Nonseptic Shock Properdin Deficiency in Murine Models Of

Home , Properdin

Properdin Deficiency in Murine Models of Nonseptic Shock Nina D. Ivanovska, Petya A. Dimitrova, Jeni C. Luckett, Rana El-Rachkidy Lonnen, Wilhelm J. Schwaeble and This information is current as Cordula M. Stover of October 3, 2021. J Immunol 2008; 180:6962-6969; ; doi: 10.4049/jimmunol.180.10.6962 http://www.jimmunol.org/content/180/10/6962 Downloaded from

References This article cites 28 articles, 6 of which you can access for free at: http://www.jimmunol.org/content/180/10/6962.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 3, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2008 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Properdin Deﬁciency in Murine Models of Nonseptic Shock1

Nina D. Ivanovska,2* Petya A. Dimitrova,2* Jeni C. Luckett,† Rana El-Rachkidy Lonnen,† Wilhelm J. Schwaeble,† and Cordula M. Stover3†

Hereditary properdin deficiency is linked to susceptibility to meningococcal disease (Neisseria meningitidis serotypes Y and W-135) with high mortality. Its relative contribution toward the outcome of nonseptic shock has not been investigated. Using properdin- deficient C57BL/6 mice and their littermates, this study examines their survival of zymosan-induced and LPS-induced shock. Properdin-deficient mice were more resistant to zymosan shock compared with wild-type mice, which showed greater impairment of end-organ function 24 h after zymosan injection, higher TNF-␣ production by alveolar and peritoneal macrophages, higher TNF-␣, and, inversely, lower IL-10 levels in peritoneal lavage and circulation and higher plasma C5a levels. Properdin-deficient mice showed significantly higher mortality in LPS shock, elevated TNF-␣, and, inversely, reduced IL-10 production by peritoneal macrophages as well as lower plasma C5a levels compared with wild-type littermates. NO production by peritoneal macrophages ␣

and plasma 1-antitrypsin levels at 24 h after the injection of LPS or zymosan were decreased in properdin-deficient mice in both Downloaded from models, and fewer histopathologic changes in liver were observed in properdin-deficient animals. This study provides evidence that properdin deficiency attenuates zymosan-induced shock and exacerbates LPS-induced shock. The Journal of Immunology, 2008, 180: 6962–6969.

omplement acts in the first line of the immune defense deficient mouse line in a model of cecal ligation and puncture that before the generation of a specific, adaptive immune re- leads to subacute polymicrobial sepsis and found that over an ob- http://www.jimmunol.org/ sponse because it recognizes and subsequently is acti- servation period of 14 days, properdin-deficient mice were signif- C 4 vated by pathogen-associated molecular patterns and immune icantly impaired in their survival compared with wild-type (WT) complexes involving preimmune Abs of the IgM type. These pat- littermates (9). Properdin participates in alternative pathway acti- terns can be heterogenous and activate one or the other or all three vation by stabilizing C3 and C5 convertases, thereby amplifying of the distinct complement activation pathways, namely the clas- ongoing complement activation initiated by any of the pathways. sical, lectin, and alternative pathways. LPS, a wall component en- Recently, the generation of another genetically engineered prop- dotoxin of Gram-negative bacteria, activates the classical and lec- erdin-deficient mouse line (on mixed background 129/C57BL/6) tin pathways of complement (1). Zymosan, a component of yeast was reported. In vitro analysis of this properdin-deficient serum

cells, initiates the alternative and classical pathways of comple- using LPS revealed a deficiency in alternative pathway activation, by guest on October 3, 2021 ment (2, 3). C3- and C4-deficient mice are significantly impaired whereas alternative pathway activation induced by zymosan was in their clearance of LPS and show greater mortality in LPS-in- only marginally impaired and classical pathway-triggered alterna- duced shock (4), mirroring the significance of intact complement tive pathway amplification remained intact (10). activation pathways for the survival of this particular type of non- The purpose of the present study was to examine, in vivo, the septic shock. By contrast, C5-deficient mice show decreased mor- course of zymosan- and LPS-induced models of shock in WT and tality in zymosan-induced shock (5), suggesting that impairment of properdin-deficient C57BL/6 mice. complement activation involving the generation of C5a, an ana- phylatoxin, and C5b, the initiating component of the membrane Materials and Methods attack complex, could be beneficial for the survival of this type of Animals nonseptic shock. Experiments were performed using properdin-deficient mice (9) that were In previous work, we have shown that an inhibitor of classical backcrossed for 9–12 generations onto C57BL/6 background and WT lit- and alternative pathways of complement (fangchinoline) improves termates (all males) weighing 30–35 g. The animals had free access to the outcome of zymosan-induced nonseptic multiorgan dysfunc- water and standard chow. The study protocols were approved by the Eth- tion syndrome (6–8). We have further characterized a properdin- ical Animal Commission of the Institute of Microbiology, Sofia, Bulgaria. These properdin-deficient mice were generated by gene-specific targeting and have recently been characterized as being completely impaired in prop- *Department of Immunology, Institute of Microbiology, Bulgarian Academy of Sci- erdin-dependent rabbit erythrocyte lysis compared with their WT litter- ences, Sofia, Bulgaria; and †Department of Infection, Immunity and Inflammation, mates; however, they do not differ in their serum levels of C3 and IgM (9). University of Leicester, Leicester, United Kingdom Zymosan- and LPS-induced nonseptic shock Received for publication June 19, 2007. Accepted for publication March 7, 2008. The costs of publication of this article were defrayed in part by the payment of page Zymosan (1 mg/kg body weight; Sigma-Aldrich) was suspended in sterile charges. This article must therefore be hereby marked advertisement in accordance water and autoclaved for 30 min. WT and properdin-deficient mice were with 18 U.S.C. Section 1734 solely to indicate this fact. injected i.p. with 1 or 0.8 (low dose) mg/g body weight of this suspension ␮ 1 This study was supported by Medical Research Council Grant G0400300 (to (0.5 ml) or with 400 g of LPS (Escherichia coli serotype 055:B5, Sigma- C.M.S.). Aldrich) per mouse, a dose that elicits severe shock in mice of this genetic 2 N.D.I. and P.A.D. contributed equally to this work. 3 Address correspondence and reprint requests to Dr. Cordula Stover, Lecturer in 4 Abbreviations used in this paper: WT, wild type; aM␾, alveolar macrophage; pM␾, Immunology, Department of Infection, Immunity and Inflammation, University of peritoneal macrophage. Leicester, University Road, Leicester LE1 9HN, U.K. E-mail address: cms13@ leicester.ac.uk Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00

