C1 Inhibitor Prevents Endotoxin Shock Via a DirectInteraction with Lipopolysaccharide1,2
Dongxu Liu,* Shenghe Cai,* Xiaogang Gu,* Jennifer Scafidi,* Xiao Wu,† and Alvin E. Davis III3*
C1 inhibitor (C1INH) is beneficial in animal models of endotoxemia and sepsis. However, the mechanism(s) of C1INH protectionremain(s) ill-defined. In this study, we demonstrated that both active C1INH and reactive center-cleaved, inactive C1INH pro-tected mice from lethal Gram-negative endotoxemia. Both forms of C1INH blocked the LPS-binding protein-dependent bindingof Salmonella typhimurium LPS to the murine macrophage cell line, RAW 264.7, and suppressed LPS-induced TNF-� mRNAexpression. Inhibition of LPS binding to RAW 264.7 cells was reversed with anti-C1INH Ab and was more efficient when C1INHwas incubated first with LPS rather than with the cells. C1INH also suppressed LPS-induced up-regulation of TNF-� mRNA inwhole human blood. The interaction of C1INH with LPS was directly demonstrated both by ELISA and by nondenaturing PAGE,but deletion of the amino-terminal 97-aa residues abrogated this binding. Therefore, C1INH, in addition to its function as a serineprotease inhibitor, has a novel anti-inflammatory function mediated via its heavily glycosylated amino-terminal non-serpindomain. The Journal of Immunology, 2003, 171: 2594–2601.
L ipopolysaccharide is a major constituent of the outermembrane of Gram-negative bacteria and is a key mole-cule in the pathogenesis of Gram-negative endotoxemia,
sepsis, and septic shock (1). Gram-negative endotoxemia is ac-companied by contact system activation, complement activation,production of cytokines, and other evidence of unregulated inflam-matory responses (2, 3). LPS activates mononuclear phagocytes toproduce and release inflammatory mediators, of which TNF-� ap-pears to be very important for the development of endotoxin shock(4). LPS interacts with the LPS-binding protein (LBP)4 and trans-fers LPS to CD14 (5–8). The formation of LPS-CD14 complexesinitiates intracellular signaling by binding to Toll-like receptorsexpressed on mononuclear phagocytes and other cells (9). Whenpure LPS or bacterial outer membrane fragments are injected intothe bloodstream, a large fraction of the LPS is cleared by the liverwithin 10 min (10, 11), whereas most of the remaining LPS bindsrapidly to plasma proteins, such as lipoproteins, which inhibit itsbiologic activity (12–15).
C1 inhibitor (C1INH) is the only inhibitor of the classical com-plement pathway proteases, C1r and C1s (16), and is the majorinhibitor of factor XII and prekallikrein of the contact system (17,18). The complement system has been implicated in both thepathogenesis of, and protection from, endotoxin shock (19). Thecontact system also appears to play a role in the mediation of septicshock (20). Levels of proteolytically inactivated C1INH are in-
creased in fatal septic shock, which suggests an increased turnoverand a relative secondary deficiency of biologically active C1INH(21). C1INH can be inactivated by limited proteolytic cleavage byelastase released from activated neutrophils (19, 22). The inacti-vation of C1INH may occur locally in inflamed tissue and therebycontribute to increased local complement activation (22). The di-rect biologic effects, if any, of inactivated C1INH remain un-known. Therapy with C1INH has been shown to improve outcomein several animal models of sepsis (19, 23–27). In addition, pre-liminary data suggest that C1INH may have beneficial effects inseptic shock in humans (19).
In this study, we demonstrated that native, active C1INH andreactive center-cleaved, inactive C1INH (iC1INH) protected micefrom lethal Gram-negative endotoxemia. This protection was as-sociated with inhibition of LPS-triggered macrophage expressionof TNF-� mRNA. Furthermore, C1INH interacts directly withLPS, and this binding appears to be a function of the amino-ter-minal mucin domain of the protein. These data provide evidencethat C1INH, in addition to its function as a serine protease inhibitor,serves as an anti-inflammatory effector via this new mechanism.
Materials and MethodsMouse endotoxemia model
C57BL/6J mice (male and female, 6–8 wk, 18–22 g; Charles River Breed-ing Laboratories, Wilmington, MA) were injected i.p with a lethal dose ofSalmonella typhimurium LPS (20 mg/kg; Sigma-Aldrich, St. Louis, MO)following treatment i.p. or i.v. with C1INH (200 �g/mouse; AdvancedResearch Technologies, San Diego, CA) or iC1INH (200 �g/mouse). Inother experiments, mice were injected i.p. with a mixture of LPS (20 mg/kg) and C1INH (200 �g/mouse) or iC1INH (200 �g/mouse). Control micewere injected with LPS (i.p.) or C1INH (i.v.) alone. Mice were monitoredfor 5 days. Each of the treated groups was compared with the control groupthat received LPS alone using the log rank test (GraphPad Prism version3.0; GraphPad Software, San Diego, CA). All experiments were performedin compliance with relevant laws and institutional guidelines and were ap-proved by the Center for Blood Research Animal Care and Use Committee.
Flow cytometry
The murine macrophage cell line RAW 264.7 (American Type CultureCollection, Manassas, VA) was incubated with FITC-conjugated S. typhi-murium LPS (175 ng/ml; Sigma-Aldrich) in the presence of active C1INH(10–150 �g/ml), C1INH-C1s complexes, iC1INH (1–150 �g/ml), or arecombinant full-length C1INH (40 �g/ml), and a truncated C1INH (50
*Center for Blood Research and †Surgical Research Laboratories, Children’s Hospi-tal, Harvard Medical School, Boston, MA 02115
Received for publication January 10, 2003. Accepted for publication June 20, 2003.
