of March 1, 2015.This information is current as

MannerDependent−Inflammation in a C5a/IL-6

Induced Intestinal−Citrobacter rodentiumProperdin Provides Protection from

StadnykWilhelm J. Schwaeble, Cordula M. Stover and Andrew W. Umang Jain, Qi Cao, Nikhil A. Thomas, Trent M. Woodruff,

ol.1401814http://www.jimmunol.org/content/early/2015/02/27/jimmun

published online 27 February 2015J Immunol 

Subscriptionshttp://jimmunol.org/subscriptions

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/ji/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/cgi/alerts/etocReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2015 by The American Association of9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Properdin Provides Protection from Citrobacter rodentium–Induced Intestinal Inflammation in a C5a/IL-6–DependentManner

Umang Jain,* Qi Cao,* Nikhil A. Thomas,*,† Trent M. Woodruff,‡ Wilhelm J. Schwaeble,x

Cordula M. Stover,x and Andrew W. Stadnyk*,{

Citrobacter rodentium is an attaching and effacing mouse pathogen that models enteropathogenic and enterohemorrhagic Esch-

erichia coli in humans. The complement system is an important innate defense mechanism; however, only scant information is

available about the role of complement proteins during enteric infections. In this study, we examined the impact of the lack of

properdin, a positive regulator of complement, in C. rodentium–induced colitis. Following infection, properdin knockout (PKO)

mice had increased diarrhea and exacerbated inflammation combined with defective epithelial cell–derived IL-6 and greater

numbers of colonizing bacteria. The defect in the mucosal response was reversed by administering exogenous properdin to PKO

mice. Then, using in vitro and in vivo approaches, we show that the mechanism behind the exacerbated inflammation of PKO mice

is due to a failure to increase local C5a levels. We show that C5a directly stimulates IL-6 production from colonic epithelial cells

and that inhibiting C5a in infected wild-type mice resulted in defective epithelial IL-6 production and exacerbated inflammation.

These outcomes position properdin early in the response to an infectious challenge in the colon, leading to complement activation

and C5a, which in turn provides protection through IL-6 expression by the epithelium. Our results unveil a previously unappre-

ciated mechanism of intestinal homeostasis involving complement, C5a, and IL-6 during bacteria-triggered epithelial injury. The

Journal of Immunology, 2015, 194: 000–000.

Gastrointestinal infections by attaching and effacing (A/E)pathogens such as enteropathogenic Escherichia coli andenterohemorrhagic E. coli are responsible for significant

mortality and morbidity in the developing world (1, 2). As thesemicrobes are specific for humans and do not cause disease inanimal models, host-protective mechanisms against entericpathogens are studied using a relevant A/E mouse pathogen,Citrobacter rodentium (3, 4). Similar to A/E infections in humans,C. rodentium is minimally invasive and colonizes the host byattaching to the colonic epithelium, resulting in diarrhea andcolonic inflammation characterized by leukocyte infiltrationand epithelial hyperplasia (5, 6). Additionally, owing to the sim-ilarities with idiopathic inflammatory bowel diseases (IBD) (7),C. rodentium–induced colitis has been used to study the patho-genic mechanisms in IBD.

Whereas host adaptive immunity has been extensively exploredin C. rodentium infection and is necessary for the host to over-come the infection, only recently have studies begun to dissect the

innate immune effectors such as TLRs and inflammasomes in the

pathogenesis of C. rodentium (8, 9). Complement is a major innatedefense mechanism and can strongly impact the local inflamma-

tory response and modify the developing adaptive immune

responses (10, 11), yet the role of complement proteins in the

response to infections of the gastrointestinal tract remains poorlyunderstood. In this regard, only one report has addressed com-

plement in C. rodentium infection where C3 deficiency was as-

sociated with increased mortality (12). Whether complement isactivated or involved in the pathology associated with the enteric

infection has not been reported.Complement can be activated via three different initiating routes,

the classical (CP), lectin (LP), or alternative pathways (AP).Properdin, a protein of the AP, is the only known positive physi-

ological regulator of complement activation. Properdin binds

C3b and through stabilizing the C3 and C5 convertases amplifies

complement activation that was initiated by other pathways (13–15). Recently, evidence has emerged showing that properdin can

directly activate complement by binding surfaces and provide

a platform for C3 convertase assembly (16). Considering thatproperdin modulates the overall amplification of complement

activation (17–19), we used properdin-deficient mice to begin to

identify the role complement may play in the pathology associated

with enteric infection. Our results demonstrate that properdinplays an important local protective role through the generation of

C5a, which in turn is critical to the stimulation of epithelial IL-6

and subsequent protection against C. rodentium infection. Thus,properdin and complement play a vital role in gut epithelial ho-

meostasis through coordinating the inflammatory response during

mucosal infection.

*Department of Microbiology and Immunology, Dalhousie University, Halifax, NovaScotia B3H 4H7, Canada; †Department of Medicine, Dalhousie University, Halifax,Nova Scotia B3H 4H7, Canada; ‡School of Biomedical Sciences, University ofQueensland, Brisbane, Queensland 4072, Australia; xDepartment of Infection, Immu-nity and Inflammation, University of Leicester, Leicester LE1 9HN, United King-dom; and {Department of Pediatrics, Dalhousie University, Halifax, Nova ScotiaB3H 4H7, Canada

Received for publication July 18, 2014. Accepted for publication January 31, 2015.

This work was supported by a partnership grant from the Canadian Institutes ofHealth Research, the Nova Scotia Health Research Foundation, and the Crohn’sand Colitis Foundation of Canada (to A.W.S.). U.J. is a recipient of an IWK graduatestudentship award.

