Carnitine and congeners as regulators of tumour necrosis factor

In the past few years the steady increase in septic shock has been fuelled by the augmented use of invasive medical technology, immunosuppressive drug therapy, and increased longevity of patients with secondary immunodeficiencies. The septic shock syndrome probably represents the eects of many mediators on vasculature, myocardium and other systems. Endotoxin has been postulated to mediate the early stages of mediator cascade during septic shock, and inhibiting its eect has become an underlying principle of certain therapies. The presently available anti-endotoxin monoclonal antibodies seem effective only in patients with Gram-negative bacterial infections. However, there is usually no time to distinguish this sub-group of individuals from others with diEerent infections before the start of therapy and one of the central problems with antibody-based therapies--the lack of understanding of how they work--is still unsolved. Tumour necrosis factor might be a direct or indirect mediator of many of the cardiovascular responses associated with sepsis. Hypotension, hypothermia, hypoglycaemia, acidosis, focal hepatic necrosis, and acute interstitial pneumonitis with neutrophil trapping have been observed following experimental injections of high doses of TNF. As a mediator of inflammation, TNF triggers the release of a large body of cytokines, interleukin 1 included, and of arachidonic acid metabolites. Experimental data have suggested that TNF could be an appropriate target for the treatment of septic shock, but work in animals up to now has shown that anti-TNF antibodies are not very effective in focal sepsis. Recently, many investigators independently from each other reached the conclusion that carnitine and congeners can behave as modulators of TNF production. The results of their studies were the subject of a symposium held in Pomezia, Rome, on 28 January, 1993 and are published in Mediators of Inqammation. The information presented in this volume must be considered as an effort to stimulate scientists’ creative thought and clinicians’ awareness of the therapeutic potential of carnitine and carnitine congeners for the treatment of septic shock.


Introduction
Endotoxins are constituents of the outer membrane of Gram-negative bacteria. Isolated endotoxin administered into experimental animals elicits a large spectrum of biological activities which are also manifested during Gram-negative septic shock.
Endotoxins are lipopolysaccharides (LPS). In Enterobacteriaceae and in many cases of other Gram-negative bacteria, LPS are found to consist of three covalently linked regions, the lipid A, the core oligosaccharide and the O-specific polysaccharide. The structure and composition of the O-polysaccharide is highly variable among Gramnegative bacteria, determining the serological specificity of the parent bacterial strain. The core oligosaccharide is less variable in its structure and composition, a given core structure being common to large groups of bacteria. Lipid A is structurally the least variable part of the LPS molecule, exhibiting a similar structure and composition among many Gram-negative bacteria (for reviews see References 1-3). All three parts of the LPS molecule are immunogenic, eliciting the formation of antibodies interacting specifically with distinct epitopes in the respective region. The biological activity of LPS resides solely in the lipid A, the polysaccharide being devoid of toxic activity. 2 Table i summarizes the large spectrum of biological activities that were found to be expressed by purified LPS or isolated free lipid A. As seen from the table the activities of endotoxin are not always harmful but some of these, such as induction of tumour necrosis and adjuvant activity, can be beneficial to the host.
The biological activities of LPS are not direct effects of the LPS molecule but are induced indirectly by endogenous mediators that are produced following interaction of endotoxin with The activity of endotoxin may be influenced (enhanced or suppressed) by a number of plasma proteins of the host, which are capable of binding LPS. These include high and low density lipoproteins (HDL and LDL), 11---14 LPS-binding protein (LBP) lb and, in addition, specific antibodies directed against the LPS serotype in question and which may be present in the individual host. In the case of endotoxin shock resulting from infection, the toxic activity of the LPS released from the infecting micro-organism may be influenced additionally by bacterial proteins (e.g. Omp A) with which the released LPS may be associated. 16 Sensitivity to endotoxin is genetically determined, rabbits, swine and humans being highly sensitive; while mice, rats or guinea-pigs are, in comparison, much less sensitive. Mice, depending on strain, usually succumb to the lethal activity of 200--400 #g of LPS. Endotoxin-resistant strains of mice have been identified which are insensitive to all LPS effects. 7'8 These are usually referred to as LPS nonresponders and are designated as lpsC, in contrast to mice with normal sensitivity (LPS responder) which are designated as lps .The high resistance of lps d mice is due to a genetic defect in the LPS gene locus present on chromosome 4.18 Endotoxin hypersensitivity _Although sensitivity to endotoxin is genetically determined it has been known for many years that the sensitivity of normal healthy animals may be considerably increased under different experimental conditions. The most important of these are listed in Table 2. Thus treatment of experimental animals with hepatotoxic agents such as D-galactosamine will increase their sensitivity to the lethal effects of endotoxin more than 100000-fold. A significant degree of sensitization may also be achieved following treatment of mice with muramyl dipeptide (MDP), a partial structure of peptidoglycan. Further, adrenalectomy hypophysectomy or exposure to a hyperthermic environment will all enhance considerably the sensitivity to endotoxin. 9 The sensitivity of mice to endotoxin was also found to be increased by a number of growing tumours. Thus Lewis lung carcinoma, and the EMT6 sarcoma growing in C57BL/6 and BALB/C mice, respectively, were shown to increase considerably the endotoxin sensitivity of the animals 2 (Table 3).
Sensitization to endotoxin also proceeds following treatment with live (infection) or killed bacteria. 21 Sensitization to LPS may be also achieved by sub-lethal infection as shown in Fig. 1. 22 Mice (C3H/Tif), infected with a sublethal inoculum of S. thyphimurium exhibit enhanced sensitivity to endotoxin. Sensitization becomes evident on day 2 after infection, reaches a maximum on days 7 to 8 and decreases again reaching normal levels several weeks later. Fig. 1 also shows that sensitization to LPS by sublethal infection is at the same time a sensitization to TNF0.

