Introduction

Top Abstract Introduction Methods Results Discussion References

Lisofylline [LSF, 1-(5-R-hydroxyhexyl)-3,7-dimethylxanthine] is a therapeutic drug candidate which is under development to decrease the incidence of neutropenic infections, to decrease the incidence and severity of mucositis and to decrease mortality after dose-intensive chemotherapy or combined chemotherapy and radiation therapy. LSF is also being developed for suppression of multiorgan dysfunction and acute respiratory distress syndrome after the systemic inflammatory response to oxidative injury or infection. Although the biomolecular target of this drug candidate is as yet undefined, it has been clearly shown to suppress cellular activation induced by cytotoxic therapies or oxidative stress, and ultimately to decrease release of inflammatory cytokines; this results in acceleration of cellular recovery after cytotoxic therapies and protection against tissue damage after oxidative injury (Abraham et al., 1995; Bursten et al., 1994, 1996; Clarke et al., 1996; Hybertson et al., 1997; Waxman et al., 1996). In addition, these activities have been tied to significant changes in lipid metabolism, including complex alterations in acyl chain handling in both in vitro and in vivo systems (Rice et al., 1994a, b; Abraham et al., 1995; Bursten et al., 1994, 1996).

Serum total FFA, as well as the component of unsaturated FFA, are increased in patients after trauma or sepsis relative to levels in healthy subjects (Bursten et al., 1996). Unsaturated FFA and their oxidative products have been implicated in the pathogenesis of several disease states, including insulin resistance (Pan et al., 1995; Storlien et al., 1996), rheumatic and other inflammatory diseases (Kremer,1996) and septic states (Keen et al., 1991; Marshall and Lands, 1986). Increased serum FFA in sepsis and related inflammatory states may result from increased activity of secretory (type II) phospholipase A₂ (Vadas et al., 1988). Both the molar contents of various polyunsaturated fatty acids in cell membranes and their relative distribution in complex lipids has been related to expression of endothelial leukocyte adhesion molecules (deCaterina and Libby, 1996), synthesis and release of IL-1 (Bursten et al., 1994; Endres, 1996) and growth-related gene expression (Sellmeyer et al., 1996). Oxidative products of the most reactive polyunsaturated fatty acids, such as HPODEs from linoleate (C18:2; -6), have been implicated as biologically active agents in inflammatory conditions resulting from sepsis and in hypoxia-reperfusion models of organ injury (Buchanan et al., 1985a, b; Kaduce et al., 1989; Reinaud et al., 1989; Folcik and Cathcart, 1994). HPODEs induce IL-1 release from macrophages (Ku et al., 1992), and serve as post-receptor mediators in the mitogenic response induced by epidermal growth factor (Glasgow et al., 1992; Cowlen and Eling, 1993). Furthermore, because these oxidative products not only induce inflammatory cytokine release, but are themselves induced by IL-1, they may function as amplifiers of inflammatory signals (Camacho et al., 1995).

Because of the previously observed effects of LSF on cellular and serum complex lipids, the participation of complex lipids, FFA, and their oxidized products in inflammation and the potent anti-inflammatory effects of LSF, we have examined the effect of LSF on FFA and other selected serum lipids as part of a pharmacokinetic study of LSF in healthy volunteers. We have demonstrated that LSF reduces serum FFA, but increases serum triglyceride levels. The significance of these observations is as follows: (1) Further evidence is provided for the unique effects of LSF on both cellular and in vivo acyl chain metabolism, where previous studies (including Rice et al., 1994a, b; Bursten et al., 1994, 1996) have implicated its activity; (2) It demonstrates that lipid metabolic effects of LSF in vivo are prolonged beyond 60 serum elimination half-lives, which indicates a possible novel compound effect, or a yet undescribed variant metabolism for LSF; (3) This effect may be an excellent surrogate marker for the pharmacodynamic activity of LSF; (4) It underlines the importance of performing careful studies on human serum lipid metabolism, as well as in vitro systems (using exhaustive techniques such as high-performance thin-layer chromatography and high-performance liquid chromatography to examine lipid mass and composition), to elucidate pharmacologic mechanisms of action such as acyl chain handling.

