Enhancement of Human Cancer Cell Motility and [email protected] doi: 10.1158/1078-0432.CCR-12-1204 2013 American Association for Cancer Research. Clinical Cancer Research

Human Cancer Biology

Enhancement of Human Cancer Cell Motility andInvasiveness by Anaphylatoxin C5a via Aberrantly ExpressedC5a Receptor (CD88)

Hidetoshi Nitta1,2, Yoshihiro Wada3, Yoshiaki Kawano3, Yoji Murakami3, Atsushi Irie4, Keisuke Taniguchi7,Ken Kikuchi6, Gen Yamada8, Kentaro Suzuki8, Jiro Honda3, Masayo Wilson-Morifuji5, Norie Araki5,Masatoshi Eto3, Hideo Baba2, and Takahisa Imamura1

AbstractPurpose: The anaphylatoxin C5a is a chemoattractant that induces leukocyte migration via C5a receptor

(C5aR). There is emerging evidence that C5a is generated in the cancer microenvironment. We therefore

sought C5aR expression and a direct influence of the C5a–C5aR axis on cancer cells.

ExperimentalDesign:C5aR expressionwas investigated in human cancer tissues and cell lines. Effects of

C5a stimulation on cancer cells were studied by cytoskeletal rearrangement, time-lapse analysis, Matrigel

chamber assay, and invasion in nude mouse in a comparison of C5aR-expressing cancer cells with control

cells.

Results:C5aRwas aberrantly expressed in various human cancers. Several cancer cell lines also expressed

C5aR. C5a triggered cytoskeletal rearrangement and enhanced cell motility three-fold and invasiveness 13-

foldofC5aR-expressing cancer cells. Such enhancement byC5awasnot observed in control cells. Cancer cell

invasion was still enhanced in the absence of C5a concentration gradient and even after the removal of C5a

stimulation, suggesting that random cell locomotion plays an important role in C5a-triggered cancer cell

invasion.C5a increased the release ofmatrixmetalloproteinases (MMP) fromcancer cells by two- to 11-fold,

and inhibition ofMMP activity abolished the C5a-enhancing effect on cancer cell invasion. Compared with

control cells, C5aR-expressing cells spread 1.8-fold more broadly at implanted nude mouse skin sites only

when stimulated with C5a.

Conclusions: These results illustrate a novel activity of the C5a–C5aR axis that promotes cancer cell

invasion through motility activation and MMP release. Targeting this signaling pathway may provide a

useful therapeutic option for cancer treatment. Clin Cancer Res; 19(8); 1–10. �2013 AACR.

IntroductionThe complement system is a biochemical cascade

involved in immune responses (1). Previous reportsshowed that the complement system is activated on cancercells in both an animal model (2) and in tissue specimens

(3). It was initially suggested that the complement systemmight be involved in cancer immune surveillance by itsdirect cytolytic effect (3) and the sensitization of cancer cellsto the immune effector cells via release of chemoattractants(4). However, cancer cells seem to evade the complementattack by expressing either soluble or membrane-associatedregulators of complement, for example, CD55, which pro-tects cancer cells from complement-dependent cytolysis(5, 6) and anticancer immune responses (7, 8). Thus, thecomplement system in cancer tissues does not seem to leadto cancer cell eradication.

Anaphylatoxin C5a is an N-terminal 74 amino acidfragment of the a-chain of the complement fifth compo-nent (C5), and is well known to act as a leukocyte chemoat-tractant and inflammatorymediator (9, 10). C5a is releasedbyC5-convertase formedduring the process of complementsystem activation (11), possibly triggered in response tocancer cells (3). Other C5a-producing pathways include C5cleavage by thrombin (12), the ultimate product of thecoagulation reaction. This cascade reaction can be triggeredby tissue factor, which is expressed in a wide range of celltypes including cancer cells (13). C5a is also generated by

Authors' Affiliations: Departments of 1Molecular Pathology, 2Gastroen-terological Surgery, 3Urology, 4Immunogenetics, and 5TumorGenetics andBiology, Faculty of Life Sciences, Kumamoto University; 6Medical QualityManagement Center, Kumamoto University Hospital, Kumamoto; 7YakultCentral Institute for Microbiological Research, Tokyo; and 8Department ofGenetics, Institute of Advanced Medicine, Wakayama Medical University,Wakayama, Japan

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Y. Wada and Y. Kawano contributed equally to this work.

Corresponding Author: T. Imamura, Department of Molecular Pathology,Faculty of Life Sciences, Kumamoto University, 1-1-1 Honjo, Kumamoto860-8556, Japan. Phone: 81-96-373-5306; Fax: 81-96-373-5308; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-12-1204

�2013 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org OF1

Research. on April 12, 2020. © 2013 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2013; DOI: 10.1158/1078-0432.CCR-12-1204

http://clincancerres.aacrjournals.org/

serine proteases from activated phagocytes (14), whichfrequently accumulate in cancer tissues. These findings leadus to the idea that C5a is also likely to be generated in thecancer microenvironment.

