Bisnaphthalimidopropyl spermidine induces apoptosis within colon carcinoma cells

Bisnaphthalimido compounds bisintercalate to DNA via the major groove and are potentially potent cancer therapeutics. We incorporated natural polyamines as linkers connecting the two-naphthalimido ring moieties to create a series of novel soluble cytotoxic bisnaphthalimidopropyl polyamines (BNIPPs). Here, we determined the cytotoxicity of bisnaphthalimidopropyl spermidine (BNIPSpd) towards Caco-2 and HT-29 colon adenocarcinoma cells revealing an IC₅₀ value of 0.15 and 1.64 μM after 48 h exposure within Caco-2 and HT-29 cells, respectively. After 4 h, ≥0.5 μM BNIPSpd treatment-induced significant DNA damage. After 24 h exposure a concentration-dependent increase in active caspase-3 expression, chromatin condensation and internucleosomal DNA fragmentation identified apoptosis as the principal manifestation for the cytotoxicity within both cell lines. By 24 h exposure, there was also a significant decline in cellular spermine and spermidine levels. It is concluded that bisnaphthalimidopropyl spermidine (BNIPSpd) toxicity primarily results from apoptosis and that BNISpd has potential to be further developed as an anti-tumour agent.     

Role of leukotriene B4 in celecoxib-mediated anticancer effect

Celecoxib, a selective cyclooxygenase-2 (COX-2) inhibitor, has anticancer effect on many cancers associated with chronic inflammation by both COX-2-dependent and COX-2-independent mechanisms. The non-COX-2 targets of celecoxib, however, are still a matter of research. Leukotriene B₄ (LTB₄) has been implicated in prostate and colon carcinogenesis, but little is known about the potential role of LTB₄ in celecoxib-mediated anticancer effect. In this study, we evaluated whether LTB₄ was involved in celecoxib-mediated inhibitory effect on human colon cancer HT-29 cells and human prostate cancer PC-3 cells. Our data showed that survival of both cell lines was obviously suppressed after celecoxib treatment for 72 h in a concentration-dependent manner. However, only in HT-29 cells, this inhibitory effect could be reversed by LTB₄, which promoted survival of HT-29 cells rather than PC-3 cells. Consistent with these results, lioxygenase (LOX) potent inhibitor nordihydroguaiaretic acid (NDGA) had a higher inhibitory effect on HT-29 cells than PC-3 cells. Additionally, ELISA results showed that celecoxib could suppress expression of LTB₄ in both cell lines, whereas, inhibition of PGE₂ was only detected in HT-29 cells. These results indicate that the anticancer effect of celecoxib is COX-2-independent in HT-29 and PC-3 cells and in HT-29 cells primarily via down-regulating LTB₄ production.     

The role of β2-glycoprotein I in liposome-hepatocyte interaction

Adsorption of serum proteins to the liposomal surface plays a critical role in liposome clearance from the blood. The aim of this study was to investigate the role of liposome-adsorbed serum proteins in the interaction of liposomes with hepatocytes. We analyzed the serum proteins adsorbing to the surface of differently composed small unilamellar liposomes during incubation with human or rat serum, and found that one protein, with a molecular weight of around 55 kDa, adsorbed in a large amount to negatively charged liposomes containing phosphatidylserine (PS) or phosphatidylglycerol (PG). The binding was dependent on the liposomal charge density. The ∼55-kDa protein was identified as β₂-glycoprotein I (β₂GPI) by Western blotting. Despite the high affinity of β₂GPI for strongly negatively charged liposomes, in vitro uptake and binding experiments with isolated rat hepatocytes, Kupffer cells or liver endothelial cells, and with HepG2 cells showed no enhancing effect of this protein on the association of negatively charged liposomes with any of these cells. On the contrary, an inhibitory effect was observed. We conclude that despite abundant adsorption to negatively charged liposomes, β₂GP1 inhibits, rather than enhances, liposome uptake by liver cells.     

