INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

In the mediation of pathophysiological changes such as fever and shock in the course of severe gram-negative bacterial infection, the involvement of a bacterial endotoxin has been postulated (21). An endotoxin is chemically a lipopolysaccharide (LPS), which is an important structural component of the outer surface membrane of gram-negative bacteria (22). LPS elicits an extraordinary variety of distinct biological effects, such as pyrogenicity, adjuvanticity, macrophage activation, B-lymphocyte mitogenicity, and tumor regression (21).

Since the biological activity of LPS depends on the chemical structure of its lipid A portion, investigation of the relationship between chemical structure and biological activity is of great importance. In fact, many investigations of endotoxins have been performed by using native and chemically synthesized lipid A (8). Although a substantial amount of data has been accumulated regarding the relationship between the structure and the activity of lipid A, many problems still remained unsolved. Owing to its chemical structure, the potency of LPS as an endotoxic agent ranges from highly toxic (e.g., LPS from enteric bacteria such as Escherichia coli or Salmonella spp.) to nontoxic (e.g., LPS from Rhodobacter sphaeroides [20, 25] or Rhodobacter capsulata [18]). In general, the potent endotoxins exhibit strong activity in all the assay systems, and nontoxic LPS expresses no activity either in vitro or in vivo. However, recent studies have revealed that cells of different species respond in different ways to specific lipid A derivatives. For example, lipid A precursor structure and Salmonella-type lipid A activate mouse cells but are inactive in human cells (6, 29, 32). These facts indicate that the cells (or receptor molecules) from different species discriminate between slight differences in the chemical structure of lipid A. Understanding the biological activity of lipid A is, therefore, not a simple task.

Recently, we have found that lipid A derived from Porphyromonas gingivalis induced splenocyte mitogenicity and tumor necrosis factor alpha (TNF-) release from peritoneal macrophages in LPS-unresponsive C3H/HeJ mice to the same extent as in LPS-responsive mice (33). Furthermore, P. gingivalis LPS induced lethal shock in galactosamine-sensitized C3H/HeJ mice and rendered them tolerant to the toxic effect of the LPS when the mice were pretreated with the same LPS (31). Lipid A from P. gingivalis is chemically characterized by the unique components of its branched and relatively longer fatty acids (15 to 17 carbon atoms) (15), which are not present in enterobacterial LPS. We have recently isolated the lipid A from Flavobacterium meningosepticum and determined its complete chemical structure (12). It was surprising that the structure of lipid A from F. meningosepticum was quite similar to that of P. gingivalis in fatty acid composition and the position of the substitution (Fig. 1). For these reasons, the study of the biological properties of F. meningosepticum lipid A is of great interest, especially in LPS-unresponsive mice. Here we report an investigation of the biological properties of lipid A. 

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Proposed chemical structure of F. meningosepticum lipid A. F. meningosepticum lipid A consists of -1,6-linked GlcN disaccharide and -1,6-linked GlcN3N-GlcN disaccharide in a molar ratio of 1.00:0.35, which carries (R)-3-hydroxy-15-methylhexadecanoic acid, (R)-3-hydroxy-13-methyltetradecanoic acid, (R)-3-0-(13-methyltetradecanoyl)-15-methylhexadecanoic acid, and (R)-3-hydroxyhexadecanoic acid at positions 2, 3, 2', and 3', respectively, and carries phosphate groups at position 1.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. F. meningosepticum strain IFO 12535 was obtained from the Institute for Fermentation (IFD) in Osaka, Japan. Recombinant TNF- standards and rabbit polyclonal antisera against murine TNF- were obtained from Asahi Kasei Kogyo, Ltd., Fuji-shi, Japan. THP-1 and J774-1 cell lines were obtained from the Japanese Cancer Research Resources Bank. Rabbit immunoglobulin G was obtained from Zymed Laboratories, Inc. (South San Francisco, Calif.). RPMI 1640 medium with glutamine and Iscove's modified Dulbecco's medium were from GIBCO Laboratories (Grand Island, N.Y.). RNase, DNase, phorbal myristate acetate, 1,25-dihydroxy vitamin D₃, and D-galactosamine were purchased from Sigma Chemical Co., St. Louis, Mo. Pyridine was purchased from Wako Chemical Co. Ltd., Tokyo, Japan. Quantitative Limulus assay reagent (Endospecy) was obtained from Seikagaku Kogyo, Tokyo, Japan. Pyrogen-free water was a product of Hikari Seiyaku, Tokyo, Japan. LPS from Salmonella enteritidis subspecies enteritidis serovar abortus equi was extracted by the aqueous phenol method (37). Lipid A was obtained as an insoluble substance after 1% acetic acid treatment of LPS at 100°C for 90 min (5). Chemically synthesized lipid A precursor 406 was a gift of Daiichi Kagaku Co., Ltd., Tokyo, Japan.