C5a activities are mediated by its binding to the mem-brane-associated C5a receptor (C5aR; CD88), which wasoriginally identified in leukocyte cell lines (15). C5aR hassince been reported to be expressed in other types of cellssuch as vascular endothelial cells, mesangial cells, and renalproximal tubular cells. Further studies have revealed thatC5aR expression is also inducible in epithelial cells byinflammatory and infection stimuli (16). About cancercells, functional C5aR expression was shown in a humanhepatoma cell line HepG2 (17), whereas normal hepato-cytes lack in its expression (16). These suggest that expres-sion of C5aR is induced in cancer cells as a consequence ofmalignant transformation.

C5a and chemokines are chemoattractants, and a body ofevidence indicates that a network of chemokines and theirreceptors influences the development of primary cancers(18–22). Recently, C5a was reported to recruit myeloid-derived suppressor cells for suppressing the antitumorCD8þ T-cell response (2, 23), suggesting its indirect rolein fostering cancer cells by protecting them from the anti-tumor CD8þ T cells. However, the direct biologic role ofC5a–C5aR system in cancer cells is largely unknown. In thisstudy, we investigated the expression of C5aR in cancer cellsof various origins and analyzed its impact on cancer cellmotility and invasiveness upon C5a stimulation.

Materials and MethodsCell lines

The human bile duct cancer cell lines MEC and HuCCT1and the human colon cancer cell lines HCT15, COLO205,

and DLD1 were provided by the Cell Resource Center forBiomedical Research Institute of Development, Aging,and Cancer, Tohoku University (Sendai, Japan). Humanbile duct cancer cell lines SSp-25 and RBE were obtainedfrom the Riken Cell Bank. Human colon cancer cell linesHCT116 and SW620 were gifts from Dr. B. Vogelstein,Johns Hopkins University (Baltimore, MD), and Dr.Kyogo Ito, Kurume University (Kurume, Japan), respec-tively. Cells were cultured in RPMI-1640 or Dulbecco’smodified Eagle’s medium supplemented with 10% FBS,penicillin (40 U/mL), and streptomycin (40 mg/mL) andwere maintained at 37�C in 5% CO2.

Tissue samples, immunohistochemistry, andretrospective analysis

Cancer tissue samples were obtained by surgical resectionor core needle biopsy in Kumamoto University Hospital(Kumamoto, Japan), and usage of those samples for thisstudy was approved by the internal ethics committee.Deparaffinized 2-mm thick sections were pretreated with0.3% H2O2 in methanol for 20 minutes, followed byProtein Block, Serum-Free (Dako Cytomation) treatmentfor 20 minutes. Sections were incubated with the primaryantibody against C5aR (2 mg/mL; Hycult Biotechnology) at4�C overnight, and subsequently stained using EnVisionþ

solution (Dako Cytomation) and 3,30-diaminobenzidinetetrahydrochloride solution containing 0.006% H2O2,according to the manufacturer’s instructions. Nuclei werecounterstained with hematoxylin. Retrospective analysiswas conducted on 42 patients with intrahepatic cholangio-carcinoma who had undergone liver resection from May2000 to November 2009. The relationship between cancercell C5aR expression and vascular invasiveness was inves-tigated and analyzed by Fisher exact test.

Real time PCRRNA was isolated from cancer cells using the Qiagen

RNAeasy Kit (Qiagen). cDNA was synthesized fromextracted RNA using the RNA PCR Kit AMV (Takara),according to the manufacturer’s instructions. PCR wasconducted using TaKaRa Ex Taq HS and primers (sense50-CGGGAGGAGTACTTTCCACC-30 and antisense 50-CTA-CACTGCCTGGGTCTTCTG-30 for human C5aR and sense50-CATCCACGAAACTACCTTCAACT-30 and antisense 50-TCTCCTTAGAGAGAAGTGGGGTG-30 for b-actin) underthe following conditions: 36 cycles for C5aR and 32 cyclesfor b-actin, 30 seconds at 94�C, 30 seconds at 58�C, and 30seconds at 72�C. PCR products were resolved by electro-phoresis using 1% agarose gels and were visualized byethidium bromide staining.

ImmunoblottingTo detect C5aR, cell lysates obtained from bile duct or

colon cancer cells were analyzed by SDS-PAGE underreducing conditions using 10% polyacrylamide gels andwere transferred to polyvinylidene difluoride (PVDF)mem-branes (Immobilon Transfer Membranes; Millipore). Afterblocking with 5% fat-free milk, the membranes were

Translational RelevanceThe anaphylatoxin C5a is a product of complement

system activation that occurs in the cancer microenvi-ronment, however, the role of C5a generated in cancertissues is largely unknown. We found C5a receptor(C5aR) expression in cancer cells of human cancertissues samples. Moreover, by Matrigel chamber assayand nude mouse skin implantation, we showed C5a-elicited enhancement of C5aR-expressing cell invasionthrough motility activation and matrix metalloprotei-nases (MMP) release. Indeed, anti-C5aR antibody andMMP inhibitor GM6001 suppressed C5aR-expressingcancer cell invasion enhanced by C5a. These resultsillustrate a novel activity of the C5a–C5aR axis in cancerand suggest that it might be beneficial to target thissignaling pathway for cancer therapy. Furthermore, asC5aR expression was seen in adenocarcinoma, squa-mous cell carcinoma, and transitional cell carcinoma,C5a–C5aR–targeting therapy may work for those differ-ent types of cancer.