www.jimmunol.org The Journal of Immunology 6963

FIGURE 1. Properdin-deficient mice are significantly impaired in their FIGURE 2. Properdin-deficient and WT mice differ in their liver his- survival of acute LPS-induced shock. Properdin-deficient mice (n ϭ 11) tology at day 7 after LPS-injection. Liver sections were stained with H&E and WT littermates (n ϭ 10) were injected with LPS (400 ␮g i.p.) and (original magnification ϫ400), and are representative for three mice of p Ͻ 0.001; two-way ,ءءء .mortality was recorded over seven days each genotype. ANOVA. background (11), and tested in a pilot experiment to achieve such pheno- NO production by macrophages type in C57BL/6 control mice. Mice injected i.p. with 0.5 ml of saline Downloaded from served as controls. Peritoneal macrophages were harvested as described above. The cells were washed once and plated at a density of 1 ϫ 106 cells/ml in 200 ␮l of RPMI Organ weights Mice were bled at various time points, culled, and liver, spleen, and kid- neys were removed and weighed. To correct for differences in body weight between animals, organ weights were calculated as a percentage of total body weight. http://www.jimmunol.org/ Histopathologic evaluation of liver Livers were fixed in formaldehyde and paraffin embedded. Sections (7 ␮m) were deparaffinized with xylene and stained with H&E. TUNEL analysis to detect apoptotic cells was performed using TdT-FragEL DNA fragmenta- tion detection kit following the manufacturer’s (Merck Chemicals) instruc- tions and evaluated blind by two assessors. Measurement of coagulation time by guest on October 3, 2021 Blood was collected in a microcapillary pipette (10 cm ϫ 1 mm) contain- ing a horse hair of 20 cm in length as described (12). Every 30 s, 5 mm of the hair was pulled out of the tube manually and formation of a clot was recorded visually. Quantification of liver and renal injury and the concentration of glucose in plasma At day 7 of zymosan inflammation heparinized blood samples were collected and plasma was separated. All samples were analyzed immediately by standard laboratory kits (Dialab), measuring the following biochemical markers of organ dysfunction: alanine aminotransferase (a specific marker for hepatic parenchymal injury), aspartate aminotransferase (a nonspecific marker for hepatic injury), bilirubin (a predictor of liver failure), creatinine (an indicator of reduced glomerular filtration ability), and glucose level. Isolation of macrophages and cytokine assays Peritoneal macrophages (pM␾) were harvested by rinsing the peritoneal cavity with 10 ml of RPMI 1640 medium (Cambrex). The cells were washed twice, resuspended at a density of 2 ϫ 106 cells/ml in RPMI 1640 medium supplemented with 5% (v/v) FCS (Sigma-Aldrich), penicillin (100 IU/ml), and streptomycin (100 ␮g/ml) (Sigma-Aldrich). A volume of 0.5 ml was dispensed in 24-well plates (Becton Dickinson), and after1hat 37°C nonadherent cells were removed by two washings and adherent cells were cultured at 37°C in the presence or absence of 100 ␮g/ml zymosan or 100 ng/ml LPS. Alveolar macrophages (aM␾) were obtained by repeated bronchopulmonary lavage with 0.6 ml of sterile PBS (pH 7.2). Macro- phages were enriched by adherence to plastic (37°C for 1 h) in the presence of 5% (v/v) FCS. The adherent cell concentration and viability were determined by trypan blue exclusion. TNF-␣ and IL-10 were determined after FIGURE 3. Properdin-deficient mice have reduced nitric oxide produc- 18 h of cultivation by quantitative ELISA (Cytolab). The detection limits tion in a model of LPS-induced shock. A, NO was determined in the peri- were Ͻ50 pg/ml for TNF-␣ and 20 pg/ml for IL-10. toneal lavage of control mice and 24 h after LPS-injection. B and C, Peri- ␣ toneal macrophages were harvested 24 h after the administration of LPS 1-Antitrypsin assay and from untreated mice. Cells (1 ϫ 106/ml) were stimulated in vitro with Plasma levels of ␣1-antitrypsin were determined by ELISA as recently LPS (100 ng/ml) and supernatants were collected after 18 h and assayed for .p Ͻ 0.001; t test ,ءءء ;p Ͻ 0.05 ,ء .described (13). NO 6964 PROPERDIN CONTRIBUTES TO OUTCOME OF ZYMOSAN AND LPS SHOCK