Prolyl oligopeptidase inhibition-induced growth arrest of human gastric cancer cells

Prolyl oligopeptidase (POP) is a serine endopeptidase that hydrolyzes post-proline peptide bonds in peptides that are <30 amino acids in length. We recently reported that POP inhibition suppressed the growth of human neuroblastoma cells. The growth suppression was associated with pronounced G₀/G₁ cell cycle arrest and increased levels of the CDK inhibitor p27^kip1 and the tumor suppressor p53. In this study, we investigated the mechanism of POP inhibition-induced cell growth arrest using a human gastric cancer cell line, KATO III cells, which had a p53 gene deletion. POP specific inhibitors, 3-({4-[2-(E)-styrylphenoxy]butanoyl}-l-4-hydroxyprolyl)-thiazolidine (SUAM-14746) and benzyloxycarbonyl-thioprolyl-thioprolinal, or RNAi-mediated POP knockdown inhibited the growth of KATO III cells irrespective of their p53 status. SUAM-14746-induced growth inhibition was associated with G₀/G₁ cell cycle phase arrest and increased levels of p27^kip1 in the nuclei and the pRb2/p130 protein expression. Moreover, SUAM-14746-mediated cell cycle arrest of KATO III cells was associated with an increase in the quiescent G₀ state, defined by low level staining for the proliferation marker, Ki-67. These results indicate that POP may be a positive regulator of cell cycle progression by regulating the exit from and/or reentry into the cell cycle by KATO III cells.     

Development of a Cell-Based Bioassay for Phospholipase A2-Triggered Liposomal Drug Release

The feasibility of exploiting secretory phospholipase A₂ (sPLA₂) enzymes, which are overexpressed in tumors, to activate drug release from liposomes precisely at the tumor site has been demonstrated before. Although the efficacy of the developed formulations was evaluated using in vitro and in vivo models, the pattern of sPLA₂-assisted drug release is unknown due to the lack of a suitable bio-relevant model. We report here on the development of a novel bioluminescence living-cell-based luciferase assay for the monitoring of sPLA₂-triggered release of luciferin from liposomes. To this end, we engineered breast cancer cells to produce both luciferase and sPLA₂ enzymes, where the latter is secreted to the extracellular medium. We report on setting up a robust and reproducible bioassay for testing sPLA₂-sensitive, luciferin remote-loaded liposomal formulations, using 1,2-distearoyl-sn-glycero-3-phosphatidylcholine/1,2-distearoyl-sn-glycero-3-phosphatidylglycerol (DSPC/DSPG) 7:3 and DSPC/DSPG/cholesterol 4:3:3 as initial test systems. Upon their addition to the cells, the liposomes were degraded almost instantaneously by sPLA₂ releasing the encapsulated luciferin, which provided readout from the luciferase-expressing cells. Cholesterol enhanced the integrity of the formulation without affecting its susceptibility to sPLA₂. PEGylation of the liposomes only moderately broadened the release profile of luciferin. The provided bioassay represents a useful tool for monitoring active drug release in situ in real time as well as for testing and optimizing of sPLA₂-sensitive lipid formulations. In addition, the bioassay will pave the way for future in-depth in vitro and in vivo studies.     

Cytotoxic benzil and coumestan derivatives from Tephrosia calophylla

A benzil, calophione A, 1-(6′-Hydroxy-1′,3′-benzodioxol-5′-yl)-2-(6″-hydroxy-2″-isopropenyl-2″,3″-dihydro-benzofuran-5″-yl)-ethane-1,2-dione and three coumestan derivatives, tephcalostan B, C and D were isolated from the roots of Tephrosia calophylla. Their structures were deduced from spectroscopic data, including 2D NMR ¹H-¹H COSY and ¹³C-¹H COSY experiments. Compounds were evaluated for cytotoxicity against RAW (mouse macrophage cells) and HT-29 (colon cancer cells) cancer cell lines and antiprotozoal activity against various parasitic protozoa. Calophione A exhibited significant cytotoxicity with IC₅₀ of 5.00 (RAW) and 2.90 μM (HT-29), respectively.     