Nitta et al.

Clin Cancer Res; 19(8) April 15, 2013 Clinical Cancer ResearchOF2




incubated with anti-human C5aR rabbit immunoglobulinG (IgG; 1,000-fold dilution; Santa Cruz Biotechnology) orpolyclonal anti-actin rabbit IgG (500-fold dilution; SantaCruz Biotechnology). This was followed by incubationwith horseradish peroxidase (HRP)–conjugated anti-rabbitIgG goat antibody (1,000-fold dilution; Amersham Bios-ciences), and bands were visualized via enhanced chemi-luminescence (ECL), according to the manufacturer’sinstructions.

Establishment of C5aR stably expressing HuCCT1 cellsFull-length human C5aR cDNA of 1,053 bp was ampli-

fied by PCR using human macrophage cDNA library andsubsequently subcloned into the pENTR/D-TOPO-vector(Invitrogen). After confirming the sequence, the cDNA wasinserted into pCAG-IRES-puro vector using the Gatewaysystem (Invitrogen). The purified plasmid was transfectedinto HuCCT1 cells using Lipofectamine 2000 (Invitrogen).After 48 hours, medium was replaced with selection medi-um supplemented with puromycin (1 mg/mL) to be cul-tured for 2 weeks. Puromycin-resistant cells were collectedand were subjected to cell sorting by fluorescence-activatedcell sorting (FACS) Vantage to isolate those cells highlyexpressing C5aR (designated HuCCT1/C5aR). HuCCT1cells transfected with empty-pCAG-IRES-puro vectors wereused as the control (HuCCT1/mock).

Flow-cytometric analysisMEC, HuCCT1/mock, or HuCCT1/C5aR cells were trea-

ted for 30 minutes with a murine monoclonal fluoresceinisothiocyanate (FITC)–conjugated anti-C5aR antibody(Serotec Ltd.), or a FITC-conjugated isotype-matched con-trol antibody (Serotec Ltd.), followed by washing with PBStwice. C5aR antigen was quantified by FACScan (BDBiosciences).

Immunofluorescence analysisFilamentous actin (F-actin) formation was visualized as

previously described (24).Cellswere seeded at a lowdensityon glass coverslips and were incubated for 24 hours. After 2hours serum starvation, cells were stimulated with 100nmol/L human C5a (Sigma) for the stated time periods.Cells were then fixed in 4% paraformaldehyde, permeabi-lized in 0.2% Triton X-100 for 5 minutes, and were incu-bated with 5 U/mL Alexa 488-phalloidin (MolecularProbes) for 40 minutes, followed by washing with PBS.Images were obtained and processed by FluoView 300 LaserScanning Confocal Microscope (Olympus).

Time-lapse video analysisCells (1� 104/well in RPMI-1640) were cultured in a 24-

well glass-bottom plate (Iwaki) for 24 hours. After additionof C5a (final concentration: 100 nmol/L), cells were main-tained at 37�C in 5%CO2 within the chamber set under thecamera during the observation. Images were obtained using�20 UPlan SApo objective (Olympus IX81). The camera,shutters, and filter wheel were controlled by MetaMorphimaging software (Molecular Devices), and images were

collected every 10 minutes with an exposure time of 50milliseconds. Cell migration distance was measured bytracing individual cells using MetaMorph imaging softwareaccording to the manufacturer’s instructions.

Invasion assay in vitroTo assess invasion of cancer cell lines in vitro, BioCoat

Matrigel invasion chambers were used (24-well plate, 8-mmpore; BDBiosciences; ref. 25).HuCCT1-derived (3.75� 104

cells), or MEC (7.5 � 104 cells) cells were suspended inserum-free RPMI-1640, then seeded into the upper cham-ber. RPMI-1640 supplemented with either C5a or carriersolution (PBS) was placed in the lower chamber. To blockC5aR-mediated signaling, anti-human C5aR rabbit IgG(10 mg/mL) or nonspecific control IgG (10 mg/mL) wasadded to the cell suspension before seeding. For analyzingthe effect of discontinuous stimulation with C5a, cells werecultured at 37�C in serum-free RPMI-1640 supplementedwith C5a for 12 hours (HuCCT1-derived cells) or 24 hours(MEC) at indicated concentrations. Cells were then washedwith serum-free RPMI-1640 andwere seeded into the upperchamber. RPMI-1640 containing 10% FBS was set in thelower chamber. Chambers were incubated for 24 hours(HuCCT1-derived cells) or 36 hours (MEC) at 37�C. Cellson the upper surface of the filter were removedwith a cottonwool swab, and cells thatmigrated to the lower surface werefixed in100%methanol andwere stainedwith 1%toluidineblue. Invaded cells were counted in 5 power fields (�20).The invasion-enhancing effect was shown as the ratio of cellinvasion by C5a stimulation versus PBS controls. To deter-mine whether matrix metalloproteinases (MMP) wereinvolved in C5a-elicited cancer cell invasion, GM6001 (5mmol/L; Merck) was added to the cell suspension when cellinvasion activity of 100 nmol/L C5a was measured. Forcheckerboard analysis for C5a cancer cell invasion activity,various concentrations of C5a were added to the HuCCT1/C5aR cell suspension in the upper chamber together withthe lower chamber, and cell invasion was assessed asdescribed earlier.