The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by National Institutes of Health Grants HD22082 andHD33727 (to A.E.D.).2 Portions of this work were presented at the 19th International Complement Work-shop, September 22–26, 2002, Palermo, Italy.3 Address correspondence and reprint requests to Dr. Alvin E. Davis III, Center forBlood Research, 800 Huntington Avenue, Boston, MA 02115. E-mail address:[email protected] Abbreviations used in this paper: LBP, LPS-binding protein; C1INH, C1 inhibitor;iC1INH, reactive center cleaved, inactive C1INH.
�g/ml) with deletion of the amino-terminal 97 aa (28) in DMEM contain-ing 10% FBS (15 min, 37°C; Life Technologies, Grand Island, NY).C1INH-C1s complexes were prepared by incubation of C1INH (150 �g/ml) with C1s (150 and 300 �g/ml, 60 min, 37°C; Advanced ResearchTechnologies) and iC1INH was generated by incubation of active C1INHwith trypsin attached to cross-linked agarose (15 min, 37°C; Sigma-Al-drich) (29, 30). In other experiments, the macrophages were incubated withFITC-conjugated LPS and C1INH that was pretreated with rabbit anti-human C1INH Ab (1 h, 37°C; DAKO, Glostrup, Denmark).
Fluorescence microscopy
Macrophages were plated on glass coverslips (Fisher Scientific, Pittsburgh,PA), treated with FITC-conjugated LPS (175 ng/ml) in the presence ofDMEM including 10% FBS without or with C1INH (50, 100, and 150�g/ml, 15 min, 37°C) (31, 32), washed with PBS, and fixed with 4%formaldehyde. Fluorescence localization was evaluated by Axiophot fluo-rescence microscopy (Zeiss, Oberkochen, Germany) with a green fluores-cence filter set.
RT-PCR
Total RNA was isolated from the RAW 264.7 macrophages induced withLPS (175 ng/ml) in the presence or absence of C1INH and iC1INH andwas reverse transcribed using Moloney murine leukemia virus reverse tran-scriptase (New England Biolabs, Beverly, MA) with oligo(dT)20 primers (1h, 37°C; Invitrogen, Carlsbad, CA). PCR primers were designed to gen-erate mouse TNF-� and �-actin fragments with lengths of 200 bp (TNF-�,sense: 5�-ATGAGCACAGAAAGCATGATCC-3� and antisense: 5�-GAGGCCATTTGGGAACTTCTC-3�; �-actin, sense: 5�-TGGATGACGATATCGCTGC-3� and antisense: 5�-AGGGTCAGGATACCTCTCTT-3�).PCR products were analyzed on 1.2% (w/v) agarose gels containing 0.5�g/ml ethidium bromide and visualized under UV light. Band density wasanalyzed and quantified using ImageQuant software (Molecular Dynamics,Sunnyvale, CA). In addition, human peripheral venous blood from a nor-mal volunteer was collected in EDTA (1 mg/ml whole blood). Aliquots ofthe whole blood were treated with LPS at a final concentration of 175ng/ml in the absence and the presence of added C1INH (5–150 �g/ml) for15 min at 37°C. Total RNA was isolated from the blood leukocytes andwas reverse transcribed using Moloney murine leukemia virus reverse tran-scriptase with oligo(dT)20 primers. PCR primers were designed for humanTNF-� (sense: 5�-ATGAGCACTGAAAGCATGATCCGGGACGTG-3�and antisense: 5�-AGGTCCCTGGGGAACTCTTCCCTCTG-3�) and human�-actin (sense: 5�-ATGGATGATGATATCGCCGCGCTCGTCGTC-3� andantisense: 5�-AGGGTGAGGATGCCTCTCTTGCTCTG-3�).
ELISA
Plates (Costar; Corning, Corning, NY) were coated with S. typhimuriumLPS (100 �l at a variety of concentrations) at 4°C overnight. Control plateswere incubated with BSA (100 �g/ml; New England Biolabs), IgG (20�g/ml; Sigma-Aldrich), or PBS in the absence of C1INH. C1INH (100 �lof 150 �g/ml) was incubated with LPS-coated plates for 1 h at room tem-perature in the presence or absence of FBS (10–100 �l) or human LBPpeptides (5–40 ng/ml; Cell Sciences, Norwood, MA). In addition, variousconcentrations of iC1INH and C1INH (100 �l) were incubated with LPS(100 �l of 87.5 ng/ml)-coated plates for 1 h at room temperature. Rabbitanti-human C1INH Ab (1/1000) was incubated for 1 h at room tempera-ture, after which plates were incubated with Immunopure goat anti-rabbitIgG (H � L) conjugated with HRP (1/100,000); Pierce, Rockford, IL). Inother experiments, recombinant truncated C1INH (100 �l of 50 �g/ml) andfull-length C1INH (100 �l of 40 �g/ml) were incubated with LPS (100 �lof 175 ng/ml)-coated plates for 1 h at room temperature. After washingwith PBS, o-phenylenediamine dihydrochloride (Sigma-Aldrich) substratewas added and the color reactions were developed for 5 min at room tem-perature and terminated with 3 N HCl. Absorbance was measured at 490nm using the Revelation Microsoft in an MRX microplate reader (DYNEXTechnologies, Chantilly, VA).