Preparation of F. meningosepticum lipid A. The method used for preparation of F. meningosepticum lipid A has been described by Kato et al. (12). LPS was extracted from the acetone-dried cells with a mixture of phenol, chloroform, and petroleum ether in a ratio of 2:5:8, (vol/vol/vol), according to the method described by Galanos et al. (4). The LPS was purified by RNase and DNase treatments and repeated ultracentrifugation at 105,000 × g for 3 h (six times). Purified LPS was hydrolyzed with 1% aqueous acetic acid at 100°C for 2 h, followed by centrifugation at 14,000 × g for 10 min. The sediment was washed three times with distilled water, and the crude lipid A was obtained after lyophilization. The crude lipid A was dissolved in 10 ml of chloroform/methanol (3:1 [vol/vol]), and purified by gel permeation chromatography using a column (2 by 100 cm) of Sephadex LH-20 (Pharmacia) with the same solvent as the eluent at a flow rate of 30 ml/h. Purified lipid A (420 mg) was thus obtained.

Succinylated lipid A precursor. Succinylation was performed according to the method described previously (26). Briefly, a suspension of 3 mg of synthesized lipid A precursor 406 (dried over P₂O₅ in a desiccator), 100 mg of succinic anhydride, and 200 µl of pyridine dried using molecular sieves (Wako Chemical Co. Ltd.) was heated in a sealed tube at 60°C for 3 h. The mixture was poured into water (2 ml; 4°C), dialyzed, and lyophilized. Five to six molecules of succinic residues were found to be substituted for the six free hydroxyl groups of lipid A precursor 406 by mass spectrometry. Succinylated precursor 406 lost all endotoxic activities and acted as an antagonist in LPS-induced TNF- production specifically in macrophages (27, 29).

Mice. Endotoxin-responsive female C3H/HeN mice (Japan SLC, Inc., Hamamatsu, Japan) and endotoxin-nonresponsive female C3H/HeJ mice (Clea Japan, Tokyo, Japan), 6 to 10 weeks old, were used for the assay of splenic mitogenicity and induction of TNF- release from peritoneal macrophages.

Induction of TNF- release from mouse peritoneal macrophages, J774-1, THP-1, and U937 cells. Mouse peritoneal macrophages were obtained by washing the peritoneal cavity with 5 ml of Iscove's medium (28). Resident and thioglycollate-elicited peritoneal macrophages were used for TNF- induction assay. Thioglycollate-elicited macrophages were harvested from mice that had been injected intraperitoneally 4 days before with 2 ml of thioglycollate medium. The cell number was adjusted to 2 × 10⁶ cells/ml. After adhesion, the cells were incubated with the stimulant for 6 h. J774-1 and THP-1 cells were grown in RPMI 1640 medium supplemented with 10% (vol/vol) heat-inactivated fetal calf serum (FCS)-50 µM 2-mercaptoethanol-5 mM HEPES-penicillin (100 U/ml)-streptomycin (100 µg/ml) in a 5% CO₂ atmosphere at 37°C. J774-1 cells were harvested by scraping with a cell scraper (Costar) and suspended in fresh medium. The cells (10⁶ cells/ml/well in 24-well dishes) were allowed to adhere to the plastic for 3 h at 37°C, washed twice with medium, and incubated an additional 4 h for TNF- induction with the stimulant. THP-1 cells (2 × 10⁵ cells/ml/well in 24-well dishes) were prepared for the experiments by adding 100 ng of phorbol myristate acetate per ml and 0.1 µM 1,25-dihydroxy vitamin D₃ to cell suspensions in RPMI 1640 medium with 10% FCS. The cell suspensions were allowed to differentiate and to adhere to plastic for 72 h at 37°C. After washing, the cells were incubated an additional 24 h with the stimulant. The cell suspensions were allowed to differentiate and to adhere to plastic for 48 h at 37°C. After washing, the cells were incubated an additional 8 h with the stimulant. The supernatant of the test samples was stored at 80°C until used to determine TNF-. Inhibition of TNF- release from the peritoneal macrophages taken from C3H/HeJ mice was performed by adding succinylated lipid A precursor 406 to the assay system prior to the addition of agonist, and TNF- production was compared with a control containing agonist alone. Inhibition was expressed as percent TNF- production, with production in the presence of agonist alone equal to 100%.