A Liposomal Drug Platform Overrides Peptide Ligand Targeting to a Cancer Biomarker,Irrespective of Ligand Affinity or Density

One method for improving cancer treatment is the use of nanoparticle drugs functionalized with targeting ligands that recognize receptors expressed selectively by tumor cells. In theory such targeting ligands should specifically deliver the nanoparticle drug to the tumor, increasing drug concentration in the tumor and delivering the drug to its site of action within the tumor tissue. However, the leaky vasculature of tumors combined with a poor lymphatic system allows the passive accumulation, and subsequent retention, of nanosized materials in tumors. Furthermore, a large nanoparticle size may impede tumor penetration. As such, the role of active targeting in nanoparticle delivery is controversial, and it is difficult to predict how a targeted nanoparticle drug will behave in vivo. Here we report in vivo studies for α_vβ₆-specific H2009.1 peptide targeted liposomal doxorubicin, which increased liposomal delivery and toxicity to lung cancer cells in vitro. We systematically varied ligand affinity, ligand density, ligand stability, liposome dosage, and tumor models to assess the role of active targeting of liposomes to α_vβ₆. In direct contrast to the in vitro results, we demonstrate no difference in in vivo targeting or efficacy for H2009.1 tetrameric peptide liposomal doxorubicin, compared to control peptide and no peptide liposomes. Examining liposome accumulation and distribution within the tumor demonstrates that the liposome, and not the H2009.1 peptide, drives tumor accumulation, and that both targeted H2009.1 and untargeted liposomes remain in perivascular regions, with little tumor penetration. Thus H2009.1 targeted liposomes fail to improve drug efficacy because the liposome drug platform prevents the H2009.1 peptide from both actively targeting the tumor and binding to tumor cells throughout the tumor tissue. Therefore, using a high affinity and high specificity ligand targeting an over-expressed tumor biomarker does not guarantee enhanced efficacy of a liposomal drug. These results highlight the complexity of in vivo targeting.     

Use of phospholipase A to compare phospholipid organization in synaptic membranes,myelin, and liposomes

Summary The pattern of fatty acid release from rat synaptic membranes in the presence of phospholipase A₂ (Vipera russelli) was compared to that from liposomes comprised of phospholipids. Phospholipase A₂ more readily attacked myelin and synaptic membranes than liposomes prepared from total phospholipids derived from myelin. Although hydrolysis of liposomal phospholipids occurred in the absence of added calcium, the presence of 2mm CaCl₂ or 2% bovine serum albumin significantly enhanced the phospholipase attack of liposomes, but not synaptic membranes or myelin. Phospholipase exhibited a marked preference for phospholipids containing docosahexaenoic acid (226) in the synaptic membranes, while with liposomes the pattern of released fatty acid reflected the fatty acid composition in the two-position of the phospholipids. Although either calcium or albumin markedly increased the phospholipase hydrolysis of liposomes, neither affected the hydrolysis of synaptic membranes or the pattern of fatty acid release from liposomes. It was concluded that the nonlipid constituents, particularly the proteins, of biomembranes were responsible for the organization of the phospholipids and accounted for the observed differences between liposomes and synaptic membranes with respect to enzymic accessibility.     

