Research Paper Mediators of Inflammation 2, S43-S50 (1993)

THE effect of L-carnitine and some of its acyl derivatives on serum TNF production and lethality in a murine experimental endotoxin shock model was investigated. In some instances, serum IL-6 production was also evaluated. In this experimental model, C57BL/6 mice received 30 mg/kg LPS (E. cell 055:B5) injected intraperitoneally, while L-carnitine and its derivatives were administered according to different schedules. Serum levels of TNF and IL-6 were evaluated I h following LPS injection. The treated animals were also monitored daily for differences in body temperature, feeding, and survival for 10 days after LPS injection. Although some derivatives were able to significantly affect TNF production, the marked decrease in serum TNF levels of LPS-treated mice was not paralleled by a substantial increase in survival.


Introduction
Septic shock is a clinical syndrome associated with a high lethality rate, and characterized by various haemodynamic and immuno-biochemical alterations. This disease is primarily caused by Gram-negative bacteria, but it may also be consequent to infections by some Gram-positive bacteria, fungi, and probably parasites and viruses. 2 The pathophysiological effects caused by Gramnegative bacteria have been ascribed to lipopolysaccharide (LPS) a component of the bacterial outer membrane, which has also been termed endotoxin because of its toxic effects. 3 LPS causes septic shock by interacting with various components of the host's immune system, and primarily with macrophages, which in turn release several endogenous mediators that are themselves the ultimate effectors of the disease. [4][5][6] Recently, the systemic release of large amounts of various cytokines has been associated with fatal outcome in human septic shock. 7,8 TNF is one of the cytokines considered to play a pivotal role as a mediator of the host's response to LPS. Therefore, blocking or antagonizing TNF in sepsis may have therapeutic potential. Furthermore, making TNF a target for intervention, rather than LPS, may be more advantageous in that TNF has also been found to be involved in the pathogenesis of shock due to Gram-positive bacteria. 1 '11 Among the drugs used in giving metabolic support to septic patients, t-carnitine has (C) 1993 Rapid Communications of Oxford Ltd been found to be a fairly good candidate to ameliorate the host's metabolic response to ep{ic [12][13][14] processes.
L-Carnitine is a drug that is not only essential for the mitochondrial oxidation of long-chain fatty acids, but also for intercompartmental shuttling of energy substrates as well as elimination of toxic metabolites, and modulation of the free CoA/Acyl-CoA ratio, is Although the information concerning the effects of carnitine on cells of the immune system is still scanty, a report showed that L-carnitine effectively inhibited chemiluminescence in PMA-stimulated human PMN leukocytes. 16 This is of some interest in the light of the findings that both chemiluminescence as well as TNF production are phenomena negatively regulated by cyclic nucleotides. 17 '18 In this report the authors investigated the effect of L-carnitine and some of its acyl derivatives on endotoxin-induced serum TNF production, lethality, and some physiological responses (feeding and body temperature) in a murine experimental endotoxin shock model. Experimental endotoxaemia has become a valuable experimental model for septicaemia in laboratory animals. Although this model does not completely reproduce all the features of clinical septic shock, the accumulated evidence shows that it reliably mimics Gram-negative sepsis. 3 Another cytokine, IL-6, has recently been implicated in the pathophysiology of septic shock, 9 and a report has also shown that it may play a role in endotoxin treated mice. 2 In some instances, therefore, the effects of L-carnitine on serum IL-6 levels in our model of endotoxin shock were also evaluated.

