| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Production by Macrophages In Vitro and In Vivo and Prevent Endotoxin-Mediated Toxic Shock
*
Institute of Medical Microbiology, Immunology and Hygiene, Technical University of Munich, Munich, Germany;
MorphoSys, Martinsried/Munich, Germany; and
Institute of Medical Immunology, Humboldt University of Berlin, Berlin, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
release of murine macrophages up to 90%. Furthermore, LL-10-H-14 not
only reduced peak serum levels of TNF- of mice when preinjected but
also reduced TNF- levels when given 15 min after the endotoxin
challenge. As compared with other peptides, only LL-10-H-14 is able to
strongly decrease endotoxin-stimulated TNF- release by human
macrophage cell lines as well as by PBMC. Furthermore, the hybrid
peptide is protective against endotoxin-provoked lethal shock. As such,
LL-10-H-14 could have prophylactic and/or therapeutic properties in
humans for the management of septic shock. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, IL-1, IL-6, and others, which in toxic concentrations are
mediators for the development of septic shock (6, 7). In
certain animal models, TNF- is prime for the lethality induced on
injection of endotoxin (8).
To prevent the cascade of endotoxin-induced events in humans, several
approaches have been explored. First, Abs directed against endotoxin or
the structurally conserved lipid A moiety were developed with the aim
to neutralize endotoxins generated by different Gram-negative bacteria.
However, these Abs appeared not to recognize natural endotoxins, i.e.,
lipid A carrying the polysaccharide component (9). mAbs
recognizing core structures of endotoxin of different
Enterobacteriaceae may be promising in that they neutralize
endotoxin-mediated toxicity in vivo (10). However, again
their use may be limited because of their lack of cross-reactivity
among different kinds of endotoxin. Second, anti-TNF- Abs or TNF
receptor antagonists have also been evaluated as very effective in
their potential to protect against lethal endotoxinemia in several
animal models (8, 11, 12). However, none of these
molecules appeared to be effective in humans. Furthermore, complete
neutralization of TNF- during an infection is probably
disadvantageous, because in physiological concentrations TNF-
orchestrates the inflammatory reaction of the infected host
(13). Third, application of endotoxin-binding molecules
has been investigated which may prevent the endotoxin-LBP-CD14 complex
formation. Several candidate molecules with endotoxin-binding and
neutralizing capability are known and bactericidal/permeability
increasing protein (BPI) may be until now the most promising candidate
(14, 15, 16). Limulus anti-LPS factor (LALF) is
a protein from the horseshoe crab with high potential to block
endotoxin-mediated activities in animals (17, 18, 19, 20). Both
BPI and LALF crystal structures are known, and the LALF analysis
revealed a detailed view of a potential endotoxin-binding site
(21). The site consists of the aa 31–52 and is
characterized by an alternating series of positively charged and
hydrophobic residues forming a positively charged amphipathic loop
(21). Two other endotoxin-binding proteins from mammals,
namely bactericidal/permeability increasing protein (BPI) and LPS
binding protein (LBP), were proposed to have a similar endotoxin
binding site (21). Although unrelated to LALF,
endotoxin-binding domains were identified in both proteins using
synthetic peptides (2, 22, 23). The endotoxin-binding
domains of all three proteins are functionally competent within LBP in
domain exchange mutant proteins (24).
Recent approaches to develop molecules that neutralize endotoxin have concentrated on characterizing lipid A-binding regions from endotoxin-binding peptides and proteins. This includes synthetic peptides derived from sequences of polymyxin B sulfate (PMB) (25), Tachypleus anti-LPS factor (26), recombinant LALF (rLALF) (27), BPI (23), CAP-18 (28, 29), and LBP (22).
Here, we further examined the endotoxin-neutralizing potential of a
variety of peptides in in vitro and in vivo studies with the aim to
optimize their endotoxin-neutralizing capacity. We developed a series
of peptides derived from LALF, BPI, and LBP as well as hybrid peptides,
comprising two endotoxin-binding domains. The peptide LL-10-H-14, a
hybrid between LALF and LBP, showed exceptional activity. It is able to
block endotoxin-induced TNF- release of murine and human macrophages
as well as peripheral blood leukocytes. Most surprisingly, LL-10-H-14
is effective in suppressing TNF- levels in mice pre- or postinjected
with endotoxin and to prevent lethality in the
endotoxin/D-galactosamine (D-GalN) mouse
model.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
C57BL/6 and BALB/c mice were bought from Harlan-Winkelmann (Borchen, Germany). All mice were ordered with an age of 8–10 wk and housed in our own animal facility.