FIGURE 4. Properdin-deﬁcient and WT mice differ in their ex vivo macrophage activation after LPS injection. Alveolar and peritoneal macrophages were obtained from properdin-deﬁcient and WT mice (n ϭ 4 each) 4 h after LPS injection and were cultivated in the presence of 100 ng/ml LPS. Supernatants were collected after 18 h and TNF-␣ (A) and IL-10 (B) production was measured. Data are expressed as percentages from unstimulated cultures (set at 100%). .p Ͻ 0.001; t test ,ءءء

1640 medium for1hat37°C. Nonadherent cells were removed by two Peritoneal NO was significantly lower in properdin-deficient washings and adherent cells were cultured at 37°C for 18 h in the presence mice than in WT mice 24 h after LPS injection (Fig. 3A). At this or absence of LPS (100 ng/ml). Concentration of the stable NO metabolite time point, peritoneal macrophages were purified and restimulated nitrite (a measure of NO activity) was assayed in culture supernatants by a standard Griess reaction as recently described (13). with LPS. Peritoneal macrophages from both groups increased their NO production after LPS treatment, but those from proper- C5a ELISA

din-deficient mice had significantly reduced NO production com- Downloaded from Plates (BD Biosciences) were coated with rat anti-mouse C5a (BD Bio- pared with WT (Fig. 3B). Peritoneal macrophages isolated from sciences catalog no. 558027) overnight at room temperature and blocked untreated properdin-deficient and WT mice and subsequently stim- with 2% BSA for 1 h at room temperature. Plasma samples (diluted 1/5) ulated with LPS showed a significant increase in NO production and serial dilutions of purified C5a (BD Biosciences catalog no. 622597) for both groups but no difference between properdin-deficient mice were added in parallel and incubated for2hatroom temperature. After washing with PBS and 0.05% (v/v) Tween 20, biotinylated rat anti-mouse and WT (Fig. 3C), demonstrating that LPS-induced NO production C5a (BD Biosciences catalog no. 558028) was added and detected using per se was not defective in properdin-deficient pM␾. http://www.jimmunol.org/ avidin-peroxidase (1/1000; Peprotech). At the onset of inflammation (4 h), properdin-deficient mice had higher TNF-␣ in plasma and peritoneum than WT mice and their Immunofluorescence analyses coagulation time was significantly shortened compared with WT Cryostat sections (7 ␮m) were fixed in formal saline (10%), washed, and, (data not shown). Alveolar and pM␾ were obtained from LPS- after application of SeroBlock FcR (BUF041A; Serotec), incubated with injected mice at this time point and restimulated with LPS in vitro. rabbit anti-mouse C5aR (1/400; BD Biosciences catalog no. 552837) and goat-anti-rabbit IgG, F(abЈ)2-tetramethylrhodamine isothiocyanate (1/400; Alveolar macrophages of properdin-deficient mice were impaired Santa Cruz Biotechnology catalog no. sc-3841). Fluorescence signals were in their responses to up-regulate TNF-␣ production, while pM␾ of analyzed using Nikon TE300 widefield epifluorescence microscope. properdin-deficient mice significantly increased their production

on LPS-stimulation (Fig. 4A). With respect to IL-10, only aM␾ of by guest on October 3, 2021 Statistical analysis WT mice responded to restimulation, whereas pM␾ from proper- Differences between groups were compared by two-way ANOVA and un- din-deficient mice showed inhibition in response to LPS stimula- Ϯ Ͻ paired t test. Data are presented as mean SD. A value of p 0.05 was tion (Fig. 4B). considered significant. Properdin-deficient mice had significantly lower plasma C5a Results levels 24 h after LPS-injection compared with WT (Table I). At Survival of properdin-deficient mice in LPS-induced shock this time point, properdin-deficient mice had half the number of peritoneal macrophages than WT (2.1 ϫ 106/ml Ϯ 0.4 vs 4.0 ϫ Intraperitoneal injection of LPS lead to signs of systemic reaction 106/ml Ϯ 0.5), whereas numbers were comparable in the unstimu- (dyspnea, diarrhea, and conjunctivitis) and a moderate reduction of lated state (4.8 ϫ 105 cells/ml vs 4.5 ϫ 105 cells/ml). The con- ϳ body weight (by 15%) over the 7-day observation period for centration of C5a in the peritoneal lavage of LPS-stimulated, prop- both properdin-deficient and WT mice. Eight of 10 WT (80%) erdin-deficient mice at 24 h did not differ significantly from that of but only three of 11 (33%) properdin-deficient animals survived control properdin-deficient mice. However, the increase was sig- (Fig. 1). nificant for LPS-stimulated WT compared with control WT mice. Histopathologic examination of livers showed periportal and sinusoidal infiltration of inflammatory cells at 24 h and 7 days for WT and properdin-deficient mice. Both showed single cell necro- Survival of properdin-deficient mice in zymosan-induced shock sis at 24 h. Intraparenchymal, granuloma-like accumulations of Over the 7 day observation period, only two of 10 WT mice (20%) mononuclear and lymphocytic cells were seen for WT only. WT but 5 of eleven properdin-deficient mice (45%) survived zymosan mice showed more focal hepatocellular necrosis and ballooning inflammation (Fig. 5). Zymosan-treated WT and properdin-defi- degeneration at 7 days than properdin-deficient mice (Fig. 2). cient mice exhibited all signs of organ dysfunction 24 h after shock

