Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Endotoxin or lipopolysaccharide (LPS), a cell wall component of Gram-negative bacteria, is responsible for the majority of septic symptoms (Zanetti et al., 1997). The common pathophysiological changes in sepsis involve altered microvascular permeability to macromolecules (McCuskey et al., 1996; Ognibene, 1997). The microvascular inflammatory response is characterized by activation of endothelium, loss of arteriolar tone, and tissue damage. LPS activates endothelial cells to a procoagulant state, increases the adhesiveness to leukocytes and platelets, and induces release of many injurious mediators including eicosanoids, cytokines, chemokines, adhesion molecules, free radicals, platelet-activating factors (PAFs), and nitric oxide (NO) (McCuskey et al., 1996). We and others have shown that s.c. or intradermal injection of LPS increases local plasma leakage in the skin of rabbits, rats, and mice (Kopaniak et al., 1980; Issekutz and Bhimji, 1982; Fujii et al., 1996a; Iuvone et al., 1998) and that the LPS-induced plasma leakage in skin is mediated by cytokines such as tumor necrosis factor (TNF)-, interleukin (IL)-1, eicosanoids produced by cyclooxygenase (COX)-2, and NO produced by inducible NO synthase (iNOS) (Fujii et al., 1996a; Muraki et al., 1996). Furthermore, the cytokine-mediated microvascular permeability change induced by LPS may be attenuated by pretreatment with low-dose LPS. This tolerance has been ascribed to enhanced serum glucocorticoid levels as a result of the effect of LPS on the adrenal gland (Fujii et al., 1996b).

Recently, Gram-positive bacteria have emerged as a cause of sepsis (Bone, 1994). In contrast with Gram-negative bacteria, Gram-positive bacteria contain lipoteichoic acid (LTA), peptidoglycan, and many toxins, all of which have been suggested to induce inflammatory responses (Murphy et al., 1998). LTA from Gram-positive bacteria was reported to induce circulatory failure in rats due to enhanced NO formation by iNOS (De Kimpe et al., 1995a,b; Kengatharan et al., 1996). The potency of LTA to induce microvascular inflammatory response has not been clarified. In this study, we compared the effect of LTA and LPS on local plasma leakage in the mouse skin. Because our previous study had shown that pretreatment with LPS induces tolerance, we also examined whether pretreatment with LTA induces tolerance as does LPS.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Animals. All protocols of the animal experiments were approved by the Animal Care Committee of the Tokyo Women's Medical University. Male ddY strain mice were obtained from Sankyo Laboratory Service (Tokyo, Japan). C57BL/6 and 129Sv mice were obtained from Jackson Laboratories (Bar Harbor, ME). Breeding pairs of iNOS knockout mice (MacMicking et al., 1995) were kindly provided by Drs. J. MacMicking and C. Nathan (Cornell University Medical College, Ithaca, NY) and J. Mudgett (Merck Research Laboratories, Rahway, NJ). F1 hybrids obtained from C57BL/6 and 129Sv crosses were used as wild-type controls for iNOS knockouts. Male ddY mice were used between 9 and 10 weeks old, and iNOS-null mice and their wild-type controls were used between 9 and 12 weeks old. The animals were housed in an air conditioned room (temperature 22 ± 2°C, humidity 55 ± 5%) with a controlled light/dark cycle (lights on 6:00 AM to 8:00 PM) and were allowed free access to food and water.