Address correspondence and reprint requests to Dr. Andrew W. Stadnyk, MucosalImmunology Research, IWK Health Centre, 5880 University Avenue, Halifax, NSB3K 6R8, Canada. E-mail address astadnyk@dal.ca

Abbreviations used in this article: A/E, attaching and effacing; AP, alternative path-way; CP, classical pathway; IBD, inflammatory bowel disease; LP, lectin pathway;PKO, properdin knockout; WT, wild-type.

Copyright� 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1401814

Published February 27, 2015, doi:10.4049/jimmunol.1401814 at D

alhousie University on M

arch 1, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

Materials and MethodsMice

Five to 6-wk-old wild-type (WT; C57BL/6), properdin knockout (PKO, theproperdin gene is on the X chromosome) mice (20) were used. Of the totalnumber of animals used in the study, 55% were males and 45% werefemales. Mice had free access to food and water and were housed on woodchip bedding under specific pathogen-free conditions on a 12 h dark/lightcycle. The mice were Helicobacter negative, assayed by PCR on stoolDNA preparations. Experiments were undertaken in the IWK HealthCentre under the approval of the University Committee on LaboratoryAnimals, Dalhousie University, who in turn adjudicates the guidelines ofthe Canadian Council on Animal Care.

C. rodentium infection

Alert WT and PKO mice were infected by gavage with C. rodentium (0.1 mlin PBS/mouse, 109 CFU) and sacrificed at indicated time points. Forbacterial burden analysis, colon, spleen, and stool samples were homog-enized in PBS and the homogenate was serially diluted and spread onMacConkey agar plates (Sigma-Aldrich, St. Louis, MO), which were in-cubated at 37˚C for 24 h. Colonies with a red center and white outer rimwere counted as C. rodentium (21). In some experiments, fresh-pooledserum from healthy WT or PKO mice was injected i.v. into PKO mice ondays 1, 3, 5, and 7 of the infection. In other experiments, infected WT micewere orally gavaged with either water (control) or PMX205 (synthesizedand purified by reversed phase HPLC and used in murine colitis) (22) at200 mg/mouse, dissolved in distilled water, daily, starting from the day ofinfection until day 10.

On the day of sacrifice, mice were anesthetized and blood was collectedby cardiac puncture. Blood was allowed to clot at room temperature andserum was collected and stored for cytokine analysis. Mice were then killedby cervical dislocation, and colons were excised and their lengths measured.Colons were then flushed with cold PBS and divided into longitudinal parts.One part was used for histology whereas others were used for colon cultureexplants and homogenates.

Histopathology

Swiss rolls were prepared from a longitudinal colon strip, fixed in bufferedformalin overnight, rehydrated in 70% (v/v) ethanol, and embedded inparaffin, from which 4-mm sections were cut and stained using H&E. Anobserver blinded to the treatments determined a colonic inflammationscore derived from the extent of epithelial hyperplasia (0–3), cellular infil-tration (0–5), crypt damage (0–5), ulceration (0–3), and whether submucosaledema was present (1 or 0, respectively). Rectal bleeding scores were basedon stool consistency and whether blood was detectable with the scaleranging between 0 (normal stool with no blood) and 4 (diarrhea with blood).

Ex vivo culture

A longitudinal strip of colon was washed in cold PBS and incubated in 1 mlhigh-glucose DMEM (Invitrogen, Burlington, ON, Canada) supplementedwith 0.5% FBS, penicillin (100 U/ml), streptomycin (100 mg/ml), 10 mMHEPES, 2 mM L-glutamine, and 50 mM 2-ME in 12-well tissue cultureplates (Costar 3513, Corning, NY) at 37˚C for 24 h, after which thesupernatants were stored at 280˚C until examined by ELISA. Spleen andmesenteric lymph node cells were isolated from infected mice on the dayof sacrifice and single-cell suspensions were prepared in RPMI 1640(Invitrogen) supplemented with 5% heat-inactivated FBS, 1% penicillin/streptomycin, and 0.1% 2-ME. Cells were then resuspended in mediumat a concentration of 6 3 106 cells/ml and cultured with or withoutC. rodentium lysate (10 mg/ml) in a round-bottom 96-well plate. Super-natants were collected after 48 or 72 h for cytokine analysis.

Cell culture

T84 and Caco-2 human colon cancer cells were cultured and stimulated withrecombinant human C5a (10 nM; Sigma-Aldrich) as described previously(23). Supernatants were collected 24 h later, stored frozen, and subse-quently assayed for IL-6 by ELISA.

ELISA measurements

IL-17A, IL-22, TNF, and IL-10 concentrations were measured usingELISAs from eBioscience (San Diego, CA). IL-6 and IL-12p40 weremeasured using ELISAs from PeproTech (Rocky Hill, NJ). C3a and C5alevels were assessed using reagents from BD Pharmingen (Mississauga,ON, Canada). All protocols were performed according to manufacturers’instructions.

Immunohistochemistry

Four-micrometer-thick paraffin-embedded sections were deparaffinized andincubated with rat anti-mouse Ly6G (1:4000; BD Pharmingen) for neu-trophils or F4/80 (1:100; Serotec, Raleigh, NC) for macrophages, overnightat 4˚C (22). Slides were then incubated with biotinylated secondary Ab(Santa Cruz Biotechnology, Santa Cruz, CA) at room temperature for 1 hand developed with diaminobenzidine and counterstained with hematox-ylin. Cells were counted in 10 different high-power fields (3400) permouse and averaged among all mice per group. The results are shown asnumber of cells per five high-power fields. For apoptosis, slides wereprocessed with a TUNEL staining kit (Millipore, Temecula, CA) accordingto the manufacturer’s instructions, and apoptotic cells were counted in 50intact crypts per mouse and averaged.