Table I. C5a concentrations in plasma and peritoneal lavage in WT and properdin-deﬁcient mice 24 h after LPS injectiona

Plasma (ng/ml) Peritoneal Lavage (ng/ml)

WT Properdin Deﬁcient WT Properdin Deﬁcient

Control (n ϭ 4) 15.0 Ϯ 1.2 18.5 Ϯ 2.0 2.5 Ϯ 0.2 2.8 Ϯ 0.3 LPS injected (n ϭ 4) 102.0 Ϯ 4.0*** 69.0 Ϯ 3.9**,*** 3.8 Ϯ 0.4*** 3.1 Ϯ 0.2

.p Ͻ 0.001, control vs LPS-injected mice; t test ,ءءء ;p Ͻ 0.01, WT vs properdin-deficient mice ,ءء ;a Data are expressed as means Ϯ SD The Journal of Immunology 6965 *** , a 0.08*** 0.03*** Ϯ Ϯ Creatinine (mg/dl) 0.04 3.64 0.05 1.54 Ϯ Ϯ 0.48 FIGURE 5. More properdin-deficient mice than WT mice survive in a model of zymosan induced-inflammation. Properdin-deficient mice (n ϭ *** ϭ ,

11) and WT littermates (n 10) were injected with zymosan (1 mg/kg test. body weight i.p.) and mortality was recorded over 7 days. t 7.6*** 0.35 6.5*** Ϯ Ϯ induction. The extent to which the level of glucose in plasma was decreased and the level of creatinine in plasma was increased was Downloaded from higher in WT than in properdin-deficient mice (Table II). Along Bilirubin (mg/dl) 4.8 260.2 5.4 112.6 Ϯ with greater increases of alanine aminotransferase, aspartate ami- Ϯ notransferase, and bilirubin in WT compared with properdin-deficient mice, this is in keeping with the outcome of the two groups in this model. There was a progressive and significant reduction of body weight in properdin-deficient and WT mice compared with http://www.jimmunol.org/ 0.08*** 65.4 untreated controls (by 18–30%; n ϭ 4). A significant increase of 0.04 56.2 Ϯ relative spleen and liver weights was observed in both groups dur- Ϯ ing zymosan inflammation, with WT showing a greater increase at day 7 of zymosan inflammation than properdin-deficient mice AST (U/L) 0.03 0.64 compared with controls (by 40%). 0.02 0.54 Ϯ Histopathologic examination of WT and properdin-deficient Ϯ mice livers showed periportal infiltration of inflammatory cells at 24 h and 7 days and sinusoidal infiltration at 4 h and 7 days. 0.44 WT showed hepatocellular necrosis and ballooning degenera- by guest on October 3, 2021 *** tion at 24 h. There was more intraparenchymal lymphocytic in- ,

filtration in WT at day 7 compared with properdin-deficient mice 0.001, zymosan-injected WT mice vs zymosan-injected properdin-deficient mice; 3.5*** 5.4*** 0.32

(Fig. 6). Ͻ Ϯ Ϯ Peritoneal NO was signiﬁcantly lower in properdin-deﬁcient p

, ءءء -mice than in WT 24 h after zymosan stimulation (Fig. 7A). Peri toneal macrophages were obtained at this time point. Eighteen ALT (U/L) 3.0 118.2 hours later, cells from properdin-deﬁcient mice showed a signiﬁ- 4.1 63.3 Ϯ cantly impaired NO secretion compared with WT cells, and this Ϯ

ﬁnding was observed after zymosan restimulation (Fig. 7B). Zy- 37.7 mosan treatment of peritoneal macrophages isolated from un-

treated WT and properdin-deficient mice lead to significant reduc- *** , tion of NO (Fig. 7C). The initiation of zymosan inflammation was followed by a 5.4*** 4.2*** 44.7 Ϯ marked decrease in coagulation time at the 4-h time point in WT Ϯ mice and remained significantly curtailed compared with the coagulation time recorded for the properdin-deficient mice until day 0.001, untreated vs zymosan-treated groups;

7 (Fig. 8). Rotary thromboelastography was used to assess, for Glucose (mg/dl) Ͻ 6.0 95.0 5.1 57.8 p Ϯ properdin-deﬁcient and WT mice at t0, the rate of initial ﬁbrin Ϯ