Materials. LTA from Staphylococcus aureus, LPS from Salmonella typhimurium, indomethacin HCl, N^G-nitro-L-arginine methyl ester (L-NAME), aminoguanidine hemisulfate, and prostaglandin (PG) E₂ were purchased from Sigma (St. Louis, MO); valeryl salicylate was from Cayman Chemical (Ann Arbor, MI); 6-amino-5,8-quinolinedione (LY83583) was from Calbiochem (San Diego, CA); diphenhydramine HCl was from Nacalai Tesque (Kyoto, Japan); and thioperamide was from ICN (Tokyo, Japan). 3-Bromo-5-[N-phenyl-N-[2-[[2-(1,2,3,4-tetrahydro-2-isoquinolyl-cabonyloxy)-ethyl]carbamoyl]ethyl]carbamoyl]-1-propylpyridinium nitrate (TCV309) was kindly provided by Takeda (Osaka, Japan); N-(2-cyclohexyloxy-4-nitrophenyl)methanesulphonamide (NS-398) was provided by Taisho (Saitama, Japan); and 3-[4-(2-chlorophenyl)-9-methyl-6H-thieno[3,2-f]-[1,2,4]triazolo-[4,3-a][1,4]-diazepine-2-yl]-1-(4-morphonilyl)-1-propanone (WEB2086) and 6-(2-chlorophenyl)-8,9-dihydro-1- methyl-3-[(4-morphonilyl)carbonyl-4H,7H-cyclopenta[4,5]thieno-[3, 5-f][1,2,4]triazolo[4,3-a][1,4]diazepine (WEB2170) were provided by Boehringer Ingelheim (Ingelheim am Rhein, Germany). Monoclonal mouse anti-human IL-1 antibody was supplied by Otsuka (Tokushima, Japan) and rabbit anti-mouse TNF- polyclonal antibody was purchased from Genzyme (Cambridge, MA). Indomethacin was first dissolved in ethanol and then in 50% propylene glycol to make a stock solution (Fujii et al., 1996a). Anti-TNF- and anti-IL-1 antibodies were diluted 400-fold with 0.9% NaCl and were injected in a volume of 10 ml/kg. Drugs were dissolved in nonpyrogenic saline (0.9% NaCl) to prevent endotoxin contamination.

Plasma Leakage Measurement in Mouse Skin. The vascular permeability to macromolecules in the skin was assessed in mice by measuring extravasation of Pontamine sky blue (PSB) as previously described (Fujii et al., 1994). Briefly, 5 min after an i.v. injection of PSB (50 mg/kg), LTA, LPS, or saline (0.1 ml/site) was administered s.c. into the dorsal skin of mice. After 5 min to 3 h, mice were sacrificed, and the stained skin of the injected site was excised, weighed, and minced. The skin specimen (approximately 1 g) was dispersed in 6 ml of 0.5% Na₂SO₄, and the dye was extracted by adding 14 ml of acetone. Dye concentration was colorimetrically determined at wavelength 590 nm.

Serum Corticosterone Assay. Serum corticosterone levels were measured in blood samples taken from naive ddY strain mice at 1:00 PM, 4 h after the injection of saline, LPS, or LTA. The assay was performed using a radioimmunoassay kit (Biotrak; Amersham, Piscataway, NJ).

Statistical Analysis. Results are expressed as means ± S.E. of more than five animals. Data were analyzed statistically using unpaired Student's t test. The dose-response effect of LTA was evaluated using Wilcoxon's direct calculation test.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Effect of LTA on Dye Leakage. LTA, like LPS, induced an increase in amount of leaked dye in the skin at the site of injection, which is an index of vascular permeability. Significant increases in dye leakage were observed 60 to 180 min after LTA injection (Fig. 1). LTA induced a comparable degree of increase in vascular permeability to that induced by LPS as reported earlier (Fujii et al., 1996a). In contrast, no change in dye leakage was observed in the control mice. In subsequent experiments, dye leakage induced by LTA and LPS was determined at 120 min. LTA (100-400 µg/site) induced a dose-dependent increase in the dye leakage (Fig. 2).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Time course of changes in dye leakage in mouse skin after LTA injection. LTA (200 µg/site, ) or saline (0.1 ml/site, ) was injected s.c. into the dorsal skin of mice 5 min after injection of PSB (50 mg/kg, i.v.). At the indicated times, local dye accumulation was determined colorimetrically. Symbols and vertical bars represent means ± S.E. of five experiments. **P < .01, *P < .05 versus saline.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Dose-response curve of LTA-induced dye leakage. Different doses of LTA were given to mice. Dye leakage induced by LTA () or saline () was assessed 2 h later. Symbols and vertical bars represent means ± S.E. of five experiments. *P < .05 versus saline.

