Toll-like receptor 4
C57BL/10ScCr (Cr) mice carry a deletion of the Toll-like receptor 4 (tlr4) gene (i.e. they are tlr40/0) and are thus refractory to LPS effects. Insertion of wild-type tlr4 transgene into the tlr40/0 Cr germ line endowed LPS susceptibility in the two transgenic lines created, indicating that TLR4 is the only limiting factor for LPS responsiveness in Cr mice. The absolute levels of tlr4 mRNA expressed by the heterozygous transgenic (tlr4Tr/0), wild-type C57BL/10ScSn (Sn) (tlr4+/+) and heterozygous F1 (Sn × Cr) (tlr4+/0) mice varied markedly. However, the pattern of distribution of expression in the different organs was the same in all strains. In different biological assays (B cell mitogenicity, cytokine induction and lethal toxicity) the degree of LPS response obtained in the different strains of mice correlated with the levels of tlr4 mRNA expression. In macrophages, investigation of the LPS-induced cytokine (IL-6) response revealed a linear relationship between the response and the logarithm of TLR4–MD-2 levels.
Because of its potent biological activity and its relative abundance on the bacterial surface, LPS constitutes a major recognition marker by which the innate immune system senses the presence of Gram-negative bacteria. The recognition of LPS by target cells leads to a number of biological reactions, including formation and release of inflammatory mediators by macrophages. The molecular basis of LPS–host interaction and the identity of the functional LPS receptor have been under study for several decades. Today it is accepted that the interaction of LPS with LPS-binding protein in circulation and its transfer to cellular CD14 is required for the optimal activation of macrophages. CD14 is an important binding component for LPS on cells. However, because of the lack of a transmembrane domain, it is unable to transduce the LPS signal across the cell membrane.
The identification of the signaling LPS receptor was facilitated by the existence of LPS-resistant (Lpsd) mouse strains. Using these strains [C3H/HeJ, C57BL/10ScCr (Cr) and C57BL/10ScN ] 1, it was shown that LPS susceptibility is controlled by a single gene locus on mouse chromosome 4, named Lps2–4. The Toll-like receptor 4 (tlr4) gene was identified as the only candidate gene within the Lps critical region 5. TLR4 is a member of an evolutionary highly conserved family of membrane proteins, the TLR molecules, which play an important role in innate immunity in mammals, insects and plants 6, 7. All above mentioned, LPS-resistant mice carry defects in the tlr4 gene. The spontaneous mutation displayed by C3H/HeJ mice results in an exchange of proline for histidine at position 712 of the TLR4 protein 5. In Cr mice and their progenitor strain C57BL/10ScN, the entire tlr4 is deleted 5, 8. In addition, a fourth LPS resistant mouse strain C57BL/6.KB2-mnd was identified recently, carrying a large insertion within exon 2 of tlr49.
Further evidence for the essential role of TLR4 in LPS responsiveness was provided by the finding that the disruption of tlr4 by molecular targeting led to an LPS resistant phenotype of the resulting knockout mice 10. Subsequent studies indicated that TLR4 participates in LPS recognition 11, 12 and transduces the LPS signal across the membrane of target cells 13, 14. TLR4-dependent signaling is at present under intensive investigation and not completely understood yet. It has been shown that for signaling TLR4 requires the help of adapter proteins.
Three such proteins, MD-2, MyD88 (myeloid differentiation marker 88) and TIRAP (Toll-IL-1 receptor domain-containing adapter protein) have been identified in monocytes/macrophages 15–18. Mice presenting a null-mutation of the MD-2 gene were recently shown to be LPS-hyporesponsive 19. Defects of LPS responsiveness have also been reported to occur in MyD88- or TIRAP-knockout mice 20–22. Activation of TLR4 by LPS induces recruitment of MyD88 to the plasma membrane 23 and initiates a signaling cascade that leads to nuclear translocation of the transcription factors NF-κB and AP-1 24. Engagement of TLR4 induces also a MyD88-independent signal transduction pathway 25, 26. In B cells, a functional cooperation of TLR4 with another TLR protein, RP105, has been reported 27. In addition to signal initiation via TLR4 on the cell membrane, intracellular cell activation after internalization of LPS has been reported in certain epithelial cells 28, 29.
In this study, the genomic insertion of a wild-type murine tlr4 transgene rendered LPS-resistant Cr mice highly sensitive to LPS. Using tlr4+/+ C57BL/10ScSn (Sn), tlr4+/0 F1 (Sn × Cr) and two transgenic Cr lines, we show a quantitative relationship between the expression level of TLR4 and LPS responsiveness.