Cell culture
Murine cell lines were cultured in medium containing <0.01 endotoxin units (EU)/ml endotoxin (very low endotoxon (VLE) RPMI 1640, Seromed, Biochrom, Berlin, Germany). The medium was supplemented with 10% (v/v) FCS (Seromed, Biochrom), 5 x 10-5 M 2-ME (Life Technologies, Karlsruhe, Germany), and antibiotics (penicillin G (100 IU/ml medium) and streptomycin sulfate (100 IU/ml medium)).
For culture of the human cell line Mono Mac 6, the VLE RPMI 1640 medium was supplemented with 5% (v/v) human serum (human serum (male), Sigma, Deisenhofen, Germany) and 5 x 10-5 M 2-ME. Human PBMC were also cultured in VLE RPMI 1640 medium but supplemented with 5% (v/v) of the donor serum and 5 x 10-5 M 2-ME.
Reagents
Peptide sequences are given in Table I
. All experiments were
performed with endotoxin from Salmonella
enteritidis (Sigma). lipid A, and PMB were purchased from
Sigma. D-GalN came from Carl Roth (Karlsruhe,
Germany). CD14 was kindly donated by Dr. C. Schütt (Institute of
Immunology, Greifswald, Germany). LALF was generously provided by Dr.
R. C. Liddington (Dana Farber Cancer Institute, Boston, MA). For
the generation of stock solutions, all reagents were dissolved in
endotoxin-free water (water for embryo transfer, Sigma).
|
Peptides were prepared on an automated peptide synthesizer (Abimed Analyes-Technik, Langenfeld, Germany) using a standard Fmoc (N-(9-fluorenyl)methoxycarbonyl) machine protocol. For detection purposes, all peptides were biotinylated using 2 eq of biotin activated with 2 eq of benzotriazole-1-yloxytripyrrolidinophosphonium hexafluorophosphate and 4 eq of N-methylmorpholine. The cleaved peptides were worked up according to the manufacturer’s protocols and purified with reversed phase HPLC to a purity of >95%. Cyclic peptides were obtained by incubating cysteine-containing peptides in 10% DMSO for 24 h at room temperature (30). The peptides were then purified by reversed phase HPLC to a purity of >95%. The characterization of the peptides was done by laser desorption-time of flight mass spectroscopy.
Competitive inhibition of rLALF/lipid A binding
Lipid A (100 µl, 0.5 µg/ml), dissolved in PBS, was coated on a microtiter plate (Polysorp U96, Nunc, Wiesbaden, Germany) for 120 min at 37°C. After blocking with PBST for 20 min (room temperature), the solid phase was incubated for 60 min with 100 µl containing increasing amounts of peptides (0.01–100 µg/ml) mixed with rLALF (0.2 µg/ml, dissolved in PBS plus Tween (PBST)). After three washings with PBST, the ELISA was developed with a rabbit antiserum against rLALF (incubation time, 60 min) and a second anti-rabbit Ab conjugated (Sigma) to alkaline phosphatase (incubation time, 60 min). p-Nitrophenyl phosphate (1 mg/ml) was used as a substrate, and the absorbance was quantitated at 405 nm with a microplate reader. Each measurement was performed in duplicate.
Competitive inhibition of CD14/lipid A binding
CD14 A (100 µl, 5 µg/ml), dissolved in PBS, was coated on a microtiter plate (Polysorp U96) for 120 min at room temperature. After blocking with PBST for 30 min (room temperature), the solid phase was incubated for 120 min with 100 µl peptides (10 µg/ml) mixed with FITC-labeled endotoxin (2.5 µg/ml, dissolved in PBST). After three washings with PBST, an alkaline phosphatase-coupled anti-FITC mAb (dissolved 1:2500 in PBST, Boehringer Mannheim, Mannheim, Germany) was added. After another wash cycle, the ELISA was developed with p-nitrophenyl phosphate (1 mg/ml) as substrate, and the absorbance was quantitated at 405 nm using a microplate reader. Each measurement was performed in duplicate.