Invasion assay in vivoHuCCT1/mock and HuCCT1/C5aR were incubated in

serum-freemedium in the presence or absence of C5a (10�7

mol/L) at 37�C for 12 hours. This was followed by labelingwith CellTracker Orange CMTMR (20 mmol/L) or Cell-Tracker Green BODIFY (25 mmol/L; Molecular Probes) forHuCCT1/C5aR and HuCCT1/mock, respectively, at 37�Cfor 45 minutes. After washing with serum-free medium,HuCCT1/mock cells and HuCCT1/C5aR cells were equallymixed to create a concentration of 3 � 107 cells/mL each.The cell mixture (50 mL) was injected intradermally into 7-week-old BALB/cA Jc1-nu/numice (CLEA Japan). After 1, 2,or 3 days, the nude mice were sacrificed by cervical dislo-cation, and the cell injection sites including surroundingtissues were excised to prepare frozen sections in liquidnitrogen. Labeled cells in 4-mmthick sectionswere observedwith a fluorescence microscope (BIOREVO; KEYENCE). Toquantify the distribution of HuCCT1-derived cells, regions

C5a–C5aR Axis Enhances Cancer Cell Motility and Invasion

www.aacrjournals.org Clin Cancer Res; 19(8) April 15, 2013 OF3




of fluorescent dots of labeled cells were encircled (Fig. 5A)then the area of each regionwasmeasured using an imagingprocessor (VH-Analyzer; KEYENCE). The ratio of the dis-tribution area of HuCCT1/C5aR versus HuCCT1/mock wascalculated. Some endogenous green fluorescence back-ground was observed in mice skin, therefore these spotswere avoided and cancer cells were specifically encircled,whichwas confirmed by observation of the adjacent sectionhematoxylin and eosin (H&E)-stained. This experimentwas carried out according to the criteria of animal experi-ments of the Kumamoto University Animal ExperimentCommittee.

Measurement of MMP concentration in culturesupernatant of cells

MMP concentration in supernatant of cancer cells stim-ulated with or without C5awasmeasured using theQuanti-body Human MMP Array 1 Kit (RayBiotech). Culturesupernatant was taken from MEC (1 � 106 cells) orHuCCT1/C5aR (5 � 105 cells) grown in a 6-well plate for24 hours in the presence or absence of C5a (100 nmol/L).The supernatant was diluted at 1:3 with PBS and thenMMPconcentrations were determined according to the manufac-ture’s instructions.

StatisticsStatistical analyses were conducted using the unpaired

Student t test. Values are expressed as means � SD andexperiments were carried out in triplicate, unless otherwisestated.

ResultsAberrant expression of C5aR in human cancer cells

We first investigated C5aR expression in human cancerspecimens from 225 patients by immunohistochemistry.C5aR expression was observed in cancer cells from all theorgans examined and in all the 3 cancer cell types, squa-mous cell carcinoma, adenocarcinoma, and transitional cellcarcinoma (Fig. 1AandSupplementary Fig. S1A).Generally,C5aR was robustly expressed in a significant proportion ofcancer samples. On the other hand, null or only faintreaction of C5aR immunohistochemistry was observed intheir normal counterparts (Fig. 1A and Supplementary Fig.S1B) except kidney tubular epithelial cells (SupplementaryFig. S1B), which is in line with the previous report (16).Percentage of C5aR-positivity in cancer cases varied amongorgans. In colon, bile duct, kidney, and prostate carcino-mas, more than 50% of cases examined were C5aR-positive(Fig. 1B). In bile duct-derived cancer in the liver, C5aR waspositive in 26 patients. Among them, vascular invasion wasfound in 18 patients, whereas vascular invasion was seen inonly 4 cases of 16 C5aR-negative patients. This resultindicates a significant relationship between cancer cell C5aRexpression and vascular invasiveness (P ¼ 0.010 by Fisherexact test). Because vascular invasion of bile duct cancer isclosely linked to metastasis and prognosis (26), C5aRexpression may correlate with those clinical endpoints.Next, we examined a panel of cancer cell lines for C5aR

expression. Real-time PCR (RT-PCR) revealed that severalhuman cancer cell lines originated from bile ducts (MECand RBE) and colon (HCT15, COLO205, and HCT116)expressed C5aR mRNA (Fig. 2A). Out of these cell lines,those except RBE also expressed C5aR protein (Fig. 2B). Thelocalization of C5aR on the cell-surface was shown by flowcytometry (Fig. 2C). These results suggest that aberrantC5aR expression observed in human cancer specimens isactually conserved in some human cancer cell lines. It isintriguing that only MEC cells express C5aR at the proteinlevel among the bile duct-derived cancer cell lines (Fig. 2B),whereas bile duct carcinomas showed the highest positiveratio of C5aR expression (Fig. 1B). This fact seems

Esophagus

Colon

A

B

0

20

40

60

80

100

Pos

itivi

ty (

%)