	Methods

Top Abstract Introduction Methods Results Discussion References

Subjects

Healthy, nonsmoking volunteers were screened with physical examinations, electrocardiography and clinical laboratory evaluations. Twelve male subjects who ranged in age from 23 to 47 years were selected to participate in the study.

Clinical Study Design

The subjects were randomly assigned to one of three groups. All were confined in the clinical research unit (CRU) from the morning of the day before dosing and throughout the study period until approximately 24 hr after the last dose. They were instructed to abstain from foods that contained xanthine, theophylline, caffeine or alcohol for 3 days before administration of the first dose and, while in the CRU, they were given a standardized diet that was low in fat and free of xanthine, caffeine, theophylline and alcohol. Vital signs, including seated blood pressure, pulse, respirations and oral temperature, were obtained at screening, at check-in and at discharge from the study. Automated seated blood pressure and pulse were measured at 0 hr (before drug dosing) and at 5, 10, 20, 30, 40, 50 and 60 min, and 2.5 hr after the start of the drug dosing. A 12-lead ECG was obtained at screening, check-in, immediately before each dosing period, 1 hr after the start of each dose, and at discharge from the study. All subjects were observed for adverse effects by experienced clinical study nurses during the study period. The subjects remained seated upright for 3 hr after the start of each drug dose administration, except as necessitated by the occurrence of an adverse event or study procedure.

Drug Administration

Unit doses of LSF were prepared from 5-ml vials containing 60 mg/ml lisofylline by a nonblinded staff member and a registered pharmacist who did not participate further in the study. On the morning of the intravenous infusion, the study drug was diluted with normal saline (0.9% NaCl) so that the calculated dose could be infused in a total volume of 50 ml. The i.v. dose was administered at a constant rate by infusion pump for a 10-min period, and this was followed by an infusion of 50 ml normal saline which was continued for at least one-half hour after the end of the i.v. drug infusion. The oral dose was prepared from the same LSF vials and mixed with 50 ml of apple juice. After the oral drug ingestion, each subject was given an additional 50 ml of apple juice to ensure delivery of the study drug to the gastrointestinal absorptive surface. Each subject, assigned to one of three equal-sized groups, was given daily doses of LSF by 10-min i.v. infusion at doses of 1, 2 and 3 mg/kg in coordinate order for each group. Thus, each group started with a different dose on day 1. Subjects receiving 1 mg/kg LSF on the first day received 2 mg/kg on the second day and 3 mg/kg on the third day. Subjects receiving 2 mg/kg LSF on the first day received 1 mg/kg on the second day and 3 mg/kg on the third day. Finally, subjects receiving 3 mg/kg LSF on the first day received 2 mg/kg on the second day and 1 mg/kg on the third day. All subjects received 6 mg/kg as the oral dose on day 4. Samples were obtained for LSF levels at 0, 5, 10, 20, 30 and 45 min, and 1.5, 2, 3, 4, 6, 8, 12 and 24 hr after the start of the 10-min infusion. Samples for serum lipid levels were obtained at 6 hr after the first i.v. infusion, and 6, 12 and 24 hr after the oral dose. All 12 subjects received the three i.v. doses, and 11 subjects also received the oral dose. One subject experienced a hypotensive episode (80/50) with light-headedness after receiving his third i.v. dose (3 mg/kg) and was not given the oral dose. Serum was available for lipid analysis using TLC as described below for 10 of 12 patients, and for individual FFA analysis using HPLC for 11 of 12 patients.

Analytical Methods for Drug Levels

The study samples were assayed for LSF by a previously validated HPLC method. The technique quantifies LSF as well as the metabolites, pentoxifylline and the antipode of LSF; lower limit of quantitation for the assay is 5 ng/ml.

Analytical Methods for Serum Lipid Levels

Extraction of serum. Two and one-half milliliters of serum were transferred to a 50-ml silanized glass centrifuge tube and 2.0 ml of 2 M KCl/0.2 M H₃PO₄ were added, followed by vigorous vortexing. Then, 5.5 ml of chloroform/methanol (2:1,v/v) was added, followed again by vortexing and then centrifugation at 350 × g for 15 min. The upper (aqueous) phase was aspirated and discarded, and the lower (organic) phase quantitatively transferred to silanized 10 × 75 ml glass tubes, followed by evaporation to dryness under a nitrogen stream.