2.1 Presence of tlr4 gene in tlr4+/+ Sn, tlr40/0 Cr, tlr4+/0 F1 (Sn × Cr), and tlr4Tg/0 transgenic-Cr mice
The presence of the tlr4 gene in Sn, Cr and F1 mice as well as in the transgenic lines (TCr-1 and TCr-5) was analyzed by Southern blotting (Fig. 1). Cr mice showed no detectable signal, as expected. Sn and F1 mice exhibited a weak signal that corresponded to two or one copies of the gene, respectively. In TCr-5 and TCr-1 mice, stronger signals were found due to multicopy insertion of the transgene. As confirmed by comparison of the density values of tlr4 in the Southern blot analysis, the approximate copy number of the transgene was 13.4 for TCr-1 and 5.7 for TCr-5.
2.2 Expression of tlr4 mRNA in organs and macrophages
The expression of tlr4 mRNA in spleen, liver, lung, kidney and brain of Sn, Cr and F1 mice, and in the two transgenic lines, was investigated by Northern blotting. As expected, Cr mice expressed no tlr4 in their organs. In the wild-type Sn mice, a relatively strong tlr4 expression was observed in the spleen. Expression decreased in lung, still further in kidney and was not detectable in liver and brain (Fig. 2).
The same pattern of expression was observed in F1 mice and in mice of both transgenic lines. However, the absolute intensity of expression varied considerably among the mouse strains and lines. In comparison with Sn mice, expression was generally found to be lower in F1 mice. tlr4 expression in TCr-5 and TCr-1 mice was successively higher when compared with Sn mice. These differences are documented for the spleen in Fig. 3, and were observed also in lung and kidney (not shown). TCr-5 and TCr-1 mice exhibited very weak tlr4 signals also in liver and brain (not shown); however, these signals, at least in part, might be accounted for by the presence of blood cells in the two organs. Similar strain-specific differences of mRNA expression were observed also in macrophages, grown in culture from bone marrow precursors (not shown).
2.3 Expression of the TLR4–MD-2 complex on macrophages
The expression of the TLR4–MD-2 complex on the cell surface has been shown to be required for optimal LPS signaling. We investigated the presence of this complex on macrophages derived from Sn, Cr, F1 (Sn × Cr) and transgenic TCr-1 and TCr-5 mice by FCM using a monoclonal anti-TLR4–MD-2 antibody 30. This antibody recognizes the complex, but neither TLR4 nor MD-2 alone. The results summarized in Fig. 4 show that TLR4–MD-2 was present on Sn, F1, TCr-5 and TCr-1 macrophages, but completely absent from Cr cells. Macrophages of F1 mice expressed approximately half as much TLR4–MD-2 than macrophages of Sn mice, whereas those of TCr-5 and TCr-1 mice expressed 2.3 and 6.6 times higher levels, respectively. The surface expression of TLR4–MD-2 correlated with the levels of tlr4 mRNA (not shown). Interestingly, however, the expression levels of md-2 mRNA, estimated by real-time PCR (LightCycler) were found to be similar in all strains [Sn, Cr, F1 (Sn × Cr), TCr-1 and TCr-5] (not shown).
2.4 Levels of TLR4 protein expression determine LPS susceptibility
2.4.1 A unique opportunity to investigate LPS susceptibility
The availability of mice differing in TLR4 expression, but otherwise genetically almost identical (Cr mice were used for the production of the transgenic mice), offered a unique possibility to study the relationship between TLR4 expression levels and LPS sensitivity. We investigated the in vivo and in vitro LPS responses of the different mice used in this study, in relation to the respective tlr4 expression levels. The results of these investigations are shown below.
2.4.2 Induction of IL-6 in macrophages
As shown in Fig. 5, macrophages of all TLR4-expressing strains produced IL-6 upon stimulation with LPS. In all cases, the threshold dose of LPS was largely independent of the level of TLR4–MD2 expression. The height of the IL-6 response, however, was proportional to the logarithm of TLR4–MD2 expressed by the respective cells (Fig. 6). Thus, both TLR4–MD2 expression and LPS susceptibility were lowest in F1 macrophages and increased successively in Sn, TCr-5 and TCr-1 cells.