Role of Autacoids in the Dye Leakage Induced by LTA. Because indomethacin, a nonspecific COX inhibitor, and NS-398, a specific inducible COX (COX-2) inhibitor, inhibited the LPS-induced dye leakage in a previous study (Fujii et al., 1996a), the effect of these inhibitors on LTA-induced dye leakage was examined. Indomethacin attenuated the LTA-induced increase in dye leakage by 40% and coadministration of PGE₂ with LTA partially reversed the inhibitory effect of indomethacin (Fig. 3). Valeryl salicylate, a selective COX-1 inhibitor, also inhibited the LTA-induced dye leakage (Fig. 4A). However, NS-398 did not inhibit the dye leakage induced by LTA (Fig. 4B).

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Effects of indomethacin and PGE2 on LTA-induced local dye leakage. Indomethacin (5 mg/kg, i.p.) was administered 35 min before LTA. LTA (200 µg/site), alone or with PGE2 (3 nmol/site), was administered s.c. and local dye accumulation was assessed 2 h later. Open columns, mice given saline (controls); hatched columns, LTA; crosshatched columns, LTA + PGE2. Columns and bars represent means ± S.E. of five experiments. *P < .05.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effect of specific COX inhibitors on LTA-induced dye leakage in mouse skin. A, valeryl salicylate (30 mg/kg, i.v.), COX-1 inhibitor, or vehicle (0.9% NaCl) was administrated 5 min before. B, NS-398 (0.1 or 1 mg/kg), COX-2 inhibitor, or vehicle (0.5% v/v propylene glycol in saline) was administered i.p. 35 min before s.c. injection of LTA (200 µg/site) or saline (0.1 ml/site); the local dye leakage was determined 2 h later. Open columns, mice given vehicle (control); hatched columns, LTA. Columns and bars represent means ± S.E. of five experiments. *P < .05.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Effect of histamine receptor antagonists on dye leakage induced by LTA and LPS in mouse skin. Diphenhydramine (H1-receptor antagonist), ranitidine (H2-receptor antagonist), or thioperamide (H3-receptor antagonist) was administered s.c. 20 min before s.c. injection of LTA (200 µg/site) or LPS (400 µg/site). Local dye leakage was determined 2 h after LTA (hatched columns), LPS (crosshatched columns), or saline (open column) injection. Columns and bars represent means ± S.E. of five experiments. **P < .01, *P < .05.

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Effect of PAF receptor antagonists on dye leakage induced by LTA and LPS in mouse skin. WEB2086 and WEB2170 were given i.p. 65 min before s.c. LTA or LPS, and TCV309 was administrated i.v. 5 min before LTA or LPS. Local dye leakage was determined 2 h after LTA/LPS. Control mice received saline. Open columns, saline controls; hatched columns, LTA-treated mice; crosshatched columns, LPS-treated mice. Columns and bars represent means ± S.E. of five experiments. *P < .05 versus LTA alone. #P < .05 versus LPS alone.

Role of NO in the Dye Leakage by LTA. In our previous study, the LPS-induced dye leakage was inhibited by L-NAME (nonspecific NOS inhibitor), aminoguanidine (specific iNOS inhibitor), and LY83583 (guanylate cyclase inhibitor) (Fujii et al., 1996a). In contrast, the effect of LTA was not altered by these inhibitors (Fig. 7). Furthermore, LTA induced dose-dependent increases in dye leakage in both iNOS-deficient mice and the wild-type controls to a similar extent (Fig. 8).

View larger version (42K): [in this window] [in a new window]   Fig. 7.   Effects of L-NAME, aminoguanidine, and LY83583 on LTA-induced dye leakage in mouse skin. L-NAME (10 mg/kg i.v.) and aminoguanidine (10 mg/kg, i.v.) were given 5 min before LTA, whereas LY83583 (10 mg/kg, i.p.) was administered 35 min before LTA. Local dye leakage was determined 2 h after LTA (200 µg/site). Open column, saline controls; hatched columns, LTA. Columns and bars represent means ± S.E. of five experiments.

View larger version (21K): [in this window] [in a new window]   Fig. 8.   LTA-induced dye leakage in iNOS-deficient mice. Dye leakage induced by increasing doses of LTA in iNOS-deficient () and wild-type control mice (F1 of C57BL/6 × 129Sv) () was assessed 2 h after LTA injection. Symbols and vertical bars represent means ± S.E. of five experiments. *P < .05, **P < .01 versus no LTA.