PAGE and Western blotting
Nondenaturing PAGE of C1INH (10 �g), recombinant full-length C1INH(40 �g/ml), and truncated C1INH (50 �g/ml) incubated with various con-centrations of LPS (37°C, 30 min in PBS) were performed as previouslydescribed (32). Proteins were electrophoretically transferred to nitrocellu-lose membranes (Invitrogen), blocked with 5% fat-free milk (Bio-Rad,Hercules, CA) in PBS (pH 7.4) containing 0.05% Tween 20 at 4°C over-night, incubated with rabbit anti-human C1INH Ab (1:1000) in 5% fat-freemilk for 2 h at room temperature, washed with 1� PBS for 20 min, andincubated with a 1/10,000 dilution of Immunopure goat anti-rabbit IgG (H
� L) conjugated with HRP (2 h, room temperature; Pierce). Developmentwas performed using a SuperSignal Chemiluminescent Substrate kit(Pierce). In other experiments, C1INH (0, 1, and 10 �g) was incubatedwith 3H-LPS (List Biological Laboratories, Campbell, CA) at 37°C for 30min as previously described (32). The gel containing 3H-LPS was fixed in10% methanol and 10% acetic acid for 1 h at room temperature, dried for1 h at 80°C, and exposed to Kodak XAR film (Kodak, Rochester, NY) for10 days. For SDS-PAGE, C1INH was incubated with LPS (175 ng/ml, 30min, 37°C), after which C1s was added for 5, 10, 15, 30, and 60 min.Loading buffer was added, and the mixture was subjected to electrophore-sis using a 10% SDS-Tris-glycine polyacrylamide gel (Invitrogen). The gelwas stained with Coomassie brilliant blue and dried (1 h, 80°C).
ResultsC1INH protection from lethal endotoxin shock
We analyzed the ability of both native, active C1INH and iC1INHderived from C1INH by trypsin treatment, to improve the survivalof mice with Gram-negative endotoxemia. The lowest dose of LPS(20 mg/kg) that resulted in 100% mortality of C57BL/6J mice (n� 16) within 48 h was selected (Fig. 1). A single dose of C1INH(200 �g) improved survival to 45% (n � 20, p � 0.0008) and 50%(n � 20, p � 0.0003) when administered via the i.p. and i.v. routes,respectively. Survival was 60% at 72 h (n � 10, p � 0.0002) inmice treated i.v. with a single dose of iC1INH (200 �g). Therefore,iC1INH, like active C1INH, protects mice from the lethal effectsof LPS.
Because iC1INH was protective, we hypothesized that C1INHmight interact directly with endotoxin. We treated mice with amixture of LPS (20 mg/kg) and C1INH (200 �g) i.p., which in-creased survival to 65% at 72 h (n � 20, p � 0.0001). Similarly,the mixture of LPS (20 mg/kg) with iC1INH (200 �g i.p.) in-creased survival to 100% (n � 10, p � 0.0001). These data indi-cate that C1INH acts to prevent the adverse biologic effects of LPSvia a mechanism unrelated to protease inhibition and is consistentwith protection secondary to a direct interaction with endotoxin.
FIGURE 1. The effect of native, active C1INH and reactive center loop-cleaved, inactive C1INH on survival of mice in Gram-negative endotoxinLPS-induced lethal endotoxemia. Mice (C57BL/6J) were injected with ei-ther LPS i.p. following treatment with C1INH i.p. (f) or i.v. (F), withiC1INH i.v. (E), or with mixtures of LPS and C1INH i.p. (Œ) or iC1INHi.p. (‚). Control mice were injected with LPS i.p. alone (�) or withC1INH i.v. (n � 4) alone (�). The indicated p values are for each treat-ment group in comparison to the group treated with LPS only.
To determine whether the C1INH used in these studies might becontaminated with high-density lipoprotein, purified C1INH wastested with anti-apolipoprotein A1 Ab using a solid-phase capturesandwich ELISA as compared with standard human plasma (1/104
dilution). Apolipoprotein A1 could not be detected in the C1INHpreparation (data not shown).
C1INH blocks LPS binding to macrophages
FITC-labeled LPS (175 ng/ml), in the presence of 10% FBS as asource of LBP, binds to the murine macrophage cell line RAW264.7, but does not bind in the absence of FBS (Fig. 2A). C1INH,at concentrations of 37.5–150 �g/ml, which are within the phys-iological concentration range in human plasma (21), completelyblocked this binding (Fig. 2A). The fluorescence intensity also wasdecreased when macrophages were pretreated with C1INH (150�g/ml) for 15 min at 37°C, after which the C1INH was removed
and LPS added. However, the effect was only observed at a lowerconcentration of LPS (40 ng/ml) and was not apparent at 175ng/ml LPS (Fig. 2B). This suggested that C1INH interacts primar-ily with endotoxin rather than with a cellular receptor. Anti-C1INHAb (175 �g/ml) completely abrogated the effect of C1INH (75�g/ml) on LPS binding to macrophages, whereas BSA (100 �g/ml) did not interfere (Fig. 2C). To visualize the binding of LPS tomacrophages, RAW 264.7 cells were cultured on microscopeslides in the presence of FITC-conjugated LPS (175 ng/ml) thathad been incubated with or without C1INH. The fluorescentsignals were decreased at a concentration of 50 �g/ml C1INHand were eliminated at 100 –150 �g/ml C1INH (Fig. 2D). LPS-induced TNF-� mRNA was detected in RAW 264.7 cells usingRT-PCR. LPS-mediated up-regulation of TNF-� mRNA wascompletely suppressed following treatment with 150 �g/mlC1INH (Fig. 3A). Dose-response analysis showed that LPS (175
FIGURE 2. The effect of C1INH on the binding of FITC-conjugated LPS to the murine macrophage cell line RAW 264.7. RAW 264.7 macrophageswere incubated in DMEM containing 10% FBS at 37°C (FITC-LPS binding, thick line; control, shaded field for A–C). A, The binding of LPS (175 ng/ml)to RAW 264.7 cells in the presence of C1INH (10, 37.5, 75, and 175 �g/ml). B, The binding of LPS (40 and 175 ng/ml) to RAW 264.7 cells after incubationof the cells with C1INH (150 �g/ml) for 60 min at 37°C. C, Anti-C1INH Ab (175 �g/ml) reversed the inhibitory effect of C1INH (75 �g/ml) on the bindingof LPS (175 ng/ml) to macrophages, but BSA (100 �g/ml) did not interfere with C1INH-mediated inhibition (150 �g/ml) of LPS (175 ng/ml) binding tomacrophages. D, The effect of C1INH (50, 100, and 150 �g/ml) on the binding of LPS (175 ng/ml) to macrophages analyzed by fluorescence microscopy.