Materials and Methods
Animals: Male C57BL/6 mice were purchased from Iffa-Credo (Lyon, France). Mice (6-7 weeks of age at the time of use) were housed in groups of eight, and were allowed food and water ad libitum. They were kept under specific pathogen-free conditions, and were randomly assigned to various treatment groups. Drugs and chemicals: L-Carnitine as well as its derivatives were all synthesized by Sigma Cell lines and tissue culture media: The TNF-sensitive L929 cell line (a murine fibrosarcoma) was grown in RPMI-1640 medium containing 25 mM HEPES and 2 mM L-glutamine (Biochrom KG, Berlin, Germany) supplemented with 10% heat-inactivated Myoclone foetal calf serum (FCS) (Gibco, Grand Island, NY, USA), and 50/ig/ml gentamicin (Sigma). The murine hybridoma cell line B9, which is dependent on IL-6 for growth 21 was maintained in the above culture medium with the further addition of 5.0 x 10 -s M 2-mercaptoethanol (Sigma) and 10 U/ml recombinant murine interleukin 6 (rmIL-6). Endotoxaemia model: C57BL/6 mice received 30 mg/kg LPS intraperitoneally (i.p.), which had been predetermined to produce approximately 60% lethality. In the majority of experiments, animals were first pre-treated i.p. with L-carnitine derivatives 1 h before LPS injection, and then treated intravenously (i.v.) 10 min following LPS administration. However, other treatment protocols were also adopted, and they are reported in the relevant tables. Some animal physiological responses to endotoxaemia were graded such as food intake, body weight, mobility, amount of ocular exudate, and rectal temperature.
In particular, temperature measurements were taken with an electronic thermometer (model Babuc/M, 1.S.I., Milan, Italy) with the appropriate thermistor (probe TM35S). Animal health state was classified as follows: very ill (body temperature S44 Mediators of Inflammation. Vol 2 (Supplement) 1993 below 32C, absence of mobility, marked body weight loss, abundant ocular exudate); fair (body temperature between 32 and 35C, poor mobility, small body weight loss, moderate ocular exudate); normal (body temperature above 35C, normal mobility, normal body weight, absence of ocular exudate). All groups of animals were examined at 24 h intervals after LPS injection to assess survival, and they were followed up to 10 days. Serum TNF and IL-6 determinations: To determine circulating TNF and IL-6 levels, ether-anaesthetized mice were bled by retro-orbital sinus puncture at 1 h following 30 mg/kg LPS injection. Blood was allowed to clot at room temperature, and then centrifuged at 1 000 x g for 15 min. The separated serum was stocked and stored at -80C until assayed.
TNF bio-assay: Concentrations of TNF in serum were determined by using the L929 cytotoxicity assay described by Flick  FCS were seeded into each well of a flat-bottomed 96-well microtitre plate (Falcon, Becton & Dickinson, Meylan Cedex, France) and incubated overnight at 37C in a humidified atmosphere of 5% COg. After incubation, spent medium was discarded and two-fold serial dilutions (carried out in RPMI-1640 1% FCS) of serum samples were added to the cells in the presence of actinomycin D-mannitol at a final concentration of 1 /lg/ml. The L929 cell cultures were then further incubated for 18h. Following incubation, the medium was discarded, and the plates were first washed twice with 0.9% NaC1 and then stained for 15 min at room temperature with 0.5% (w/v) Crystal Violet in 20% ethanol and 8% formaldehyde. After discarding the stain, the wells were gently washed in running tap water; the dye which had been taken up was eluted with 33% acetic acid and the relative absorbance at 540 nm was measured with a Multiskan MCC/340 ELISA reader (Flow Laboratories, Mclean, VA, USA). TNF activity is expressed in units/ml, and is defined as the reciprocal of the dilution necessary to cause death of 50% of the actinomycin-treated control cells. Final TNF concentrations for each sample were calculated by interpolating the values of four to six sample dilutions. IL-6 bio-assay: Levels of IL-6 in serum were measured by its ability to induce proliferation of the murine B-cell hybridoma line B9. 21 B9 cells were washed twice prior to use in the assay in order to remove the exogenous rmIL-6 added to sustain their growth.