Stimulation protocol
To analyze the inhibitory effect of the peptides on TNF-
release in vitro, the murine macrophage cell line ANA-1 (1 x
105 cells/well), the human monocyte cell line
Mono Mac 6 (1 x 105 cells/well), as well as
PBMC (5 x 105 cells/well) were incubated
with different peptides (5 µg/ml), and 5 min later the cells were
stimulated with endotoxin (40 ng/ml for ANA-1 and Mono Mac 6 cells, 5
ng/ml for PBMC) for 30 min (37°C, 5% CO2). To
remove unbound endotoxin and peptide, the cells were washed twice and
incubated for another 4 h (37°C, 5% CO2).
At this time point, supernatants were harvested and stored at -20°C
until the TNF- content was measured by ELISA. To analyze cell
viability a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) assay was performed as described below. Cell cultures
were performed in 96-well, flat-bottom microtiter plates (Nunclon
surface, Nunc).
Injection protocol
To induce endotoxin shock, mice were preinjected with
D-GalN (i.p., 4 mg/mouse in 100 µl) and after 1 h
challenged with endotoxin (i.p., 1 µg/mouse in 50 µl). For
prevention of shock or systemic TNF- release mice were pretreated 15
min before the endotoxin injection with peptides (i.p., 100 µg/mouse
in 50 µl), PMB (i.p., 100 µg/mouse in 50 µl), or solvent
(endotoxin-free water, i.p., 50 µl).
Serum preparation
Treated and control animals were killed and blood was drawn from
the heart. Subsequently, the samples were centrifuged at 10,000 rpm (7
min), and the supernatant was collected and stored at -20°C until
used in the TNF- ELISA.
Determination of TNF-
Murine TNF- serum levels were determined using commercially
available TNF- ELISA kits (DuoSeT, Genzyme, Cambridge, MA). The
assay was performed as described by the manufacturer except that the
capture Ab was used in a concentration of 3 µg/ml, experimental
samples and standards were incubated overnight (4°C), and the
detecting Ab was added in a concentration of 2.1 µg/ml. Serum samples
and culture supernatants were used at a dilution of 1:2 and measured
twice. Human TNF- was monitored using the TNF- ELISA (OptEIA)
from PharMingen (San Diego, CA). The assay was performed as described
by the manufacturer.
MTT assay
To measure cell viability, the MTT assay was performed. Briefly, the cultured cells (1 x 105 ANA-1 cells/well, 1 x 105 Mono Mac 6 cells/well, or 5 x 105 PBMC/well) were incubated with MTT (420 µg/ml, Sigma), which is metabolized by living cells in 3 h. After solubilization of MTT crystals with HCl-isopropanol (150 µl/well), the OD was determined at 570/690 nm.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To analyze the potential of peptides derived from LBP, BPI, and
LALF to neutralize endotoxin, we compared their ability to block the
binding of lipid A to rLALF. The cyclic LALF-derived peptide LALF-14
was chosen in that it has been described as the smallest unit still
possessing full endotoxin-binding capacity (Ref. 31 and
Table I
). Analogues peptides derived from
BPI (B-14) and LBP (H-14) were prepared based on the alignment as
proposed by Hoess et al. (21). Both, H-14 and B-14 were
derived from amino acid regions 90–101 of LBP or BPI, respectively,
and comprise 12 aa from the endotoxin-binding domain as well as 2
cysteines at both ends to form a cyclic peptide after oxidation (Table I). The cyclic peptide LALF-22 consists of the complete
endotoxin-binding domain of LALF (Table I). As controls, we used a
linear peptide comprising the minimal LALF/endotoxin-binding domain
(LL-10) and NEU-10 with a neutral net charge. In agreement with Ried et
al. (31), we observed that the cyclic peptides LALF-14
(Fig. 1, 
) and H-14 (Fig. 1, 
)
efficiently blocked the rLALF/lipid A interaction in contrast to the
linear peptide LL-10 (Fig. 1, •) and the cyclic peptide B-14 (Fig. 1, 
). The control peptide NEU-10 was inactive as expected (Fig. 1, 
). The cyclic peptide LALF-22 (Fig. 1, 
) was as active as the
positive control PMB (Fig. 1, 
). At least 10 µg/ml of an active
peptide were required to display full activity. Similar results were
obtained when the peptides LALF-14, H-14, and B-14 were analyzed for
their ability to bind to lipid A or to block the LBP/lipid A
interaction (data not shown). Based on these results, the peptides
derived from LALF and LBP were further investigated as candidates to
block release of TNF- by endotoxin-stimulated macrophages.