2596 C1INH INTERACTS DIRECTLY WITH GRAM-NEGATIVE ENDOTOXIN
ng/ml)-induced expression of TNF-� mRNA was completelyinhibited with 37.5 �g/ml of intact active C1INH and with 10�g/ml iC1INH (Fig. 3, B and C). Similarly, C1INH at concen-trations of 37.5–150 �g/ml suppressed LPS-induced TNF-�mRNA expression by cells in whole human blood (Fig. 3D).
Interaction of C1INH with LPS
We immobilized LPS in microtiter wells at a variety of concen-trations and measured the amount of C1INH binding in the pres-ence or absence of BSA (100 �g/ml) or IgG (20 �g/ml). Thebinding of C1INH (150 �g/ml) was maximal to 175 ng/ml LPSand neither BSA nor IgG interfered with this binding (Fig. 4A).Dose-response analysis of C1INH and iC1INH binding to LPS(87.5 ng/ml) showed similar binding curves, although maximalbinding may occur with a somewhat lower concentration ofiC1INH (37.5 �g/ml) than of intact C1INH (75 �g/ml; Fig. 4B).FBS (10–100 �l) and the human LBP peptide (5–40 ng/ml) re-duced C1INH binding by �80 and 75%, respectively (Fig. 4, Cand D). Binding of the negatively charged LPS to proteins resultsin a characteristic anodal shift in the mobility of the protein on
native PAGE (31, 32). C1INH-LPS mixtures, when analyzed bynative PAGE/Western blot with anti-C1INH Ab (Fig. 4E) and bynative PAGE following incubation of C1INH with 3H-LPS (Fig.4F), also demonstrated an anodal shift, which increased with in-creasing amounts of LPS. LPS does not alter the electrophoreticmobility of a variety of other proteins, including OVA, �2-mac-roglobulin, and catalase (32). C1INH incubated with LPS for 30min at 37°C before the addition of C1s had no effect on either therate or extent of complex formation with C1s in comparison toC1INH incubated with C1s in the absence of endotoxin (Fig. 4G).Therefore, although Gram-negative endotoxin LPS binds directlyto C1INH, it neither enhances nor suppresses the ability of C1INHto complex with target protease.
The C1INH amino-terminal domain is responsible for theinteraction with LPS
To investigate whether an intact reactive center loop is required forinhibition of LPS binding to macrophages, we prepared C1INH-C1s complexes and iC1INH (1, 5, 10, 37.5, 75, and 150 �g/ml) by
FIGURE 3. C1INH-mediated inhibition of LPS-induced TNF-� mRNA expression in the murine macrophage cell line RAW 264.7. Total RNA from mac-rophages was isolated after treatment with LPS (175 ng/ml). RT-PCR was performed using mouse TNF-� cDNA and �-actin cDNA primers. A, RAW 264.7 cellswere induced with LPS for 0, 5, 15, 30, 60, and 120 min at 37°C in the presence of C1INH (150 �g/ml). B, RAW 264.7 cells were induced with LPS for 30 minat 37°C in the presence of C1INH (150, 75, 37.5, 10, 5, 1, and 0 �g/ml). C, RAW 264.7 cells were induced with LPS for 30 min at 37°C in the presence of iC1INH(150, 75, 37.5, 10, 5, 1, and 0 �g/ml). D, Total RNA from whole human blood cells was isolated after treatment with LPS (175 ng/ml) in the presence of C1INH(150, 75, 37.5, 10, 5 and 0 �g/ml). RT-PCR was performed using human TNF-� and �-actin oligonucleotide primers.
cleaving C1INH at P1-P1� with trypsin. Both C1INH-C1s com-plexes (Fig. 5A) and iC1INH (Fig. 5B) retained the ability to blockFITC-LPS (175 ng/ml) binding to RAW 264.7 cells. Therefore, anintact reactive center loop is not required for this inhibition.
Binding of a truncated C1INH molecule consisting of aminoacid residues 98–478 (28) to LPS, as assessed by ELISA, wasreduced by 85% compared with full-length C1INH (Fig. 6A). Thetruncated C1INH also did not block the binding of LPS to RAW264.7 cells (Fig. 6B), but full-length C1INH in the presence of thetruncated protein was able to block binding (data not shown). Mix-tures of recombinant full-length C1INH and truncated C1INH withLPS were analyzed using native PAGE followed by Western blot.