TNF- assay. The supernatant of each culture obtained was transferred to a plastic tube, the cells were centrifuged at 600 × g, and the supernatant was stored at 80°C until used to determine TNF-. The TNF- produced was measured by cytotoxicity assay against L929 murine fibroblast cells. L929 cells were grown in tissue culture flasks in RPMI 1640 medium supplemented with 10% FCS-50 µM 2-mercaptoethanol-5 mM HEPES-penicillin (100 U/ml)-streptomycin (100 µg/ml). Cells were detached with trypsin, washed, resuspended in medium at 4 × 10⁵ cells/ml, and 100-µl aliquots were plated in 96-well flat-bottomed plates (Corning Glassworks, Corning, N.Y.). After incubation for 3 to 5 h at 37°C in 5% CO₂, 50 µl of actinomycin D (4 µg/ml) in RPMI 1640 medium was added to each well, and 50 µl of test sample was then added to the wells (final volume, 200 µl/well). The results are expressed as means plus or minus the standard deviation of triplicate wells.

Mitogenicity. Mitogenicity was tested using spleen cells obtained from C3H/HeN and C3H/HeJ mice according to the method of Tanamoto et al. (35). The isolated spleen cells were mashed gently on a stainless steel mesh, passed through the mesh, and suspended in Iscove's medium. The cells were washed with the medium three times and adjusted to 4 × 10⁶ cells/ml in Iscove's medium. Two hundred microliters of the cell suspension was seeded into each well of the 96-well microplate (No. 24850-96; Corning Glassworks), followed by addition of the test sample and cultivation at 37°C for 48 h in the presence of 5% CO₂. After addition of [³H]thymidine (0.2 mCi), the suspensions were incubated further for 24 h. The cells were harvested on glass fiber filters, and radioactivity incorporated into the cells was measured in toluene-based scintillation fluid (5 ml) using a liquid scintillation counter. Data were expressed as mean counts per minute from triplicate determinations.

Induction of nitric oxide synthesis. Generation of nitric oxide was tested using J774-1 cells (7), which were prepared in the same manner as described in the TNF- assay. In the NO assay, the cells were cultured for 3 days with test samples. The centrifugal supernatant was immediately used for the NO assay. NO was measured as a stable form of NO₂ by using Griess reagent. One-hundred-microliter quantities of test samples were mixed with the same volume of Griess reagent (1% sulfanilamide/0.1% N-[1-naphthyl] ethylenediamine dihydrochloride at a ratio of 1:1 [vol/vol]) in a 96-well plate at room temperature. After 10 min, the absorbance was read at 570 nm with a microplate reader.

LAL gelation activity. Limulus amoebocyte lysate (LAL) gelation activity of test samples was estimated colorimetrically by measuring the absorbance of p-nitroaniline released from a synthetic substrate (Endospecy) in a quantitative assay. The assay was performed in 96-well flat-bottom plates (Costar) at 37°C for 30 min, and the chromogen was measured at 405 nm with a microplate reader (Thermo max; Molecular Devices) taking the absorbance at 490 nm as background. Pyrogen-free water was used for the dilution of test samples. All glassware used for the test was heated at 250°C for 3 h before use.