TLC. Each dried lower phase was resuspended in 200 µl chloroform/methanol (2:1,v/v) and 10 µl of this solution was applied 2 cm from the bottom of the plate in 1.0-cm lanes, each separated by 1.0 cm, to a 20 × 20 cm HPTLC plate (Analtech, Newark, DE). To attain maximum separation of extracted serum lipids, five sequential TLC solvent systems were applied to the plates, with drying by a hair dryer between each application. The solvent systems were: (a) chloroform/methanol/ammonium hydroxide, 45:45:3.5 (solvent front run to 8 cm from the bottom of the plate); (b) chloroform/methanol/ammonium hydroxide/water, 60:35:4:1 (10 cm); (c) chloroform/methanol/ammonium hydroxide, 90:10:1 (13 cm); (d) hexane/ether/acetone, 60:40:5 (16 cm); and (e) hexane/ether, 97:3 (run to the top of the plate and continued for an additional 5 min). Standards for neutral lipids and phospholipids (Avanti; Matreya, Inc., Pleasant Gap, PA), including monoacylglycerols, DG, TG, CE, cholesterol, FFA, phosphatidylethanolamine, phosphatidylcholine, phosphatidic acid, phosphatidylinositol, phosphatidylserine, phosphatidylglycerol, cardiolipin, lysophosphatidylcholine and lysophosphatidic acid, were applied to TLC both separately and after mixing to determine separation of the various species. Lipids were visualized after spraying the TLC plate with a 0.1% solution of Primulin in 4:1 acetone/water (Sigma, St. Louis, MO), and then placing the plate under a long wavelength UV hand-held light for assessment. From the mixture of neutral lipids and phospholipids, bands corresponding to CE (cholesteryl linoleate), TG (tripalmitin), and FFA (linoleate, palmitate) were scraped, recovered from silica by resolubilizing in 4 ml chloroform/methanol/water 1:1:0.1 and dried again. The structures of these materials were then verified by FAB-PI and FAB-NI (cf. below). Standards and reference sera were found to yield only ions consistent with each indicated species (cf. Results). FFA were also structurally identified and quantitated by HPLC as described below.

Quantitative analysis. Individual lipid standards (Avanti, Alabaster, AL) including CE (cholesteryl linoleate), TG (tripalmitin) and FFA (linoleate, palmitate) were applied to the same HPTLC plate as the experimental lipids derived from sera, and linear standard curves were obtained, with a Storm Phosphorimager in the blue fluorescence-mode to determine UV absorption of lipid-associated Primulin. Corresponding amounts of lipid in serum samples were obtained by interpolating within the standard curves. The final values were expressed as micrograms of lipid per milliliter of sera.

Individual FFA analysis. Individual FFA were analyzed by HPLC as described previously (Bursten et al., 1996). After lipids were extracted from 2.5 ml serum, as described above, FFA were separated from phospholipids by normal-phase HPLC on a Gilson system 45 apparatus and a 4.6 mm × 30 cm Waters µ-Porasil (10-micron) silica column. The mobile phase consisted of a gradient of hexane/isopropanol/water from 3:4:0.07, v/v/v, to 3:4:0.7, v/v/v, for 10 min, at a flow rate of 1 ml/min, collecting 0.5-ml fractions (Bursten et al., 1994, 1996; Bursten and Harris, 1994). Collected FFA with retention times (R_T) of 3 to 4.5 min were dried down, resuspended in 200 µl acetone, diluted with 200 µl methanol and then derivatized by addition of 400 µl of 9-anthroyl diazomethane (Nakaya et al., 1967; Yoshida et al., 1988). Twenty microliters of this solution were then analyzed by a reverse-phase HPLC method which separates and quantifies individual derivatized anthroyl FFA. A reverse-phase C8 column (4.6 cm × 25 cm, 5 micron Spherisorb C8, Alltech Associates, Inc., Deerfield, IL) was placed in series with and preceding a reverse-phase C18 column (4.6 mm × 15 cm, 3 micron Spherisorb ODS2, Alltech Associates, Deerfield, IL). Anthroyl FFA were separated using a gradient from 70:30 acetonitrile/water (v/v) to 100% acetonitrile (Aveldano et al., 1983) for 60 min at a flow rate of 1.0 ml/min. Separated anthroyl FFA were detected by fluorescence with excitation between 305 and 395 nm and emission at 430 to 470 nm (Bursten et al., 1991, 1994; Bursten and Harris, 1994). Anthroyl FFA were quantitated against fatty acid standards (Matreya, Inc.) derivatized by the same method, with efficiencies of recovery quantitated in comparison with that of anthroyl C17:0 generated from the added internal standard heptadecanoate (C17:0). The amount of each fatty acid recovered was a linear function of the amount added in a range from 30 ng to 10 µg, with r values between 0.96 and 0.995; the efficiency of recovery was 85 to 95%.