, ءءء formation, the time taken to reach standard clot ﬁrmness, the rate of clot growth, and the clot strength, and did not reveal signiﬁcant Untreated Zymosan Treated Untreated Zymosan Treated Untreated Zymosan Treated Untreated Zymosan Treated Untreated Zymosan Treated SD;

differences (data not shown). Ϯ 4) 137.7 In WT mice there was significantly more TNF-␣ in plasma and ϭ peritoneum at 4 h and significantly less IL-10 at 4 h (plasma and peritoneum) and 24 h (peritoneum) compared with properdin-deficient mice (Fig. 9), consistent with the overall outcome of the two groups. 4) 147.2 ϭ Alveolar and pM␾ macrophages were harvested at different Plasma concentrations of glucose, alanine aminotransferase (ALT), aspartate aminotransferase (AST), bilirubin, and creatinine in WT and properdin-deficient mice 24 h after zymosan injection points of zymosan inflammation and cultivated for 18 h in the Data are expressed as means Properdin deficient ( n ␾ WT ( n absence or presence of zymosan. Although aM from properdin- a ␣ deficient and WT mice produced comparable TNF- levels up to Table II. 6966 PROPERDIN CONTRIBUTES TO OUTCOME OF ZYMOSAN AND LPS SHOCK

FIGURE 8. Properdin-deficient and WT mice differ in their coagulation time profile in zymosan-induced inflammation. Mice were injected with FIGURE 6. Properdin-deficient and WT mice differ in their liver his- zymosan and coagulation time was measured at0h(n ϭ 4),4h(n ϭ 6), .p Ͻ 0.001; t test ,ءءء ;p Ͻ 0.01 ,ءء .(tology of at day 7 after zymosan (Zy) injection. Liver sections were stained 24h(n ϭ 4), and 168 h (n ϭ 4 with H&E (original magnification ϫ400) and are representative of three mice for each genotype. Controls are the same as in Fig. 2.

niﬁcantly lower TNF-␣ production compared with WT over all

␣ Downloaded from 24 h, there was a significant reduction of TNF- by properdin- time points (Fig. 10B). Peritoneal macrophages isolated from prop- ␾ deficient aM isolated at 168 h (Fig. 10A). After in vitro stimu- erdin-deficient mice without (Fig. 10C) and with (Fig. 10D) sub- ␾ lation with zymosan, aM of properdin-deficient mice had a sig- sequent stimulation with zymosan showed a significant reduction in TNF-␣ secretion over all time points compared with WT pM␾. With regard to IL-10 production by aM␾ and pM␾ with or without restimulation with zymosan, there was no significant difference between WT and properdin-deficient mice over time (data not http://www.jimmunol.org/ shown). A lower dose of zymosan was used to investigate further the cellular responses in properdin-deficient and WT mice (n ϭ 4 each). At 0.8 mg/g body weight, no mortality was recorded for either group. Peritoneal macrophages from these lower dose zymosan-treated mice were purified on day 7 and analyzed for their NO production. At this time point, pM␾ from properdin-deficient mice produced significantly less NO than pM␾ from WT mice by guest on October 3, 2021 (0.065 Ϯ 0.005 ␮g/ml vs 0.173 Ϯ 0.006 ␮g/ml), consistent with the higher dose model. After restimulation with zymosan for 24 h, however, contrary to the higher dose zymosan treatment, NO production was not increased further (0.083 Ϯ 0.004 ␮g/ml for properdin-deficient mice and 0.139 Ϯ 0.005 ␮g/ml for WT), but the difference between WT and properdin-deficient mice remained significantly different. Severity of inflammation and extent of organ damage was greater in WT than in properdin-deficient mice also at this nonlethal dose. Seven days after zymosan-injection, WT and properdin-deficient mice again showed a significant increase in relative spleen and liver weights compared with untreated controls, with zymosan-treated WT mice significantly exceeding zymosan- treated, properdin-deficient mice (data not shown). Liver histology at day 7 of zymosan-induced inflammation showed the qualitative difference observed for the higher dose model in terms of more necrosis and more lymphocytic infiltration in WT mice. The serum creatinine level measured at 24 h was significantly more elevated in WT than in properdin-deficient mice (0.35 Ϯ 0.04 mg/dl creatinine (control WT) and 0.48 Ϯ 0.05 mg/dl (control properdin- deficient) vs 2.04 Ϯ 0.05 mg/dl (zymosan-treated WT, n ϭ 4) and 1.14 Ϯ 0.03 (zymosan-treated properdin-deficient, n ϭ 4), con- firming the observation made in the higher dose model. C5a was significantly increased in plasma and peritoneal lavage of WT compared with properdin-deficient mice 7 days after zymosan-in- FIGURE 7. Properdin-deficient mice have reduced nitrite oxide produc- jection (Table III). At this time point, WT mice showed signifi- tion in a model of zymosan-induced shock. A, NO was determined in the ␾ peritoneal lavage of control mice and 24 h after zymosan (Zym) injection. cantly fewer pM than properdin-deficient mice (control WT, ϫ 5 ϫ 5 B and C, Peritoneal macrophages were harvested 24 h after the adminis- 4.5 10 cells/mouse; control properdin-deficient, 4.8 10 6 tration of zymosan and from untreated mice. Cells (1 ϫ 106/ml) were cells/mouse; zymosan-treated WT, 2.4 ϫ 10 cells/mouse Ϯ 0.6; stimulated in vitro with zymosan (50 ␮g/ml) and supernatants were col- zymosan-treated properdin-deficient, 4.2 ϫ 106 cells/mouse Ϯ 0.5; .(p Ͻ 0.001; t test. 2–4 mice each ,ءءء ;p Ͻ 0.05 ,ء .lected after 18 h and assayed for NO The Journal of Immunology 6967

