1-palmitoyl-2-O-methyl-sn-glycero-3-phosphorylcholine. Hydrolytic rates with the corresponding isomer, 1-O-methyl-2-palmitoyl-sn-glycero-3-phosphorylcholine, are about 3-4 orders of magnitude higher. The similarities in Ca2+ requirement and inactivation profiles suggest that deacylation, albeit with different rates, from both sn-1 and sn-2 positions is catalyzed by the same catalytic site of phospholipase A2. Furthermore, evidence is provided that phospholipase A2-catalyzed 1-acyl lysophospholipid deacylation is mediated by sn-1-directed action, but above pH 7 acyl migration with subsequent enzyme-catalyzed hydrolytic cleavage from the sn-2 position contributes to the overall deacylation of monoacylphosphoglycerides, acyl migration becoming eventually the rate-limiting factor.

Induction of resistance to hexadecylphosphocholine in the highly sensitive human epidermoid tumour cell line KB

Hexadecylphosphocholine (HePC, Miltefosine) is a representative of the group of alkyl-lysophosphocholines showing remarkable antitumoral activity in in vitro experiments and in experimental animal tumour models. The epidermoid tumour cell line KB, which is highly sensitive to HePC (half-maximal growth inhibiting concentration, IC50: 1.2 microM; half lethal concentration, LC50: 2.8 microM), was slowly adapted to increasing concentrations of HePC. After 14 months, the adaptation process was stopped at a concentration of 10 micrograms/ml (23.5 microM). At this point, the KB cells tolerated high doses of HePC (IC50: 41.2 microM; LC50: 87.1 microM). The resistant cells (KBr) also showed crossresistance to the other well studied ether-lysophospholipids, Edelfosine (1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine, OMG-3PC; ET18OCH3) and Ilmofosine (1 S-hexadecyl-2-methoxymethyl-rac-(1-thio-3-hydroxy)propyl-3-phosphocho lin e, BM 41.440). Comparison of the KB and KBr cells showed that total lipid phosphate, ether-lipid content, vinyl-ether-lipid content, protein content as well as cholesterol content were unchanged. Furthermore, no changes were observed in the lipid composition between KB and KBr cells. Uptake of choline was also unchanged in both cells, but the uptake of D-myo-inositol was lower by a factor of two in the KBr cells. However, in KB cells, the addition of HePC induced a 50% reduction of D-myo-inositol-uptake, whereas in KBr cells inositol uptake was unchanged. Differences in HePC uptake and HePC metabolism were apparent between the KB and KBr cell lines. KBr cells showed a 3-fold lower uptake for HePC and a 3- to 4-fold faster metabolism of HePC than KB cells. However, the amount of non-metabolised HePC after 2 days of incubation with 1 microgram/ml HePC (LC50: 1.2 microgram/ml) in KB cells was 3- to 4-fold lower than the amount of HePC in KBR cells at 10 micrograms/ml (LC50: 37 micrograms/ml), indicating that KBr cells can incorporate higher amounts of HePC than KB cells without adverse effects for cell growth and viability. This seems to indicate that mechanisms other than slower uptake and faster metabolism are involved in the induction of resistance to HePC in KBr cells.

Analogs of alkyllysophospholipids: chemistry, effects on the molecular level and their consequences for normal and malignant cells

In the search for new approaches to cancer therapy, the first alkyllysophospholipid (ALP) analogs were designed and studied about two decades ago, either as potential immunomodulators or as antimetabolites of phospholipid metabolism. In the meantime, it has been demonstrated that they really act in this way. However, their special importance is based on the fact that, in addition, they interfere with key events of signal transduction, such as hormone (or cytokine)-receptor binding or processing, protein kinase C or phospholipase C function and phosphatidylinositol and calcium metabolism. There are no strict structural requirements for their activity. Differences in the cellular uptake or the state of cellular differentiation seem to be mainly responsible for higher or lower sensitivities of cells towards ALP analogs. Consequences of the molecular effects mentioned on the cellular level are cytostasis, induction of differentiation (while in contrast the effects of known inducers of differentiation such as 12-O-tetradecanoylphorbol-13-acetate are inhibited, probably as a consequence of protein kinase C inhibition) and loss of invasive properties. Already in sublytic concentrations, alterations in the membrane structure were observed, and lysis may begin at concentrations not much higher than those causing the other effects described. Few ALP analogs have already entered clinical studies or are in clinical use. ALP analogs are the only antineoplastic agents that do not act directly on the formation and function of the cellular replication machinery. Therefore, their effects are independent of the proliferative state of the target cells. Because of their interference with cellular regulatory events, including those failing in cancer cells, ALP analogs, beyond their clinical importance, are interesting model compounds for the development of new, more selective drugs for cancer therapy.