FAB-MS of HPTLC-isolated fractions. FAB-MS spectra were acquired with a VG 70 SEQ tandem hybrid instrument of EBqQ geometry (VG Analytical, Altrincham, UK). The instrument was equipped with a standard unheated VG FAB ion source and a standard saddle-field gun (Ion Tech Ltd., Middlesex, UK) that produced a beam of xenon atoms at 8 KeV and 1 mA. The mass spectrometer was adjusted to a resolving power of 1,000 and spectra were obtained at 8 kV and at a scan speed of 10 sec/decade. 2-Hydroxyethyl disulfide was used as matrix in positive-ion mode FAB-MS (FAB-PI), and triethanolamine was used as a matrix in the negative ion mode FAB-MS (FAB-NI) (Bursten et al., 1994; Rice et al., 1994a).

Statistical Analysis

Data are presented where indicated as averages as means ± S.D. Paired or unpaired Student's t tests were used to assess differences within groups of individuals with repeated measurements. Differences between multiple groups (determined for each set of individuals receiving a different LSF dosing regimen as described above) were studied by a single-factor analysis of variance using the Student-Neuman-Keuls analysis. P values of less than 0.05 indicated statistical significance.

	Results

Top Abstract Introduction Methods Results Discussion References

Demographic data for the 12 patients enrolled in the study are listed in table 1. All measured laboratory parameters remained within normal limits in each patient and did not change from base-line values, including glucose (89 ± 6 mg/dl; P > .50 for all time points), serum cholesterol (198 ± 43 mg/dl; P > .50 for all time points), electrolytes, blood urea nitrogen and creatinine.

Pharmacokinetic parameters. Pharmacokinetics of LSF were characterized by maximal plasma concentrations obtained at the end of a 10-min i.v. infusion, a rapid distribution phase and a terminal t_1/2 of approximately 0.75 hr (fig. 1). The terminal t_1/2 values of the two principal metabolites of LSF were 0.78 and 1.17 hr, respectively. Dose-related parameters such as C_max and AUC increased linearly from 1 to 3 mg/kg (correlation coefficient = 0.9998: fig. 1; table 2), whereas t_1/2 remained constant. The mean ± S.D. for clearance was 1116 ± 206.5 ml/hr/kg and ranged between 1093 and 1135 ml/hr/kg among all i.v. dose groups. LSF proved to be poorly bioavailable by the oral route of administration, most likely because of extensive first pass metabolism. LSF was apparently well absorbed as metabolites of LSF were measured (data on file) in quantities that fell on the line extrapolated from 1 to 3 mg/kg i.v. LSF doses. Bioavailability of LSF was calculated to be 5.9% and was thus assumed to contribute minimally to the effects on lipid metabolism.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Plasma concentrations of LSF after 10-min intravenous infusions of 1 (), 2 () and 3 () mg/kg doses. The concentrations shown are the mean values at each time point for the 12 subjects. The lower limit of quantitation for the assay is 5 ng/ml. The period of infusion, between 0 and 10 min, is indicated on the abscissa ().

Serum lipid determinations. The results of a representative example of the HPTLC lipid separation experiments on one of the 10 patients studied (for whom serum was available, as indicated under "Methods") are depicted in figure 2. With this methodology, there is excellent separation between phospholipids and FFA, with phosphatidylethanolamine being the most difficult to separate. The neutral lipids, including monoacylglycerol, cholesterol, diacylglycerol, cholesteryl esters and triacylglycerols, also separate well from each other and from FFA. LSF was not expected to be present in measurable quantities 6 hr after dosing, when the samples were taken for lipid analysis. Notwithstanding, it was noted that, in this TLC system, the LSF standard separates from the lipids of interest, migrating between monoacylglycerols and cholesterol (cf. fig. 2).

