| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Institut für Immunologie und Transfusionsmedizin, Klinikum der Universität Greifswald, Greifswald, Germany; and
Institut für Pathologie und Tumorimmunologie, Universitätsklinikum Regensburg, Regensburg, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
LPS forms micelles in aqueous solution, and the concentration of the biologically active monomer form is extremely low (7). Because of this, the interaction between LPS and CD14 is slow unless catalyzed by the serum protein LPS-binding protein (LBP)3 (8, 9, 10). In vitro experiments have established a number of additional possible roles for LBP. Under appropriate conditions it can transfer LPS not only to CD14, but also to high density lipoprotein (11, 12) and can take part in transfer reactions involving a number of phospholipids and membrane structures (13, 14). These interactions between LBP, LPS, lipoproteins, and cells in vitro are complex (15), but their relevance to LBP function in vivo is uncertain. In vivo, two functions for LBP have been established in mice. First, it is required for the release of proinflammatory mediators such as TNF (16) in response to a challenge with purified bacterial LPS. These mediators initiate a pleiotropic cascade of events, including activation of the coagulation system and recruitment of phagocytes (17). Second, LBP is an acute phase protein, and in high concentrations it has an anti-inflammatory effect that is thought to be due to LPS sequestration (18).
Given this plethora of suggested functions we sought to narrow the in vivo role of LBP during the course of an infection by tracing the initial steps that require LBP for the control of a Gram-negative pathogen in mice. For this purpose we have generated an LBP-deficient mouse line (LBP-/-) and shown that these animals are defective in their ability to mount an inflammatory response, as judged by the serum titer of TNF when challenged with purified LPS (16). To investigate the response of these animals to a pathogen we have used a peritoneal infection with Salmonella. In this model system a lethal infection can be initiated with a very low number of bacteria.
Salmonella, like all pathogens, must solve problems at three quite different levels. First, it must cross the physical barriers that constitute our first line of defense. Having done this it must survive whatever the innate defense system undertakes against it, and finally, any survivors must face up to the specific immune system (19). There are complex interactions taking place at each of these three levels. However, since our interest centers on the role of the innate defense system, we avoid the complexities associated with the breaching of the physical barriers by injecting the Salmonella directly into the peritoneum and ignore the contribution from the specific immune system by concentrating on early events in the infection.
To determine the role of LBP in this infection, we have made use of mice in which LBP function has been specifically ablated by gene targeting. We showed that the LBP-deficient animals are incapable of controlling the Salmonella infection and succumb to a dose of the pathogen that their heterozygous litter mates survive. Here we show that the inability of these mice to control an i.p. Salmonella infection is a direct result of the loss of LBP function by demonstrating rescue of the animals by supplementing them with recombinant mouse LBP. We then show that in wild-type mice the infection results in the rapid elaboration of an inflammatory response, as evidenced by the release of detectable levels of IL-6 within the first hours of infection, and that this inflammatory response is deficient in LBP-/- animals. Finally, we show that the requirement for LBP in these animals can be replaced by injection of the principle proinflammatory mediator TNF. The ability to substitute TNF for LBP indicates that a major nonredundant role for LBP in this infection model involves the generation of an inflammatory response.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Recombinant murine TNF was purchased from Life Technologies (Karlsruhe, Germany), and recombinant human TNF was obtained from BASF Bioresearch (Ludwigshaven, Germany). Histidine-tagged recombinant murine LBP was prepared as previously described (20) from supernatants of transfected Chinese hamster ovary cells by metal chelate chromatography and was stored frozen in PBS at -20°C until required. Samples for injection were diluted in pyrogen-free PBS (Sigma, Diesenhofen, Germany).
Mice
BALB/c and CBA/J mice were purchased from Charles River Breeding Laboratories (Wilmington, MA). The generation of LBP-deficient mice on a strain 129 background has been previously described (16). The animals used in the experiments described here are progeny of LBP+/- heterozygotes obtained from a second backcross generation of the LBP-/- animals to strain CBA. All of the animals carried the resistance allele of the NRAMP 1 gene (21).