Synthesis of 1-palmitoyl-2-hexadecyl-sn-glycero-3-phosphocholine (PHPC)

A general method for the chirospecific synthesis of 1-acyl-2-alkyl-sn-glycero-3-phosphocholines is described. 1-Palmitoyl-2-hexadecyl-sn-glycero-3-phosphocholine (PHPC) was synthesized in 18% overall yield in ten steps via five new synthetic intermediates, and 1-acetyl-2-hexadecyl-sn-glycero-3- phosphocholine (AHPC) was also synthesized. 1-Acyl-2-alkyl-sn-glycero-3-phosphocholines, which have not been found to exist in nature, are ether lipid analogs of 1,2-diacyl-sn-glycero-3-phosphocholines, which are important components of cell membranes. Biophysical studies of hydrated bilayers of PHPC will be of interest in probing the critical importance of the central region of these amphiphilic molecules to the molecular assemblies that are formed.

Debenzylation and detritylation by bromodimethylborane: synthesis of mono-acid or mixed-acid 1,2- or 2,3-diacyl-sn-glycerols

A novel method of deprotecting primary alcohols protected with either benzyl or trityl groups by using bromodimethylborane under mild reaction conditions (dichloromethane, -20 to 5 degrees C) has been applied to the synthesis of optically pure mono-acid or mixed-acid 1,2- or 2,3-diacyl-sn-glycerols. This method was particularly useful for the synthesis of long saturated acyl (C12 to C24) as well as unsaturated diacyl-sn-glycerols since little or no acyl migration occurred during deprotection. Diacylation of 3-benzyl-sn-glycerol or 1-benzyl-sn-glycerol followed by bromodimethylborane debenzylation gave mono-acid 1,2- or 2,3-diacyl-sn-glycerols, respectively. The mixed-acid 1,2- or 2,3-diacyl-sn-glycerols were prepared from 1-acyl-sn-glycerols or 3-acyl-sn-glycerols, respectively, by tritylation, acylation with a different fatty acid, followed by detritylation with bromodimethylborane.

Distribution of hexadecylphosphocholine and octadecyl-methyl-glycero-3-phosphocholine in rat tissues during steady-state treatment

The distribution of the alkylphosphocholine hexadecylphosphocholine (He-PC) and the (alkyl)lysophospholipid 1-0-octadecyl-2-0-methyl-rac-glycero-3-phosphocholine (ET18-OCH3) was analyzed in rats. The compounds were given orally at a daily dose of 75 mumol/kg body weight. After 6, 11, and 18 days, three rats in each treatment group were killed and the drug concentration in various tissues and fluids was determined. With the exception of the kidney (He-PC) and brain (He-PC and ET18-OCH3), steady-state levels of the drugs could be achieved in all organs investigated and in serum. Maximal concentrations of He-PC were found in the kidney, adrenal glands, and spleen, whereas the highest concentrations of ET18-OCH3 were detected in the adrenal glands, spleen, and small intestine. The concentrations of He-PC exceeded those of ET18-OCH3 in most tissues by a factor of about 2-25. Since samples of urine and feces did not contain detectable amounts of the compounds, the absorption of both lipid analogues was assumed to be complete. The total amount of He-PC recovered after 6, 11, and 18 days was 15%, 12%, and 6%, respectively, and that of ET18-OCH3 was 1.3%, 0.8%, and 0.3%, respectively. This indicates that the bioavailability of He-PC and ET18-OCH3 is not controlled by differences in the uptake of the two drugs, but by differences in their metabolism. The results could explain the differing efficacy of these two compounds in their antitumor action in animal models.

