INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Lipid A from Porphyromonas gingivalis is chemically characterized by its unique components of branched and relatively long fatty acids (15 to 17 carbon atoms) (11), which are not present in enterobacterial lipopolysaccharide (LPS), which consists mainly of 3-hydroxytetradecanoic acid (17, 18). Although the biological activity of P. gingivalis lipid A is moderate (26), it has been found to induce splenocyte mitogenicity and tumor necrosis factor alpha (TNF-) release by peritoneal macrophages to the same extent in non-LPS-responsive C3H/HeJ mice as in LPS-responsive mice (27). The participation of LPS in the activation of non-LPS-responsive mice was suggested in a previous study by the following observations. De-1-O-phosphorylated P. gingivalis lipid A showed partial loss of TNF- induction activities in the peritoneal macrophages of both LPS-responsive and non-LPS-responsive mice, and de-O-acylated lipid A showed complete loss. The TNF- induction activity was suppressed by an LPS-specific antagonist, succinylated lipid A precursor. Furthermore, the relative TNF- induction activity of the intact and treated lipid A compounds was similar to the relative Limulus gelation activity of these preparations (27).

In the present study, the action of P. gingivalis LPS on mice was examined in vivo by using galactosamine sensitization to the lethality of LPS and the induction of tolerance to it in order to obtain further evidence that the LPS and not a contaminating protein is the portion responsible for the activation of non-LPS-responsive mice at the whole-body level. For this reason, a protein-LPS complex derived from Pseudomonas aeruginosa was used as a control.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Endotoxins. LPSs from Salmonella abortus equi (Salmonella choleraesuis subsp. choleraesuis pathovar abortus equi) and P. gingivalis were extracted from the acetone-dried cells with hot phenol-water (28); digested with RNase A (Sigma Chemical Co., St. Louis, Mo.), DNase I (Sigma), and proteinase K (Sigma) (20); and then purified by repeated ultracentrifugation (105,000 × g, 12 h, 15°C) (11). A protein-LPS complex obtained from P. aeruginosa was prepared from bacterial autolysate with toluene and was purified by DEAE-Sephadex A-50 ion-exchange column chromatography (1, 24, 25). The partially purified protein-LPS complex contained approximately 80% protein.

Mice. Female C3H/HeJ mice (Nihon Kurea, Tokyo, Japan) and C57BL/6 mice (Nihon SLC, Shizuoka, Japan), more than 10 weeks old, were used for the lethality test.

Lethal toxicity test. The lethality test was performed as previously described (6). Since the activity of P. gingivalis LPS was not very strong, 30 mg of D-galactosamine per mouse was used to induce greater sensitization in non-LPS-responsive C3H/HeJ mice. Test samples in 0.1 ml of pyrogen-free water were injected intravenously into the mice immediately after intraperitoneal administration of D-galactosamine (Sigma) in 0.5 ml of pyrogen-free phosphate-buffered saline.

Induction of tolerance to the lethal toxicity of LPS in galactosamine-sensitized mice. Tolerance to the lethal toxicity of LPS in galactosamine-sensitized mice was induced by treatment with LPS prior to challenge with LPS and galactosamine. A 30-mg dose of D-galactosamine was injected intraperitoneally 90 min after the treatment with LPS, and the mice were subsequently intravenously challenged with LPS.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Lethal toxicity of P. gingivalis on galactosamine-sensitized LPS-responsive and non-LPS-responsive mice. The lethal toxicity of P. gingivalis LPS was tested in galactosamine-sensitized LPS-responsive C57BL/6 mice and non-LPS-responsive C3H/HeJ mice. LPS from S. abortus equi was used as a control. In the absence of galactosamine, 33 and 100% lethalities were obtained in C57BL/6 mice each intravenously injected with 100 and 1,000 µg of S. abortus equi LPS, respectively, and 0% lethality was obtained with injections of 1,000 µg of P. gingivalis LPS (Table 1). Treatment of C57BL/6 mice with 30 mg of D-galactosamine dramatically increased their sensitivity to the lethal toxicities of S. abortus equi and P. gingivalis LPS, resulting in 100 and 33% lethalities with 10 and 1 ng of S. abortus equi LPS, respectively, and in 100 and 17% lethalities with 1 and 0.1 µg of P. gingivalis LPS, respectively (Table 1). Treatment with galactosamine, therefore, increased the sensitivity of the mice to S. abortus equi and P. gingivalis LPS by factors greater than 100,000 and 10,000, respectively.