2.4.3 B cell mitogenicity
The proliferative responses of splenic B lymphocytes of the different strains of mice to LPS are shown in Fig. 7. Splenocytes of all TLR4-expressing strains exhibited mitogenic responses, the magnitude of which correlated with the levels of TLR4 mRNA expressed. Lowest responses were obtained in F1 splenocytes and increased successively in Sn, TCr-5 and TCr-1 cells. In this assay the increase of TLR4 expression was paralleled by a decrease of the threshold of LPS responsiveness. As in Fig. 5, the heterozygous F1 mice exhibited a lower reactivity to LPS than the parent homozygous Sn animals (five independent experiments, documented in Fig. 7).
2.4.4 Lethal effects of LPS in unsensitized mice
The results of lethality tests in mice of Sn, Cr, F1 strains and of the transgenic lines are summarized in Table 1. Like in the two in vitro assays described above, also here we found a significant correlation between degree of LPS susceptibility and levels of TLR4. As expected, Cr mice were highly resistant, surviving very high LPS doses (up to 400 μg / g body weight), with no apparent symptoms of illness. Mice of all other strains exhibited a varying degree of susceptibility to the lethal effects of LPS which was lowest in F1 mice and increased successively in Sn, TCr-5 and TCr-1 animals. When compared with the wild-type Sn mice, the differences in LPS susceptibility of the other strains were found to be of high statistical significance (F1 p=0.0003; TCr-5 p=0.0006; TCr-1 p=0.0001).
As previously shown, Lpsd mice carry mutations in tlr4, which in Cr mice encompasses a deletion of 74723 bp, spanning tlr4 but no other recognizable gene 8. By the insertion of a tlr4-expressing transgene into Cr mice, we generated two distinct transgenic lines exhibiting different levels of TLR4 expression.
Southern blot analysis of the heterozygous transgenic animals revealed that the number of copies of the transgene inserted into the genome was higher than the amount of tlr4 copies in the wild type. Furthermore the transgene copy-number varied substantially between the two lines created. The transgene insertion technology (microinjection into fertilized oocytes) leads to a random incorporation of the transgene into the genome of the recipient. It is therefore remarkable that, as in the homozygous wild-type Sn and heterozygous F1 (Sn × Cr), the level of tlr4 mRNA corresponded to the level of genomic tlr4 DNA. Furthermore, the pattern of tlr4 mRNA expression in different organs was the same as in wild-type mice and was therefore normal in both transgenic lines. This indicates that all the regulatory elements required for tlr4 gene expression in cells and tissues were present in the tlr4 transgene, and that these elements must be located in close proximity to the coding region of the tlr4 gene.
Constitutive expression of TLR4 analyzed by Northern blotting was below the detection limit in liver of wild-type mice. The absence of TLR4 detectability in liver is in accordance with our recent finding 31. The very low expression of tlr4 that is detected in the liver of the transgenic mice may partly originate from blood leukocytes. Thus, in this respect liver differs from other organs belonging to the reticuloendothelial system, such as spleen or lung, which exhibit readily detectable tlr4 expression. Similarly, CD14, another LPS-recognition molecule intimately involved in the LPS-receptor complex, was also not detectable in liver of healthy mice 31, 32. Thus, under physiological conditions, the LPS-receptor and accessory molecules of the receptor complex seem to be strongly down-regulated in mouse liver. This is conceivably because of the special role of this organ. The liver, with its large reticuloendothelial system, is the main organ of LPS clearance and excretion 33. Therefore the down-regulation or absence of LPS-receptor in liver of healthy animals would be of physiological advantage because it precludes a persistent inflammatory reaction caused by LPS of gut origin, entering this organ via the portal vein.
The association of TLR4 with the adapter protein MD-2 is pre-requisite for activation of cells by LPS 19. In this study, TLR4 that was present on macrophages of the different mouse strains and lines used was measured as a TLR4–MD-2 complex, the levels of which correlated with the levels of tlr4 mRNA. This indicates that a post-translational control of TLR4 protein expression is not exercised under the conditions studied. Interestingly, we found no significant differences in expression levels of md-2 mRNA, showing that the MD-2 transcription is not influenced by the amount of TLR4 present. This may suggest that MD-2 protein is either produced in excess or its production is influenced by the amount of TLR4 at the translational level.
We show here that in three classical activities of LPS — cytokine (IL-6) induction, B-cell mitogenicity and lethal toxicity — the height of the responses obtained in the different mouse strains clearly correlated with the respective TLR4 expression levels. The study disclosed further that F1 (Sn × Cr) mice exhibit a lower TLR4 expression and lower LPS susceptibility than the parent tlr4-homozygous Sn strain. This finding is in contradiction to an earlier study in which the LPS-induced mitogenic responses of splenocytes from F1 hybrids and parent tlr4-homozygous Sn mice were indistinguishable 3. The reason for the discrepancy between our study and the earlier study is not known at present.