EP

ST

Col Liv

BD

Pan Kid

UB

PT

Lun

Mam

Figure 1. Expression of C5aR in humanprimary cancers. A, human cancertissueswere immunohistochemically stained with an anti-C5aR antibodyor control IgG (inset), and representative examples are shown. Coloncancer (top left) and normal colon epithelium (top right); esophagusnormal epithelium (bottom, encircled) and cancer (bottom, not encircled).Scale bars, 50 mm. B, cancer cell C5aR-positive case ratios (positivecases/total cases examined; *) in primary organs. The vertical scale barshows the 95% confidence interval that is estimated from the binominalproportion based on the b distribution. EP, esophagus (n ¼ 20); ST,stomach (n ¼ 20); Col, colon (n ¼ 40); Liv, liver (n ¼ 20); BD, bile ducts(n ¼ 46); Pan, pancreas (n ¼ 10); Kid, kidney (n ¼ 11); UB, urinarybladder (n¼ 45); Lun, lung (n¼ 12); PT, prostate (n¼ 25);Mam,mammarygland (n ¼ 21).

Nitta et al.





paradoxical to the result of immunohistochemistry at aglance (Fig. 1A and B), implying that C5aR expression waslost during the process of cell line establishment fromprimary culture of human cancer cells, as those cells canprioritize expression of other essential proteins for clonaldevelopment in the context of 2-dimensional culture invitro, which is usually in the absence of C5a.

Cytoskeletal rearrangement and enhanced motility ofC5aR-expressing cancer cells by C5a stimulusTo analyze the biologic effects of C5aR expression in

cancer cells under C5a stimuli, we chose C5aR-negativeHuCCT1 cells derived from bile duct carcinomas of thehighest C5aR expression ratio (Fig. 1B). HuCCT1/C5aRcells but not HuCCT1/mock cells expressed C5aR (Fig.2A and B). The cell surface expression of C5aR inHuCCT1/C5aR cells was confirmed by flow-cytometricanalysis (Fig. 2C), which was comparable with that in MEC

cells (Fig. 2C). Because the chemoattractant C5a causescytoskeletal rearrangement and stimulates migration ofleukocytes (27, 28), we hypothesized that cancer cells mayexploit this mechanism to gain the ability of migration andinvasion by activationof aberrantly expressedC5aRon theircell surface. To test this, the effect of C5aR activation onactin rearrangement was analyzed by F-actin immunofluo-rescence labeling. The majority of cells at the outer edges ofHuCCT1/C5aR cell clusters clearly showed strong filopodiaformation 30 minutes after C5a treatment (Fig. 3A), whichwas followed by development of membrane ruffling anddissolution of stress fibers (Fig. 3A). Three hours after thetreatment, such ruffles disappeared and formation of stressfibers became evident again (Fig. 3A). Some cells at theperiphery of the cluster even changed their morphology tospindle-like shape and protruded from the cluster to thevacant area (Fig. 3A). On the other hand, HuCCT1/mockcells did not show any remarkable changes in both cellmorphology and actin cytoskeleton at any time points afterC5a stimulation (Fig. 3A). The time-lapse video analysis ofHuCCT1/C5aR cell movement showed that C5a activatedmotility of the cells (Fig. 3B and Supplementary Fig. S2video). Tracing of cell movement revealed that C5aenhanced motility of HuCCT1/C5aR cells in a dose-depen-dentmanner, increasingmotility 3-fold at 100 nmol/L (Fig.3C), whereas motility of HuCCT1/mock cells was notsignificantly affected by C5a (Fig. 3B and C). This experi-ment confirmed that C5a enhances motility of cancer cellsin a C5aR-dependent fashion.

Enhanced invasiveness of C5aR-expressing cancer cellsby C5a stimulus in vitro

Experiments using Matrigel chambers revealed that C5astimulated invasion of HuCCT1/C5aR cells through thematrix layer in a C5a concentration-dependentmanner andenhanced approximately 13-fold over carrier control (PBS)at 100 nmol/L (Fig. 4A). The enhancing effect of C5a oninvasion of HuCCT1/mock cells was not seen (Fig. 4A).Intriguingly, cancer cells pretreated with C5a also showedenhanced invasiveness even in the absence of a C5a con-centration gradient (Fig. 4C). This result indicates thatstimulation with C5a during pretreatment is sufficient forenhancing invasion of C5aR-expressing cancer cells in vitro,and suggests that neither concentration gradient nor con-tinuous exposure to C5a is required for activating invasionof C5aR-expressing cancer cells. Similarly, C5a enhancedinvasion ofMEC cells that endogenously express C5aR (Fig.2C), in a C5aR-dependent manner, as this enhancementwas abrogated by a C5aR antagonist (Supplementary Fig.S3) and by the neutralizing antibody against C5aR, but notby nonspecific IgG (Fig. 4B and D). Dose-dependent effectof C5a on invasion was also the case observed in MEC cells(Fig. 4B and D). To determine whether C5a-elicited cancercell migration is dependent on the C5a concentrationgradient, we conducted the checkerboard analysis. In addi-tion to C5a concentration-dependent invasion in theabsence of C5a in the upper chamber, HuCCT1/C5aR cellinvasion by 100 nmol/L C5a in the lower chamber was