Induction of tolerance by endotoxin to lethality in galactosamine-sensitized LPS-responsive C57BL/6 mice. Pretreatment with LPS is known to render mice tolerant to a lethal combination of LPS and galactosamine within 1 h, (3, 5). This phenomenon was used to test the induction of tolerance to the lethality of the LPS-galactosamine combination described above by pretreatment with P. gingivalis LPS. The results obtained with LPS-responsive C57BL/6 mice are shown in Table 3. When 0.1 µg of Salmonella LPS was administered to C57BL/6 mice 90 min before galactosamine was administered, no lethality was caused, even at 10-µg doses of both Salmonella and P. gingivalis LPS; lethality (33%) was first observed at a 100-µg dose of S. abortus equi LPS, again confirming that the degree of tolerance produced in this model is dose dependent (3). When mice were each pretreated with 1,000 µg of S. abortus equi LPS, none died, even when challenged with 100 µg of Salmonella LPS. Furthermore, when the mice were pretreated with 10 µg of P. gingivalis LPS, tolerance to a subsequent challenge with a lethal dose of LPS and galactosamine was induced, and no lethality was observed when mice were challenged with 10 µg of both Salmonella and P. gingivalis LPS plus galactosamine. However, when mice were pretreated with 1,000 µg of Salmonella LPS or 10 µg of P. gingivalis LPS 30 min before galactosamine sensitization, all of the mice died at the doses of 1 ng and 1 µg of Salmonella and P. gingivalis LPS, respectively, showing that tolerance was not induced under these conditions. A protein-LPS complex from P. aeruginosa also induced tolerance by pretreatment and protected galactosamine-sensitized mice from the lethal actions of both Salmonella and P. gingivalis LPS.

P. gingivalis LPS induction of tolerance to lethality in galactosamine-sensitized non-LPS-responsive C3H/HeJ mice. The ability of P. gingivalis LPS to induce tolerance to its own lethal effects in non-LPS-responsive C3H/HeJ mice was tested in the galactosamine-sensitized model. The results are shown in Table 4. When non-LPS-responsive C3H/HeJ mice were pretreated with 10 µg of P. gingivalis LPS, they became tolerant to challenge with a lethal dose of 10 µg of P. gingivalis LPS in the galactosamine sensitization system, showing that tolerance to lethal toxicity of the LPS was induced by P. gingivalis LPS. No mice died, even when challenged with 100 µg of LPS. Pretreatment with 1 µg of P. gingivalis LPS was found to be adequate to induce tolerance to its own lethality in these mice. When LPS and galactosamine challenge was performed 30 min after pretreatment with P. gingivalis LPS, none of the mice were protected from death, indicating that tolerance had not yet been induced at this stage, which was also the case for the LPS-responsive mice described above.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, LPS from P. gingivalis was found to show marked lethal toxicity to the same extent in galactosamine-sensitized C3H/HeJ mice as in LPS-responsive C57BL/6 mice. The lethal dose of P. gingivalis LPS in C3H/HeJ mice treated with 30 mg of galactosamine (80% lethal at 1 µg/mouse) was almost the same as that in LPS-responsive C57BL/6 mice (17 and 100% lethal at 0.1 and 1 µg/mouse, respectively). Furthermore, pretreatment of C3H/HeJ mice with P. gingivalis LPS induced tolerance in the mice to the lethal action of the same LPS when it was administered together with galactosamine 90 min later, and they were completely protected from death, even at a dose that was 100-fold greater than the lethal dose of LPS. Tolerance was not induced when P. gingivalis LPS was injected into mice 30 min before the LPS-galactosamine challenge, showing the typical mode of action of LPS-induced tolerance observed in galactosamine-sensitized LPS-responsive mice (3, 5). Neither sensitization nor induction of tolerance in galactosamine-sensitized mice was expressed by Salmonella LPS. These findings in the in vivo experiments, in addition to the in vitro activities observed in the previous study (27), provided further strong evidence that P. gingivalis LPS plays an essential role in the activation of non-LPS-responsive C3H/HeJ mice. The findings, however, were not in accordance with a previous study in which P. gingivalis LPS was found to induce anaphylactoid but not lethal shock in galactosamine-administered C3H/HeJ mice (22). The discrepancy is partly due to the different doses of galactosamine used. In the present study, a relatively higher dose was used to obtain higher sensitization than that (16 mg) in the previous study. Furthermore, LPS preparations used in these two studies may be different as a result of culture conditions. The chemical structure of P. gingivalis lipid A was observed in another study (15), in which the major lipid A structure was described as a partial structure, whereas it was not recognized in our preparation monitored by mass spectrometry (11).