Cr mice were refractory to all in vitro and in vivo LPS effects investigated in this study. Since the genomic insertion of a wild-type tlr4 transgene reversed the LPS-resistant phenotype of these mice, it is obvious that the tlr4 mutation of Cr mice is the primary reason for their LPS unresponsiveness. Furthermore it confirms that TLR4 is the limiting factor for LPS responsiveness in vivo and that Lps and tlr4 are identical genes. It also demonstrates the existence of a quantitative relationship between the expression level of TLR4 and the height of the LPS response.
4 Materials and methods
A highly pure preparation of Salmonella abortus equi LPS in the uniform triethylamine salt form was prepared as described previously 34. A sterile aqueous stock solution (20 mg/ml) was prepared and stored at 4°C.
The tlr40/0 Cr, tlr4+/+ Sn, tlr40/+ F1 (Cr × Sn) mice, and tlr4 transgenic mouse lines (TCr-1 and TCr-5) were bred under conventional clean conditions in the animal facilities of the Max-Planck-Institut für Immunbiologie.
4.3 Generation of tlr4Tg/0 Cr mice
To obtain transgenic mice, a purified and vector-free tlr4 mouse genomic DNA fragment was injected into the pronuclei of fertilized oocytes of tlr40/0 Cr mice at a concentration of 2–4 ng/μl, according to standard protocols 35. The DNA was obtained by a NotI digest (150,000 base pairs) of BAC 152C16 (Research Genetics, Huntsville, AL, USA). A large part of the DNA was sequenced (91,748 bp; Genebank accession No. AF177767) and the presence of the whole murine wild-type tlr4 gene (129 Sv background) was confirmed. In the course of Lps gene cloning no other genes were found in this area 8. Two founder mice (No. 1 and 5) carrying the transgene were identified by PCR analysis of tail DNA using the primers CAGTCGGTCAGCAAACGCCTTCTTC and CAAGGCAGGCTAGCAGGAAAGGGTG at an annealing temperature of 68°C. The transgenic founders were mated with Cr mice and their offspring were used to perpetuate the TCr-1 and TCr-5 lines. All transgenic animals were born and developed normally, without apparent defects.
4.4 Lethal toxicity test
The susceptibility of the different strains and transgenic lines of mice to the lethal activity of LPS was tested by administering different amounts of LPS to 7–10-week-old mice, i.p. Mortality of the mice was recorded up to 96 h. Statistical significance for the degree of LPS sensitivity of the different mouse strains in the lethal toxicity test was calculated using binary logistic regression analysis with outcome death as dependent variable and type of mouse strain and log (LPS dose) as independent variables.
Cultured macrophages derived from bone marrow precursor cells of the various mouse strains were grown in the presence of L-cell-conditioned medium in teflon bags as described previously 36. After 10 days of culture the cells were washed twice with a serum-free, high-glucose formulation of Dulbecco's modified Eagle medium (DMEM).
For FCM analysis, macrophages were resuspended at a concentration of 106 cells/25 μl in PBS containing 3% FCS. For the isolation of total RNA, macrophages (pellet) were extracted as described below.
For induction of IL-6, macrophages were resuspended in serum-free DMEM (105 cells/0.2 ml/well), placed in 96-well plates (Nunc, Roskilde, Denmark) and cultured for 24 h at 37°C in a humidified atmosphere containing 8% CO2. There-after macrophage supernatants were removed and fresh DMEM (0.2 ml) added. The macrophages were then stimulated in triplicates with different amounts of LPS (in 10 μl/well) and culture supernatants for IL-6 measurements were collected 24 h later. They were stored in aliquots at –80°C until use. IL-6 levels were estimated by ELISA using the MP5–20F3 rat anti-mouse-IL-6 antibody (PharMingen, San Diego, USA) as capturing reagent and the MP5–32C11 biotinylated rat anti-mouse-IL-6 antibody (PharMingen) as detection reagent for IL-6, according to the instructions of the supplier. The detection limit of the assay was 60 pg of IL-6/ml cell supernatant.