Isolation of epithelial cells and quantitative real-time PCR

Murine colonic epithelial cells were isolated as described previously (24).Briefly, colons were washed with cold PBS, opened longitudinally, andincubated with 1 mM DTT/PBS at room temperature for 10 min. The colonwas then cut into 5- to 10-mm pieces and incubated for 15 min with HBSSbuffer (calcium- and magnesium-free; Life Technologies, Carlsbad, CA)containing 5% FBS, 2 mM EDTA, 1 mM DTT, and 10 mM HEPES (LifeTechnologies). The HBSS incubation step was performed twice. Sampleswere then passed through a 100-mm pore size filter and the epithelial cells(supernatant) were centrifuged at 300 3 g for 10 min at 4˚C then resus-pended in TRIzol (Invitrogen). RNA was isolated according to the man-ufacturer’s instructions. First-strand cDNAwas generated then analyzed byreal-time PCR using SYBR Green technology (Roche, Mississauga, ON,Canada). GAPDH was used as an internal control for the amount of cDNA,and relative expression over uninfected WT mice was calculated using theDDCt method. Oligodeoxynucleotide primers used are as follows: IL-6,sense, 59-TAGTCCTTCCTACCCCAATTTCC-39, antisense, 59-TTGGTCC-TTAGCCACTCCTTC-39; and GAPDH, sense, 59-GAAGGTCGGTGTG-AACGGATT-39, antisense, 59-TTGATGTTAGTGGGGTCTCGC-39.

Statistical analyses

Statistical analysis was performed using GraphPad prism version 5(GraphPad Software, La Jolla, CA). Parametric data are shown as means6SEM and were compared using a two-tailed t test. Bacterial numbers,rectal bleeding, and inflammation scores were compared using the Mann–Whitney U test. A significant difference was defined as p # 0.05.

ResultsProperdin-deficient mice experience exacerbated colitis toC. rodentium

To identify the contribution of complement to the colonic in-flammation induced by C. rodentium, we orally infected WT andPKO mice with 109 CFU/mouse and then euthanized the mice early(day 3) or late (day 10) during the course of the infection. Un-infected strains had normal fecal pellet consistency and did notshow any rectal bleeding. In response to C. rodentium, the rectalbleeding score was comparable to WT at day 3 but was signifi-cantly higher in day 10 infected PKO mice (Fig. 1A). Consistentwith the bleeding score data, colon shortening, an objectivemeasure of inflammation, was significantly more severe in day 10infected PKO mice compared with WT controls (Fig. 1B).To further characterize the colitis we examined H&E-stained

colon sections for signs of inflammation. Uninfected WT and PKO

mice did not show any overt signs of inflammation (Fig. 1C).Consistent with the macroscopic indications, C. rodentium in-duced mild pathology in WT mice, including patchy hyperplasiaand cellular infiltration (Fig. 1C). In contrast, day 10 infected PKO

mice suffered significantly exacerbated pathology compared toWT mice, characterized by severe hyperplasia, edema, and strik-ing cellular infiltration and ulceration (Fig. 1C). Ulceration isuncommon in C. rodentium infection but was frequent and ex-tensive in PKO mice (arrows in Fig. 1C), and the ulcer incidencewas almost 4-fold higher in day 10 infected PKO mice comparedwith WT controls (12 of 22 PKO versus 3 of 20 WT mice). Indeed,the colonic inflammation score was significantly higher in PKO

2 ROLE OF COMPLEMENT IN INFECTIOUS COLITIS

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

mice compared with WT controls (Fig. 1D). Ulceration is veryclosely associated with apoptosis of epithelial cells during IBD(25, 26), and we thought to examine the infected colons for ap-optosis. Infected PKO mice showed more TUNEL+ colonic epi-thelial cells compared with WT controls (Fig. 2A). Neutrophil/macrophage infiltration is another hallmark of C. rodentium–in-duced colon pathology. Immunohistochemistry of colon sectionsrevealed increased numbers of Ly6G+ neutrophils and F4/80macrophages in day 10 infected PKO mice compared with infectedWT controls (Fig. 2B, 2C). Overall, these outcomes indicate thatthe absence of properdin rendered hosts more susceptible toC. rodentium–mediated colonic injury, suggesting that properdinand complement play a protective role in this model.

Exacerbated colitis in properdin-deficient mice is associatedwith defective epithelial IL-6 production

Cytokines shape the inflammatory response against microbialpathogens. To understand the mechanism of exacerbated pathologyin PKO mice, we assessed the colonic levels of IL-6, IL-12, IL-17A, IL-22, IL-10, and TNF, cytokines that have been mecha-nistically implicated in the host response to C. rodentium infection(24, 27–31). Uninfected strains had comparable colon explant

cytokine levels (Fig. 3A–D). Following infection, IL-6 levels weremarkedly increased in WT mice; however, no increase was ob-served in PKO mice over uninfected controls. Moreover, IL-6remained 68% lower in infected PKO compared with infectedWT controls (p = 0.02, Fig. 3A). In contrast to the differenceobserved in IL-6, levels of IL-12, IL-17A, and IL-22 increasedwith infection but to comparable levels between WT and PKO mice(any apparent difference did not reach statistical significance,Fig. 3B–D). TNF and IL-10 were below the detection limit in allthe samples tested.Considering the profound deficit in the IL-6 response, we next

addressed whether IL-6 production throughout the host was im-paired in infected PKO mice. To this end, total spleen and lymphnode cells from each infected strain were stimulated ex vivo withC. rodentium lysate and supernatants were assayed for IL-6. IL-6was undetectable in any cell culture lacking bacterial lysate at 48h, and only unstimulated spleen cells made detectable levels by72 h. In cultures with lysate added, IL-6 production after 48 and72 h incubation was increased, although to similar extents be-tween WT and PKO groups (Fig. 3E, 3F). Collectively, these out-comes indicate that defective IL-6 production in PKO mice islimited to the colon.Mouse colonic epithelial cells have been identified as a signifi-

cant source of IL-6 during C. rodentium infection (24). Consid-

FIGURE 1. Lack of properdin exacerbates infectious colitis. WT and

PKO mice were infected with C. rodentium and then sacrificed on day 3 or

day 10. (A) Rectal bleeding scores during the course of infection and (B)

average colon lengths at necropsy. (C) Representative figures of the colon

prepared from uninfected and infected WT and PKO groups. Black arrows

identify the extent of ulceration. Compared to WT infected group, PKO

mice showed exacerbated inflammation characterized by extensive epi-

thelial hyperplasia, edema, increased cellular infiltration, and ulceration.