Genotypic characterization
Genotypes were determined by PCR analysis of genomic DNA prepared from the tail tip. Tail tips were digested overnight with 100 µg/ml proteinase K (Roche, Mannheim, Germany) in 100 mM Tris (pH 8.5), 200 mM NaCl, 5 mM EDTA, and 0.02% SDS. The digest was extracted three times with phenol, precipitated with an equal volume of isopropanol, and redissolved in 100 µl of TE buffer (10 mM Tris (pH 8.0) and 1 mM EDTA). LBP-/- and LBP+/- animals were distinguished by PCR. The LBP-/- NRAMP+/+ progeny were identified by PCR using DNA prepared from the tail tip. The LBP genotype was determined using the LBP-specific primers 5'-GCGGCCAAGGAGGGGCTGGTG-3' and 5'-CCAGTCACCTATGAAACTCG-3' and the neoR primers 5'-TCGCCGCCAAGCTCTTCAGC-3' and 5'-GTCAAGACCGACCTGTCCGG-3'. This PCR yields a band of 125 bp from the wild-type LBP gene and 559 bp from the neoR insert in the disrupted LBP gene. NRAMP genotype was determined in a PCR using the 5' primer 5'-CCCCCATCTATGTTATCACCC-3'. In the 3' primer (5'-CAATGGTGATCAGTACAGCG-3') substitution of C instead of G in the third base from the 3' end results in a product from the NRAMPR allele that contains an HhaI site that is absent from the product amplified from the NRAMPS allele. The PCR products were digested with HhaI and analyzed on a 1.5% agarose gel. The wild-type allele gives a fragment of 266 bp, which is readily distinguished from the 285-bp fragment from the null allele.
ELISA determinations
ELISAs for murine IL-6 and murine TNF-
were purchased from BD
PharMingen (San Diego, CA). The ELISA for mouse LBP was obtained from
Biometec (Greifswald, Germany).
Salmonella infection
Salmonella enterica serotype typhimurium (ATCC 14028s, American Type Culture Collection, Manassas, VA; Salmonella typhimurium) transfected with the plasmid pFGV25.1 coding for green fluorescent protein was a gift from Dr. M. Hensel (Max von Pettenkofer Institut, Munich, Germany). A single colony was picked and grown overnight in Trypto-C Yeast Extract broth containing 100 µg/ml ampicillin. The overnight culture was diluted 1/100 into fresh Trypto-C Yeast Extract without antibiotics and grown at 37°C with vigorous shaking for 2.5 h. Cells were collected by centrifugation and were washed in PBS. The number of bacteria was estimated from the OD at 550 nm by reference to a standard curve. The actual number of CFU used in each experiment was determined by plating samples of the injected suspension. For determination of TNF and IL-6, animals were bled 2 h after infection, and serum was stored at -80°C before analysis by ELISA.
Determination of bacterial load in peritoneum and lung early in infection
Mice were infected with Salmonella and 24 h later were sacrificed by cervical dislocation, and the peritoneal cavity was washed with 10 ml of cold sterile PBS. Appropriate dilutions of the peritoneal wash were plated in duplicate. Plates were incubated for 16 h at 37°C, and colonies were counted. For determination of the bacterial load in the lungs, animals were infected for 48 h, then sacrificed, and the lungs were homogenized in 2 ml of sterile PBS containing 0.2% saponin. Lung homogenates were incubated on ice for 30 min, and suitable dilutions of this extract were plated and counted as described for the peritoneal washes.
Reconstitution of LBP-/- mice with recombinant LBP
LBP-/- mice were injected i.p. with 20 µg of recombinant murine LBP in a final volume of 200 of µl sterile pyrogen-free PBS. After 1, 2, 4, and 6 h, groups of three mice were anesthetized with diethyl ether and bled from the orbital complex. The animals were then sacrificed by cervical dislocation, and the peritoneal cavity was washed with 10 ml of cold sterile PBS. The PBS wash was centrifuged at 1500 x g to remove cells. Serum and cleared peritoneal wash were assayed by ELISA for LBP content. Peritoneal wash data are expressed as total LBP recovered from the peritoneum. LBP levels in serum are expressed as nanograms of LBP per milliliter of serum.
Substitution with recombinant mouse or human TNF
Animals were injected with recombinant mouse TNF (Life Technologies) or with clinical trial grade recombinant human TNF (gift from BASF Bioresearch). TNF (800 ng) was injected i.p. along with the Salmonella, and an additional 800 ng was administered i.p. 90 min later. Infected mice were examined daily over a period of 3 wk.