Influence of Phospholipids Structure on the Physicochemical Properties and In Vitro Digestibility of Lactoferrin-Loaded Liposomes

The instability of liposomal delivery system during passaging through the gastrointestinal tract (GIT) stimulates a demand to find a stable liposome. This research studied the implications of different types of phospholipids (different fatty acid chain length and saturation, various head group) on liposomal physiochemical properties and stability in the human GIT. The micropolarity of liposomal membrane increased with the decrease of chain lengths of phospholipids, while the morphology observation revealed that the liposomes formed by different phospholipids showed similar in appearance and shapes. The liposomes formed by C_20:0 deformed more severely in simulated gastric fluid, while others exhibited slight changes in the membrane structure. In simulated intestinal fluid, pancreatic lipase and phospholipase A₂, synergized with bile salts, damaged the bilayers structure of all liposomes, with the entrapped lactoferrin release and hydrolysis. Although the various phospholipid structures lead to some difference on the physicochemical properties (size and micropolarity), the enzymic influence displayed more significance during in vitro digestion compared to the types of wall materials. Current results could provide valuable information for the development of more stable and reliable food-grade liposomes in the GIT.

     

Prostaglandin phospholipid conjugates with unusual biophysical and cytotoxic properties

The synthesis of two secretory phospholipase A₂ IIA sensitive 15-deoxy-Δ^12,14-prostaglandin J₂ phospholipid conjugates is described and their biophysical and biological properties are reported. The conjugates spontaneously form particles in the liposome size region upon dispersion in an aqueous buffer and both phospholipids are hydrolyzed by phospholipase A₂, but with different conversion rates and extent of hydrolysis. The cytotoxicity was evaluated in HT-29 and Colo205 cells and the conjugates induced cell death in the presence of phospholipase A₂ and surprisingly also in the absence of the enzyme.     

Effect of Macrophage Elimination Using Liposome-Encapsulated Dichloromethylene Diphosphonate on Tissue Distribution of Liposomes

Abstract

Effect of macrophage elimination using liposomal dichloromethylene diphosphonate (C1₂MDP)1 on tissue distribution of different types of liposomes was examined in mice. Intravenously administration into mice with CI2MDP encapsulated in liposomes composed of phosphatidylcholine, cholesterol and phosphatidylserine exhibits a temporary blockade of liver and spleen function for liposome uptake. At a low dose of 90 (ig/mouse, the liposome uptake by the liver was significantly decreased. Such decrease was accompanied by an increase in liposome accumulation in either spleen or blood depending on liposome composition and size. Direct correlation between the administration dose of liposomal CI2MDP and the liposome circulation time in blood was also obtained even for liposomes with an average diameter of more than 500 nm. These results indicate that temporary elimination of macrophages of the liver and spleen using liposomal CI2MDP may prove to be useful to enhance the drug delivery efficiency of liposomes.     

Grafting of different glycosides on the surface of liposomes and its effect on the tissue distribution of 125I-labelled γ-globulin encapsulated in liposomes

Different glycosides were grafted on the surface of liposomes containing ¹²⁵I-labelled γ-globulin by two ways: (1) by using glycolipid and (2) by covalent coupling of $\text{p-}\text{aminophenyl-}\text{d}\text{-glycosides}$ to phosphatidylethanolamine liposomes using glutaraldehyde. The distribution of ¹²⁵I-labelled γ-globulin was determined in mouse tissues from 5–60 min after a single injection of these liposomes. The liver uptake of encapsulated ¹²⁵I-labelled γ-globulin was highest from liposomes having galactose and mannose on the surface. Competition experiments and cross-inhibition studies indicate that this uptake are mediated by specific recognition of the surface galactose and mannose residues of liposomes by the receptors present on the plasma membrane of liver cells. Stearylamine-containing liposomes were found to be more efficient in mediating the uptake of ¹²⁵I-labelled γ-globulin by the lung, whereas in the case of spleen, phosphatidylethanolamine liposomes were more efficient. The extent of uptake of ¹²⁵I-labelled γ-globulin from all types of liposome decreases as the amount of given liposomes increases. The uptake of ¹²⁵I-labelled γ-globulin from liposomes containing asialogangliosides depends upon the phospholipid/ glycolipid ratio. These experiments clearly demonstrate that enhanced liposome uptake by liver cells could be achieved by grafting galactose and mannose on the liposomal surface.     