Lethal toxicity. Lethality testing was carried out according to the method described by Galanos et al. (2). Female 10- to 15-week-old C57BL/6 mice (Nihon SLC) were injected intraperitoneally with 12 mg of D-galactosamine-HCl in 0.5 ml of pyrogen-free phosphate-buffered-saline. The test samples in pyrogen-free water were injected by the intravenous route immediately after the administration of galactosamine. Death of the mice was confirmed on the next day of the test, and the result was expressed as the number of dead mice of the total number tested.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Induction of TNF- release from peritoneal macrophages of C3H/HeN and C3H/HeJ mice by F. meningosepticum lipid A. TNF- induction by F. meningosepticum lipid A in both resident and thioglycollate-elicited peritoneal macrophages of LPS-responsive and LPS-unresponsive mice was examined. TNF- released into the medium after lipid A stimulation was estimated by cytotoxicity against actinomycin D-sensitized L929 murine fibroblasts. As shown in Fig. 2A, F. meningosepticum lipid A as well as LPS started to induce TNF- release from thioglycollate-elicited peritoneal macrophages of C3H/HeN mice at a concentration of 100 ng of lipid A or LPS per ml, increased its activity dose-dependently, and induced maximum TNF- production (27 and 34 ng/ml, respectively) at the highest concentration tested (10 µg/ml). However, the activity of F. meningosepticum lipid A or LPS was significantly lower than that of control Salmonella serovar abortus equi LPS, which stimulated TNF- production at a concentration as low as 1 ng/ml and induced maximum TNF- production (65 ng/ml) at a concentration of 10 µg/ml. Similar results were obtained by using resident peritoneal macrophages, although the activity of F. meningosepticum LPS and lipid A was relatively weak in both the minimum stimulation dose and the maximum TNF- production (Fig. 2C). F. meningosepticum LPS and lipid A also significantly stimulated TNF- production in both resident and thioglycollate-elicited peritoneal macrophages of LPS-unresponsive C3H/HeJ mice. These LPS-unresponsive mice showed a similar minimum-lipid A stimulatory dose and similar amounts of TNF- production to that in LPS-responsive mice (Fig. 2B, D). On the other hand, no induction of TNF- release was observed with Salmonella serovar abortus equi LPS in peritoneal macrophages from C3H/HeJ mice (Fig. 2B, D).

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Induction of TNF- release from peritoneal macrophages of C3H/HeN and C3H/HeJ mice by F. meningosepticum lipid A. Thioglycollate-induced (A, B) and resident (C, D) peritoneal macrophages from C3H/HeN (A, C) and C3H/HeJ (B, D) mice were used for TNF--inducing assays. Macrophages (2 × 106 per ml) were incubated in serum-free Iscove's medium with various concentrations of LPS or lipid A. After 6 h of incubation at 37°C, the supernatants were examined for TNF-. The results are expressed as means ± SD of triplicate wells. Symbols: , S. serovar abortus equi LPS; , F. meningosepticum LPS; , F. meningosepticum lipid A.

Inhibition of TNF--inducing activity of F. meningosepticum lipid A from peritoneal macrophages of C3H/HeJ mice by LPS-specific antagonist. The inhibition of TNF- induction from macrophages of LPS-unresponsive mice by succinylated lipid A precursor 406, a specific inhibitor of LPS activity, was tested using thioglycollate-induced peritoneal macrophages. TNF- induction in thioglycollate-induced peritoneal macrophages of C3H/HeJ mice by 1 µg of F. meningosepticum lipid A per ml (TNF- induction, 15 ng/ml) was almost completely inhibited in the presence of 10 µg of succinylated precursor 406 per ml (TNF- induction, 1.7 ng/ml).