|
secretion in vitro
TNF- is a key molecule in the pathophysiology of septic shock
(8, 32). It has been shown that Abs against TNF-
protect experimental animals in endotoxin- as well as
superantigen-induced lethal shock models, which are considered as model
systems for Gram-negative and Gram-positive sepsis, respectively
(8, 33, 34). To explore the potential of the
endotoxin-binding peptides to inhibit TNF- release by
endotoxin-stimulated macrophages we first evaluated cells of the murine
macrophage cell line ANA-1. Accordingly, the peptides LALF-14, H-14,
B-14, LALF-22, NEU-10, and PA-10 (another control peptide with a
randomized LL-10 peptide sequence) were compared with Polymyxin B (PMB)
in their ability to neutralize endotoxin. Most of the cyclic peptides
tested failed to block endotoxin-induced TNF- secretion by the cell
line (Fig. 2A). However, the
cyclic peptide LALF-22 (which consists of the complete
endotoxin-binding domain of LALF) was effective in that it blocked up
to 65% of the endotoxin-induced TNF- secretion (Fig. 2A). In fact, the inhibitory activity of LALF-22 was
comparable with that of PMB (Fig. 2A). As expected, the
control peptides NEU-10 and PA-10 did not influence TNF- production
in this in vitro system (Fig. 2A). To verify that the
peptides were not toxic for ANA-1 cells, the metabolic activity of the
cells was analyzed using the MTT assay. As demonstrated in Fig. 2B, the peptides did not diminish the metabolism of MTT. The
statistical significance of the data are presented in Table II. Whereas the extent of LALF-22
mediated inhibition of TNF- release by ANA-1 cells differed
significantly from inhibition by the negative control peptide PA-10,
this was not the case for the peptides H-14, B-14, and LALF-14. Because
peptides LALF-14 and H-14 bind lipid A and block the LBP/lipid A
interaction but failed to neutralize natural endotoxin-induced cytokine
production, we conclude that effective peptides, such as LALF-22, need
to consist of the complete endotoxin-binding domain.
|
|
Endotoxin has been described to form multimeric complexes
(35). In an attempt to increase the biological potential
of endotoxin-binding peptides, we reasoned that two endotoxin-binding
domains encoded by one peptide might increase the binding activity to
endotoxin because of an avidity effect. Therefore, we developed
peptides consisting of two endotoxin-binding domains ("dimeric"
peptides). It was important for the design of the dimeric peptides to
select these peptides with high endotoxin-binding capacity and with as
few amino acids as possible because the peptide synthesis efficiency
decreases with the growing length of the peptide chain. The peptides
LL-10-H-14 and LL-10-L-14 are based on the linear LALF-derived peptide
LL-10 fused with the cyclic LBP-derived peptide H-14 or the cyclic
LALF-derived peptide LALF-14, respectively (Table I). Peptide B-10-H-14
consists of the linear BPI-derived peptide B-10 plus the cyclic
LBP-derived peptide H-14 (Table I). We used the combination of a linear
with a cyclic peptide, as the combination of two cyclic peptides with
the need of four cysteines would have been to complicated to be
prepared. We then compared the ability to block the CD14/endotoxin
interaction of the dimeric peptides with the cyclic peptides LALF-14
and H-14. As detailed in Fig. 3,
LL-10-H-14, LL-10-L-14, and B-10-H-14 inhibited binding of endotoxin to
CD14 to a higher degree than LALF-14 or H-14. Importantly, the potency
of LL-10-H-14 to neutralize endotoxin and thus to suppress the release
of TNF- by stimulated ANA-1 cells was higher than that of LALF-22
and was at least as effective as PMB (Fig. 4A). The magnitude of
suppression of the TNF- release could not be enhanced by the
addition of the single peptides LL-10 and H-14 to the cell culture,
indicating that the two endotoxin-binding domains of LL-10-H-14 must be
in close proximity for efficient neutralization of endotoxin (Fig. 4A). The dimeric peptide LL-10-L-14, in which the cyclic
LBP-derived domain was substituted with a LALF-derived cyclic domain,