View larger version (172K): [in this window] [in a new window]   Fig. 2.   HPTLC lipid separations for subject 2811, which was representative of findings for 10 subjects for whom serum was available for TLC analysis. Lanes 1, 2 and 9 to 12 contain authentic standards of indicated lipids. Lane 3 is the lipid extracted from serum obtained from the subject at time zero; materials in lanes 4 to 8 are, respectively, from 6 hr after the first i.v. dose of LSF, 6 hr after the second i.v. dose, 6 hr after the third i.v. dose, 30 hr after the third dose and 48 hr after the third dose of LSF. The closed arrow indicates FFA in each lane, and the open arrow indicates TG. The serum samples obtained 36 hr after the third i.v. dose were run on separate plates. Each set of subject samples was run twice.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Absolute and relative masses of serum FFA and TG for the 10 subjects analyzed. (A) Serum levels at time zero and at the time of maximal LSF effect. The separate subjects are identified in this and in subsequent panels of this figure as:  2801;  2802;  2803;  2804;  2807;  2808;  2809;  2810;  2811;  2812. Statistical significance by Student's t test is denoted by: oP < .001; *P < .05 (B) Ratios of the mass of FFA for each subject at each sampling point to the corresponding mass at time zero. Arrows indicate the times of i.v. LSF dosing. The inset depicts the mean (±S.D.) data. Statistical significance by t test is denoted by: oP < .01 vs. time zero; *P < 103 vs. time zero. (C) Ratios of the mass of TG for each subject at each sampling point to the corresponding mass at time zero. Arrows indicate the times of i.v. LSF dosing. The inset depicts the mean (±S.D.) data. Statistical significance by t test is denoted by: #P < .05 vs. time zero.

Determination of individual FFA in serum. The ranges of individual FFA found in study sera were equivalent to those found previously in other healthy subjects (Bursten et al., 1996). Individual FFA values after LSF treatment were compared as a ratio with the pretreatment value. The mean ± S.D. values of the ratios for all subjects at each time point are depicted in figure 4, A to D, for three common serum FFA (palmitate, oleate, linoleate) and an unusual pair of polyunsaturated serum FFA (linolenate: C18:3, + ). Each individual fatty acid (or pair) follows the pattern obtained for total FFA as determined by HPTLC. For example, palmitate (fig. 4A) declines by a mean ± S.D. of 70.8 ± 10% after the first dose and to a value of 71 ± 12% after the third dose. Serum oleate, linoleate and linolenate follow nearly identical patterns.

View larger version (37K): [in this window] [in a new window]   Fig. 4.   Time course for masses of four serum FFA, depicted with respect to the administered i.v. doses of LSF. For each FFA, the mass from serum of each subject at the specified time has been converted to a ratio with the mass at time zero, and the results from 11 subjects with serum available for analysis are calculated as means ± S.D. (A) Palmitate; (B) Linoleate; (C) Oleate; (D) ( + ) Linolenate.

	Discussion

Top Abstract Introduction Methods Results Discussion References

LSF caused a marked decrease in serum FFA, as well as a later, more modest increase in serum TG. This effect was observed within 6 hr after intravenous drug administration and appeared to be maximal after a single dose of 1 mg/kg given by 10 min i.v. infusion. The effect of LSF on serum FFA persisted for 48 hr (64 elimination half-lives) after an i.v. dose, although this was complicated by administration of a minimally bioavailable, oral dose of LSF 24 hr after the last i.v. dose. That there were no significant differences between the decreases seen after the first, second or third doses indicates that the maximal suppression of serum FFA occurred after a single dose and that the threshold for the maximal effect occurred at a dose below 1 mg/kg. Although serum C_max and AUC levels of LSF obtained after the oral dose of 6 mg/kg given the fourth day were each a modest fraction of those after any i.v. dose, because of first pass metabolism (table 2), it is nevertheless possible, given that maximal suppression of FFA might occur at a dose below 1 mg/kg, that the low level of LSF or one of its metabolites could have maintained suppression of FFA. In addition, LSF appears to nonselectively lower all major serum FFA; thus a selective effect of LSF on a particular tissue phospholipase is not suggested. The rise in serum TG was either a delayed response to the initial dose or was a cumulative effect of the three LSF doses.