Comparison of selective cytotoxicity of alkyl lysophospholipids

Alkyl lysophospholipids have been shown to be cytooxic to a number of neoplastic tissues. One, ET-18-OCH3, has been used to selectively purge leukemic cells from mixtures with normal marrow progenitor cells, in vitro and in vivo. We have measured the 50% inhibitory (IC50) effect of a series of ether lipids (EL) on leukemic cells (HL60, K562, Daudi, KG-1, KG-1a) and normal marrow progenitor cells. Cells were incubated with varying concentrations of EL for 4 hr and assayed for viability, [3H]thymidine incorporation and clonogenicity in semi-solid media. The effect on protein kinase C (PKC) activity was assayed for each compound. Compounds tested included three glycerophosphocholine analogs--ET-18-OCH3, ET-16-NHCOCH3, and BM 41.440. In addition, a lipoidal amine, CP 46665, an ethyleneglycolphospholipid, AEPL, and four single chain alkylphosphocholine analogs, HePC2, HePC3, HePC4 and HePC6 were also tested. During the period of incubation, the cells remained viable (greater than 70%) as judged by trypan blue dye exclusion. The glycerophosphocholines were the most active and showed the highest therapeutic index. The lipoidal amine was active, but toxic to normal marrow progenitor cells. The ethyleneglycolphospholipid was active against HL60, but not against the other cell lines. The single chain alkylphosphocholine analogs were less active. All of the compounds inhibited PKC activity; however, the glycerophosphocholines were the most inhibitory.

Regio- and stereoselective enzymatic esterification of glycerol and its derivatives

A methodology for regio- and stereoselective preparation of acyl glycerol derivatives is presented. It offers easy access to specific 1,2-, 1,3-diglycerides and triglycerides as well as alkyl glycerol esters, phospholipids and glycolipids. These compounds are prepared by esterification of the corresponding glycerol derivatives such as 2-monoglycerides, alkyl glycerols, glyceryl glycosides, glyceryl phosphate esters, or unsubstituted glycerol. The regio- and stereoselectivity in the esterification is achieved by using fatty acid anhydrides and an enzymatic catalyst, 1,3-specific lipase. NMR methods for determining the regio- and stereoselectivity of esterification are discussed.

Active-site-directed specific competitive inhibitors of phospholipase A2: novel transition-state analogues

More than 100 amphiphilic phosphoesters, possible tetrahedral transition-state analogues capable of coordinating to the calcium ion at the active site of phospholipase A2, were designed, synthesized, and tested as inhibitors for the hydrolysis of 1,2-dimyristoyl-sn-glycero-3-phosphomethanol vesicles in the scooting mode. This assay system permits the study of structurally diverse inhibitors with phospholipase A2S from different sources, and it is not perturbed by factors that change the quality of the interface. As a prototype, 1-hexadecyl-3-trifluoroethylglycero-2-phosphomethanol (MJ33) was investigated in detail. Only the (S)-(+) analogue of MJ33 is inhibitory, and it is as effective as the sn-2 phosphonate or the sn-2 amide analogues of sn-3 phospholipids. The inhibitory potencies of the various phosphoesters depended strongly on the stereochemical and structural features, and the mole fractions of inhibitors required for 50% inhibition, X1(50), ranged from more than 1 to less than 0.001 mole fraction. The affinity of certain inhibitors for enzymes from different sources differed by more than 200-fold. The inhibitors protected the catalytic site residue His-48 from alkylation in the presence of calcium but not barium as expected if the formation of the EI complex is supported only by calcium. The equilibrium dissociation constant for the inhibitor bound to the enzyme at the interface was correlated with the XI(50) values, which were different if the inhibition was monitored in the pseudo-zero-order or the first-order region of the progress curve. These results show that the inhibitors described here interfered only with the catalytic turnover by phospholipase A2's bound to the interface, their binding to the enzyme occurred through calcium, and the inhibitors did not have any effect on the dissociation of the enzyme bound to the interface.

Structural requirements of lysophospholipid-regulated mitochondrial Ca2+ transport

Analogues of lysophosphatidylcholine, including PAF (platelet-activating-factor) and HePC (an experimental anticancer drug), were studied for their influence on mitochondrial Ca2+ transport and membrane potential. Lysophospholipids released Ca2+ from mitochondria and reduced the maximal Ca2+ uptake. The structure-activity relations indicate that deprotonated head groups like phosphocholines yield active compounds while partially protonated head groups like phosphoethanolamines are essentially inactive. Structural requirements for the apolar part of the molecules were acyl or alkyl chain lengths of less than 18 carbon atoms at the C1-position of the glycerol backbone and residues of small size and/or low polarity at the C2-position. Choline lysophospholipids, but not ethanolamine lysophospholipids, may therefore induce mitochondrial Ca2+ efflux and become mediators of ischaemic tissue damage where dysregulated phospholipase A2 activity and an impairment of mitochondrial function are supposed to play a crucial role.