was incapable of blocking TNF- release (Table I, Fig. 4B). Substitution of the linear part of LL-10-H-14 with a
linear peptide derived from BPI resulting in the peptide B-10-H-14
(Table I) also diminished its capacity to inhibit TNF- release (Fig. 4B). In addition, the two dimeric proteins LL-10-L-11
(consisting of a linear LALF-derived part plus a shorter cyclic
LALF-derived part) and B-10-L-14 (consisting of a linear BPI-derived
part plus a cyclic LALF-derived part) displayed no capacity to inhibit
TNF- release (data not shown). Statistical analysis of the data
revealed that LL-10-H-14 differed significantly in its potential to
block the release of endotoxin-mediated TNF- release in comparison
to the negative control peptide PA-10 (Table II). There was no
difference when PMB and LL-10-H-14, or LL-10-H-14 and LALF-22 were
compared (Table II). These data suggested that LL-10-H-14 and LALF-22
display the highest potency for the neutralization of endotoxin among
the peptides examined.
|
|
secretion of the human macrophage cell
line Mono Mac 6 as well as peripheral blood leukocytes is reduced by
endotoxin-binding peptides
To address the question of whether endotoxin-binding peptides are
also able to reduce TNF- release of human cells on endotoxin
stimulation, we used the human macrophage cell line Mono Mac 6 and
tested the peptides in the same system as described for the murine cell
line ANA-1. As demonstrated in Fig. 5A, LL-10-H-14 reduced TNF-
levels by 
66%. The observation that LALF-22 was ineffective in
blocking TNF- release from the human Mono Mac 6 cell line contrasted
to the results obtained with the murine ANA-1 cell line. As described
for murine cells, the combination of peptides LL-10 plus H-14 was also
incompetent to block the release of TNF- by endotoxin-stimulated
Mono Mac 6 cells (Fig. 5A). A statistical analysis of the
data showed that LL-10-H-14 differed significantly from the control
peptide PA-10 in its potential to inhibit TNF- release by Mono Mac 6
cells (Table II). This was not the case for LALF-22. Comparison of
LL-10-H-14 with PMB showed no difference; however, LL-10-H-14 differed
significantly from LALF-22 (Table II). Finally, we analyzed human PBL.
Fig. 5B shows that LL-10-H-14 was also able to diminish the
endotoxin-stimulated TNF- release. Again LALF-22 was ineffective.
These data show that LL-10-H-14 is the most active peptide on human
cells within the series of peptides analyzed here.
|
release in vivo
To address the issue whether endotoxin-binding peptides are
functional in vivo we analyzed whether LL-10-H-14 blocks systemic
TNF- release in mice challenged with endotoxin. To this C57BL/6 and
BALB/c mice were pretreated with LL-10-H-14 and 15 min later challenged
with endotoxin. TNF- serum levels were determined at the peak
response, i.e., 75 min after the endotoxin challenge. Fig. 6A demonstrates that
LL-10-H-14 inhibited the endotoxin-induced TNF- serum level by
75% in both strains of mice and was as effective as PMB. When
compared with the linear peptide LL-10 and the cyclic peptides LALF-14
and LALF-22, LL-10-H-14 appeared to be more effective, although there
were no differences to LALF-22 (Fig. 6B). Statistical
evaluation of the data revealed that only LL-10-H-14 and LALF-22 were
significantly different to the control peptide PA-10 (Table II). As
described for the experiments with murine ANA-1 cells, there was no
difference between LL-10-H-14 and LALF-22 (Table II). To evaluate how
long the peptide LL-10-H-14 is effective in vivo the time interval
between peptide injection and endotoxin challenge was varied from
1 h pre- to 1 h post-endotoxin challenge. TNF- serum
levels were suppressed maximally if peptide was injected in an interval
15 min before or after endotoxin challenge (Fig. 6C). Longer
intervals of pretreatment or a further delayed injection were less or
not effective (Fig. 6C). Obviously, LL-10-H-14 is functional
in vivo for 15 to 30 min.
|
It is most critical to determine whether a synthetic peptide is
able to prevent endotoxin-induced lethality and therefore might be a
candidate for treatment of septic shock. We therefore used the murine
endotoxin/D-GalN shock model (36).