The recombinant full-length C1INH-LPS mixture showed an anodalshift (Fig. 6C) very similar to that of native C1INH (Fig. 4E). Theelectrophoretic mobility of the recombinant truncated C1INH, on theother hand, was unchanged when mixed with LPS (Fig. 6D). There-fore, the amino-terminal domain appears to be required for the inhi-bition of LPS-mediated macrophage activation that is a result of theinteraction of C1INH with Gram-negative endotoxin LPS.
DiscussionC1INH is an acute phase protein with a mean plasma level of�250 mg/L. It may increase up to 2.5-fold during inflammation(33). Antigenic levels of C1INH tend to be normal in patients with
FIGURE 4. Interaction of C1INH with LPS. A, Binding of C1INH (150 �g/ml) to immobilized LPS (0, 43.75, 87.5, 175, 350, 700, 1400, and 2800ng/ml) either alone (�) or in the presence of added BSA (f) or IgG (Œ). Controls consisted of BSA (�), IgG (�), or PBS (�) added to plates coatedwith LPS in the absence of C1INH. B, Binding of C1INH (F; 1, 10, 37.5, 75, 150, 300 �g/ml) and iC1INH (f; 1, 10, 37.5, 75, 150, 300 �g/ml) toimmobilized LPS (87.5 ng/ml) assessed by ELISA. C, FBS (0, 10, 50, and 100 �l) competed with C1INH (150 �g/ml) for binding to LPS (175 ng/ml),as detected by ELISA. D, Human LBP peptide (0, 5, 10, 20 and 40 ng/ml) competed with C1INH (150 �g/ml) for binding to LPS (175 ng/ml), as analyzedby ELISA. E, LPS (0.1, 0.5, 1, 5, 10, and 20 �g) alters the electrophoretic mobility of C1INH (10 �g; lanes 2–7 vs lane 1) as assessed by native PAGEand Western blot. F, The electrophoretic mobility of 3H-LPS (0.2 �g) shifts in native PAGE (lane 3 vs lane 1) after incubation with C1INH (10 �g). G,LPS (175 ng/ml) has no effect on the rate or extent of C1INH-C1s complex formation (lanes 3–7 vs lanes 8–12) as assessed by SDS-PAGE stained withCoomassie brilliant blue.
2598 C1INH INTERACTS DIRECTLY WITH GRAM-NEGATIVE ENDOTOXIN
fatal septic shock but their levels of proteolytically inactivatedC1INH are increased (21). This results in a deficiency of biolog-ically active C1INH during sepsis (33), which in turn results inunregulated activation of the complement and contact systems. Inanimal models of sepsis, exogenous C1INH administration pro-vides protection which is associated with reduced production ofcytokines. Therefore, C1INH may have a beneficial effect on theclinical course and outcome of severe sepsis through modulationof cytokine release. This is consistent with the observation thatactivation of complement and/or contact system proteases is asso-ciated with organ injury and impending lethality (27). It has notbeen established whether the improved survival that results fromC1INH administration is a product of its regulation of complementpathway activation, of contact system activation, or whetherC1INH may have another unrelated function that results in atten-uation of the cytokine response. Treatment with C1INH also pro-tects C3- and C4-deficient mice from endotoxin shock (26), whichsuggests that regulation of complement system activation is not themechanism by which C1INH mediates an improved outcome.
LBP transports LPS to macrophage CD14 to form LPS-CD14complexes that initiate intracellular signaling reactions by bindingto Toll-like receptors. This results in the release of a variety ofinflammatory products, including TNF-�. LPS also can interactwith a wide variety of structurally diverse proteins in plasma, in-cluding proteases, complement proteins, lipoproteins, MD-2, L-selectin, P-selectin, and CD14 (31, 34–36). The binding of LPS toRAW 264.7 cells is dependent on the presence of LBP in serum;in the absence of serum, little binding is observed (37). The bind-ing of LPS by other proteins may affect the binding activity of LPSto LBP or the transfer of LPS to macrophages. The data presentedhere suggest the possibility that, in addition to protection via in-hibition of complement and contact system activation, C1INH alsomay protect from endotoxin shock via a direct interaction withendotoxin. We demonstrated that plasma-derived and recombinantfull-length C1INH (Fig. 6, A and C) bind with relatively high af-finity to purified bacterial endotoxin LPS from S. typhimurium(Fig. 4, A, B, E, and F) and that this binding inhibits the bindingof LPS to RAW 264.7 cells (Figs. 2, A and D, and 6B), probablyby preventing the interaction of endotoxin with LBP. This inhibi-tion prevents macrophage activation as shown by suppression ofTNF-� mRNA synthesis by the RAW 264.7 cells (Fig. 3, A–C) andby leukocytes in whole human blood (Fig. 3D).