Mitogenicity of F. meningosepticum lipid A. The mitogenic activities of F. meningosepticum LPS and lipid A were tested on murine splenic cells of LPS-responsive C3H/HeN and LPS-unresponsive C3H/HeJ mice. As shown in Fig. 3, Salmonella serovar abortus equi LPS exhibited the activity even at a dose of 3.7 µg/ml, and the activity increased in a dose-dependent manner in the dose range tested. The maximum incorporation of [³H]thymidine was 7,832 cpm at 11.1 µg/ml, while the minimum stimulation doses of F. meningosepticum lipid A and LPS for mitogenicity were 3.7 and 33.3 µg/ml, respectively. Thus, the mitogenic activity of both LPS and lipid A from F. meningosepticum was 10- to 100-fold weaker than that of Salmonella serovar abortus equi LPS. Moderate but significant mitogenicity was observed in the splenic cells of LPS-unresponsive mice treated with either F. meningosepticum lipid A or LPS. On the other hand, no mitogenicity was exhibited with control Salmonella serovar abortus equi LPS even at a concentration of 100 µg/ml (Fig. 3).

View larger version (18K): [in this window] [in a new window]   FIG. 3.   Mitogenic responses of spleen cells from C3H/HeN and C3H/HeJ mice to F. meningosepticum lipid A. Spleen cells from C3H/HeN (A) and C3H/HeJ (B) mice were suspended in serum-free Iscove's medium at 4 × 106 cells/ml, and 200-µl aliquots were plated in 96-well tissue culture dishes, and mitogen that was reciprocally diluted in 10 µl of water was added. After culturing the cells for 48 h, [3H]thymidine (0.2 µCi per well) was added. After additional culturing for 24 h, the cells were harvested, and the radioactivity incorporated was measured. The results are expressed as mean counts per minute ± SD of triplicate wells. Significance and P values were obtained in panel B by paired t-test: *, P < 0.05, **, P < 0.01 versus background. Symbols are as defined in the legend to Fig. 2.

Induction of TNF- and NO release by F. meningosepticum lipid A from J774-1 cells. The ability of F. meningosepticum lipid A as well as LPS to induce TNF- and NO from mouse macrophage-like J774-1 cells was compared with that of Salmonella serovar abortus equi LPS, used as a control. The J774-1 cells are very sensitive to stimulation by the control LPS, and significant production of both TNF- (92 ng/ml) and NO (15.6 µM) was observed at a concentration of 1 ng/ml (Fig. 4). F. meningosepticum lipid A and LPS started to induce secretion of both TNF- and NO from J774-1 cells at a concentration of 100 to 1,000 ng/ml, showing that the activity is about 100 to 1,000 times lower than that of the control LPS.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Induction of TNF- and NO release from J774-1 cells by F. meningosepticum lipid A. J774-1 cells (106 cells/ml per well of a 24-well dish) were incubated in RPMI 1640 medium supplemented with 10% (vol/vol) FCS with the stimulant. After 4 and 72 h of incubation at 37°C, the supernatants were examined for TNF- (A) and NO (B), respectively. The results are expressed as means ± SD of triplicate wells. Symbols are as defined in the legend to Fig. 2.

Induction of TNF- release by F. meningosepticum lipid A from THP-1 cells. It has been suggested that human cells respond to LPS in a different manner than murine cells, as seen in their response to lipid A precursor or Salmonella-type lipid A (6, 29, 32). To determine the ability of F. meningosepticum lipid A to activate human cells, human monocyte-macrophage cell line THP-1 was examined for TNF- production in response to F. meningosepticum lipid A. As shown in Fig. 5, the cells were stimulated with Salmonella serovar abortus equi LPS at a concentration of 10 ng/ml and produced 10.3 ng of TNF- per ml. F. meningosepticum lipid A and LPS started to induce secretion of TNF- from J774-1 cells at concentrations of 100 and 1,000 ng/ml, respectively, showing that their activity was about 10 to 100 times lower than that of the control LPS.