Hexadecylphosphocholine: a new and selective antitumor drug

Interdisciplinary cooperation between basic and clinical research has resulted in the discovery and development of alkylphosphocholines, a new class of substances for the treatment of breast cancer. In contrast to most antitumor substances, the alkylphosphocholines do not attack the cell nucleus, but the cell membrane. This report presents a systematic study which, for the first time, provides a correlation between their chemical structure, antitumor efficacy and selectivity. Through an understanding of the metabolism of tumor growth inhibiting (ether)-lysolecithins, the minimal structural requirements for the antineoplastic efficacy of these substances have been obtained. This knowledge was used to identify molecular structures which are more effective and less toxic for the organism. The active principle derived from a study of (ether)-lysolecithins active as antitumor agents represents a new class of compounds: the alkylphosphocholines. As reported here, hexadecylphosphocholine is the most promising candidate of this group of compounds. It has an extremely selective action against chemically induced, autochthonous rat mammary carcinomas. No loss of activity was observed when comparing oral and intravenous administration. Particularly striking (and favorable for long-term therapy) is the fact that immunosuppression and hematotoxicity were not found at drug concentrations which lead to complete tumor remissions. Results obtained from animal experiments have been confirmed by preliminary clinical investigations.

Chirospecific syntheses of 2H- and 13C-labeled 1-O-alkyl-2-O-alkyl'-sn-glycero-3-phosphoethanolamines and 1-O-alkyl-2-O-alkyl'-sn-glycero-3-phosphocholines

A convenient sequence for the synthesis of 1-O-alkyl-2-O-alkyl'-sn-glycero-3-phospholipids was demonstrated starting from 2,3-O-isopropylidene-sn-glycerol, which was first alkylated with 1-bromohexadecane, then converted to the corresponding benzylidene analog. Other less convenient methods to prepare 2,3-O-benzylidene-1-O-hexadecyl-sn-glycerol were also investigated. The key step in the synthesis was the reduction of 2,3-O-benzylidene-1-O-hexadecyl-sn-glycerol with lithium aluminum hydride-aluminum chloride to give 3-O-benzyl-1-O-hexadecyl-sn-glycerol as the major product in 79% yield. The syntheses of 1-O-hexadecyl-2-O-hexadecyl-(1',1'-d2,-sn-glycero-3-phosphoethanolamine and 1-O-hexadecyl-2-O-hexadecyl-(1'-13C)-sn-glycero-3-phosphoethanolamine as well as the correspondingly labeled sn-glycero-3-phosphocholine analogs were then performed. The optical purities of the synthetic intermediates and the ether lipids were established by a novel 1H-NMR method.

Substrate specificity for interfacial catalysis by phospholipase A2 in the scooting mode

Action of pig pancreatic phospholipase A2 on vesicles and micelles of homologous anionic phospholipids is examined in the absence of additives. As shown elsewhere (Jain et al. (1986) Biochim. Biophys. Acta 860, 435-447), hydrolysis of anionic vesicles occurs by interfacial catalysis in the scooting mode, i.e., the catalytic turnover is fast relative to the off-rate of the enzyme from the interface. When the rate of intervesicle exchange of the enzyme is negligibly slow, it hydrolyses only the substrate molecules in the outer monolayer of the vesicle to which it is bound. Interfacial catalysis in the scooting mode with a high processivity occurs on vesicles of anionic phospholipids, and under these conditions the dynamics and order of the substrate in the interface influences the catalytic turnover only moderately, i.e., about 2- to 10-fold. Similarly, anomalous kinetic effects of the thermotropic gel-fluid phase transition or of a change in the general disorder of the bilayer organization (fluidity) has a minor effect on the kinetics of hydrolysis in the scooting mode. Similarly, higher unsaturation and shorter acyl chains in the substrate modestly increase the rate of catalytic turnover by the low-calcium form of the enzyme without noticeably influencing the affinity of the enzyme for the interface. On the other hand, perturbation of the charge distribution in the substrate interface can shift the proportion of the bound enzyme by several orders of magnitude. For example, the membrane perturbing amphiphiles (e.g., mepacrine, indomethacin, compound 48/80, aristolochic acid, local anesthetics, and the products of hydrolysis) do not influence the catalytic turnover of the bound enzyme but the proportion of the bound enzyme. Short-chain anionic phospholipids are readily hydrolyzed by phospholipase A2. Now no anomalous increase in the rate of hydrolysis is observed at the critical micelle as is the case with the zwitterionic analogs. This is because with anionic (but not with zwitterionic) substrates the enzyme forms an aggregated complex below the cmc of the monomer. The stability of these micellar complexes does not appear to change noticeably with the acyl chain length of the monomers. These observations show that the factors regulating the quality of interface substantially influence the binding of the enzyme, but not the catalytic turnover in the interface.