Effects of trehalose-loaded liposomes on red blood cell response to freezing and post-thaw membrane quality

We are investigating the use of liposomes, which are synthetic, microscopic vesicles, for the intracellular delivery of trehalose into mammalian cells. This study focuses on the effects trehalose-containing liposomes improve the recovery and membrane quality of human RBCs following cryopreservation. Unilamellar liposomes consisting of a lipid bilayer composed of DPPC, PS and cholesterol (60:30:10 mol%) were synthesized using an extrusion method. Liposome-treated RBCs (l-RBCs) were resuspended in either physiological saline, 0.3 M trehalose or liposome solution, then cooled with slow (0.95 ± 0.02 °C/min), medium (73 ± 3 °C/min) and fast (265 ± 12 °C/min) cooling rates and storage in liquid nitrogen, followed by a 37 °C thawing step. RBC post-thaw quality was assessed using percent recovery, RBC morphology, PS and CD47 expression. Liposome treatment did not adversely affect the RBC membrane. Post-thaw recovery of l-RBCs was significantly higher (66% ± 5% vs 29% ± 4%) compared to control RBCs (c-RBC, p = 0.003). Medium and high cooling rates resulted in significantly higher cell recovery compared to a slow cooling rate (p = 0.039 and p = 0.041, respectively). The recovery of l-RBCs frozen in liposome solution and trehalose solution was significantly higher than that of l-RBCs frozen in NaCl solution for all three cooling rates (p = 0.021). Flow cytometry and morphology assessment showed that liposome treatment resulted in improved post-thaw membrane quality. There was no statistically significant difference in the post-thaw recovery between RBCs treated with liposomes containing trehalose in their aqueous core and RBCs treated with liposomes containing saline in their aqueous core (p = 0.114). Liposome treatment significantly improves the recovery and membrane integrity of RBCs following low temperature exposure.     

Interaction of 10-(octyloxy) decyl-2-(trimethylammonium) ethyl phosphate with mimetic membranes and cytotoxic effect on leukemic cells

10-(Octyloxy) decyl-2-(trimethylammonium) ethyl phosphate (ODPC) is an alkylphospholipid that can interact with cell membranes because of its amphiphilic character. We describe here the interaction of ODPC with liposomes and its toxicity to leukemic cells with an ED-50 of 5.4, 5.6 and 2.9 μM for 72 h of treatment for inhibition of proliferation of NB4, U937 and K562 cell lines, respectively, and lack of toxicity to normal hematopoietic progenitor cells at concentrations up to 25 μM. The ED-50 for the non-malignant HEK-293 and primary human umbilical vein endothelial cells (HUVEC) was 63.4 and 60.7 μM, respectively. The critical micellar concentration (CMC) of ODPC was 200 μM. Dynamic light scattering indicated that dipalmitoylphosphatidylcholine (DPPC) liposome size was affected only above the CMC of ODPC. Differential calorimetric scanning (DCS) of liposomes indicated a critical transition temperature (T_c) of 41.5 °C and an enthalpy (?H) variation of 7.3 kcal mol⁻¹. The presence of 25 μM ODPC decreased T_c and ?H to 39.3 °C and 4.7 kcal mol⁻¹, respectively. ODPC at 250 μM destabilized the liposomes (36.3 °C, 0.46 kcal mol⁻¹). Kinetics of 5(6)-carboxyfluorescein (CF) leakage from different liposome systems indicated that the rate and extent of CF release depended on liposome composition and ODPC concentration and that above the CMC it was instantaneous. Overall, the data indicate that ODPC acts on in vitro membrane systems and leukemia cell lines at concentrations below its CMC, suggesting that it does not act as a detergent and that this effect is dependent on membrane composition.     