Somewhat surprisingly, the ratio of C1INH to LPS that is pro-tective in vivo is lower than that required to inhibit LPS binding toRAW 264.7 cells or to bind to LPS in the gel shift experiments.Treatment with C1INH at a C1INH:LPS molar ratio of 1:3 reducedmortality from 100 to 45–50% (Fig. 1). LPS is present in aqueoussolution in micellar form with an effective molecular mass �1million. Upon interaction with proteins to which it binds, it disso-ciates to individual subunits with Mr �4000 (32). In the in vitroexperiments, the molar ratio of C1INH (Mr �100,000) to LPS was�2:1 (Fig. 4F) and �4:1 (Fig. 4E) in the gel shift experiments,which probably provide the best estimate of stoichiometry. A ratioof 9:1 C1INH:LPS was required to inhibit binding of LPS to RAW264.7 cells (Fig. 2A) or to inhibit TNF-� expression by these cells(Fig. 3B). The reason for the difference between the in vivo and invitro experiments is not obvious. However, in vivo, multiple phys-iologic mechanisms contribute to protection from endotoxin in-cluding, among others, the binding of endotoxin to lipoproteins(12–15) and complement-mediated clearance (26). In addition, theanimal also has circulating C1INH that would contribute. There-fore, because of these other mechanisms, the quantity of C1INHrequired for an effect in vivo might be less than would be predictedbased on the in vitro experiments in which only the direct bindingby C1INH would be operative.
Another important question is whether cleaved iC1INH, in fact,provides a greater degree of protection than does the intact activeprotein. A single dose of i.v. iC1INH increased survival from 50 to60%, a difference that was not statistically significant (Fig. 1).However, mixing the iC1INH with LPS before i.p. administrationresulted in an improvement in survival from 65 to 100%, a differ-ence that was statistically significant ( p � 0.0074). These findingsare consistent with the in vitro studies that demonstrated that alower concentration of iC1INH was required to inhibit TNF-� pro-duction by RAW 264.7 cells (Fig. 3, B and C), to bind to LPS inELISA (Fig. 4B), and to inhibit LPS binding to RAW 264.7 cells(Figs. 2A and 5B).
Because LPS can interact with a number of different proteins inplasma, it is possible that a contaminant in C1INH preparationscould be responsible for the results observed here. This seemsextremely unlikely for the following reasons. First, the inability todetect apolipoprotein A1 suggested that the C1INH preparationsare not contaminated with high-density lipoproteins, which also
FIGURE 5. The effect of C1INH-C1s complexes and of reactive centerloop-cleaved, inactivated C1INH on LPS binding (LPS binding, thick line;control, shaded field). A, C1INH was incubated with C1s to form C1INH-C1s complexes. RAW 264.7 cells were incubated with LPS (175 ng/ml) inthe presence of C1INH-C1s complexes. B, C1INH was cleaved at P1-P1�with trypsin to form iC1INH. RAW 264.7 macrophages were incubatedwith LPS (175 ng/ml) in the presence of iC1INH (1, 5, 10, 37.5, 75, and150 �g/ml).
can bind LPS and prevent macrophage activation (12–15). Second,the inhibition of binding of LPS to RAW 264.7 cells by the C1INHpreparation was reversed with an Ab to C1INH. Third, several ofthe effects (inhibition of binding of LPS to RAW 264.7 cells, bind-ing of C1INH to LPS in an ELISA, complex formation demon-stration by gel shift analysis) were duplicated with a recombinantfull-length C1INH protein. It is unlikely that the recombinantC1INH would contain the same contaminating proteins as the plas-ma-derived C1INH. It therefore seems most likely that the findingsdescribed here are the result of a property of C1INH and not of acontaminant.
The primary recognition element for serpin-protease associationis a 15-aa residue-exposed segment known as the reactive centerloop. The specificity of serpins is determined by the amino acidsequence within the reactive center loop, particularly the aminoacids from approximately P5 through P5� (38). As might be ex-pected, this region therefore reveals little homology among differ-ent serpins. Reactive center loop cleavage by either target or non-
target proteases results in a structural rearrangement with completeinsertion of the loop into the five-stranded � sheet A. In the caseof nontarget proteases, the active protease is released, while targetproteases remain covalently bonded to the serpin via the P1 residue(39, 40). In either case, the ability to inhibit additional protease islost. We demonstrated that C1INH-C1s complexes or C1INH in-activated by cleavage of the reactive center loop with trypsin re-tained the ability to block the binding of LPS to macrophages.C1INH has an amino-terminal heavily glycosylated mucin-like do-main (aa 1–120) that contains seven repeats of the tetrapeptidesequence Glx-Pro-Thr-Thr or variants thereof (28, 41). This do-main does not influence complex formation with target proteases(28). The functional significance of the amino-terminal extensionof C1INH is unknown, although several possibilities have beensuggested (28). Deletion of the amino-terminal 97 aa residues ab-rogated the ability of C1INH to bind to LPS. Therefore, one po-tential role for this domain is to participate in host protection fromGram-negative sepsis via this direct binding activity.
FIGURE 6. The effect of recombinant truncated C1INH on LPS binding. A, Recombinant truncated C1INH (50 �g/ml) had greatly reduced binding toLPS (175 ng/ml) as assessed by ELISA. B, FITC-conjugated LPS (175 ng/ml) binding to RAW 264.7 macrophages was not inhibited with recombinanttruncated C1INH (50 �g/ml; LPS binding, thick line; control, shaded field). C, LPS (10 and 20 �g) clearly altered the electrophoretic mobility ofrecombinant full-length C1INH (40 �g/ml) as measured by native PAGE/Western blot (lanes 1 and 2). D, LPS (10 and 20 �g) did not change theelectrophoretic mobility of recombinant truncated C1INH (50 �g/ml) as detected by native PAGE/Western blot (lanes 1 and 2).