D-GalN sensitizes mice to the toxic effects of TNF- by
several orders of magnitude by an ill-defined, UTP-depleting,
hepatotoxic mechanism (37). Pretreatment of
D-GalN-sensitized C57BL/6 mice with the peptide 15 min
before the endotoxin challenge increased the survival rate to 100%
(Fig. 7). In this experiment, mice were
sensitized with 4 mg D-GalN/mouse. Higher concentrations of
D-GalN (which renders mice more susceptible to TNF-
(36)) reduced the ability of LL-10-H-14 to increase
survival of endotoxin challenged mice. Accordingly, sensitization with
8 mg D-GalN/mouse resulted in a survival rate of only 40%
of LL-10-H-14 pretreated mice (data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and Ref. 31) and the LBP/lipid A
interaction (data not shown). LBP catalyzes the transfer of endotoxin
from the solubilized to the membrane-bound form; i.e., it promotes the
binding of endotoxin to the endotoxin receptor CD14 on macrophages, and
this process is thought to represent the starting point of the shock
cascade (1, 3, 42). In view of the ability of the peptides
to block this interaction, we anticipated that they inhibit TNF-
secretion by endotoxin-stimulated macrophages. However, it turned out
that the cyclic peptides H-14, B-14, and LALF-14 failed to be effective
(Fig. 2A). In contrast, the cyclic peptide LALF-22 which
consists of the complete endotoxin-binding domain of LALF was more
effective (Fig. 2A). Thus, inhibition of endotoxin-induced
TNF- release may require larger structures than the minimal
endotoxin-binding domains, which could be explained by the decreased
affinity of the peptides as compared with the wild-type proteins.
Alternatively, the different requirements are related to the structural
differences between lipid A used for the binding to rLALF or LBP and
natural endotoxin which was taken for the stimulation of macrophages.
However, this seems unlikely because H-14 and LALF-14 bind to natural
endotoxins of different Enterobacteriaceae (Ref. 31 and
data not shown). The reason why LALF-22 was inactive on human cells is
at present unclear (Fig. 5). In both in vitro settings, the identical
endotoxin was used. Furthermore, LALF-22 was also ineffective when the
human cells were stimulated by endotoxin in the presence of FCS, which
was also used in the murine system, ruling out the possibility that
different LBPs might be responsible for the differences observed (data
not shown). The sequences of mouse and human CD14 show 66% amino acid
identity (43). Therefore, it is possible that endotoxin is
recognized by mouse and human CD14 in such a different way that LALF-22
is able or unable to interfere with this process, respectively.
To further increase the efficiency of endotoxin-binding peptides, we
created a series of dimeric peptides, i.e., peptides containing two
endotoxin-binding domains. The idea for the development of dimeric
peptides was based on the rationale to increase the avidity of the
peptide for endotoxin and that natural endotoxin-binding proteins like
LBP may bind more than one endotoxin molecule per molecule of LBP
(44). The dimeric peptide LL-10-H-14 consisting of the
linear LALF-derived peptide LL-10 and the cyclic LBP-derived peptide
H-14 displayed the highest activity of the series of peptides tested
here to block the release of TNF- by human and murine cells in
repeated experiments (Figs. 4A and 5). The combination of
the single peptides LL-10 plus H-14 did not result into the same
effectiveness (Figs. 4A, 5, and 6B), emphasizing
the importance of two endotoxin-binding domains being in close
proximity. The exchange of the cyclic part H-14 with different
LALF-derived cyclic peptides as well as substitution of LL-10 with B-10
destroyed the functional capacity of the dimeric peptide (Fig. 4B). Thus, endotoxin-binding domains from different
endotoxin proteins cannot be exchanged on the peptide level, in
contrast to mutant LBP proteins where the endotoxin-binding domain of
LBP can be substituted with the domain from LALF as well as BPI without
destroying the function of the protein (24).