Different modes of membrane permeabilization by two RTX toxins: HlyA from Escherichia coli and CyaA from Bordetella pertussis

This study clarifies the membrane disruption mechanisms of two bacterial RTX toxins: αhemolysin (HlyA) from Escherichia coli and a highly homologous adenylate cyclase toxin (CyaA) from Bordetella pertussis. For this purpose, we employed a fluorescence requenching method using liposomes (extruded through filters of different pore size — 1000 nm, 400 nm or 100 nm) with encapsulated fluorescent dye/quencher pair ANTS/DPX. We showed that both toxins induced a graded leakage of liposome content with different selectivities α for DPX and ANTS. In contrast to HlyA, CyaA exhibited a higher selectivity for cationic quencher DPX, which increased with vesicle diameter. Large unilamellar vesicles (LUV₁₀₀₀) were found to be more suitable for distinguishing between high α values whereas smaller ones (LUV₁₀₀) were more appropriate for discriminating an all-or-none leakage (α = 0) from the graded leakage with low values of α. While disrupting LUV₁₀₀₀, CyaA caused a highly cation-selective leakage (α ~ 15) whereas its mutated form with decreased channel K⁺/Cl⁻ selectivity due to two substitutions in a predicted transmembrane segment (CyaA-E509K + E516K) exhibited much lower selectivity (α ∼ 6). We concluded that the fluorescence requenching method in combination with different size of liposomes is a valuable tool for characterization of pore-forming toxins and their variants.     

Discovery of thiazolidin-4-one urea analogues as novel multikinase inhibitors that potently inhibit FLT3 and VEGFR2

A series of novel thiazolidine-4-one urea analogues were designed, synthesized and biologically evaluated. The structure-activity relationship (SAR) at several positions of the scaffolds was investigated and its binding mode was analyzed by molecular modeling studies. Compound 17b proved to be the most potent one, and IC₅₀ values against A549 and HT-29 cancer cell lines were 0.65?μM and 0.11?μM, respectively. The results of kinase profile demonstrated that compound 17b is a multikinase inhibitor that potently inhibits FLT3 (IC₅₀?=?8.6?nM) and VEGFR2 (IC₅₀?=?18.7?nM). The results of real-time live-cell imaging indicated that compound 17b showed excellent cytotoxicity and anti-proliferative activity against HT-29 cancer cells in a time- and dose-dependent manner, which was significantly potent than that of Cabozantinib. In addition, in vitro antitumor activity was associated with inducing cancer cell apoptosis and suppression of cancer cell migration.     

ATP binding and hydrolysis steps of the uni-site catalysis by the mitochondrial F1-ATPase are affected by inorganic phosphate

The effect of inorganic phosphate (P_i) on uni-site ATP binding and hydrolysis by the nucleotide-depleted F₁-ATPase from beef heart mitochondria (ndMF₁) has been investigated. It is shown for the first time that P_i decreases the apparent rate constant of uni-site ATP binding by ndMF₁ 3-fold with the K_d of 0.38 ± 0.14 mM. During uni-site ATP hydrolysis, P_i also shifts equilibrium between bound ATP and ADP + P_i in the direction of ATP synthesis with the K_d of 0.17 ± 0.03 mM. However, 10 mM P_i does not significantly affect ATP binding during multi-site catalysis.     

Effects of equatorial ligand bulk on the orientation of axial ligands in methyl organocobalt B12 models assessed by X-ray and NMR methods

We prepared two new analogues of ([CH₃Co((DO)(DOH)pn)L]⁺) [(DO)(DOH)pn = N²,N^2′-propane-1,3-diylbis(2,3-butanedione-2-imine-3-oxime)] B₁₂ models but with an O-BF₂-O unit replacing the O-H?O unit as follows: [CH₃Co((DO)(DOBF₂)pn)L]PF₆ with L = pyridine (py) and 1,5,6-trimethylbenzimidazole (Me₃Bzm). Our goal was to compare the properties of these new O-BF₂-O complexes with the well-established O-H?O analogues. The Co-CH₃¹H NMR shifts indicate that the BF₂ group makes the Co(III) less electron rich. The X-ray crystal structures determined for the new compounds were compared to the one known structure with L = imidazole (Im). With increasing size of L, in the series Im < py < Me₃Bzm, the plane of L orients so as to avoid the bulky BF₂ group. This orientation effect becomes apparent in the L ¹H NMR shifts, which are not sensitive to Co(III) electronic properties. Thus, in the O-BF₂-O versus the O-H?O analogue, the Me₃Bzm H4 signal shifts 0.41 ppm upfield from the anisotropic effect of the equatorial ligand double bonds. We advance the concept (applicable to a broad series of complexes) that steric interactions between L and the equatorial ligands are alleviated by a combination of Co-N_ax bond elongation and opening of the N_eq-Co-N_ax angles.     