Substrate specificity of Staphylococcus aureus (TEN5) lipases with isomeric oleoyl-sn-glycerol ethers as substrates

For the first time fully protected substrates with only one hydrolyzable ester bond have been used to analyze the substrate specificity of microbial lipases. In these substrates the ester is attached to the glycerol molecule in a precisely defined position. The use of three different substituents generates chirality and thus allows the analysis of positional specificities of individual lipases. Therefore, these new substrates have been used to study the enzymatic activities of two closely related lipases isolated from Staphylococcus aureus (TEN5) designated the 44 and 43 kDa lipase. The lipases, especially the 44 kDa molecule, show a high specificity for the hydrolysis of the ester in the sn-1 position (S-configuration), which is hydrolyzed by a factor of ten faster than that in the sn-3 position. In addition, the study demonstrates for the first time that the rate of hydrolysis of a fatty acid ester attached to the sn-2 position of glycerol by microbial lipases depends on the configuration of the substrate molecule.

Metabolism of ether phospholipids and analogs in neoplastic cells

The ether phospholipid 1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine (OM-GPC) is known to be a potent inhibitor of cell growth. Metabolic studies in both Raji and L1210 leukemic cells on OM-GPC, 3H-labeled in the methyl groups of the choline moiety, showed a (diacyl)-phosphatidylcholine as the only labeled metabolite. Since the formation of radiolabeled (diacyl)-phosphatidylcholine showed a direct correlation with cell death, we tested other lipid analogs. One of these compounds, hexadecylphosphocholine (He-PC), which was 3H-labeled in the methyl-choline groups, showed a formation of labeled (diacyl)-phosphatidylcholine similar to that found with OM-GPC. Again, there was a direct linear correlation between the formation of the labeled product and cell death. He-PC was found to be a potent cell toxin in in vitro experiments on cell cultures. However, analogs with an elongated phosphor to trimethylammonium distance showed no toxicity towards the cells in in vitro experiments. From the data, we conclude that the ether phospholipids are substrates for a phospholipase C or related enzyme. This substrate property may be responsible for the toxicity of the compounds in neoplastic cells.

A continuous assay for O-alkylglycerol monooxygenase (E.C. 1.14.16.5)

The antitumor activity of alkyl lysophospholipids has raised some questions concerning the degradation of O-alkyl bonds in naturally occurring ether lipids. In this report, we describe the first continuous assay for O-alkylglycerol monooxygenase (AGMO), the only enzyme known to cleave the O-alkyl bond in saturated ether lipids and ether phospholipids. AGMO activity was monitored at 340 nm by coupling the NADH redox reaction to the tetrahydropteridine cofactor of the rat liver microsomal enzyme. Turnover rates as low as 0.6 nmol/min could be measured. Using radiolabeled substrates, the products were identified with a TLC-Linear-Analyzer. The only interference with this assay can arise from other reducing agents, e.g. dithiothreitol. The assay was used to develop protocols for the solubilization of AGMO from membrane preparations in the presence of detergents.

Substrate specificity of O-alkylglycerol monooxygenase (E.C. 1.14.16.5), solubilized from rat liver microsomes

Synthetic alkyl lysophospholipids (ALP) show antineoplastic activity. In the present discussion, the cytotoxicity of ALP is attributed to their accumulation in tumor cells. Some neoplastic cell species, in contrast to normal cells, cannot metabolize ALP because of a lack of O-alkylglycerol monooxygenase (AGMO) activity. To understand the metabolic fate of ether lipids and ether-linked phospholipids, AGMO substrate specificity studies were undertaken. Thirty-five different natural and synthetic ether lipids and their metabolites (including a thioether) were tested as substrates for AGMO. The study revealed that the potent cytostatic substance, 1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine is not a substrate for AGMO. Therefore, its selective toxicity to tumor cells cannot be based on the differences in direct detoxification of 1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine by AGMO in normal and malignant cells. However, 1-O-octadecyl-2-O-methyl-rac-glycerol, which can be formed by phospholipase C hydrolysis, is a good substrate for AGMO.