There is evidence that peptides derived from LBP, BPI and LALF
containing the endotoxin-binding domains may inhibit the
endotoxin-mediated release of TNF- in vivo (45).
However, the interpretation of these data is clouded because in these
experiments the peptides were mixed with endotoxin in vitro and the
mixture used for challenge in vivo. Hence, we separated the injection
of the peptide and endotoxin in time to mimic more closely the clinical
situation. Using this injection protocol, we show that the peptide
LL-10-H-14 was able to inhibit endotoxin-triggered TNF- release in
vivo (Fig. 6). Although LL-10-H-14 did not completely impair TNF-
production, mice were protected from endotoxin-mediated lethal shock
(Fig. 7). A complete blockade of TNF- production is probably not
desirable because physiological concentrations of TNF- are required
to orchestrate inflammatory reactions aimed to localize bacterial
invaders. For example, using the sublethal model of cecal ligation and
puncture, Echtenacher et al. (13) showed that mice treated
with a neutralizing anti TNF- Ab succumbed to peritonitis, whereas
untreated mice were able to control the infection. The ability of
LL-10-H-14 to protect mice from lethal endotoxin shock was confined to
a narrow window of D-GalN concentrations. Thus, mice
sensitized with higher concentrations of D-GalN were only
partially protected, and although the TNF- levels were reduced they
were presumably still sufficient to induce lethal shock.
The peptide LL-10-H-14 was as effective as PMB. Although in some experiments, LL-10-H-14 seemed to be more active, the statistical analysis revealed no differences between the two compounds. The biggest disadvantage of PMB is its well known toxicity which impairs severely its clinical use. Adverse reactions include nephrotoxicity, which is the most serious toxic effect, neurotoxicity, and hypersensitive reactions. We observed no toxic side effects in mice bolus-injected with LL-10-H-14. On the other, we did not analyze the potential toxicity of the compound after prolonged or repeated administration. Clearly, additional experiments need to be done to show that the peptide LL-10-H-14 is less toxic than PMB.
A limitation of the peptide LL-10-H-14 is the fact that the peptide is
only active in vivo for 30 min (Fig. 6C). Because of its
low m.w., the agent is presumably rapidly cleared via the kidneys.
Because endotoxin is given as a bolus injection in the model analyzed
here, TNF- serum levels are increased for only a short period of
time (4 h). In this case, the short period where the peptide
LL-10-H-14 is effective in vivo is sufficient to inhibit systemic
TNF- secretion. However, in experimental models of bacterial
infection as well as in an infectious clinical situation, a continuous
supply of endotoxin is generated. In this situation, peptides like
LL-10-H-14 must be either administered continuously or coupled to,
e.g., protein carriers to increase their half-life in vivo. Experiments
addressing this issue will be performed.
In comparison with murine macrophages, human macrophages are much more
sensitive toward endotoxin; i.e., picogram amounts of endotoxin are
sufficient to trigger TNF- secretion in these cells. Therefore, the
effectiveness of the peptide LL-10-H-14 to block endotoxin-mediated
TNF- secretion by human macrophages could be much lower. However, as
compared with all other peptides, only LL-10-H-14 was able to block
release of TNF- by endotoxin-stimulated human macrophages and PBMC
to an extent similar to that observed with murine macrophages (Fig. 5).
We think that the peptide LL-10-H-14 could be the basis for the
development of endotoxin-neutralizing drugs for clinical use.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Abbreviations used in this paper: LBP, endotoxin-binding protein; BPI, bactericidal/permeability increasing protein; LALF, Limulus anti-LPS factor; D-GalN, D-galactosamine; PMB, PMB sulfate; rLALF, recombinant LALF; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; VLE, very low endotoxin.