FIGURE 9. Properdin-deficient and WT mice differ in their profile for TNF-␣ and IL-10 in the acute phase of zymosan-induced inflammation. Mice were injected with zymosan and their plasma and peritoneal lavages assessed for TNF-␣ (A and B) and IL-10 (C and D)at4h p Ͻ ,ء .(n ϭ 10), 24 h (n ϭ 5), and 168 h (n ϭ 2) .p Ͻ 0.001; t test ,ءءء ;p Ͻ 0.01 ,ءء ;0.05 Downloaded from

Discussion loop toward classical and lectin pathway activation triggered http://www.jimmunol.org/ Endotoxin- and zymosan-induced shock are murine models of a through the binding of C1 and mannan-binding lectin to LPS. This nonseptic, systemic inflammatory response syndrome with high extends the findings of a previous study, which concluded that the mortality in man. The present study shows the significance of pro- outcome of LPS-induced shock is dependent on IgM, C3, and C4 perdin in the survival of LPS-induced shock and reveals that pro- (14). LPS, a TLR4 agonist, activates macrophages, resulting in the perdin has a contributory role in organ damage during zymosan- release of proinflammatory cytokines such as TNF-␣, IL-1, IL-6, induced inflammation. and IFN-␥. Tissue injury in endotoxemia depends not only on Properdin is able to bind directly to LPS (E. coli 026:B6, Ref.10; proinflammatory cytokine production but also on the release of and E. coli 055:B5, the serotype used for the in vivo model in this counter-regulatory cytokines such as IL-10. LPS induces IL-10 by guest on October 3, 2021 study; data not shown). The lack of properdin resulted in signifi- production by macrophages and, in septic patients, its action cor- cantly elevated mortality in LPS-induced shock and is likely to relates with increased survival (15). Unexpectedly, properdin-de- reflect the contribution of the alternative pathway amplification ficient and WT mice did not differ significantly in their plasma

FIGURE 10. Properdin-deﬁcient and WT mice differ in their ex vivo macrophage activation after zymosan (Zy) injection. Alve- olar and peritoneal macrophages were obtained at different time points (0, 24, and 168 h, n ϭ 4;4h,n ϭ 6) and were cultivated in the absence of zymosan (A and C) or with 50 ␮g/ml zymosan (B and D). Supernatants were collected after 18 h and TNF-␣ produc- p Ͻ ,ءءء ;p Ͻ 0.01 ,ءء .tion was measured 0.001; two-way ANOVA. 6968 PROPERDIN CONTRIBUTES TO OUTCOME OF ZYMOSAN AND LPS SHOCK

Table III. C5a concentrations in plasma and peritoneal lavage in WT and properdin-deﬁcient mice at day 7 after zymosan injectiona

Plasma (ng/ml) Peritoneal Lavage (ng/ml)

WT Properdin Deﬁcient WT Properdin Deﬁcient

Control (n ϭ 4) 15.0 Ϯ 1.2 18.5 Ϯ 2.0 2.5 Ϯ 0.2 2.8 Ϯ 0.3 Zymosan injected (n ϭ 4) 51.0 Ϯ 3.0*** 29.6 Ϯ 2.5**,*** 7.8 Ϯ 0.4*** 3.5 Ϯ 0.2**,***

.p Ͻ 0.001, control vs zymosan-injected mice; t test ,ءءء ;p Ͻ 0.01, WT vs properdin-deﬁcient mice ,ءء ;a Data are expressed as means Ϯ SD