View larger version (15K): [in this window] [in a new window]   FIG. 5.   Induction of TNF- release from human THP-1 cells by F. meningosepticum LPS and lipid A. Human THP-1 cells (2 × 105 cells/ml per well) were incubated with 100 ng of PMA per ml and 0.1 µM 1,25-dihydroxyvitamin D3 for 72 h in RPMI medium containing 10% (vol/vol) FCS at 37°C. After an additional 24 h of incubation with 10 µl of test sample, the supernatants were assayed for TNF-. Values represent the mean concentration of TNF- ± SD for triplicate experiments. Symbols are as defined in the legend to Fig. 2.

LAL gelation activity. LAL gelation activity of F. meningosepticum LPS and lipid A was estimated colorimetrically by measuring the absorbance of p-nitroaniline released from a synthetic substrate. Salmonella serovar abortus equi LPS was used as a control. As shown in Fig. 6, an increase in the activity occurred in a dose-dependent manner in the range of concentrations tested (12.5 to 100 pg/ml). F. meningosepticum LPS and lipid A exhibited slightly weaker activity in this assay compared to that of the control. The doses of samples required to give an optical density of 0.15 were 29 pg of the lipid A per ml, 42 pg of the LPS per ml, and 15 pg of Salmonella serovar abortus equi LPS per ml.

View larger version (17K): [in this window] [in a new window]   FIG. 6.   Limulus gelation activity of F. meningosepticum LPS and lipid A. The LAL was incubated with test samples for 30 min, and the released chromogen was measured. Symbols are as defined in the legend to Fig. 2. O.D., optical density.

Lethal toxicity. Lethal toxicity of F. meningosepticum LPS and lipid A was tested using galactosamine-sensitized C57BL/6 mice. As shown in Table 1, a 100% death rate in mice was obtained at a dose of 10 ng of Salmonella serovar abortus equi LPS per mouse, whereas F. meningosepticum LPS and lipid A exhibited no lethality at that dose level. F. meningosepticum lipid A showed a 25% death rate at a dose of 100 ng/mouse, and 100% lethality was first observed at a dose of 1.0 µg/mouse. The LPS did not give a 100% death rate even at the highest dose tested (1.0 µg/mouse). Thus, the lethal toxicity of F. meningosepticum LPS and lipid A was at least 100-fold weaker than that of Salmonella serovar abortus equi LPS.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, free lipid A isolated from F. meningosepticum LPS was characterized biologically. The lipid A as well as the parent LPS from F. meningosepticum exhibited 100- to 1000-fold-lower activity than the active compounds of Salmonella serovar abortus equi LPS used as a control in lethal toxicity assays against galactosamine-sensitized mice; TNF- induction in J774-1 and THP-1 cells and peritoneal macrophages of LPS-responsive C3H/HeN mice; NO induction in J774-1 cells; and mitogenicity in spleen cells of C3H/HeN mice. The LPS and lipid A, however, exhibited strong Limulus gelation activity, which was nearly comparable to that of Salmonella LPS. Such variations among the biological assay systems of endotoxin were formerly observed with derivatized lipid A part-structure, acetylated precursor 406, which displays divergent action with regard to Limulus gelation activity and the other endotoxic activities of LPS. (30)