Received for publication August 2, 1999. Accepted for publication February 17, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
/cachectin in mice. J. Exp. Med. 167:1987. antibody protects against lethal meningococcaemia. Mol. Microbiol. 6:591.[Medline]
secretion in vitro. Surgery 118:318.[Medline]
This article has been cited by other articles:
|
|
 |
|
  J. Andra, T. Gutsmann, P. Garidel, and K. Brandenburg Invited review: Mechanisms of endotoxin neutralization by synthetic cationic compounds Innate Immunity, October 1, 2006; 12(5): 261 - 277. [Abstract] [PDF]
|
|
|
|
 |
|
 D. Okuda, S. Yomogida, H. Tamura, and I. Nagaoka Determination of the Antibacterial and Lipopolysaccharide-Neutralizing Regions of Guinea Pig Neutrophil Cathelicidin Peptide CAP11. Antimicrob. Agents Chemother., August 1, 2006; 50(8): 2602 - 2607. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. McMichael, A. Roghanian, L. Jiang, R. Ramage, and J.-M. Sallenave The Antimicrobial Antiproteinase Elafin Binds to Lipopolysaccharide and Modulates Macrophage Responses Am. J. Respir. Cell Mol. Biol., May 1, 2005; 32(5): 443 - 452. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Santamaria, S. Larios, S. Quiros, J. Pizarro-Cerda, J.-P. Gorvel, B. Lomonte, and E. Moreno Bactericidal and Antiendotoxic Properties of Short Cationic Peptides Derived from a Snake Venom Lys49 Phospholipase A2 Antimicrob. Agents Chemother., April 1, 2005; 49(4): 1340 - 1345. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Geetha, S.G. Venkatesh, L. Bingle, C.D. Bingle, and S.-U. Gorr Design and Validation of Anti-inflammatory Peptides from Human Parotid Secretory Protein Journal of Dental Research, February 1, 2005; 84(2): 149 - 153. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Matsuzaki, H. Kobayashi, T. Yagyu, K. Wakahara, T. Kondo, N. Kurita, H. Sekino, K. Inagaki, M. Suzuki, N. Kanayama, et al. Bikunin Inhibits Lipopolysaccharide-Induced Tumor Necrosis Factor Alpha Induction in Macrophages Clin. Vaccine Immunol., November 1, 2004; 11(6): 1140 - 1147. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Silverstein Review: D-Galactosamine lethality model: scope and limitations Innate Immunity, June 1, 2004; 10(3): 147 - 162. [Abstract] [PDF]
|
|
|
|
|
|
M. d. J. Arana, M. G. Vallespi, G. Chinea, G. V. Vallespi, I. Rodriguez-Alonso, H. E. Garay, W. A. Buurman, and O. Reyes Inhibition of LPS-responses by synthetic peptides derived from LBP associates with the ability of the peptides to block LBP-LPS interaction Innate Immunity, October 1, 2003; 9(5): 281 - 291. [Abstract] [PDF]
|
|
|
|
 |
|
 E. S. Van Amersfoort, T. J. C. Van Berkel, and J. Kuiper Receptors, Mediators, and Mechanisms Involved in Bacterial Sepsis and Septic Shock Clin. Microbiol. Rev., July 1, 2003; 16(3): 379 - 414. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. G. Scott, D. J. Davidson, M. R. Gold, D. Bowdish, and R. E. W. Hancock The Human Antimicrobial Peptide LL-37 Is a Multifunctional Modulator of Innate Immune Responses J. Immunol., October 1, 2002; 169(7): 3883 - 3891. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Nagaoka, S. Hirota, F. Niyonsaba, M. Hirata, Y. Adachi, H. Tamura, S. Tanaka, and D. Heumann Augmentation of the Lipopolysaccharide-Neutralizing Activities of Human Cathelicidin CAP18/LL-37-Derived Antimicrobial Peptides by Replacement with Hydrophobic and Cationic Amino Acid Residues Clin. Vaccine Immunol., September 1, 2002; 9(5): 972 - 982. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Augusto, J. Li, M. Synguelakis, J. Johansson, and R. Chaby Structural Basis for Interactions between Lung Surfactant Protein C and Bacterial Lipopolysaccharide J. Biol. Chem., June 21, 2002; 277(26): 23484 - 23492. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Nagaoka, S. Hirota, F. Niyonsaba, M. Hirata, Y. Adachi, H. Tamura, and D. Heumann Cathelicidin Family of Antibacterial Peptides CAP18 and CAP11 Inhibit the Expression of TNF-{alpha} by Blocking the Binding of LPS to CD14+ Cells J. Immunol., September 15, 2001; 167(6): 3329 - 3338. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