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Fate of recombinant murine LBP in vivo
We first determined the fate of recombinant LBP injected i.p. into
LBP-/- animals. Groups of
LBP-/- mice were injected i.p. with 20 µg of
recombinant murine LBP. One, 2, 4, and 6 h later animals were
bled, then sacrificed by cervical dislocation, and the peritoneum was
washed with PBS. The amount of LBP present in serum and peritoneal wash
was determined by ELISA. Despite considerable individual variation in
the titers it is clear that the injected recombinant LBP is quite
rapidly lost from the peritoneum, with a half-life of 
1 h (Fig. 1
a).Concomitantly LBP starts
to appear in the circulation and reaches a maximal concentration of
about 5 µg/ml at 2 h, after which it decays with a half-life of
1 h, reaching a value of about 0.3 µg/ml at 6 h (Fig. 1b). For comparison, LBP is normally present in serum at a
concentration of 1 µg/ml and rises to 5 µg/ml during an acute phase
response (22). These data indicate that the reconstitution
that we achieve is modest, and it is unlikely that a significant
concentration of the injected LBP is available for more than a brief
part of the 7–12 days covered by the lethal infection.
|
To determine whether the injected LBP is able to rescue
LBP-deficient animals from a lethal Salmonella infection, we
infected the mice i.p. with 50 CFU of S. typhimurium and at
the same time injected 2 or 20 µg of recombinant murine LBP. Control
animals given Salmonella alone died, while those
supplemented with LBP showed a dose-dependent rescue (Fig. 2). Given the brief half-life of the
recombinant protein in vivo (Fig. 1
), the protection given by the
recombinant LBP indicates that one crucial LBP-dependent step in the
host response to the pathogen takes place within the first few hours of
the infection. The outcome of this LBP-dependent process determines
whether the animal will succumb to the infection several days
later.
|
Our LBP-deficient animals fail to respond to a bolus injection of
LPS by releasing TNF. In contrast, a second LBP knockout that was
constructed using an alternative strategy does respond to LPS with the
production of TNF (23). The reason for this discrepancy
between the two knockout strains is unclear. One possibility is
differences in the LPS preparation used. A second possible explanation
lies in differences between the two knockout strains. The LBP-deficient
mouse of Wurfel et al. (11) does produce an LBP
message, and although LBP is not detectable in the blood, the authors
speculate that low levels of a functional truncated form of the protein
may be sequestered elsewhere, perhaps in the liver. In the case of our
mice, in which removal of an internal exon causes a frame shift, the
capacity to respond to LPS with the production of TNF is completely
abolished (16). To determine whether our
LBP-/- mice failed to initiate an inflammatory
response on infection with S. typhimurium we measured the
serum concentrations of the proinflammatory cytokines TNF and IL-6
following infection. TNF, which is released from activated macrophages,
is clearly involved in innate resistance to S. typhimurium,
since the TNFR1-deficient mice are highly susceptible to this infection
(24). However, this cytokine, which has a very short
half-life in serum, is not present at detectable levels in the serum of
Salmonella-infected mice. In contrast, the more stable
proinflammatory cytokine IL-6, which, like TNF, is produced by
activated macrophages, is readily detected in serum 2 h after
initiating the infection. Individual mice show a broad scatter in the
titer, but all animals carrying a wild-type allele of LBP
have readily detectable IL-6 in their serum 2 h after infection.
The scatter in the IL-6 values is unlikely to be due to variability in
the genetic background of the individual animals, since pure-bred
BALB/c mice show a similar individual diversity of response. Despite
the diversity of the response it is clear that all animals tested to
date that carry at least one wild-type LBP allele produce
readily detectable levels of IL-6 early in a Salmonella
infection. In contrast, the LBP-/- animals
produce no detectable IL-6 (Table I). By
this criterion they are deficient in their ability to mount a rapid
inflammatory response with the concomitant release of proinflammatory
mediators. This inability to mount an effective inflammatory response
is the earliest defect we have observed in this infection model. It is
important to note that the injection of LBP (20 µg/mouse i.p.) does
not in itself cause an inflammatory response (Table II). This shows that the LBP-dependent
rescue from the lethality caused by the Salmonella infection
documented in Fig. 2 is indeed dependent on a proinflammatory LBP
function that is triggered by the bacterial infection.
|
|
In the early stages of this Salmonella infection, the
bacteria first multiply in the peritoneum and then spread out to
peripheral organs, including liver, spleen, and lungs.