Characterization of a novel pH-sensitive peptide that enhances drug release from folate-targeted liposomes at endosomal pHs

Although liposomes have proven useful for the delivery of drugs and gene therapy vectors, their potencies are often compromised by poor unloading following uptake into their target cells. We have consequently explored the properties of a novel 29-residue amphipathic peptide that was designed by arrangement of hydrophobic and hydrophilic residues to disrupt liposomes at lower peptide concentrations than previously tested peptides. The peptide was indeed found to promote pH-dependent liposome unloading with improved efficiency. A peptide of the same sequence, but half the length, however, promoted pH-dependent permeabilization only at much higher concentrations. Further characterization of the longer peptide revealed that release of liposome contents (i) occurred at a pH of ∼6, (ii) became less efficient as the size of the encapsulated cargo increased, and (iii) was moderately suppressed in cholesterol-containing liposomes. Use of this peptide to enhance the cytotoxicity of cytosine arabinoside encapsulated in folate-targeted liposomes demonstrated an increase in drug potency of ∼30-fold. Gene expression by a serum-stable folate-targeted liposomal vector was also measurably enhanced by inclusion of the peptide. We conclude that intracellular unloading of liposomal contents can be significantly improved by co-encapsulation of an optimally designed, pH-sensitive peptide.     

Therapy with interleukin-2 induces the systemic release of phospholipase-A2

Therapy with interleukin-2 (IL-2) induces remissions in some forms of cancer. This treatment however, is accompanied by side-effects which, in part, may be mediated by the formation of eicosanoids and plateletactivating factor. We investigated the systemic release of phospholipase A₂ (PLA₂), a rate-limiting enzyme in the formation of these lipid mediators, in patients receiving IL-2. In a pilot study of 4 patients we observed an increase in PLA₂ activity in serial plasma samples obtained during the first day after a bolus infusion of IL-2, which increase closely correlated with that of antigen levels of secretory phospholipase A₂ (sPLA₂) as measured by enzyme-linked immunosorbent assay (r=0.92;P<0.001). In 20 patients, receiving 12×10⁶–18×10⁶ IU IL-2/m², we then investigated the course of antigenic levels of sPLA₂ in relation to those of the cytokines tumour necrosis factor (TNF) and interleukin-6 (IL-6) (both cytokines may induce sPLA₂ in vivo). From 4 h on, sPLA₂ levels significantly increased, reaching a peak 24 h after the IL-2 infusion. Subsequent IL-2 infusions even induced a further increase of sPLA₂. This increase of sPLA₂ was presumably not due to a direct effect of IL-2 on, for example, hepatocytes, since this cytokine, in contrast to IL-1, IL-6, TNF and interferon , was not able to induce the synthesis of sPLA₂ by Hep G2 cells in vitro. Consistent with this, plasma levels of TNF and IL-6 in the patients rose, reaching peak levels before a zenith of sPLA₂ occurred, i.e at 2 h and 4 h after the start of the IL-2 infusion respectively. sPLA₂ levels significantly correlated with the development of the side-effects increase in body weight (r=0.49;P<0.0001) and decrease in mean arterial blood pressure (r=0.40;P<0.0001). Moreover, maximum sPLA₂ levels induced by IL-2 were higher in patients who had progressive disease after therapy than in patients who had stable disease or a partial response.