Original magnification 3100, stained with H&E. (D) A colon inflamma-

tion score was determined based on these criteria. For (A), dots represent

data from a single mouse and the line represents the median score. For (B)

and (D), the data are shown as means6 SEM; WT mice (uninfected, n = 4;

day 3, n = 4; day 10, n = 14) and PKO (uninfected, n = 5; day 3, n = 4; day

10, n = 18). *p # 0.05, ***p # 0.001 versus WT infected controls.

FIGURE 2. Lack of properdin leads to exacerbated hallmarks of colitis

during C. rodentium infection. Colonic sections from uninfected or day 10

infected mice were analyzed for apoptosis and cellular infiltration. (A)

Numbers of TUNEL+ colonic epithelial cells. Representative immuno-

histochemistry and quantification of stained cell numbers for (B) Ly6G

staining for neutrophils (original magnification 3400) and (C) F4/80

staining for macrophages in the mucosa and submucosa (original magni-

fication 4003) of infected mice on day 10 are shown. Data are shown as

means 6 SEM (n = 4–5 mice/group). *p # 0.05, **p # 0.01 versus

infected WT controls. HPF, high-power field (3400).

The Journal of Immunology 3

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

ering that the IL-6 deficit was detected in colon explant super-natants, we thought to examine epithelial cell preparations forIL-6. Indeed, colon epithelium isolated from infected PKO micepossessed significantly less IL-6 mRNA compared with infectedcontrols (Fig. 3G). Because IL-6–deficient mice display increasedC. rodentium colonization compared with controls (24), we hy-pothesized that impaired IL-6 production in PKO mice might resultin increased bacterial colonization. To test our hypothesis wefollowed infection in both strains until day 21. As predicted, asignificantly higher bacterial burden was observed in PKO micecompared with infected WT controls (Fig. 4). Furthermore, on day21, whereas bacteria were detected in 100% of the PKO mice, tissuehomogenates of 5 of 13 WT mice had no detectable bacteria. Insummary, exacerbated pathology and increased bacterial coloniza-tion in PKO mice was associated with defective IL-6 productionfrom colonic epithelial cells.Based on these findings, we next hypothesized that properdin is

required for epithelial IL-6 production; hence, restoring properdinin PKO mice should restore the IL-6 response and reduce the se-verity of colitis in infected mice. One strategy that has been usedto restore properdin is injections of WT serum into deficientanimals (32, 33). Thus, we injected C. rodentium–infected PKO

mice with either WT or properdin-deficient serum. WT serum–reconstituted PKO mice displayed significantly higher colonic IL-6levels (Fig. 5A) and reduced inflammation and bacterial burden

compared with PKO mice reconstituted with properdin-deficientserum (Fig. 5B–D). Importantly, we did not detect any IL-6 inthe serum used for reconstitution, eliminating the possibility thatwe were adding IL-6 in the reconstitution experiment. In sum-mary, after C. rodentium infection, properdin mediates the pro-duction of mucosal IL-6, which protects mice from C. rodentium–induced pathology.

Properdin-dependent C5a generation mediates epithelial IL-6production and protection during C. rodentium–induced colitis

Complement activation leads to the generation of the anaphyla-toxins C3a and C5a. To examine whether defective IL-6 productionand heightened susceptibility in PKO mice could be attributed toanaphylatoxins, we measured colonic explant levels by ELISA. InWT mice, mucosal C3a and C5a levels were significantly higherdue to the infection (Fig. 6). However, in infected PKO mice, onlyC3a was significantly higher whereas C5a levels were not elevatedcompared with uninfected mice (Fig. 6). Considering the deficit inepithelial IL-6 and impaired generation of C5a in PKO mice, wehypothesized that C5a might be responsible for IL-6 productionfrom epithelial cells. We have previously reported that humancolon epithelial cell lines express the C5a receptor (23); hence, totest our hypothesis, these cells (T84 and Caco-2) were stimulatedwith C5a and supernatants assessed for IL-6. As predicted, C5a led toa significant increase in IL-6 secretion from both cell types (Fig. 7A).

FIGURE 3. Exacerbated colitis in PKO mice is as-

sociated with impaired production of IL-6 from epi-

thelial cells. (A) IL-6, (B) IL-12p40, (C) IL-17A, and

(D) IL-22 concentrations in colon explant culture

supernatants from uninfected or day 10 infected WT

and PKO mice (n = 3–12 mice/group). (E) Total spleen

and (F) lymph node cells were isolated from day 10

infected mice cultured with (stim.) or without (unstim.)

bacterial lysate and the supernatants assessed for IL-6

after 48 and 72 h in culture (n = 4 mice/group). (G)

Epithelial cells were isolated from the colons of un-

infected and day 10 infected strains and IL-6 expres-

sion was analyzed by quantitative PCR. GAPDH was

used as internal control and values are represented as

fold induction over uninfected WT mice (n = 3 mice/

group). Data are shown as means 6 SEM. *p # 0.05.

4 ROLE OF COMPLEMENT IN INFECTIOUS COLITIS

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

Finally, to directly examine the impact of C5a on the epithelialIL-6 response during C. rodentium colitis, we treated infected WTmice with a selective C5a receptor (C5aR1) antagonist, PMX205(22). Oral PMX205 treatment resulted in a reduction .60% incolonic IL-6 concentrations compared with controls (Fig. 7B).Furthermore, PMX205-treated mice, similar to properdin-deficientmice, displayed exacerbated colonic pathology (Fig. 7C) andsignificantly higher bacterial colonization (Fig. 7D). Altogether,these data suggest that complement activation generating C5a in aproperdin-dependent manner is necessary for epithelial IL-6 pro-duction and subsequent epithelial protection against C. rodentiumcolitis.