A

B

C

C5aR(+)HuCCT1HuCCT1/C5aR

100 101 102 103 104

200

160

120

80

40

0

Cel

l cou

nts

C5aR expression

HuCCT1/mock

100 101 102 103 104

200

160

120

80

40

0

Cel

l cou

nts

C5aR expression

MEC

100 101 102 103 104

200

160

120

80

40

0

Cel

l cou

nts

C5aR expression

C5aR

β-Actin

β-ActinM

EC

HuC

CT

1

SS

p-25

RB

E

moc

k

C5a

R

DLD

1

SW

620

HC

T15

CO

LO20

5

HC

T11

6

Bile duct ColonHuCCT1

C5aRM

EC

Bile duct Colon

HuC

CT

1

SS

p-25

RB

E

moc

k

DLD

1

SW

620

HC

T15

CO

LO20

5

HC

T11

6

HuCCT1

C5a

R

Figure 2. Expression of C5aR on cancer cell lines including HuCCT1/C5aRandHuCCT1/mock.A, expression ofC5aRmRNA in several cancercell lines shown by RT-PCR. B, expression of C5aR protein in bile ductand colon cancer cell lines, shownby immunoblotting using an anti-C5aRantibody.b-actinmRNAandproteinwere used ascontrols. C, expressionofC5aRon the cellmembrane inMECandHuCCT1/C5aR, shownby flowcytometry using FITC-conjugated anti-C5aR antibody (gray line) orcontrol antibody (black line).






inhibited byC5a in the upper chamber in a dose-dependentmanner (Fig. 4E). Moreover, invasion by 10 nmol/L C5a inthe lower chamberwas completely inhibited by 100nmol/LC5a in the upper chamber (Fig. 4E). These results suggestthat C5a-induced cancer cell invasion is explained partly bychemotaxis, particularly in the presence of 100 nmol/L C5a.However, when C5a concentration in the upper chamberwas equal to that in the lower chamber, C5a was still able toinduce HuCCT1/C5aR invasion to the significant extent.The invasion by 10 nmol/L C5a in the lower chamber wasnot affected by addition of 10 nmol/L C5a in the upperchamber (Fig. 4E). Furthermore, when 100nmol/LC5awas

added to the cell suspension in the upper chamber, ittriggered significant cancer cell migration even in theabsence of C5a in the lower chamber (Fig. 4E). These resultssuggest that enhanced random cell locomotion plays a vitalrole in the C5a-elicited cancer cell invasion.

C5a elicitsMMPsecretion fromC5aR-expressing cancercells

Degradation of extracellularmatrix (ECM) byMMPs is anessential process for cancer cell invasion (29). The interac-tion of the chemokine CXCL12 and its receptor CXCR4 hasbeen shown to increase MMP expression and invasion of

HuCCT1/C5aR

HuCCT1/mock

0 min 30 min 60 min 180 min 360 minA

C5a concentration (nmol/L)

Cel

l mig

ratio

n (µ

m)

200

400

600

800

1,000

0 10 100

≈≈

≈≈

*

n.s.

*

B

C

HuCCT1/C5aR

HuCCT1/mock

C5a(100 nmol/L)

(+)

(+)

( )

( )

0 3 6 9

Time lapse (h)Figure 3. C5a elicits C5aRexpressing cancer cells byinducing cytoskeletalreorganization and changes incellular morphology. A, HuCCT1/C5aR and HuCCT1/mock cellswere incubated with C5a (100nmol/L) and fixed at indicated timepoints. F-actin was visualized byimmunofluorescence staining withAlexa 488–conjugated phalloidin.Scale bars, 20 mm. Arrowhead andan arrow indicate filopodia andmembrane ruffling, respectively. B,time-lapse analysis of cell motility.Cell images taken at 0, 3, 6, and 9hours are shown. Broken circlesindicate the initial cell positionshown at 0 hours. C, cell migrationdistance was measured by tracinga cell. Open and closed circlesindicate HuCCT1/C5aR andHuCCT1/mock, respectively.�, P < 0.01 (n ¼ 6). n.s., notsignificant.

Nitta et al.





prostate cancer cells (30).WhileC5a inducesMMP-9 releasefrom human neutrophils (31), such C5a-elicited MMPrelease from cancer cells has not been reported. Interest-ingly, C5a-enhanced invasion of C5aR-expressing cancercells in the transwell chamberswas significantly hinderedbyan MMP inhibitor GM6001 (32; Fig. 4A and B), indicatingthat enzymatic activity of MMPs plays a crucial role in theC5a-enhanced invasion of C5aR-expressing cancer cells.Hence, we explored the possibility of C5a provoking MMPsecretion from C5aR-expressing cancer cells. MMP expres-sion array analysis showed that C5a significantly increasedrelease of MMP-1, 3, 9, 10, and 13 from MEC cells, andMMP-8 and 10 from HuCCT1/C5aR cells (Table 1). TheseMMPs are known tobe associatedwithboth cancer invasionand patient prognosis (33, 34). Together with inhibition ofC5a-enhanced invasion by GM6001 (Fig. 4A and B),increased secretion of MMPs by C5a (Table 1) indicatesthat MMPs contribute to the C5a-enhanced invasion ofC5aR expressing cancer cells. Among specific MMP inhibi-

tors, an MMP-8 inhibitor exhibited the most significanteffect to impede the MEC cell invasion enhanced by C5(Supplementary Fig. S3). Intriguingly, about a 3.2-foldincrease in MEC cell invasion induced by C5a at 10nmol/L (Fig. 4B) seemed to correlate with a 3.7-foldincrease in MMP-8 (Table 1). These results suggest thatMMP-8 is the most responsible MMP for C5a-elicited MECcell invasion.