LBP-/- mice on a strain 129 background show an
increased rate of growth of Salmonella in the peritoneum
compared with their heterozygous littermates (16). This
reduced capacity to control the rate of bacterial growth is also
evident in LBP-/- animals on a CBA background.
Animals infected with 50 CFU Salmonella i.p. have a greater
bacterial burden in the peritoneum (Fig. 3a). In addition, we have
compared the rate at which the bacteria spread from the peritoneum to
the lungs in LBP-/- animals and their
LBP+/- littermates. As shown in Fig. 3b infection of LBP-/- animals i.p.
with 50 CFU results in substantially greater bacterial load in the
lungs than is evident in their LBP+/- litter
mates. Thus, the phenotype of Salmonella-infected
LBP-deficient animals includes a decreased inflammatory response 2
h after infection, an increase in bacterial load in the peritoneum and
lungs at 48 h, and lethality evident 5–7 days after
infection.
|
To determine whether the sensitivity of LBP-deficient mice to
Salmonella is causally related to their inability to mount
an early inflammatory response, we attempted to bypass the requirement
for LBP by supplementing the LBP-deficient animals early in the
infection with the proinflammatory mediator TNF. LBP-deficient animals
infected with 50 CFU of S. typhimurium died within the first
week. However, animals that were supplemented with exogenous
recombinant murine TNF along with the Salmonella survived
(Fig. 4). There are two murine receptors
for TNF, both of which bind murine TNF with high affinity. However,
only the 55-kDa TNFR1 binds human TNF (25). We therefore
repeated the TNF supplementation experiment using recombinant human
TNF. As shown in Fig. 5b,
recombinant human TNF was also able to rescue these mice, and hence an
initial interaction requiring only TNFR1 is sufficient to reverse the
lethality associated with lack of LBP. These results show that the
requirement for LBP can be entirely replaced by exogenously added
TNF.
|
|
). They
are nevertheless sensitive to the infection and die within 5–7 days
because of a mutation in the NRAMP-1 gene.
NRAMP-1 encodes a protein that is expressed in macrophages
in the endosomal compartment (21, 26). This protein is
required to pump divalent metal ions out of the endosome, thus creating
an environment hostile to the pathogen’s survival (27).
NRAMP 1 is an essential component of the innate defense system and is
required so that mice can control the growth and survival of S.
typhimurium, which enter host macrophages early in the infection
and hide out in the endosome compartment.
Two groups of BALB/c mice were injected with 50 CFU of S.
typhimurium, and one group was also treated with recombinant
murine TNF as described for the LBP-/- mice.
As shown in Fig. 5a, the susceptibility of these animals to
Salmonella is not alleviated by this treatment. Exogenous
TNF thus protects the LBP-/- animals, which
have a demonstrated defect in their ability to produce an inflammatory
response (Fig. 1), but fails to protect animals that are sensitive to
the infection due to a defect in a different part of the innate defense
system. This rules out the possibility that the injected TNF rescues
the LBP-deficient animals by a nonspecific direct bactericidal
mechanism unrelated to normal LBP function.
Although numerous studies have shown that an interaction of LPS with LBP leads to macrophage activation and the induction of an inflammatory reaction, it has been unclear how important this pathway may be in combating an infection in vivo. Innate defense is both multifaceted and redundant, and Gram-negative bacteria such as Salmonella contain many components other than LPS, some of which contribute to cell activation (28). Our data make it clear that the LBP-initiated pathway is essential for defense, that it is required for the induction of an inflammatory reaction, and that this is a major nonredundant function of LBP in this Gram-negative peritoneal infection.
The control of acute and chronic inflammatory disorders is currently a major problem in clinical medicine (29, 30), and much effort has been invested in developing therapeutic strategies that block appropriate parts of the inflammatory network. Our results emphasize that any procedure that blocks the LBP-initiated inflammatory cascade will disable an essential defense pathway. Any anti-inflammatory protection that may be achieved must be balanced against the risks inherent in blinding the innate system to the presence of Gram-negative pathogens.