The persistence of serum FFA suppression for the observed length of time may nonetheless argue for the possibility of either a prolonged effect of LSF, or a persistent stable metabolite. Because the major metabolites of LSF, including pentoxifylline and the S-enantiomer, have equally short half-lives in all systems studied heretofore (cf. above; Lillibridge et al., 1996; Lee and Slattery, in press, 1997), it is difficult to attribute the duration of serum FFA level suppression purely to a combination of these compounds. We are currently examining in vivo systems (Abraham et al., 1995) for possible variant and persistent metabolites that may account for the observed duration of LSF serum effects. Ongoing clinical studies suggest that, in any case, the measured reduction in serum FFA may serve as a surrogate marker for the pharmacodynamics of LSF.

The ultimate regulation of serum FFA levels remains poorly understood (Coppack et al., 1994; Aarsland et al., 1996). The regulation of serum FFA and TG in healthy subjects involves tissue triglyceride lipases operating on serum lipoproteins and tissue triglycerides in adipose tissue, intestine, muscle and liver. FFA absorbed in the intestine are re-esterified into TG and, as chylomicrons and VLDL, are transported through the portal circulation to the liver. There, depending on the metabolic state (i.e., levels of insulin, glucagon, epinephrine, somatotropin and the insulin-like growth factors), they are -oxidized or repackaged and transported to adipose tissue in VLDL or LDL for storage as TG. Release of FFA from adipose tissue depends on activation of the hormone-sensitive lipase, which is also controlled by insulin. Metabolism of serum TG occurs by heparin-sensitive lipoprotein lipase, which allows release and uptake of FFA into adipose tissue and muscle, followed by re-esterification into TG in fat or -oxidation in muscle. In homeostatic circumstances, there is little contribution of serum or tissue phospholipase A₂ to serum FFA levels.

The molecular site of action of LSF in FFA metabolism is currently unknown. There are a variety of potential target sites for the observed effect of LSF on serum FFA, including inhibition of one or more tissue triglyceride lipases, alteration in the effect of insulin on FFA uptake or alteration of the rate of hepatic -oxidation. It is likewise possible that the drug could regulate acyl transfer, reacylation or even (although least likely) fatty acid synthesis. We have examined the effect of LSF on heparin-regulated lipoprotein lipase (LPL) in differentiated 3T3 L1 cells by a published method (Beutler et al., 1985) and found that LSF neither directly inhibits nor stimulates the activity of this lipolytic enzyme; in addition, preincubation of 3T3 L1 cells with LSF does not result in changes in LPL activity, nor alter the reduction in activity observed with tumor necrosis factor- preincubation of the cells. Similarly, we have examined the effect of LSF on DG acyltransferase and monoacylglycerol acyltransferase in these cells (Lehner and Kuksis, 1995) and have found neither stimulation nor inhibition of these enzymes (B. Meengs, T. White and S. L. Bursten, unpublished data).

The lack of effect of LSF on serum phospholipids (particularly phosphatidylcholine and lysophosphatidylcholine), cholesterol and cholesteryl esters also argues against specific inhibitory effects on serum lecithin:cholesterol acyltransferase activity. A possible mode of action involving an increase in consumption of FFA via -oxidation could involve interaction with PPAR, particularly PPAR-, but this is unlikely because of the diffuse acute hypolipidemic (particularly hypotriglyceridemic) effect associated with PPAR activation (Moody and Reddy, 1982; Moody, 1984), as opposed to the elevation in triglycerides observed after treatment with LSF. In addition, it has recently been shown that peroxisome-proliferators are potent inducers of cyclooxygenase-2 expression, and hence promote elevation of PGE₂ (Ledwith et al., 1997); LSF, in an ex vivo system, did not by itself induce PGE₂, nor did it act synergistically or additively with IL-1 to induce PGE₂ synthesis in this system (Rice et al., 1994b), further militating against possible LSF activity as a peroxisome proliferator. It is more likely, given the feedback effects of lipids on the insulin receptor (Randle et al., 1964; Field et al., 1988; Storlien et al., 1991, 1996; Pan et al., 1995), and the time course observed, that the described decrease in FFA and increase in serum TG by LSF occur through modulation of specific insulin receptor effects on lipids, such as those described by Farese et al. (1987). This would also account for previously observed modulation of phosphatidic acid and diacylglycerol levels in selected but not all cell types (Rice et al., 1994a, b; Bursten et al., 1994). Preliminary results studying insulin and insulin-depleted systems such as cultured Islet of Langerhans cells in inflammation and diabetic Zucker rats reveal that FFA, oxidized FFA and unsaturated phosphatidic acid species are generated under these inflammatory conditions and are decreased by LSF, with amelioration of dysfunction (Bleich et al., 1996). Further studies of the effect of LSF on insulin-induced changes in hydroperoxy FFA and peroxidized phospholipids, as well as on the insulin receptor, insulin receptor phosphorylation changes induced by FFA and peroxidized FFA, are ongoing. We have found on a preliminary basis that LSF modulates oxidized FFA and glucose-induced phosphorylation of the insulin receptor in insulin-insensitive states, and that this is associated with reversal of insulin insensitivity (J. Nadler and S. L. Bursten, unpublished results).