Enhanced invasiveness of C5aR-expressing cancer cellsby C5a in vivo

To evaluate the effect of C5a stimulation on the invasive-ness of C5aR-expressing cancer cells in vivo, HuCCT1/mockand HuCCT1/C5aR cells were pretreated with C5a, labeledwith green or orange fluorescent dyes, respectively, thenmixed to be injected into nude mice skin. This assay systemenables direct comparison of spreading in situ between 2different sublines. This assay revealed that C5a-treatedHuCCT1/C5aR cells spread more broadly than C5a-treated

Figure 4. Enhancement of C5aR-expressing cancer cell invasion byC5a in vitro. Cancer cell invasion wasmeasured by the Matrigel invasionchamber assay. HuCCT1-derived (A)or MEC cells (B) were placed in theupper chamber and mediumsupplemented with variousconcentrations of C5a was placed inthe lower chamber. HuCCT1-derived(C) or MEC cells (D) were pretreatedwith various concentrations of C5afor 12 and 24 hours, respectively.These cells were washed and wereplaced in the upper chamber, andmedium was placed in the lowerchamber. Incubation time for assay:24 hours (A and C) and 36 hours (BandD). (&), HuCCT1/mock (AandC);(&), HuCCT1/C5aR (A and C); cross-hatched and closed bars (B and D)indicate MEC cells treated with anti-C5aR antibody or nonspecific IgG,respectively. Gray bars (A and B)indicate cells incubated in the upperchamber in the presence of GM6001(5 mmol/L). E, checkerboard analysisfor C5a cancer cell invasion activity.Closed, cross-hatched, and openbars indicate C5a concentration 0,10, or 100 nmol/L, respectively, in thelower chamber. �, P < 0.01; n.s., notsignificant.

0

5

10

15

0 1 10 100

*

*

**


Mig

rate

d ce

ll ra

tio (

fold

)

0

5

10

0 1 10 100

**

*


Mig

rate

d ce

ll ra

tio (

fold

)

A B

C D

Mig

rate

d ce

ll ra

tio (

fold

)

0

1

2

3

4

5

0 1 10 100

**

**n.s.


Mig

rate

d ce

ll ra

tio (

fold

)

0

1

2

3

0 1 10 100

** *

n.s.

E

0

5

10

15

Mig

rate

d ce

ll ra

tio (

fold

)

0 10 100C5a in the upper chamber (nmol/L)

** *

**

n.s.

n.s.


+ G

M

+ G

M

+ a

nti-

C5a

R+

ant

i- C

5aR

+ n

onsp

ecifi

c Ig

G

+ n

onsp

ecifi

c Ig

G






HuCCT1/mock cells in nude mice skin tissue (Fig. 5A).The most evident induction of the spreading of HuCCT1/C5aR by C5a was observed at day 2, which was a 1.8-foldincrease over HuCCT1/mock (Fig. 5B). In contrast, therewas no difference in spreading between those 2 sublines of

HuCCT1 when they were not treated with C5a (Fig. 5Aand B). This result suggests that C5a enhances invasion ofC5aR-expressing cancer cells in vivo as well as in vitro, andagain stimulation with C5a before injection is sufficient forC5aR-expressing cancer cells to show such enhancedinvasiveness.

DiscussionC5-derived fragments were reported to enhance cancer

cell locomotion, however, these were not identified asC5a when such reports were made. Rather, they seemednot to be C5a because of its molecular weight and lack ofchemotactic activity for leukocytes (35, 36). When thesestudies were conducted, the C5aR molecule had not beenidentified, and neither recombinant C5a nor C5aR-spe-cific antibodies were available to prove the activity of C5ato enhance cancer cell migration. In the present study, wehave shown several lines of evidence indicating a crucialrole of C5a–C5aR interaction in cancer cell invasion: (i)C5aR expression was observed in cancer cells frompatients’ tissues and in various human cancer cell lines(Figs. 1 and 2), (ii) recombinant C5a enhanced cancer cellmotility (Fig. 3) and invasion both in vitro (Fig. 4) and invivo (Fig. 5), and (iii) enhanced cancer cell invasivenesselicited by the C5a–C5aR axis was dependent on theincreased release of MMPs (Fig. 4 and Table 1), proteasesthat are indispensable for cancer cell invasion towardsurrounding tissues (29). C5aR expression is essential forcancer cells of epithelial origin to enhance motility andinvasiveness by C5a, given that C5aR expression isrequired for any remarkable changes in cell morphology(Fig. 3A) and enhanced invasiveness (Fig. 4A and C) inHuCCT1 cells after C5a stimulation (Fig. 5A and B). Inaddition, a C5aR antagonist (Supplementary Fig. S3) andby a neutralizing antibody against C5aR (Fig. 4B and D)abrogated C5a enhanced invasion of MEC cells. Thesedata are consistent with the phenomenon that C5a