2600 C1INH INTERACTS DIRECTLY WITH GRAM-NEGATIVE ENDOTOXIN
The above-described results indicate that protease inhibitory ac-tivity is not required for the binding of LPS to C1INH and that thebinding site(s) is(are) almost certainly contained within the amino-terminal domain. Therefore, it is likely that C1INH contributes toprotection from Gram-negative endotoxin shock via three mecha-nisms: inhibition of excessive complement activation which wouldlimit the amount of C5a generated (42); inhibition of contact sys-tem activation which would limit the amount of activated plasmakallikrein, factor XIIa, and bradykinin generated (19, 20); and bydirect inhibition of endotoxin binding to macrophages whichthereby suppresses macrophage activation.
AcknowledgmentsWe thank Dr. Chester Alper and Dr. Eileen Remold-O’Donnell for criticalreview of this manuscript.
References1. Bone, R. C. 1991. The pathogenesis of sepsis. Ann. Intern. Med. 115:457.2. Dietch, E. A. 1997. Animal models of sepsis and shock: a review and lessons
learned. Shock 9:1.3. Bone, R. C. 1996. Toward a theory regarding the pathogenesis of the systemic
inflammatory response syndrome: what we do and do not know about cytokineregulation. Crit. Care Med. 24:163.
4. Glauser, M. P., G. Zanetti, J. D. Baumgartner, and J. Cohen. 1991. Septic shock:pathogenesis. Lancet 338:732.
5. Poltorak, A., X. He, I. Smirnova, M. Y. Liu, C. van Huffel, X. Du, D. Birdwell,E. Alejos, M. Silva, C. Galanos, et al. 1998. Defective LPS signaling in C3H/HeJand C57BL/10ScCr mice: mutations in the Tlr4 gene. Science 282:2085.
6. Schumann, R. R., and E. Latz. 2000. Lipopolysaccharide-binding protein. Chem.Immunol. 74:42.
7. Kitchens, R. L. 2000. Role of CD14 in cellular recognition of bacterial lipopoly-saccharides. Chem. Immunol. 74:61.
8. Fenton, M. J., and D. T. Golenbock. 1998. LPS-binding proteins and receptors.J. Leukocyte Biol. 64:25.
9. Aderem, A., and R. J. Ulevitch. 2000. Toll-like receptors in the induction of theinnate immune response. Nature 406:782.
10. Munford, R. S., C. L. Hall, J. M. Lipton, and J. M. Dietschy. 1982. Biologicalactivity, lipoprotein-binding behavior, and in vivo disposition of extracted andnative forms of Salmonella typhimurium lipopolysaccharides. J. Clin. Invest.70:877.
11. Mathison, J. C., and R. J. Ulevitch. 1979. The clearance, tissue distribution, andcellular localization of intravenously injected lipopolysaccharide in rabbits. J. Im-munol. 123:2133.
12. Read, T. E., C. Grunfeld, Z. L. Kumwenda, M. C. Calhoun, J. P. Kane,K. R. Feingold, and J. H. Rapp. 1995. Triglyceride-rich lipoproteins preventseptic death in rats. J. Exp. Med. 182:267.
13. Munford, R. S., C. L. Hall, and J. M. Dietschy. 1981. Binding of Salmonellatyphimurium lipopolysaccharides to rat high-density lipoproteins. Infect. Immun.34:835.
14. Pajkrt, D., J. E. Doran, F. Koster, P. G. Lerch, B. Arnet, T. van der Poll,J. W. ten Cate, and S. J. van Deventer. 1996. Anti-inflammatory effects of re-constituted high-density lipoproteins during human endotoxemia. J. Exp. Med.184:1602.
15. Parker, T. S. 1995. Reconstituted high-density lipoprotein neutralizes Gram-neg-ative bacterial lipopolysaccharides in human whole blood. Infect. Immun. 63:253.
16. Sim, R. B., A. Reboul, G. J. Arlaud, C. L. Villiers, and M. G. Colomb. 1979.Interaction of 125I-labelled complement components C1r and C1s with proteaseinhibitors in plasma. FEBS Lett. 97:111.
17. de Agostini, A., H. R. Lijnen, R. A. Pixley, R. W. Colman, and M. Schapira.1984. Inactivation of factor-XII active fragment in normal plasma: predominantrole of C1-inhibitor. J. Clin. Invest. 93:1542.
18. Schapira, M., C. F. Scott, and R. W. Colman. 1982. Contribution of plasmaprotease inhibitors to the inactivation of kallikrein in plasma. J. Clin. Invest.69:462.
19. Caliezi, C., W. A. Wuillemin, S. Zeerleder, M. Redondo, B. Eisele, andC. E. Hack. 2001. C1-esterase inhibitor: an anti-inflammatory agent and its po-tential use in the treatment of diseases other than hereditary angioedema. Phar-macol. Rev. 52:91.
20. Colman, R. W. 1999. Biologic activities of the contact factors in vivo: potenti-ation of hypotension, inflammation, and fibrinolysis, and inhibition of cell adhe-sion, angiogenesis and thrombosis. Thromb. Haemostasis 82:1568.
21. Nuijens, J. H., A. J. M. Eerenberg-Belmer, C. C. M. Huijbregts, W. O. Schreuder,R. J. F. Felt-Bersma, J. J. Abbink, L. G. Thijs, and C. E. Hack. 1989. Proteolyticinactivation of plasma C1 inhibitor in sepsis. J. Clin. Invest. 84:443.
22. Brower, M. S., and P. C. Harpel. 1982. Proteolytic cleavage and inactivation of�2 plasmin inhibitor and C1 inactivator by human polymorphonuclear leukocyteelastase. J. Biol. Chem. 257:9849.