DiscussionComplement is critical to our host defenses yet through unre-strained inflammation can be harmful (34). Despite that comple-ment activation in the mucosal environment has been reported incolitis including IBD, a limited number of studies have addressedthe role of complement in IBD (reviewed in Ref. 35). Whethercomplement is locally activated or involved in the pathologyduring enteric infections has not been reported. Because properdinis the only known positive regulator of complement and canamplify and/or initiate complement activation (17, 18), we usedproperdin-deficient mice to broadly study complement in thepathogenesis of C. rodentium. We discovered that mice lacking

properdin are more susceptible to infection-induced inflammationdue to a failure to upregulate C5a and subsequent epithelial IL-6production beyond uninfected levels. That C5a is protective in thiscontext was confirmed directly by the discovery that infected WTmice treated with a C5aR1 antagonist exhibited defective IL-6production and exacerbated inflammation. Taken together, thesedata support the notion that properdin-dependent generation ofC5a mediates epithelial IL-6 production, which impacts intestinaldefense.Previous studies have suggested an important in vivo role of

properdin in defense against pathogens, including Streptococcuspneumonia, Listeria monocytogenes, and Neisseria meningitidis(19, 36). Additionally, properdin-deficient mice injected with LPShad a higher mortality rate compared with injected WT controls(37). Moreover, properdin is also reported to be involved innoninfectious diseases, including arthritis, glomerulonephritis, andabdominal aneurysm (33, 38, 39). To the best of our knowledge,the present study is the first one implicating properdin in the in-flammatory response and host protection in the intestinal tractagainst an enteric pathogen. We attribute its protective role to theability to induce epithelial IL-6 production. Defective IL-6 pro-duction during infection is of mechanistic importance, as a lack ofIL-6 results in exacerbated disease in infectious colitis (24). IL-6can provide protection by inducing IL-17 and IL-22 responses (30,31, 40); however, that was not the case in the present study, as IL-17 and IL-22 levels were similar between the groups. IL-6 pro-

FIGURE 4. Exacerbated colitis and defective IL-6 production are as-

sociated with increased bacterial colonization in infected PKO mice. WT

and PKO mice were gavaged with C. rodentium and colonization was

assessed on days 10 and 21 in (A) stool, (B) colon, and (C) spleen by

culturing serial dilutions of tissue homogenates on MacConkey agar plates.

Of note, whereas all infected PKO mice harbored bacteria on day 21,

C. rodentium was not detected in the tissue homogenates of 5 of 13 WT

mice. Any WT mice with undetectable bacteria are not shown on the graph

and were not included in the statistical analysis. Each dot on the plot

represents a single animal and the line represents mean value. *p # 0.05,

**p # 0.01, ***p # 0.001 versus WT infected controls.

FIGURE 5. Exogenous properdin (WT serum) induces IL-6 and pre-

vents exacerbation in pathology in PKO mice. C. rodentium–infected PKO

mice were injected with either WT (WT→PKO) or properdin-deficient

serum (PKO→PKO) and analyzed for inflammation on day 10. (A) IL-6

production, (B) H&E stained histological images of serum-injected PKO

mice (original magnification 3100), and (C) colonic inflammation score

are shown (means 6 SEM, n = 5/group). (D) Stool colonization, with each

dot representing a single animal and the line showing the mean value. *p#

0.05, **p # 0.01 versus PKO→PKO group.

The Journal of Immunology 5

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

vides protection during C. rodentium–triggered intestinal inflam-mation by reducing apoptosis of epithelial cells and the conse-quential development of ulcers (24). Indeed, impaired IL-6production was associated with excessive apoptosis and increasedulceration in properdin-deficient mice. Ulceration in the colonprovides a favorable environment for the growth of C. rodentiumand possibly other luminal microbes, which in turn further pro-mote pathology (24). Consistent with this understanding, ulcera-tion in IBD is associated with increased colonization and invasionby luminal microbes (41, 42).It was previously reported that during C. rodentium infection,

colonic epithelial cells serve as a major source of IL-6, with hostfactors and not bacterial contact with the epithelium inducingthe IL-6 production (24). However, the mediator responsible forstimulating the epithelial cells remained unknown. We close thisgap in the understanding by identifying C5a as the amplifying hostfactor that directly stimulates IL-6 secretion from colonic epi-thelial cells. Furthermore, the use of PMX205, a selective C5aR1antagonist (22), confirms that the effects of C5a in this model aremediated through C5aR1 and not C5L2 (43).The finding that C5a is protective in this intestinal inflamma-

tory model is in contrast to its presumed proinflammatory nature,including in chemical-induced colitis. We showed earlier thatPMX205 protects from trinitrobenzene sulfonic acid– and dextransulfate sodium–induced colitis in rats and mice, respectively (22,44, 45), and others had the same outcome using a C5a blockingAb in mouse trinitrobenzene sulfonic acid colitis (46). This dif-ference from the chemical colitis models may be explained by thedifferent effects of IL-6 on infectious versus noninfectious mod-els; for example, IL-6 is proinflammatory in chemical (47) andT cell–dependent models of colitis (48) but is protective againstC. rodentium–induced colonic inflammation (24). Additionally,there may be redundant mediators to IL-6 generated in chemicallyinduced colitis. Nevertheless, considering our findings in this in-fection model, the use of C5a inhibitors as a therapeutic treatmentfor intestinal diseases (22, 35) needs to be closely assessed foreach particular circumstance, particularly with regard to microbialconfounders during the disease.A further novel finding from our study is that an enteric pathogen

