NK T cells stimulated with a ligand for TLR2 at least partly contribute to liver injury caused by Escherichia coli infection in mice

2.1 Contribution of Fas L on NK T cells to liver injury after E. coli infection

We have previously reported that Fas L expression was selectively induced on NK T cells in liver 7 days after Salmonella infection 23. To determine whether Fas L expression was induced on NK T cells in vivo by E. coli infection, we analyzed the expression of Fas L on intrahepatic lymphocytes of C57BL/6 mice 12 h after intraperitoneal inoculation with E. coli at a dose of 1.0×10⁹ CFU/mouse by flow cytometry. As shown in Fig. 1, the level of Fas L expression on NK T cells was apparently increased after E. coli infection, although the relative number of NK T cells in whole liver lymphocytes was decreased (Table 1). On the other hand, the expression level on NK cells or T cells did not significantly increase after infection. Thus, Fas L was preferentially up-regulated on NK T cells after E. coli infection.

To investigate the involvement of Fas L on liver NK T cells in liver injury following E. coli infection, we next compared the serum glutamic-pyruvic transaminase (GPT) levels in Jα281^–/– and gld/gld mice with non-functional Fas L expression and those in B6 control mice after infection of E. coli. The serum GPT levels in Jα281^–/– and gld/gld mice were significantly lower than those in control B6 mice at 12 h after inoculation with 1.0×10⁹ CFU of E. coli (Fig. 2, p<0.05).

The tissues of liver of E. coli-infected Jα281^–/–, gld/gld, and control B6 mice were obtained at 12 h after infection. Histological examination stained with hematoxylin and eosin confirmed that necroinflammatory foci, cell death of hepatocytes characterized by cell shrinkage and chromatin condensation, and lymphocyte infiltration were obviously less in Jα281^–/– and gld/gld mice than in control mice after infection (Fig. 3A). Apoptotic cells were next identified using the TUNEL technique. After E. coli infection, higher numbers of cells undergoing apoptosis could be observed in liver of control mice. The apoptotic cells showed characteristic shrinking, chromatin clumping, and nuclear fragmentation. Most of the events in the liver were located in the cells lining the sinusoids. In contrast, the amount of apoptotic cells remained low in Jα281^–/– and gld/gld mice (Fig. 3B). These results suggest that the Fas/Fas L system mediated by liver NK T cells could contribute to liver injury following E. coli infection.

2.2 Similar levels of bacterial burden in livers and cytokines in serum of Jα281^–/– mice and control mice after E. coli infection

It is possible that the difference in susceptibilities of the liver to E. coli-induced injury is due to a difference in bacterial burdens in Jα281^–/–, gld/gld, and +/+ mice. To test this possibility, we examined the numbers of bacteria in organs after infection with a sublethal dose (1×10⁸ CFU: 1/5LD₅₀) of E. coli. The bacteria in organs were eliminated in the similar kinetics in the liver, spleen and peritoneal cavity of Jα281^–/– and gld/gld mice after infection with 1×10⁸ CFU of E. coli. The numbers of bacteria in organs were similar in each group of mice 12 h after E. coli infection (Fig. 4). These results suggest that the degree of liver injury may not be due to a difference in bacterial loads in Jα281^–/–, gld/gld, and +/+ mice.

Inflammatory cytokines are known to be involved in liver injury induced by LPS or Gram-negative bacterial infection 3, 5. Thus, we next examined kinetics of serum cytokines (TNF-α, IL-10, IL-12, and IL-1β) after E. coli infection. Jα281^–/– and C57BL/6 mice were inoculated with E. coli at a dose of 1.0×10⁸ CFU/mouse. Cytokine levels in the serum were determined by ELISA. As shown in Fig. 5, there were no significant differences in kinetics and peak levels of serum TNF-α, IL-10, and IL-1β concentrations between Jα281^–/– mice and those in C57BL/6 mice, but the serum level of IL-12 in Jα281^–/– mice was significantly lower than that in B6 mice at 3 h after E. coli infection (p<0.05). Serum IL-4 or IFN-γ was not detected in either Jα281^–/– or B6 mice at any stage after E. coli infection (data not shown).

2.3 Expression of functional TLR2 by NK T cells in the liver

We previously reported that intrahepatic NK T cells freshly isolated from naive mice expressed significant levels of TLR2 messenger RNA (mRNA) and TLR2 signaling is suggested to contribute to liver injury induced by Salmonella infection via Fas L induction on NK T cells 23. To determine whether NK T cells express functional TLR on their surface, we studied the in vitro responses of intrahepatic lymphocytes to a ligand for TLR2 (lipoprotein) or for TLR4 (lipid A) by assessing induction of Fas L on the cell surface. Intrahepatic lymphocytes of naive C57BL/6 mice were treated with 10 μg/ml of synthetic lipoprotein or lipid A for 12 h, and the expression of Fas L on intrahepatic lymphocytes was analyzed using a flow cytometer. As shown in Fig. 6, the Fas L expression on NK T cells apparently increased after stimulation with a ligand for TLR2 (lipoprotein), whereas the levels of Fas L on NK T cells increased to a lesser extent after lipid A stimulation. In contrast, Fas L expression on NK and NK1.1^– T cells did not increase in response to either lipid A or lipoprotein. Thus, TLR2 expressed on NK T cells may function to induce Fas L expression on the cell surface in response to its ligand.