Role of TNF- and IL-1 in LTA-Induced Dye Leakage. The antibody against TNF- or IL-1 was administered before LTA treatment. As shown in Fig. 9, LPS-induced dye leakage was significantly attenuated by both TNF- and IL-1 antibodies. However, dye leakage induced by LTA was not affected by any of these antibodies.

View larger version (41K): [in this window] [in a new window]   Fig. 9.   Effect of antibodies against TNF- and IL-1 on LTA- or LPS-induced dye leakage in mouse skin. Anti-TNF- antibody (dilution 1:400; 10 ml/kg) or anti-IL-1 antibody (dilution 1:400; 10 ml/kg) was administered s.c. 24 h before the experiment. Mice received LTA (200 µg/site, hatched columns), LPS (crosshatched columns), or saline (0.1 ml/site, open columns), and local dye leakage was determined 2 h later. Columns and bars represent means ± S.E. of five experiments. *P < .05 versus no antibodies.

Effect of Preconditioning with Systemic Administration of LTA or LPS. Preconditioning with systemic LPS administration 4 and 24 h before topical LPS treatment attenuated LPS-induced but not LTA-induced dye leakage (Fig. 10). In contrast, preconditioning with LTA did not confer tolerance against subsequent topical challenge with either LTA or LPS (Fig. 11). Instead, the LTA-induced dye leakage was enhanced in mice pretreated with systemic LTA.

View larger version (33K): [in this window] [in a new window]   Fig. 10.   Effect of LPS pretreatment on dye leakage induced by LTA or LPS. Four (A) and 24 h (B) after a single injection of LPS (0, 0.05, 0.1, and 0.15 mg/kg, i.p.), mice received LTA (200 µg/site, hatched columns), LPS (400 µg/site, crosshatched columns), or saline (0.1 ml/site, open columns). The local dye leakage was determined 2 h later. Columns and bars represent means ± S.E. of five experiments. #P < .05.

View larger version (38K): [in this window] [in a new window]   Fig. 11.   Effect of LTA pretreatment on dye leakage induced by LTA or LPS. Four (A) and 24 h (B) after a single injection of LTA (0, 0.05, 0.1, and 0.2 mg/kg i.v.), mice received LTA (200 µg/site, hatched columns), LPS (400 µg/site, crosshatched columns), or saline (0.1 ml/site, open columns). The local dye leakage was determined 2 h later. Columns and bars represent means ± S.E. of five experiments. *P < .05, **P < .01.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Subcutaneous injection of LTA, a cell wall component of Gram-positive bacteria, caused local dermal dye leakage. The onset of LTA-induced dye leakage followed a slow course as in LPS-induced dye leakage. Because the dose response of LTA was similar to LPS, the potency of LTA in inducing increased microvascular permeability seems to be comparable to that of LPS. The rather slow onset of dye leakage induced by LTA indicates that LTA may increase the vascular permeability through production of secondary mediators. This study demonstrated that different mediators were involved in LTA- and LPS-induced changes in vascular permeability. This and our previous studies (Fujii et al., 1996a) indicated that LPS-induced dye leakage was mediated by eicosanoids, NO, histamine, PAF, and cytokines and that induction of COX-2 and iNOS was required. However, in LTA-induced dye leakage, major contribution of eicosanoids, PAF, and histamine but not NO, TNF-, and IL-1 was demonstrated.

Of the two types of COX, COX-2 is generally considered to be a major isoform in inflammatory cells (DeWitt et al., 1993). The contribution of COX-1 in inflammation is controversial (Langenbach et al., 1995; Smith et al., 1998). We have shown previously that both indomethacin and NS-398 inhibit the dye leakage induced by LPS, suggesting a causal role of COX-2 in the microvascular permeability change induced by LPS in mouse skin (Fujii et al., 1996a). In this study, we demonstrated that in contrast with LPS, the effect of LTA was not inhibited by NS-398. Therefore, despite the previous report of COX-2 induction by LTA in bovine endothelial cells (Abate et al., 1996), COX-2 may not be involved in the LTA-induced dye leakage in our experimental settings. Because valeryl salicylate inhibited dye leakage induced by LTA, COX-1 may play a role in the dermal microvascular permeability change induced by LTA.