The effects observed in this study are also consistent with those observed in prior studies of patients with systemic inflammation treated with LSF. Serum FFA is increased in conditions associated with development of acute respiratory distress syndrome, such as sepsis and trauma; the unsaturated FFA, linoleate and oleate, are increased to an even greater degree than is the saturated FFA, palmitate. This increase is believed to be mediated by a series of metabolic events, including activation of phospholipase A₂, that do not occur under the homeostatic conditions present in healthy subjects, such as those in the present study (Vadas et al., 1988). In a 13-patient phase II pharmacokinetic study of LSF in patients with septic shock, LSF treatment prevented the rise in FFA that was observed in patients treated with placebo (Bursten et al., 1996). These data suggest that the effect of LSF on circulating FFA may be capable of overriding the release of FFA observed during acute inflammatory conditions. We have examined at least two isoforms of cytosolic phospholipase A₂ with described inhibitors (Balsinde and Dennis, 1997), as well as at least one form of secretory (type II) phospholipase A₂, without finding either inhibitory or stimulatory effects of LSF (T. White, T. Wilson and S. L. Bursten, unpublished data). Exhaustive examination of phospholipase A₂ families is ongoing, but again, a unified explanation of the effects of LSF on inflammatory metabolism may center on LSF-induced utilization of acyl chains because of interaction with insulin metabolism and decreased production of peroxidized acyl chains as a result (Bursten et al., in preparation).

Unsaturated FFA and their oxidation products have been implicated in the pathophysiology of acute organ dysfunction during acute inflammation (Keen et al., 1991; Marshall and Lands,1986; Rao et al., 1995). It is possible that the oxidized linoleate derivatives hydroxy-octadecadienoic acid(s) and HPODE are of importance in inflammatory conditions such as sepsis or hypoxia-reperfusion (Zimmermann,1995; Goode et al., 1995). Lipid peroxidation has been closely correlated with secondary multiple organ dysfunction in sepsis (Goode et al., 1995). Therefore, it is possible that one mechanism through which LSF protects against tissue damage during inflammation is through reduction in oxidizable circulating FFA, thus preventing amplification of oxidative injury. Insofar as formation of hydroxy and hydroperoxy derivatives of linoleic acid depend on concentrations of free linoleate (Claeys et al., 1988; Camacho et al., 1995), the action of LSF in diminishing available oxidizable FFA substrate may be one means by which its protective effects are exerted. Recently it has been suggested that linoleate peroxidation may be responsible for the impairment of insulin action observed during fatty acid infusion (Paolisso et al., 1996). It will be of interest to examine the effect of LSF in models of insulin resistance to determine whether decreasing FFA results in decreased resistance.

These findings suggest that LSF has a unique profile of effects on serum lipids, possibly related to previously observed effects on acyl chain metabolism. The resulting modulation of bioactive lipids such as FFA and (potentially) FFA oxidation products may explain some of the protective effects observed for LSF in both preclinical and clinical studies. Moreover, this study raises the possibility that LSF and related compounds may be able to modulate the biological consequences associated with high levels of unsaturated FFA. Finally, the study also suggests the cogency of detailed examination of serum lipids when studying compounds and drugs with observed lipid effects in vitro.