2.4 Lipoprotein and LPS act additively in inducing liver injury

To investigate whether in vivo administration of LP could induce liver injury via a Fas/Fas L-dependent manner in mice, C57BL/6 and gld/gld mice were given an i.p. injection of saline, 10 μg of LP, 10 μg of LPS, or 10 μg of LP plus 10 μg of LPS in 200 μl PBS. Twelve hours after injection, sera were collected and serum GPT activities were determined (Fig. 7A). The serum GPT levels were moderately increased at 12 h after inoculation of LP or LPS alone. Notably, LP in combination with LPS synergistically elevated serum GPT levels. This effect of LP and LPS on induction of liver injury was not seen in gld/gld mice. Interestingly, the synergistic effects of LP and LPS on the serum TNF-α level was not seen at any stage after injection (data not shown) These results indicate that LP and LPS act additively in inducing liver injury in vivo in a Fas/Fas L-dependent manner.

We next investigated whether administration of LP and/or LPS act in Fas L expression on liver NK T cells. The expression of Fas L on intrahepatic lymphocytes was analyzed using a flow cytometer 12 h after stimulation with 10 μg/ml of synthetic lipoprotein, LPS, or LP plus LPS. Analysis gates were set on NK T cells. As shown in Fig. 7B, the Fas L expression on NK T cells apparently increased after stimulation with LP and further increased after LP plus LPS. In contrast, Fas L expression on NK and NK1.1^– T cells did not increase in response to any stimulation (data not shown).

We have previously reported that TLR2 gene transcription was up-regulated by TLR4 signaling via NFκB activation 24, 25. To address the mechanisms for additive effect of LP and LPS on induction of liver injury in Fas/Fas L-dependent manner, we examined the expression of TLR2 mRNA in intrahepatic lymphocyte subsets, including NK, NK T, and T cells. Intrahepatic lymphocyte subsets were purified from these mice 12 h after inoculation with LP, or LP plus LPS by cell sorting. Total RNA was extracted from each population, and the expression of TLR2 mRNA was analyzed by semiquantitative RT-PCR. As shown in Fig. 7C, TLR2 mRNA was expressed more abundantly in NK T cells of mice challenged with combination of LP and LPS than in NK T cells with LP, whereas TLR4 mRNA level remained constant in each subset (data not shown). These results indicate that LP and LPS additively induce enhanced liver injury with increased Fas L expression on NK T cells via up-regulation of TLR2 genes.

4.1 Animals and microorganisms

Jα281^–/– mice with the C57BL/6 (B6) background were kindly provided by Dr. Taniguchi (Chiba University). Age- and sex-matched B6 wild-type mice (as normal controls) and gld/gld (gld) mice with non-functional Fas L expression with the B6 background were purchased from Japan SLC (Hamamatsu, Japan). These mice were bred in our institute under specific pathogen-free conditions. Six-to-eight-week-old female mice were used for the experiments. All experiments were carried out according to the criteria in the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Science. E. coli (American Type Culture Collection No.26; Manassas, VA) was inoculated intraperitoneally in all experiments.

4.2 Antibodies and reagents

PE-conjugated anti-NK1.1 mAb, biotin-conjugated anti-TCRα β mAb and Cy-chrome-conjugated streptavidin were purchased from PharMingen (San Diego, CA). Anti-Fas L hamster IgG mAb was purchased from Medical & Biological Laboratories (Nagoya, Japan). FITC-conjugated mAb to hamster IgG was purchased from ICN Pharmaceuticals (Aurora, Ohio). Synthetic E. coli-type lipid A, ONO4007, was kindly provided by Ono Pharmaceutical (Tokyo, Japan) and has been described 29. Synthetic lipoprotein, Palmitoyl-Cys((RS)2,3-di(palmitoyloxy)-propyl)-Ala-Gly-OH, was purchased from Bachem (Bubendorf, Switzerland). LPS (purified from E. coli 055:B5) was purchased from Sigma (St. Louis, MO).

4.3 Preparation of liver lymphocytes

Fresh liver was immediately perfused with sterile HBSS through the portal vein to wash out all remaining peripheral blood and then passed through stainless-steel mesh. After the coarse pieces had been removed by centrifugation at 50 g for 1 min, the cell suspensions were again centrifuged, resuspended in 8 ml of 45% Percoll (Sigma Chemical Co., St. Louis MO), and layered on 5 ml of 67.5%Percoll. The gradients were centrifuged at 600 g for 20 min at 20°C. Lymphocytes at the interface were harvested and washed twice with HBSS.