Freudenberg and Galanos found that galactosamine-treated non-LPS-responsive mice exhibited high sensitivity to the lethal toxicity of whole killed bacteria (4), indicating that gram-negative bacteria contain other components in addition to LPS that are capable of inducing lethal shock. However, they also showed that gram-negative bacteria are much more toxic to LPS-responsive mice than to non-LPS-responsive mice and that the lethality of gram-negative bacteria is much greater than that of gram-positive bacteria, suggesting the additive effect of unknown substances in combination with LPS. The results of the present study differ in this regard. The lethality of P. gingivalis LPS was almost the same in both LPS-responsive and non-LPS-responsive mice. Furthermore, the relative lethal activities of Salmonella and P. gingivalis LPS were similar to those seen in splenocyte mitogenicity, TNF- induction activity in macrophages, and Limulus gelation activity, as shown in the previous study (27). These facts also corroborate the possibility that the activity of P. gingivalis LPS in non-LPS-responsive mice was due to the LPS itself.

Since the most likely candidate for the stimulator of non-LPS-responsive mice in bacteria is protein (2, 9, 14, 21), the protein contaminating LPS preparations had always made interpretation of LPS actions in mice complicated. In order to directly demonstrate the participation of bacterial protein or the LPS of P. gingivalis in the activation of C3H/HeJ mice, it is desirable to isolate pure protein or LPS from the bacteria. However, it is impossible to obtain LPS-free protein, just as it is impossible to obtain protein-free LPS. Instead of the component from P. gingivalis, a protein-LPS complex derived from P. aeruginosa consisting of about 80% protein and 20% LPS was used in the present study to examine the action of bacterial protein in LPS-responsive mice. It displayed both strong lethality and induction of tolerance to the lethal action of Salmonella LPS in galactosamine-sensitized LPS-responsive mice, but it exhibited neither lethal toxicity nor induction of tolerance to the P. gingivalis LPS challenge in galactosamine-sensitized C3H/HeJ mice. These results indicate that the bacterial protein does not play a central role in the lethality for galactosamine-sensitized C3H/HeJ mice and that the unidentified bacterial component predicted by Freudenberg and Galanos that induces lethal shock in LPS-galactosamine model mice is not a protein.

Lipid A from P. gingivalis has been found to consist of unique branched fatty acids with longer carbon atom chains (15 to 17) than those of Salmonella minnesota and other enterobacteria (11). The fact that P. gingivalis LPS stimulates C3H/HeJ mice suggests that unique fatty acid components of lipid A and the position of their substituents may be involved in the stimulation of C3H/HeJ mice and that the cells, including B cells and macrophages, discriminate between subtle differences in the chemical structure of lipid A.

Differential recognition of the chemical structure of lipid A by human and murine cells has also been observed. A lipid A precursor (lipid IV_A, or 406), which acts as an agonist in murine cells and exhibits strong lethality in mice (7, 23), has no endotoxicity in human cells and antagonizes LPS action (8, 10, 12). Although we still do not know how the stimulation pathway differs in LPS-responsive and non-LPS-responsive mice or in humans and mice, there must be a special event that discriminates the response or lack of response in these mice and regulates the response, depending on the chemical structure of lipid A.

Recently, interesting studies directly related to the understanding of LPS action have been reported. Poltorak et al. 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), predicted to result in a replacement of proline with histidine at position 712 of the protein (16). Furthermore, they also found that C57BL/10ScCr mice are homozygous for a null mutation of TLR4.

In humans, five homologues of TLRs have been identified. Among them, TLR4, a constitutively active version, has been found to activate NF-B and induce expression of inflammatory cytokines and costimulatory molecules (13, 19). Yang et al. have demonstrated that TLR2 is a signalling receptor in human cells that is activated by LPS in a response that depends on LPS-binding protein and is enhanced by CD14 (29). These studies indicate that the TLR is closely involved in signal transduction in the myeloid cells of both humans and mice by LPS stimulation. The present study should be developed further, taking these points into consideration. However, in order to establish the nonresponsiveness of the mice to endotoxin, including a B-cell response, further studies on the genetic and biochemical bases of this problem are required. P. gingivalis LPS may help in the elucidation of specific cell determinants of endotoxin activity.

Here, I have shown that P. gingivalis LPS plays an essential role in the stimulation of C3H/HeJ mice. However, the possible participation of a low level of contaminating protein in the activation of the mice cannot completely be ruled out, and the possibility still remains that it acts as a cofactor with LPS in activating mice and not as an active center. Further experiments are required to clarify the above points, including the total chemical synthesis of lipid A.