The finding that F. meningosepticum lipid A stimulates the peritoneal macrophages from LPS-nonresponsive C3H/HeJ mice and induces significant amounts of TNF- production from the cells was of great interest. Several protein molecules derived from bacteria are known to be the activators in C3H/HeJ mice (1, 9, 16, 17, 24). To minimize the effects of contaminating protein, the preparation was purified with several procedures (e.g., separation with phenol-chloroform-petroleum ether and gel permeation chromatography). As a result, the protein content of the lipid A was almost negligible (0.36% according to the results of amino acid analysis). The purified lipid A still retained strong activity in the peritoneal macrophages of C3H/HeJ mice. Although the lipid A used in the present study contained a trace amount of protein, the possibility of its participation in the macrophage activity seems to be low. First, the activation of macrophages from both C3H/HeJ and C3H/HeN cells occurred in almost the same manner. Moreover, the lipid A exhibited significant mitogenicity in spleen cells from the mice. Furthermore, we tested the effect of the LPS antagonist, succinylated lipid A precursor. It has been shown to specifically suppress the LPS action in TNF- induction from macrophages but not to affect TNF- induction by zymosan (27, 29). The activity of F. meningosepticum lipid A in TNF- induction from macrophages was effectively suppressed by the antagonist. Taking all this into consideration, the action of F. meningosepticum lipid A in C3H/HeJ mice is thought to be a result of endotoxic stimulation of the lipid A portion but not to be caused by the contaminated protein. These biological properties of F. meningosepticum are quite similar to those observed in P. gingivalis lipid A. The chemical structures of the lipid A from these two organisms resemble each other remarkably in their fatty acid composition, number, and position of substitution. Compared to enterobacterial lipid A from organisms such as E. coli and Salmonella spp. (10, 11) F. meningosepticum lipid A and P. gingivalis lipid A contain a relatively longer chain (15 to 17 carbon atoms), fewer numbers of fatty acids, and isoform fatty acids (12, 15). In addition, the phosphate group at position 4' is almost or completely lacking in lipid A. Fatty acids are known to play an essential role in the activity of endotoxin. Their number, binding site, and kind appear to be critical determinants of the capacity for activity (3, 11, 13). In fact, some of the nontoxic lipid A preparations so far found have the characteristic fatty acids with the usual diglucosamine backbone and phosphates (14, 16, 23). Therefore, the reason for the low endotoxicity of F. meningosepticum lipid A as well as that of P. gingivalis lipid A is thought to depend partially on the defect of phosphate at position 4' and primarily on the unique fatty acid composition of these lipids and their position of substitution. Moreover, considering the similarity of both the chemical structure and novel action on LPS-nonresponsive mice of the lipid A from these two organisms, the unique fatty acids possibly also play an essential role in activation of LPS nonresponsive mice. F. meningosepticum lipid A contains a hybrid backbone consisting of the usual GlcN-GlcN disaccharide and additional -1,6-linked GlcN3N-GlcN disaccharide. The unusual lipid A backbone detected in the lipid A, however, does not seem to control the biological activity, since P. gingivalis lipid A has no GlcN3N-GlcN heterodimer background. Furthermore, no such novel action on LPS-nonresponsive mice has been reported so far in the lipid A from Brevundimonas diminuta, Brevundimonas vesicularis, and Legionella pneumophila, having GlcN3N as a constituent of the lipid A backbone.

The fact that F. meningosepticum lipid A actually stimulates activation of macrophages in C3H/HeJ mice suggests that activation in the mice can be stimulated by the lipid A analogues with special chemical structures and that macrophages discriminate between the fine differences in the chemical structures of lipid A. However, the endotoxin recognition system of macrophages still has not been clarified. Recently, Poltorak et al. (19) found that the codominant LPS^d allele of C3H/HeJ mice corresponds to a missense mutation in the third exon of the Toll-like receptor-4 gene (TLR4), which is predicted to result in a replacement of proline with histidine at position 712 of the protein (19). Possibly, lipid A that has a unique fatty acid as a constituent may be recognized by a mutated TLR4 molecule or there may be some other molecules which transduce the endotoxic signal in the cells, including other TLR. The activation of macrophages was suppressed by succinylated precursor 406, which is known to suppress specifically the LPS action on macrophages. Therefore, inhibition seems to occur through the usual pathway.

Such discrimination between slight chemical differences in lipid A has also been reported between human and murine macrophages; That is, a lipid A precursor (lipid IV, or 406) and Salmonella-type lipid A (516) are agonists in murine cells and exhibit strong lethality for mice (36) whereas they express no endotoxicity in human cells and antagonize LPS action (6, 32). Although we still do not know the difference in the stimulation pathway between LPS-responsive and LPS-nonresponsive mice nor the difference between the pathways of humans and mice, there must be a special event which discriminates between the responsive and unresponsive pathways and regulates the response depending on the chemical structure of the lipid A. The problem must be studied further, including the genetic and biochemical issues, in order to establish the concept of unresponsiveness to endotoxin. For this research the F. meningosepticum lipid A serves as a promising tool.