Briefly, two-fold serial dilutions of thawed sera in RPMI-1640 supplemented with 10% FCS, 5.0 x 10-5M 2-mercaptoethanol were added to flat-bottomed 96-well microtitre plates containing 5.0 x 103 cells per well. Cells were then incubated for 96 h at 37C in a humidified atmosphere of 5% CO2, and pulsed with 0.5/Ci/well of 3H-TdR during the last 18 h of incubation. Afterwards, the cells were harvested onto glass fibre filtermats (Pharmacia, Turku, Finland) by using an automated cell harvester (Pharmacia), and 3H-TdR incorporation (an index of cell proliferation) was assessed by a /g-plate counter (Pharmacia). IL-6 activity is expressed in B9 hybridoma growth units per ml, and one unit is defined as the reciprocal of the sample dilution required to produce one halfmaximal proliferation. Values for each sample were calculated by interpolating the sample cpm of four to six dilutions by means of linear regression analysis.
Statistical analysis: Statistical analysis of TNF and IL-6 values was performed by using the Student's t-test, while survival data were assessed by Fisher's exact test.

Results
Physiological responses to endotoxin: LPS intraperitoneal injection in C57BL/6 mice (30 mg/kg body weight) provoked a considerable serum TNF release ( Fig.  1). Systemic TNF secretion started as early as 0.5 h after LPS administration reaching a peak at I h, and then declined to undetectable levels within the next 4 h. Twenty-four hours following LPS challenge, animals exhibited decreased body weights paralleled by concomitant hypothermia, which lasted up to 3 days before they started to recover (Fig. 2). Furthermore, the animals given endotoxin showed an abundant ocular exudate, a dramatic decrease in food intake as well as a loss of mobility and, when touched, a comatose-dazed behaviour.        (Table  5).
Isovaleryl-L-carnitine and six derivatives: Among the compounds belonging to this class, only two derivatives, namely ST 687 and ST 1037, were effective in significantly reducing (p < 0.05) serum TNF levels (  Fig. 3.

Discussion
Septic shock is an increasingly serious health problem in hospitals, especially in intensive care units, despite the use of multiple antibiotics, surgical drainage, and intervention including vasopressor and metabolic support. It was estimated that this disease causes 100000 deaths annually in the United States. 23 It is generally accepted that the primary mediator of the pathophysiological changes occurring in Gramnegative sepsis is LPS, which in turn stimulates host macrophages to produce a cascade of endogenous mediators responsible for many alterations in host physiology. There are many reports showing that Treatments were performed by injecting LPS 30 mg/kg i.p. at time zero and administering the compounds, at the dose indicated, as follows: 1) -60 min i.p. and + 10 min i.v., 2) from day -1 to day + 3 i.p. twice a day (once at day 0); 4) -60 min i.p. Statistical significance (p) was evaluated over-production of several cytokines such as IL-1, IL-6, TNF, and IFN-2 is associated with severe sepsis. Among these cytokines, however, TNF has been shown to play a key role in many metabolic derangements occurring during septic shock, 24 and its abnormally high serum levels are often correlated with poor prognosis. 2s'26 Moreover, TNF can also be found in the serum of animals experimentally intoxicated with LPS and additionally causes, when injected in animals, a toxic syndrome indistinguishablegrom endotoxaemia. 2v-29 Experimentally induced endotoxin shock in laboratory animals, although it does not exactly reproduce all the alterations of septic shock, has become a valuable and very convenient model for studying sepsis. In fact, it reliably mimics many physiological as well as immunological dysregulations occurring in this disease. 24 By using this model, the authors found that LPS injection in mice caused a dramatic reduction in body temperature. This decrease was evident no sooner than 24 h following LPS administration and 16  argued that hypothermia is a typical physiological response of mice to LPS and appears to be independent of the strain used. In the present model, moreover, smaller decreases in body temperature were observed when lower doses of LPS (down to 0.3 mg/kg) were employed, while a dose as low as 0.03mg/kg was completely ineffective (data not shown). Furthermore, the decrease in body temperature was paralleled by a concomitant body weight loss, which appeared to be related to the dose of LPS (data not shown). The maximum decrease of body weight was observed 48 h after LPS challenge, then the mice slowly began to recover their normal weight. Taken together, these data indicate that body temperature and body weight are valuable parameters to monitor health conditions of mice during endotoxaemia. In a preliminary experiment, performed to ascertain the time course of TNF production, it was found that serum TNF peaked as early as 1 h following LPS injection, and became undetectable by 4 h. This result is consistent with similar studies reported by other investigators in different animal models sand in humans. 