TNF-␣ profile. This may be explained by an acute bone marrow- (25), which was found more elevated in WT mice at day 7. There- suppressive effect exerted by doses of LPS lower than the one used fore, increased coagulation may be a reflection of changes in other in this model (16). Properdin-deficient and WT mice compare in parameters and not necessarily an expression of the beginning of their differential blood leukocyte counts at t0 (data not shown), but disseminated intravascular coagulopathy. Additional experiments a differential susceptibility to LPS-induced bone marrow suppres- are needed to investigate the contribution of properdin to zymosan- sion cannot be excluded. In support of this possibility, we find that initiated, alternative pathway convertase-mediated prothrombin the pM␾ count at 24 h after LPS injection was significantly less in cleavage (26). properdin-deficient mice compared with WT and that properdin- Systemic inflammation and organ injury were present even after deficient mice did not increase their peritoneal C5a as much as WT a nonlethal dose of zymosan, but properdin deficiency attenuated Downloaded from mice at this time point compared with unstimulated control mice. this effect. WT mice had significantly elevated C5a levels in WT macrophages produced more NO at 24 h and had greater liver plasma as well as peritoneal lavage. At day 7, WT showed greater pathology at the 24-h and 7-day time points compared with prop- hepatic C5aR expression than properdin-deficient mice (by immu- erdin-deficient mice, consistent with the understanding that over- nofluorescence; data not shown). A detrimental role for C5a in production of NO contributes to superoxide-mediated liver injury zymosan-induced inflammation has been demonstrated (27, 28). in the LPS model (17). Properdin-deficient mice were severely Our study of properdin-deficient mice compared with littermate http://www.jimmunol.org/ compromised in their survival of LPS-induced shock and did not WT mice suggests that pathology in zymosan inflammation is me- mount a plasma C5a response as the WT mice did. Further studies diated by alternative pathway-dependent C3 as well as by C5 split are needed to investigate the possibility that high levels of C5a in product generation. Given that the number of pM␾ is lower in WT, WT mice are able to suppress TLR4-mediated responses (18). further studies are needed to clarify whether their recruitment is Zymosan-induced inflammation differs from acute endotoxemic C5aR independent (22). shock in that it causes multiple organ injury. Properdin appeared In conclusion, this study demonstrates for the first time that detrimental in the course of zymosan inflammation based on in- properdin deficiency attenuates the symptoms of zymosan-induced creased NO in peritoneal exudate, NO production by pM␾, relative inflammation and exacerbates LPS-induced shock. by guest on October 3, 2021 spleen and liver weights, and plasma parameters (hypoglycemia, plasma alanine aminotransferase, aspartate aminotransferase, bili- Acknowledgments rubin, and creatinine) of WT compared with properdin-deficient We thank Dr. Mariya Hristova (University College London) for helpful mice. Properdin apparently is involved in the pathogenesis of acute discussions, the Transgenic Unit team and Biomedical Services of the Uni- liver dysfunction and renal failure. In terms of mortality, however, versity of Leicester for excellent maintenance of the mouse colony, Aline additional determining factors seem to be necessary because sur- Dupont and Dr. Chris Jones (University of Leicester) for thromboelastog- vival for properdin-deficient mice was only marginally improved raphy analyses, and Irina Elliott for TUNEL analysis. compared with that of WT mice. The relative abundance of TUNEL-positive cells was assessed and found not to differ be- Disclosures tween WT and properdin-deficient mice (data not shown). TNF-␣ The authors have no financial conflict of interest. and IL-10 levels are determinants in the outcome of zymosan inflammation (19, 20). In vitro, mainly zymosan-bound, activated References ␣ 1. Dumestre-Pe´rard, C., E. Doerr, M. G. Colomb, and M. Loos. 2007. Involvement complement, leads to macrophage activation and TNF- release, of complement pathways in patients with bacterial septicemia. Mol. Immunol. 44: while uncoated zymosan particles have little effect (21). In vivo, 1631–1638. TLR2- and complement receptor-dependence of effector mecha- 2. Fearon, D. T., and K. F. Austen. 1977. Activation of the alternative complement pathway due to resistance of zymosan-bound amplification convertase to endog- nisms are thought to be time dependent (22). The more severe enous regulatory mechanisms. Proc. Natl. Acad. Sci. USA 74: 1683–1687. course of zymosan inflammation observed for WT mice, compared 3. Schenkein, H. A., and S. Ruddy. 1981. The role of immunoglobulins in alterna- with properdin-deficient mice, correlated with the elevated TNF-␣ tive pathway activation by zymosan: the effect of IgG on the kinetics of the alternative pathway. J. Immunol. 126: 11–15. level in peritoneal lavages and circulation and simultaneously with 4. Fischer, M. B., A. P. Prodeus, A. Nicholson-Weller, M. Ma, J. Murrow, a lower IL-10 level in WT at the early phase of developing mul- R. R. Reid, H. B. Warren, A. L. Lage, F. D. Moore, F. S. Rosen, and M. C. Carroll. 1997. Increased susceptibility to endotoxin shock in complement tiorgan dysfunction. In general, in zymosan-induced inflammation C3- and C4-deficient mice is corrected by C1 inhibitor replacement. J. Immunol. properdin-deficient aM␾ and pM␾ released less TNF-␣ spontane- 159: 976–982. 5. Miller, C. G., D. N. Cook, and G. J. Kotwal. 1996. Two chemotactic factors, C5a ously and after zymosan restimulation but were able to up-regulate ␣ ␣ and MIP-1 , dramatically alter the mortality from zymosan-induced multiple their TNF- production on subsequent LPS stimulation (data not organ dysfunction syndrome (MODS): C5a contributes to MODS while MIP-1␣ shown). A significant decrease in coagulation time was observed at has a protective role. Mol. Immunol. 33: 1135–1137. 4, 24, and 168 h in WT mice. TNF-␣, elevated in WT mice, has 6. Hristova, M., M. Yordanov, and N. Ivanovska. 2003. Effect of fangchinoline in murine models of multiple organ dysfunction syndrome and septic shock. In- been described as a procoagulant (23). Furthermore, ␣1-antitryp- flamm. Res. 52: 1–7. sin, which was significantly more elevated in WT at 24 h and day 7. Ivanovska, N., M. Hristova, and S. Philipov. 1999. Complement modulatory activity of bisbenzylisoquinoline alkaloids isolated from Isopyrum thalictroides. II. 7 compared with properdin-deficient mice (data not shown), has Influence on C3–9 reactions in vitro and antiinflammatory effect in vivo. Int. procoagulant activity (24). The same argument holds true for C5a J. Immunopharmacol. 21: 337–347. The Journal of Immunology 6969