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Robert S. Jack, Institut für Immunologie und Transfusionsmedizin, Klinikum der Universität Greifswald, Sauerbruchstrasse, D-17489 Greifswald, Germany. E-mail address: jack{at}mail.uni-greifswald.de
3 Abbreviation used in this paper: LBP, LPS-binding protein.
Received for publication February 2, 2001. Accepted for publication May 21, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
p55 receptor. Infect. Immun. 66:3355.
treatment. Nat. Med. 3:678.[Medline]This article has been cited by other articles:
|
|
 |
|
  T. Sterns, N. Pollak, B. Echtenacher, and D. N. Mannel Divergence of Protection Induced by Bacterial Products and Sepsis-Induced Immune Suppression Infect. Immun., August 1, 2005; 73(8): 4905 - 4912. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Hamann, C. Alexander, C. Stamme, U. Zahringer, and R. R. Schumann Acute-Phase Concentrations of Lipopolysaccharide (LPS)-Binding Protein Inhibit Innate Immune Cell Activation by Different LPS Chemotypes via Different Mechanisms Infect. Immun., January 1, 2005; 73(1): 193 - 200. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Branger, S. Florquin, S. Knapp, J. C. Leemans, J. M. Pater, P. Speelman, D. T. Golenbock, and T. van der Poll LPS-binding protein-deficient mice have an impaired defense against Gram-negative but not Gram-positive pneumonia Int. Immunol., November 1, 2004; 16(11): 1605 - 1611. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Frank, K. Witte, W. Schrodl, and C. Schutt CHRONIC ALCOHOLISM CAUSES DELETERIOUS CONDITIONING OF INNATE IMMUNITY Alcohol Alcohol., September 1, 2004; 39(5): 386 - 392. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Knapp, A. F. de Vos, S. Florquin, D. T. Golenbock, and T. van der Poll Lipopolysaccharide Binding Protein Is an Essential Component of the Innate Immune Response to Escherichia coli Peritonitis in Mice Infect. Immun., December 1, 2003; 71(12): 6747 - 6753. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Xu, R. Lucas, M. Schuchmann, S. Kuhnle, T. Meergans, A. P. Barreiros, A. W. Lohse, G. Otto, and A. Wendel GM-CSF Restores Innate, But Not Adaptive, Immune Responses in Glucocorticoid-Immunosuppressed Human Blood In Vitro J. Immunol., July 15, 2003; 171(2): 938 - 947. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. K. Yang, B. G. Dorner, U. Merkel, B. Ryffel, C. Schutt, D. Golenbock, M. W. Freeman, and R. S. Jack Neutrophil Influx in Response to a Peritoneal Infection with Salmonella Is Delayed in Lipopolysaccharide-Binding Protein or CD14-Deficient Mice J. Immunol., October 15, 2002; 169(8): 4475 - 4480. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.M. Monick and G.W. Hunninghake Activation of second messenger pathways in alveolar macrophages by endotoxin Eur. Respir. J., July 1, 2002; 20(1): 210 - 222. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Fierer, M. A. Swancutt, D. Heumann, and D. Golenbock The Role of Lipopolysaccharide Binding Protein in Resistance to Salmonella Infections in Mice J. Immunol., June 15, 2002; 168(12): 6396 - 6403. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Inagawa, T. Honda, C. Kohchi, T. Nishizawa, Y. Yoshiura, T. Nakanishi, Y. Yokomizo, and G.-I. Soma Cloning and Characterization of the Homolog of Mammalian Lipopolysaccharide-Binding Protein and Bactericidal Permeability-Increasing Protein in Rainbow Trout Oncorhynchus mykiss J. Immunol., June 1, 2002; 168(11): 5638 - 5644. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Li, V. M. Holers, S. A. Boackle, and C. M. Blatteis Modulation of Mouse Endotoxic Fever by Complement Infect. Immun., May 1, 2002; 70(5): 2519 - 2525. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Viguerie, L. Millet, S. Avizou, H. Vidal, D. Larrouy, and D. Langin Regulation of Human Adipocyte Gene Expression by Thyroid Hormone J. Clin. Endocrinol. Metab., February 1, 2002; 87(2): 630 - 634. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Echtenacher, M. A. Freudenberg, R. S. Jack, and D. N. Mannel Differences in Innate Defense Mechanisms in Endotoxemia and Polymicrobial Septic Peritonitis Infect. Immun., December 1, 2001; 69(12): 7271 - 7276. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or 20 µg (
) recombinant murine LBP.