2 In this experimental model L-carnitine and some acyl derivatives were tested for possible protective effects in LPS-challenged mice. A role of L-carnitine in the treatment of sepsis was first suggested by Border et al., who hypothesized that some septic processes might be associated with an impairment of lipid oxidation due to systemic depletion of I-carnitine. 3u Support to this hypothesis was lent by different investigators, who observed a significant improvement in survival of LPS-challenged animals following treatment with t-carnitine. 39'4 In this LPS shock model, L-carnitine and its derivative ST 784 were able to reduce serum TNF levels, although this decrease was not always statistically significant. Similarly, the compounds derived from propionyl-L-carnitine and isobutyryl--carnitine were effective in reducing serum TNF release in LPS-challenged mice. In contrast, only two isovaleryl-L-carnitine derivatives, namely ST 687 and ST 1037, were found to significantly decrease TNF levels. However, L-carnitine and isovaleryl-t.-carnitine, which differently affected serum TNF levels, were both able to improve health conditions of LPS-injected mice and to lightly reduce lethality as well. Different treatment schedules and different dosages may be necessary to ascertain the real therapeutic potential of these compounds in counteracting LPS intoxication.
The authors found that serum TNF and IL-6 levels were modulated differently by treatment with L-carnitines. In fact, in most cases where serum TNF was significantly decreased, concomitant significantly high levels of serum IL-6 were observed. This is of interest in the light of the suggestion that IL-6 may act as a negative modulator of TNF production. '4 In addition, IL-6 was found to induce hyporesponsiveness to endotoxin in mice, 42 and to inhibit the release of TNF in mice as well as in human U937 cells and monocytes in vitro. 4 Furthermore, it is known that TNF and IL-6 synthesis is regulated differently by various agents. High c-AMP levels, for example, were reported to inhibit TNF synthesis while inducing IL-6 production. 44-4s Additionally, Marcinkiewicz 46 observed that the production of these cytokines by murine peritoneal macrophages was affected differently by prostaglandins PGE2 and PGI2. In fact, while both prostaglandins inhibited the release of TNF, they increased IL-6 production at the same time.
It would be tempting to speculate, therefore, that -carnitines might increase prostaglandin synthesis and c-AMP levels, the latter either directly or via prostaglandins, thereby decreasing TNF synthesis and increasing IL-6 levels. IL-6 in turn would further dampen the production of TNF. '4 Although additional investigations are needed, reports showing that -carnitine and some of its congeners were able to positively affect prostaglandins release from macrophages 4 support this hypothesis.
Although TNF release has been .reported to be a prominent event during endotoxaemia, 2's'6'4 the authors' results also point out that TNF by itself cannot be the only crucial mediator of lethality. In fact, it was found that the significant reductions in TNF levels brought about by several L-carnitine derivatives were not correlated with a significant reduction of lethality, though prolongation of mean survival time (data not shown) and ameliorated health conditions of treated animals were observed. The involvement of other critical factors in addition to TNF is supported by several studies showing that TNF by itself is necessary, although not sufficient, to cause lethality during endotox-aemia4, 8,49, s0 Even though many questions concerning the ultimate molecular mechanism of action of t-carnitines in endotoxaemia are yet to be answered, Mediators of Inflammation-Vol 2 (Supplement) 1993 S49 it is envisaged that these compounds may be helpful, when associated with conventional therapy, in that they can effectively reduce TNF levels and ameliorate the host's metabolic processes. In conclusion, these results would justify clinical trials with L-carnitine in the treatment of sepsis. Indeed, preliminary reports 12,51 indicate that early supplementation of L-carnitine in the therapy of septicaemia appears to counteract the onset of the metabolic cascade leading to septic shock.