8. Ivanovska, N., P. Nikolova, M. Hristova, S. Philipov, and R. Istatkova. 1999. 19. Jansen, M. J., T. Hendriks, R. Hermsen, J. W. van der Meer, and R. J. Goris. Complement modulatory activity of bisbenzylisoquinoline alkaloids isolated 1998. A monoclonal antibody against tumour necrosis factor-␣ improves survival from Isopyrum thalictroides. I. Influence on classical pathway in human serum. in experimental multiple organ dysfunction syndrome. Cytokine 10: 904–910. Int. J. Immunopharmacol. 21: 325–336. 20. Jansen, M. J., T. Hendriks, B. M. de Man, J. W. van der Meer, and R. J. Goris. 9. Stover, C., J. Luckett, B. Echtenacher, A. Dupont, S. Figgitt, J. Brown, 1999. Interleukin 10 mitigates the development of the zymosan-induced multiple D. Ma¨nnel, and W. Schwaeble. 2008. Properdin plays a protective role in polymi- organ dysfunction syndrome in mice. Cytokine 11: 713–721. crobial septic peritonitis. J. Immunol. 180: 3313–3318. 21. von Asmuth, E. J., J. G. Maessen, C. J. van der Linden, and W. A. Buurman. 10. Kimura, Y., T. Miwa, L. Zhou, and W. C. Song. 2008. Activator-specific re- 1990. Tumour necrosis factor ␣ (TNF-␣) and interleukin 6 in a zymosan-induced quirement of properdin in the initiation and amplification of the alternate pathway shock model. Scand. J. Immunol. 32: 313–319. of complement. Blood 111: 732–740. 22. Mullaly, S., and P. Kubes. 2007. Mast cell-expressed complement receptor, not 11. Tzung, S. P., T. C. Mahl, P. Lance, V. Andersen, and S. A. Cohen. 1992. Inter- TLR2, is the main detector of zymosan in peritonitis. Eur. J. Immunol. 37: ␣ feron- prevents endotoxin-induced mortality in mice. Eur. J. Immunol. 22: 224–234. 3097–3101. 23. Echtenacher, B., K. Weigl, N. Lehn, and D. N. Mannel. 2001. Tumor necrosis 12. Addis, T. 1908. The coagulation time of the blood in man. Q. J. Exp. Physiol. 1: factor-dependent adhesions as a major protective mechanism early in septic peri- 305–334. tonitis in mice. Infect. Immun. 69: 3550–3555. 13. Ivanovska, N., R. Kalfin, M. Lazarova, and P. Dimitrova. 2007. Exogenous VIP 24. Dhainaut, J. F., N. Marin, A. Mignin, and C. Vinsonneau. 2001. Hepatic response limits zymosan-induced generalized inflammation (ZIGI) in mice. Immunol. Lett. to sepsis: interaction between coagulation and inflammatory processes. Crit. Care 110: 126–132. Med. 29: S420–S7. 14. Reid, R. R., A. P. Prodeus, W. Khan, T. Hsu, F. S. Rosen, and M. C. Carroll. 1997. Endotoxin shock in antibody-deficient mice: unraveling the role of natural 25. Markiewski, M., B. Nilsson, K. N. Ekdahl, T. E. Mollnes, and J. D. Lambris. antibody and complement in the clearance of lipopolysaccharide. J. Immunol. 2007. Complement and coagulation: strangers or partners in crime? Trends Im- 159: 970–975. munol. 28: 184–192. 15. Marchant, A., J. Deviere, B. Byl, D. de Groote, J. L. Vincent, and M. Goldman. 26. Ozge-Anwar, A. H., J. J. Freedman, A. F. Senyi, A. L. Cerskus, and 1994. Interleukin-10 production during septicaemia. Lancet 343: 707–708. M. A. Blajchman. 1984. Enhanced prothrombin-converting activity and factor Xa 16. Kotani, J., N. J. Avallone, E. Lin, M. Goshima, S. F. Lowry, and S. E. Calvano. binding of platelets activated by the alternative complement pathway. 2006. Tumor necrosis factor receptor regulation of bone marrow cell apoptosis Br. J. Haematol. 57: 221–228. during endotoxin-induced systemic inflammation. Shock 25: 464–471. 27. Mahesh, J., J. Daly, W. G. Cheadle, and G. J. Kotwal. 1999. Elucidation of the Downloaded from 17. Han, X., M. P. Fink, T. Uchiyama, R. Yang, and R. L. Delude. 2004. Increased early events contributing to zymosan-induced multiple organ dysfunction syn- iNOS activity is essential for hepatic epithelial tight junction dysfunction in en- drome using MIP-1␣, C3 knockout, and C5-deficient mice. Shock 12: 340–349. dotoxemic mice. Am. J. Physiol. 286: G126–G36. 28. Nieuwenhuijzen, G. A., M. P. Meyer, T. Hendriks, and R. J. Goris. 1995. Defi- 18. Hawlisch, H., Y. Belkaid, R. Baelder, D. Hildeman, C. Gerard, and J. Ko¨hl. 2005. ciency of complement factor C5 reduces early mortality but does not prevent C5a negatively regulates Toll-like receptor 4-induced immune responses. Immu- organ damage in an animal model of multiple organ dysfunction syndrome. Crit. nity 22: 415–426. Care Med. 23: 1686–1693. http://www.jimmunol.org/ by guest on October 3, 2021