Similar to LPS (Fujii et al., 1997), LTA-induced dye leakage was inhibited by diphenhydramine but not by ranitidine or thioperamide. Therefore, H1 receptors may be involved in the LTA-induced increase in vascular permeability. With this regard, Nissen et al. (1997) demonstrated that histamine was released from basophils when they were incubated with LTA in vitro. Our results suggest that LTA as well as LPS acts on dermal mast cells to release histamine. A PAF-mediated mechanism was also suggested in dye leakage induced by both LTA and LPS because a PAF antagonist (WEB2086) inhibited the vascular permeability change induced by LTA and LPS. The other two PAF antagonists inhibited the effect of LPS but not that of LTA. Because WEB2086 inhibits intracellular PAF receptors, activation of intracellular PAF receptors may be required for dye leakage induced by LTA. De Kimpe et al. (1995b) showed that WEB2086 inhibited LTA-induced circulatory and renal failure in anesthetized rats, whereas other PAF antagonists were not effective.

Experimental manipulations to determine the role of NO or cyclic GMP (e.g., pretreatment with L-NAME, aminoguanidine, or LY83583 and use of iNOS-deficient mice) did not affect the effect of LTA on microvascular permeability. These results suggest that NO, cyclic GMP, or iNOS does not play a major role in LTA-induced vascular permeability change in mouse skin. Previous study showed that the acute renal failure induced by LTA and peptidoglycan is not mediated by iNOS, whereas aminoguanidine was effective in preventing circulatory, respiratory, and hepatic failures (Kengatharan et al., 1996). Although LTA, like LPS, has been shown to induce iNOS expression in vivo and in vitro (De Kimpe et al., 1995a; Kengatharan et al., 1998), the regulation and effect of iNOS by LTA may be different among tissues.

Our results indicate that cytokines such as TNF- or IL-1 are not involved in dye leakage induced by LTA. The Staphylococcus aureus-derived LTA used in this study does not induce monokine production (Bhakdi et al., 1991; Takada et al., 1995), whereas LTA derived from Staphylococcus epidermidis has been reported to induce TNF- and IL-1 production (Wakabayashi et al., 1991). Thus, LTA from other bacterial species may activate monokine production, which may potentially induce vascular permeability change.

We have previously reported that a single pretreatment with LPS transiently inhibits LPS-induced plasma extravasation and that LPS induces cross-tolerance against topical TNF--, IL-1- and IL-6-induced increase in vascular permeability (Fujii et al., 1996b). In this study, however, LPS pretreatment did not induce cross-tolerance against LTA-induced increase in vascular permeability. Moreover, LTA pretreatment failed to confer tolerance against subsequent challenge with LTA or LPS. These results suggest that systemic administration of LTA does not stimulate mechanisms associated with LPS-induced preconditioning such as endogenous glucocorticoid and NO (Fujii et al., 1996b). Activation of the pituitary-adrenocortical axis is proposed as a mechanism of tolerance induction (Evans and Zuckermann, 1991; Szabó et al., 1994; Ziegler-Heitbrock, 1995; Fujii et al., 1996b). Exogenous dexamethasone prevents the effect of LPS on vascular permeability (Fujii et al., 1996a). We found that LPS increased serum levels of corticosterone in mice, whereas LTA did not. Failure of LTA to induce the release of endogenous corticosteroids may partially explain the inability of LTA to induce tolerance. Instead, pretreatment with LTA potentiated plasma extravasation elicited by subsequent LTA challenge due to unknown mechanisms. Thus, LTA may not be useful as a prophylactic therapy of sepsis.

Our results showed a difference in mediators to induce vascular permeability change and in the ability to induce tolerance between LTA and LPS. LPS appears to activate cells via CD14 and Toll-like receptor (TLR) 2 and TLR4 (Yoshimura et al., 1999). TLR2 has been identified as a signal transducer for LTA (Schwandner et al., 1999). Both TLR2 and TLR4 are expressed in dermal endothelial cells and monocytic cells (Zhang et al., 1999). Such differential activation of these TLRs and subsequent signaling cascades by LTA may be responsible for the pharmacological characteristics of LTA on skin microvasculature.