induces local complement activation, evident by increased gen-

eration of colonic C3a and C5a in the WT strain of host. Using thePKO strain, we show that C5a but not C3a generation is properdin-dependent, therefore indicating that C3a generation is likely dueto CP or LP activation. This outcome is compatible with theemerging understanding that the CP or LP mediate complementactivation up to the C3 cleavage step but beyond which the APtakes over, and properdin stabilizes the C5 convertase (17, 18, 49).Another possibility is the presence of proteases that directly cleaveC3 in an AP- and LP-independent manner (50). Further studies areneeded to understand the mechanism of complement activation upto the C3 cleavage step as well as the contribution the split C3products may make during infection in the gastrointestinal tract.Finally, we interpret these outcomes to have relevance in un-

derstanding the pathogenesis of IBD, as an abnormal host responseto enteric organisms is an important factor contributing to thesediseases. Both Crohn’s disease and ulcerative colitis patients withextraintestinal complications have reduced serum properdin levelscompared with healthy control donors (51). The reduction wasassociated with a significant reduction in AP activation in plasmain response to cobra venom factor, suggesting properdin is con-sumed during complement activation in IBD. Considering ourdata, one interpretation of the findings in patients is that normallevels of properdin and its functional activity are limiting in the

FIGURE 6. Complement is activated during C. rodentium infection,

although C5a generation is defective in infected PKO mice. (A) C3a and (B)

C5a were measured by ELISA in colon explant culture supernatants of

uninfected or day 10 infected WT and PKO mice. Data are shown as

means6 SEM (uninfected, n = 4; day 10, n = 9) and PKO (uninfected, n = 4;

day 10, n = 9). *p # 0.05, **p # 0.01.

FIGURE 7. C5a is required for epithelial IL-6 production and protection

during C. rodentium infection. To assess whether C5a induces IL-6 pro-

duction, (A) Caco-2 and T84 cells were stimulated with 10 nM C5a and

24 h culture supernatants were analyzed for IL-6. Shown are the mean

concentrations 6 SEM from three independent experiments that were

performed in triplicates. *p # 0.05 compared with nontreated (NT) cells.

(B) C. rodentium–infected WT mice were gavaged daily with either water

(control) or 200 mg PMX205 and then sacrificed on day 10. C5a inhibition

reduced IL-6 production and (C) exacerbated colonic inflammation. Data

are shown as means 6 SEM (n = 4–5 mice/group). (D) Bacterial burden in

stool, with each dot representing a single animal and the line showing the

mean value. *p # 0.05, **p # 0.01 compared with controls.

6 ROLE OF COMPLEMENT IN INFECTIOUS COLITIS

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

face of inflammation in the gut. Our study indicates that in thiscircumstance exogenous properdin might be a therapeutic option.In conclusion, we discovered that properdin mediates the gut

mucosal inflammatory response through the generation of C5a,which in turn is critical for host protection against epithelial injury.

AcknowledgmentsWe thank Hana James and Maria Vaci for expert technical assistance, Helen

Dranse for assistance with quantitative PCR, and Dr. Jun Wang (Dalhousie

University) for helpful suggestions.

DisclosuresThe authors have no financial conflicts of interest.

References1. Kaper, J. B., J. P. Nataro, and H. L. Mobley. 2004. Pathogenic Escherichia coli.

Nat. Rev. Microbiol. 2: 123–140.2. Nataro, J. P., and J. B. Kaper. 1998. Diarrheagenic Escherichia coli. Clin.

Microbiol. Rev. 11: 142–201.3. Borenshtein, D., M. E. McBee, and D. B. Schauer. 2008. Utility of the Cit-

robacter rodentium infection model in laboratory mice. Curr. Opin. Gastro-enterol. 24: 32–37.

4. Mundy, R., T. T. MacDonald, G. Dougan, G. Frankel, and S. Wiles. 2005. Cit-robacter rodentium of mice and man. Cell. Microbiol. 7: 1697–1706.

5. Croxen, M. A., and B. B. Finlay. 2010. Molecular mechanisms of Escherichiacoli pathogenicity. Nat. Rev. Microbiol. 8: 26–38.

6. Eckmann, L. 2006. Animal models of inflammatory bowel disease: lessons fromenteric infections. Ann. N. Y. Acad. Sci. 1072: 28–38.

7. Barthold, S. W. 1979. Autoradiographic cytokinetics of colonic mucosal hy-perplasia in mice. Cancer Res. 39: 24–29.

8. Lebeis, S. L., B. Bommarius, C. A. Parkos, M. A. Sherman, and D. Kalman.2007. TLR signaling mediated by MyD88 is required for a protective innateimmune response by neutrophils to Citrobacter rodentium. J. Immunol. 179:566–577.

9. Liu, Z., M. H. Zaki, P. Vogel, P. Gurung, B. B. Finlay, W. Deng, M. Lamkanfi,and T. D. Kanneganti. 2012. Role of inflammasomes in host defense againstCitrobacter rodentium infection. J. Biol. Chem. 287: 16955–16964.

10. Klos, A., A. J. Tenner, K. O. Johswich, R. R. Ager, E. S. Reis, and J. Kohl. 2009.The role of the anaphylatoxins in health and disease. Mol. Immunol. 46: 2753–2766.

11. Dunkelberger, J. R., and W. C. Song. 2010. Complement and its role in innateand adaptive immune responses. Cell Res. 20: 34–50.

12. Belzer, C., Q. Liu, M. C. Carroll, and L. Bry. 2011. The role of specific IgG andcomplement in combating a primary mucosal infection of the gut epithelium.Eur. J. Microbiol. Immunol. (Bp) 1: 311–318.