4.6 B cell proliferation assay
Spleens of three mice were teased through a sterile stainless-steel wire mesh. Pooled cells were washed three times in DMEM containing 2.5% FCS, 100 μg/ml streptomycin, 100 U/ml penicillinand 2 mM L-glutamine, resuspended in the same medium and 4×105 cells / 0.2 ml per well placed in 96-well round-bottom plates. Duplicates of cells were stimulated with different amounts of S. abortus equi LPS and cultivated for 66 h at 37°C in a humidified atmosphere containing 5% CO2. [3H]thymidine (0.2 μCi/well) was added during the last 18 h of stimulation. For each test two identical plates were prepared as replicates. The cells were harvested on GF/A-filters (NR: 20–182–70, Dunn Labortechnik GmbH, Asbach, Germany) and [3H]thymidine incorporation was measured by an automated β-counter (Inotech, Asbach, Germany).
4.7 RNA extraction
Total RNA was isolated from freshly removed organs or cultured macrophages by a guanidinium isothiocyanate-phenol-choloroform-isoamyl alcohol procedure 37 as described in detail in 38. The RNA concentration was determined by absorbance at 260 nm.
4.8 Real-time RT-PCR with LightCycler
Total spleen RNA (1 μg) was reverse transcribed with M-MuLV reverse transcriptase and oligo-p(dT) primers (Expand reverse transcriptase kit, Roche, Mannheim, Germany) according to the manufacturer's instructions.
Real-time hot-start PCR was performed with the LC FastStart DNA Master SYBR Green I Kit (Roche Diagnostics, Mannheim, Germany) in a LightCycler instrument (Roche Diagnostics) according to the manufacturer's instructions. Primers for murine MD-2 were: sense 5′ CCC ATA TTG ACT GAA TCT GAG AA; antisense 5′ AGC TTC TGC AAC ACA TCT GTA AT. HGPRT expression (primers: sense, 5′ GCT CGT GAA AAG GAC GTC; antisense, 5′ CAC AGG ACT AGA ACA CCT GC) was used to normalize the cDNA levels. An annealing temperature of 58°C was used for both primer pairs. RT-reactionsin which reverse transcriptase was replaced with water were used as negative controls to exclude products derived from contaminating genomic DNA. After amplification, products were analyzed by melting curve and agarose gel electrophoresis analysis, in order to ascertain that only a single DNA fragment of the expected size was produced.
4.9 Northern blot analysis
RNA samples (5–18 μg) were fractionated on 1.2% denaturing agarose-formaldehyde gels and transferred to Nylon filters (Nytran, Schleicher & Schuell, Keene, NH, USA) as described previously31. RNA were hybridized overnight at 60°C with a random primed [32P]-labeled probe as previously described 39. The TLR4 probe was a 1.8 kbp cDNA fragment of the murine wild-type TLR4 as described earlier 31. The amount of total RNA per sample applied to the electrophoresis gel was visualized by the intensity of the ethidium bromide (EtBr)-stained 18 S rRNA band on the Nylon Filter used for hybridization. Autoradiography was performed for 6 h to 10 days at –80°C with an intensifying screen (Cronex Lightening Plus, Dupont) using Biomax MS Films (Eastman Kodak Co., Rochester, NY, USA).
4.10 Southern blot analysis
Genomic DNA (10 μg) was digested with EcoRI and separated on a 1% agarose gel. Thereafter the DNA was capillary blotted onto Nylon membrane (Amersham, Piscataway, NJ, USA). Samples were probed with α-[32P]dCTP random-labeled fragments corresponding to nucleotides 512–1500 of mouse tlr4 cDNA sequence (accession number AF095353) using standard conditions of hybridization and washing. After exposing the membrane, the intensities of the bands were determined by Phosphorimager (Molecular Dynamics, USA) analysis.
TLR4 was detected on macrophages as a TLR4–MD2 complex. The cells were incubated with an anti-TLR4–MD2 antibody (Clone MTS510, Rat IgG2a, Mo Bi Tec, Göttingen, Germany) and stained with a FITC-conjugated goat F(ab′)2 anti-rat-IgG secondary reagent (Immunotech, Marseille, France). Ten thousand cells were acquired for each sample. Dead cells were excluded after staining with propidium iodide. All incubation steps were carried out on ice. Nonspecific binding was blocked by preincubation with normal goat serum. Cells were analyzed using a FACSCalibur (Becton Dickinson, Mountain View, CA, USA).
This study was partly supported by DFG, SP "Angeborene Immunität" (FR 448/4–1). We are grateful to Bruce Beutler for helpful comments and discussions, Viktor Steimle for support with the LightCycler and Jürgen Schulte-Mönting for the statistical analysis. We thank H. Kochanowski, H. Stübig, J. Kühnle and N. Goos for excellent technical assistance.