23. Triantaphyllopoulos, D. C., and M. S. Cho. 1986. Effect of injection of C1 in-activator on the platelet count and blood coagulation in rabbits infused withendotoxin. Thromb. Haemostasis 55:293.
24. Guerrero, R., F. Velasco, M. Rodriguez, A. Lopez, R. Rojas, M. A. Alvarez,R. Villalba, V. Rubio, A. Torres, and D. del Castillo. 1993. Endotoxin-inducedpulmonary dysfunction is prevented by C1-esterase inhibitor. J. Clin. Invest. 91:2754.
25. Scherer, R. U., R. M. Giebler, U. Schmidt, D. Paar, and W. J. Kox. 1996. Theinfluence of C1-esterase inhibitor substitution on coagulation and cardiorespira-tory parameters in an endotoxin-induced rabbit model of hypercoagulability. Se-min. Thromb. Hemostasis 22:357.
26. Fischer, M. B., A. P. Prodeus, A. Nicholson-Weller, M. Ma, J. Murrow,R. R. Reid, H. B. Warren, A. L. Lage, F.D. Moore, Jr., F. S. Rosen, andM. C. Carroll. 1997. Increased susceptibility to endotoxin shock in complementC3- and C4-deficient mice is corrected by C1 inhibitor replacement. J. Immunol.159:976.
27. Jansen, P. M., B. Eisele, I. W. de Jong, A. Chang, U. Delvos, J. Taylor, F. B., andC. E. Hack. 1998. Effect of C1 inhibitor on inflammatory and physiologic re-sponse patterns in primates suffering from lethal septic shock. J. Immunol.160:475.
28. Coutinho, M., K. S. Aulak, and A. E. Davis III. 1994. Functional analysis of theserpin domain of C1 inhibitor. J. Immunol. 153:3648.
29. Zahedi, R., J. Wisnieski, and A. E. Davis III. 1997. Role of the P2 residue ofcomplement 1 inhibitor (Ala443) in determination of target protease specificity.J. Immunol. 159:983.
30. Zahedi, R., R. C. MacFarlane, J. J. Wisnieski, and A. E. Davis III. 2001. C1inhibitor: Analysis of the role of amino acid residues within the reactive centerloop in target protease recognition. J. Immunol. 167:1500.
31. Viriyakosol, S., P. S. Tobias, R. L. Kitchens, and T. N. Kirkland. 2001. MD-2binds to bacterial lipopolysaccharide. J. Biol. Chem. 276:38044.
32. Hailmen, E., H. S. Lichenstein, M. M. Wurfel, D. S. Miller, D. A. Jounson,M. Kelly, L. A. Busse, M. M. Zukowski, and S. D. Wright. 1994. Lipopolysac-charide (LPS)-binding protein accelerates the binding of LPS to CD14. J. Exp.Med. 179:269.
33. Kirschfink, M., and W. Nurnberger. 1999. C1 inhibitor in anti-inflammatory ther-apy: from animal experiment to clinical application. Mol. Immunol. 36:225.
34. Ding, A., N. Thieblemont, J. Zhu, F. Jin, J. Zhang, and S. Wright. 1999. Secretoryleukocyte protease inhibitor interferes with uptake of lipopolysaccharide by mac-rophages. Infect. Immun. 67:4485.
35. Malhotra, R., R. Priest, and M. Bird. 1996. Role for L-selectin in lipopolysac-charide-induced activation of neutrophils. Biochem. J. 320:589.
36. Shimazu, R., S. Akashi, H. Ogata, Y. Nagai, K. Fukudome, K. Miyake, andM. Kimoto. 1999. MD-2, a molecule that confers lipopolysaccharide responsive-ness on Toll-like receptor 4. J. Exp. Med. 189:1777.
37. Nagaoka, I., S. Hirota, F. Niyonsaba, M. Hirata, Y. Adachi, H. Tamura, andD. Heumann. 2001. Cathelicidin family of antibacterial peptides CAP18 andCAP11 inhibit the expression of TNF-� by blocking the binding of LPS toCD14� cells. J. Immunol. 167:3329.
38. Djie, M. Z., B. F. Le Bonniec, P. C. R. Hopkins, K. Hipler, and S. R. Stone. 1996.Role of the P2 residue in determining the specificity of serpins. Biochemistry35:11461.
39. Huntington, J. A., R. J. Read, and R. W. Carrell. 2000. Structure of a serpin-protease complex shows inhibition by deformation. Nature 407:923.
40. Loebermann, H., R. Tokuoka, J. Deisenhofer, and R. Huber. 1984. Human �1-proteinase inhibitor: crystal structure analysis of two crystal modifications, mo-lecular model and preliminary analysis of the implications for function. J. Mol.Biol. 177:531.
41. Bock, S. C., K. Skriver, E. Nielsen, H. C. Thogersen, B. Wiman,V. H. Donaldson, R. L. Eddy, J. Marrinan, E. Radziejewska, R. Huber, et al.1986. Human C1 inhibitor: primary structure, cDNA cloning, and chromosomallocalization. Biochemistry 25:4292.
42. Czermak, B. J., V. Sarma, C. L. Pierson, R. L. Warner, M. Huber-Lang,N. M. Bless, H. Schmal, H. P. Priedl, and P. A. Ward. 1999. Protective effects ofC5a blockage in sepsis. Nat. Med. 5:788.
![C1 [Modo de compatibilidad]](https://static.documents.pub/doc/80x56/631cc065a1cc32504f0ca2a0/c1-modo-de-compatibilidad.jpg)