13. Medicus, R. G., O. Gotze, and H. J. M€uller-Eberhard. 1976. Alternative pathwayof complement: recruitment of precursor properdin by the labile C3/C5 con-vertase and the potentiation of the pathway. J. Exp. Med. 144: 1076–1093.

14. Pillemer, L., L. Blum, I. H. Lepow, O. A. Ross, E. W. Todd, and A. C. Wardlaw.1954. The properdin system and immunity. I. Demonstration and isolation ofa new serum protein, properdin, and its role in immune phenomena. Science 120:279–285.

15. Smith, C. A., M. K. Pangburn, C. W. Vogel, and H. J. M€uller-Eberhard. 1984.Molecular architecture of human properdin, a positive regulator of the alternativepathway of complement. J. Biol. Chem. 259: 4582–4588.

16. Spitzer, D., L. M. Mitchell, J. P. Atkinson, and D. E. Hourcade. 2007. Properdincan initiate complement activation by binding specific target surfaces and pro-viding a platform for de novo convertase assembly. J. Immunol. 179: 2600–2608.

17. Harboe, M., and T. E. Mollnes. 2008. The alternative complement pathwayrevisited. J. Cell. Mol. Med. 12: 1074–1084.

18. Lutz, H. U., and E. Jelezarova. 2006. Complement amplification revisited. Mol.Immunol. 43: 2–12.

19. Ali, Y. M., A. Hayat, B. M. Saeed, K. S. Haleem, S. Alshamrani, H. I. Kenawy,V. P. Ferreira, G. Saggu, A. Buchberger, P. J. Lachmann, et al. 2014. Low-doserecombinant properdin provides substantial protection against Streptococcuspneumoniae and Neisseria meningitidis infection. Proc. Natl. Acad. Sci. USA111: 5301–5306.

20. Stover, C. M., J. C. Luckett, B. Echtenacher, A. Dupont, S. E. Figgitt, J. Brown,D. N. Mannel, and W. J. Schwaeble. 2008. Properdin plays a protective role inpolymicrobial septic peritonitis. J. Immunol. 180: 3313–3318.

21. Bergstrom, K. S., V. Kissoon-Singh, D. L. Gibson, C. Ma, M. Montero,H. P. Sham, N. Ryz, T. Huang, A. Velcich, B. B. Finlay, et al. 2010. Muc2protects against lethal infectious colitis by disassociating pathogenic and com-mensal bacteria from the colonic mucosa. PLoS Pathog. 6: e1000902.

22. Jain, U., T. M. Woodruff, and A. W. Stadnyk. 2013. The C5a receptor antagonistPMX205 ameliorates experimentally induced colitis associated with increasedIL-4 and IL-10. Br. J. Pharmacol. 168: 488–501.

23. Cao, Q., S. M. McIsaac, and A. W. Stadnyk. 2012. Human colonic epithelialcells detect and respond to C5a via apically expressed C5aR through the ERKpathway. Am. J. Physiol. Cell Physiol. 302: C1731–C1740.

24. Dann, S. M., M. E. Spehlmann, D. C. Hammond, M. Iimura, K. Hase, L. J. Choi,E. Hanson, and L. Eckmann. 2008. IL-6-dependent mucosal protection preventsestablishment of a microbial niche for attaching/effacing lesion-forming entericbacterial pathogens. J. Immunol. 180: 6816–6826.

25. Di Sabatino, A., R. Ciccocioppo, O. Luinetti, L. Ricevuti, R. Morera,M. G. Cifone, E. Solcia, and G. R. Corazza. 2003. Increased enterocyte apoptosisin inflamed areas of Crohn’s disease. Dis. Colon Rectum 46: 1498–1507.

26. G€unther, C., H. Neumann, M. F. Neurath, and C. Becker. 2013. Apoptosis, ne-crosis and necroptosis: cell death regulation in the intestinal epithelium. Gut 62:1062–1071.

27. Simmons, C. P., N. S. Goncalves, M. Ghaem-Maghami, M. Bajaj-Elliott,S. Clare, B. Neves, G. Frankel, G. Dougan, and T. T. MacDonald. 2002. Impairedresistance and enhanced pathology during infection with a noninvasive,attaching-effacing enteric bacterial pathogen, Citrobacter rodentium, in micelacking IL-12 or IFN-g. J. Immunol. 168: 1804–1812.

28. Goncalves, N. S., M. Ghaem-Maghami, G. Monteleone, G. Frankel, G. Dougan,D. J. Lewis, C. P. Simmons, and T. T. MacDonald. 2001. Critical role for tumornecrosis factor alpha in controlling the number of lumenal pathogenic bacteriaand immunopathology in infectious colitis. Infect. Immun. 69: 6651–6659.

29. Dann, S. M., C. Le, B. K. Choudhury, H. Liu, O. Saldarriaga, E. M. Hanson,Y. Cong, and L. Eckmann. 2014. Attenuation of intestinal inflammation ininterleukin-10-deficient mice infected with Citrobacter rodentium. Infect.Immun. 82: 1949–1958.

30. Geddes, K., S. J. Rubino, J. G. Magalhaes, C. Streutker, L. Le Bourhis,J. H. Cho, S. J. Robertson, C. J. Kim, R. Kaul, D. J. Philpott, and S. E. Girardin.2011. Identification of an innate T helper type 17 response to intestinal bacterialpathogens. Nat. Med. 17: 837–844.

31. Rubino, S. J., K. Geddes, and S. E. Girardin. 2012. Innate IL-17 and IL-22responses to enteric bacterial pathogens. Trends Immunol. 33: 112–118.

32. Pagano, M. B., H. F. Zhou, T. L. Ennis, X. Wu, J. D. Lambris, J. P. Atkinson,R. W. Thompson, D. E. Hourcade, and C. T. Pham. 2009. Complement-dependent neutrophil recruitment is critical for the development of elastase-induced abdominal aortic aneurysm. Circulation 119: 1805–1813.