Mechanisms of endotoxin hypersensitivity induced by bacteria
The property of bacteria to enhance endotoxin sensitivity is not confined to S. typhimurium, but is a general phenomenon observed with dierent live or killed Gram-negative and Gram-positive bacteria. An example of this is shown in Table 4.
Mice made hypersensitive to the lethal eects of I,PS by bacteria are found, on LPS challenge, to produce considerably more TNF0 than do normal animals. 25 This is shown in Table 5  Thus, administration of anti-IFNy monoclonal antibodies to mice pre-treated with bacteria inhibited the overproduction of TNF0 (Fig. 3) and abolished the development of sensitization to the lethal activity of LPS (Table 7). 6000 C57BL/10 ScSn mice were treated with heat killed P. acnes (500/g i.v., 7 days before challenge), or infected with S. typhimurium (50 CFU i.p., 3 days before challenge). Human TNF was administered i.v.
LPS, in bacteria sensitized animals, would alone explain hypersensitivity. However, the high endotoxin sensitivity of bacteria treated mice is not due only to an overproduction of TNF0. Mice sensitized to the lethal effects of LPS by bacteria are also found to be hypersensitive to the lethal activity of TNF0. This is shown in Table 6. Therefore the hypersensitivity to the lethal effects of endotoxin seen in bacteria sensitized mice is based on (a) an overproduction of TNF0, and (b) a higher sensitivity to the lethal effects of TNF. The induction of hypersensitivity by Gramnegative bacteria is of special interest since these micro-organisms also produce endotoxin. The present results make it evident that Gram-negative bacteria not only produce endotoxin but also sensitize the infected organism to its toxic action, and therefore enable a better understanding of the hazardous consequences of Gram-negative infections.
Mechanism of the sensitization to endotoxin by bacteria Interferon gamma, a mediator of the bacteria-induced sensitization: Very recently a breakthrough in the understanding of the mechanism by which bacteria sensitize the organism to endotoxin was achieved. When mice are infected with live, or treated with killed, Gram-negative or Gram-positive bacteria, they are found to contain in their serum significant amounts of interferon gamma (IFN). 2s The production of IFN2 following treatment with bacteria is true for all strains of mice that are sensitive to endotoxin. The observation was made however that a similar treatment of LPS resistant (lpsd) strains of mice (see introduction) with bacteria does not lead to IFNy production. 26 This observation suggested that the inability of lps d mice to be sensitized by bacteria might be due to their inability to produce IFNy. This possibilit} was investigated closely, and evidence could be obtained sooo-  Protection against endotoxin shock Even since LPS was recognized as the main toxic component of Gram-negative bacteria, research groups all over the world have been searching for eective ways for treating and preventing endotoxin shock. One approach investigated extensively has been the use of antibodies to different regions of the LPS molecule. Of special interest have been antibodies directed towards parts of the LPS molecule that are common among clinically relevant Gram-negative micro-organisms. Other approaches include the use of cortisone, antibodies to the LPS receptor and the use of LPS receptor antagonists.
Recently, the authors investigated whether carnitine congeners might exhibit a protective effect against the lethal action of LPS. In these experiments both -carnitine and acetyl4-carnitine were used. The lethality models used include mice sensitized to the lethal activity of LPS by D-GaIN and by Propionibacterium acnes, as well as normal mice. In approximately 50% of the experiments a protection was seen in both sensitization models. Even where no protection was found in terms of survival, a prolongation of survival was always evident. The following tables show the results of typical experiments in which protection was found. Table 8 shows that administration of 5 mg acetyl--carnitine/mouse, 30 min prior to a lethal challenge with LPS and D-GaIN afforded complete protection to the animals. A similar protection was seen when instead of LPS, recombinant hTNF0 and D-GaIN were used for challenge (Table 9). An investigation of the time of acetyl--carnitine administration that yields optimal protection   (C57BL/6) mice were treated with 500 #g killed Propionibacterium acnes i.v. 7 days before challenge. Acetyl-L-carnitine was administered i.p. h before LPS (i.v.). revealed that the drug afforded maximum protection when administered 1-2 h before LPS/GalN challenge (Table 10). A protection by acetyl4carnitine was also seen in mice sensitized by P. acnes and challenged with LPS (Table 11). More experiments are being carried out in order to confirm the protection seen so far and to make a preliminary identification of the possible mechanisms involved. Mice were C57B'L/6 strain; LPS was from S. abortus equi; all injections were administered i.v.; acetyl-L-carnitine was administered 30 min prior to LPS/D-GalN Mice were C57BL/6 strain; LPS was from S. abortus equi; all injections were administered i.v." acetyl-L-carnitine was administered 30 min prior to LPS/D-GalN