4.4 Flow cytometry analysis

For Fas L-staining, liver mononuclear cells isolated from C57BL/6 mice were first stained with hamster IgG anti-Fas L mAb and successively stained with FITC-conjugated mAb to detect hamster IgG. The cells that had been stained with anti-Fas L mAb were stained with PE-conjugated anti-NK1.1 mAb and biotin-conjugated anti-TCRβ mAb. To detect biotin-conjugated mAb, cells were stained with Cy-Chrome-conjugated streptavidin. To block Fc receptor (FcR)-mediated binding of the mAb, 2.4G2 (anti-FcR mAb) was added. All incubation steps were performed at 4°C for 30 min. The stained cells were analyzed with a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA). The cells were carefully gated by forward and side light scattering for the live lymphocytes. The data were analyzedusing FACSCalibur research software (Becton Dickinson).

4.5 Assay for GPT activity

Blood samples were collected at the time of sacrifice. Serum GPT activity was determined using a serum transaminase C II test kit (Wako Pure Chemical Industries, Osaka, Japan) according to themanufacturer's instructions. GPT activities (international units per liter) were calculated from the standard curve.

4.6 Histological studies

Liver specimens were removed and fixed with 10% buffered formalin, paraffin-embedded, and stained with hematoxylin and eosin for light microscopic examination. In situ terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay was performed using an in situ apoptosis detection kit (Apoptag, Intergen, Purchase, NY). All steps were performed according to the instructions of the manufacturer. Briefly, paraffin-embedded sections were deparaffinized and rehydrated and then permeabilized by incubation with 20 mg/ml proteinase K (Sigma, St. Louis, MO) for 15 min at 37^oC. Inactivation of endogenous peroxidases was accomplished by immersing the tissue sections in 3% hydrogen peroxide diluted in methanol for 10 min at room temperature. Then, the sections were incubated in equilibration buffer for 5 min. The sections were then incubated with the labeling solution containing terminal deoxynucleotidyl transferase in a humidified chamber for 1 h at 37^oC. The reactions were terminated by rinsing the sections in a stop/wash buffer. The incorporated digitonigen-dUTP was detected by incubation with anti-digoxigenin peroxidase at room temperature for 30 min and positive reactions were revealed using 3,3′ diaminobenzidine. Methyl green was used for counterstaining for nuclei.

4.7 Bacterial growth in organs

Peritoneal exudates were obtained from the peritoneal cavity by lavage with 5 ml of HBSS at various times after E. coli infection. The liver was then perfused with 8 ml of sterile HBSS to wash out bacteria in the blood vessels. The bacterial counts in the homogenates of the liver and spleen or peritoneal exudates were established by plating serial 10-fold dilutions in sterile distilled water on tryptic soy agar (Nissui Laboratories, Detroit. MI). Colonies were counted 24 h later, after incubation at 37°C.

4.8 Cytokine assays

Serum was collected at various times after E. coli infection and TNF-α, IL-10, IL-12, IL-1β, IFN-γ, and IL-4 levels in serum were determined by enzyme-linked immunosorbentassay (ELISA) using Genzyme mAb according to the manufacturer's instructions (Genzyme, Cambridge, MA).

4.9 Sorting of intrahepatic lymphocyte subsets

The intermediate intensities of TCR (TCR^int) NK1.1⁺ T cells, TCR^– NK1.1⁺ cells and TCR⁺ NK1.1^– T cells were purified as NK T cells, NK cells and T cells, respectively, in the B6 strain from isolated intrahepatic lymphocytes by cell sorting using a FACSVantage (Becton Dickinson) electric cell sorter. The purity of sorted cells was morethan 97% in all experiments.

4.10 Expression of TLR genes

Messenger RNA of liver tissue, purified NK cells, NK1.1⁺ T cells, NK1.1^– T cells, and J774A.1 cells (a macrophage cell line as a positive control for TLR2) was extracted by using TRIZOL reagent (Gibco BRL, Rockville, MD), and cDNA synthesis was performed as described 22. Synthesized cDNA was amplified by polymerase chain reaction (PCR) using 20 pmol of each of the primers specific for TLR2 or β-actin. The primers were as follows: TLR2 sense, 5′-AAAGTCTAAAGTCGATCCGC-3′; antisense, 5′-ATATGCAACCTCCGGATAGT-3′; β-actin sense, 5′-TGGAATCCTGTGGCATCCATGAAAC-3′; antisense, 5′-TAAAACGCAGCTCAGTAACAGTCCG-3′. The PCR cycles were run for 1 min at 94°C followed by 1 min at 54°C and 1 min at 72°C. Adenaturing step for 5 min at 94°C was included before the first cycle. The PCR product was subjected to electrophoresis on a 1.8% agarose gel (Life Technologies Laboratories, Grand Island, NY) and visualized by ethidium bromide staining.

4.11 Statistical analysis

Data were analyzed by Student's t-test, and a Bonferroni correction was applied for multiple comparison. The value of p<0.05 was considered statistically significant.