33. Zhou, H. F., H. Yan, C. M. Stover, T. M. Fernandez, S. Rodriguez de Cordoba,W. C. Song, X. Wu, R. W. Thompson, W. J. Schwaeble, J. P. Atkinson, et al.2012. Antibody directs properdin-dependent activation of the complement al-ternative pathway in a mouse model of abdominal aortic aneurysm. Proc. Natl.Acad. Sci. USA 109: E415–E422.

34. Ricklin, D., G. Hajishengallis, K. Yang, and J. D. Lambris. 2010. Complement:a key system for immune surveillance and homeostasis. Nat. Immunol. 11: 785–797.

35. Jain, U., A. R. Otley, J. Van Limbergen, and A. W. Stadnyk. 2014. The com-plement system in inflammatory bowel disease. Inflamm. Bowel Dis. 20: 1628–1637.

36. Dupont, A., F. Mohamed, N. Salehen, S. Glenn, L. Francescut, R. Adib,S. Byrne, H. Brewin, I. Elliott, L. Richards, et al. 2014. Septicaemia modelsusing Streptococcus pneumoniae and Listeria monocytogenes: understanding therole of complement properdin. Med. Microbiol. Immunol. (Berl.) 203: 257–271.

37. Ivanovska, N. D., P. A. Dimitrova, J. C. Luckett, R. El-Rachkidy Lonnen,W. J. Schwaeble, and C. M. Stover. 2008. Properdin deficiency in murine modelsof nonseptic shock. J. Immunol. 180: 6962–6969.

38. Kimura, Y., L. Zhou, T. Miwa, and W. C. Song. 2010. Genetic and therapeutictargeting of properdin in mice prevents complement-mediated tissue injury. J.Clin. Invest. 120: 3545–3554.

39. Ruseva, M. M., K. A. Vernon, A. M. Lesher, W. J. Schwaeble, Y. M. Ali,M. Botto, T. Cook, W. Song, C. M. Stover, and M. C. Pickering. 2013. Loss ofproperdin exacerbates C3 glomerulopathy resulting from factor H deficiency. J.Am. Soc. Nephrol. 24: 43–52.

40. Li, L., Q. G. Shi, F. Lin, Y. G. Liang, L. J. Sun, J. S. Mu, Y. G. Wang, H. B. Su,B. Xu, C. C. Ji, et al. 2014. Cytokine IL-6 is required in Citrobacter rodentiuminfection-induced intestinal Th17 responses and promotes IL-22 expression ininflammatory bowel disease. Mol Med Rep 9: 831–836.

41. Cartun, R. W., H. J. Van Kruiningen, C. A. Pedersen, and M. M. Berman. 1993.An immunocytochemical search for infectious agents in Crohn’s disease. Mod.Pathol. 6: 212–219.

42. Liu, Y., H. J. van Kruiningen, A. B. West, R. W. Cartun, A. Cortot, andJ. F. Colombel. 1995. Immunocytochemical evidence of Listeria, Escherichia coli,and Streptococcus antigens in Crohn’s disease. Gastroenterology 108: 1396–1404.

43. Klos, A., E. Wende, K. J. Wareham, and P. N. Monk. 2013. International Unionof Basic and Clinical Pharmacology. LXXXVII. Complement peptide C5a, C4a,and C3a receptors. Pharmacol. Rev. 65: 500–543.

44. Woodruff, T. M., T. V. Arumugam, I. A. Shiels, R. C. Reid, D. P. Fairlie, andS. M. Taylor. 2003. A potent human C5a receptor antagonist protects againstdisease pathology in a rat model of inflammatory bowel disease. J. Immunol.171: 5514–5520.

45. Woodruff, T. M., S. Pollitt, L. M. Proctor, S. Z. Stocks, H. D. Manthey,H. M. Williams, I. B. Mahadevan, I. A. Shiels, and S. M. Taylor. 2005. Increasedpotency of a novel complement factor 5a receptor antagonist in a rat model ofinflammatory bowel disease. J. Pharmacol. Exp. Ther. 314: 811–817.

46. Chen, G., Y. Yang, X. Gao, Y. Dou, H. Wang, G. Han, R. Wang, J. Wang,L. Wang, X. Li, et al. 2011. Blockade of complement activation product C5aactivity using specific antibody attenuates intestinal damage in trinitrobenzenesulfonic acid induced model of colitis. Lab. Invest. 91: 472–483.

47. Gay, J., E. Kokkotou, M. O’Brien, C. Pothoulakis, and K. P. Karalis. 2006.Interleukin-6 genetic ablation protects from trinitrobenzene sulfonic acid-induced colitis in mice. Putative effect of antiinflammatory cytokines. Neuro-immunomodulation 13: 114–121.

The Journal of Immunology 7

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

48. Yamamoto, M., K. Yoshizaki, T. Kishimoto, and H. Ito. 2000. IL-6 is requiredfor the development of Th1 cell-mediated murine colitis. J. Immunol. 164: 4878–4882.

49. Harboe, M., G. Ulvund, L. Vien, M. Fung, and T. E. Mollnes. 2004. Thequantitative role of alternative pathway amplification in classical pathway in-duced terminal complement activation. Clin. Exp. Immunol. 138: 439–446.

50. Amara, U., D. Rittirsch, M. Flierl, U. Bruckner, A. Klos, F. Gebhard,J. D. Lambris, and M. Huber-Lang. 2008. Interaction between the coagulationand complement system. Adv. Exp. Med. Biol. 632: 71–79.

51. Lake, A. M., A. E. Stitzel, J. R. Urmson, W. A. Walker, and R. E. Spitzer. 1979.Complement alterations in inflammatory bowel disease. Gastroenterology 76:1374–1379.

8 ROLE OF COMPLEMENT IN INFECTIOUS COLITIS

at Dalhousie U

niversity on March 1, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from