Table 1. C5a-stimulated MMP release from C5aR-expressing cancer cells

MEC HuCCT1/C5aR

MMP C5a (�)a C5a (þ)a Ratiob C5a (�)a C5a (þ)a Ratiob

1 149,332.6 � 17,039 291,140.4 � 5,360.4� 1.95 88,906.6 � 27,436.8 129,408 � 3,485 1.462 32.5 � 11.3 116.9 � 67.8 3.59 1,220.4 � 184.3 870.7 � 73 0.713 26.8 � 12.0 134.7 � 42.1� 5.03 971.6 � 224.6 927.5 � 162.9 0.958 20.1 � 23.6 74.9 � 29.9 3.72 335.6 � 77.7 1,128.3 � 120.9� 3.369 < 0.1c 12.3 � 3.6� — 6,105.1 � 2,011.3 6,213.4 � 4,357.7 1.0210 857.6 � 740.5 9,301.4 � 641.8� 10.84 57,641.4 � 5,197.5 111,666.4 � 7,911.1� 1.9413 359.2 � 118.7 1,367.3 � 309.3� 3.80 1,411.9 � 154.6 1,642.8 � 287 1.16

aMMP concentrations (pg/mL) in culture supernatants of cancer cells cultured for 24 hours in the presence (þ) or absence (�) of C5a(100 nmol/L).bRatio of C5a(þ) vs. C5a(�) in mean values.cBelow the detection limit.�, P < 0.01 (n ¼ 4).

B

1

2

Cel

l dis

trib

utio

n ar

ea r

atio

C5a

R(+

)/C

5aR

(–)

(fol

d)

21 3 (d)

**

C5aR

(+) (–)Merge

C5a

(+)

(–)

A

Figure 5. Enhanced invasiveness ofC5aR-expressing cancer cells byC5ain vivo. HuCCT1/C5aR and HuCCT1/mock cells were incubatedseparately in the presence or absence of 100 nmol/L C5a [C5a(þ) or C5a(�), respectively] for 12 hours. After washing, cells were labeled, mixed,and injected together into nude mouse skin. A, distribution of HuCCT1/C5aR (orange) and HuCCT1/mock (green) cells in nude mouse skin 2days after implantation. Right, corresponding sections stained with H&E.Bars, 300 mm. B, cell distribution square ratio (HuCCT1/C5aR cells vs.HuCCT1/mock cells). Cells were incubated in the presence (&) orabsence (&) of C5a. �, P < 0.01 (n ¼ 5).

Nitta et al.





enhances cancer cell invasion via C5aR, and to ourknowledge, this is the first report that shows the biologicrole of the C5a–C5aR axis in human cancer cell invasion.C5aR being expressed in cancerous cells but not normal

epithelial cells except kidney proximal tubular epithelialcells in human tissue specimens (Fig. 1A and Supplemen-tary Fig. S1A) may indicate C5aR expression to be aconsequence of malignant transformation. A similarexample is CXCR4: the receptor of CXCL12 that is apotent chemoattractant such as C5a and is produced inthe cancer microenvironment (20). This receptor is com-monly found in cancer cells and its expression is inducedby factors such as hypoxia, VEGF, and estrogen in thecancer microenvironment. CXCR4 expression is also acti-vated by mutations in genes that alter levels of hypoxia-inducible factors, and gene fusion events. Interleukin(IL)-6–induced C5aR expression in rat hepatocytes (37)suggests that C5aR can be expressed in response to spe-cific cytokines that are rich in the cancer microenviron-ment (20, 38). However, IL-6 and IFN-g did not induceC5aR expression in HuCCT1 cells (unpublished data).This may suggest that C5aR expression is dependent ongenetic events characteristic in individual cancers. Suchdifferences might reflect variation in C5aR-positivity indifferent primary organs (Fig. 1B).Leaky cancer vasculature facilitates the supply of the

complement system components from the bloodstreamto cancer tissues (39), in which as shown in an animalcancer model (2), C5a is generated through activation ofthe complement system in response to cancer cells (3),although they are protected from complement attack bycomplement regulators (6). Indeed, C5a is detectable inhuman plasma incubated with MEC or HuCCT1 cells invitro (Supplementary Fig. S4). Besides this pathway, C5ais possibly generated directly from C5 through thrombin-dependent cleavage (12), following the coagulation reac-tion initiated by tissue factor that can be expressed oncells in cancer tissues including cancer cells, fibroblastsand activated leukocytes (13). C5a can also be generatedfrom C5 by a serine protease from activated phagocytes(14) recruited to the cancer tissue. In addition, comparedwith tightly adhering noncancerous epithelial cells, theloose cell-to-cell contact of cancer cells allows the gener-ated C5a to access the cancer cell membrane, enabling itto bind to C5aR. Thus, C5a produced in the cancermicroenvironment can be predicted to activate C5aR-positive cancer cells to promote migration from theprimary site.C5a induced dynamic sequential reorganization of

actin cytoskeleton in C5aR-expressing cancer cells, name-ly, filopodia formation, membrane ruffling then forma-tion of stress fibers (Fig. 4A). These processes have beenreported to be provoked by activation of Cdc42, Rac1,and RhoA, respectively (40). Such sequential activation ofthese small G proteins, which is a robust driving force forcell movement, has been documented previously (41). Infact, C5a induces activation of Cdc42 and Rac1 in neu-trophils, leading to actin reorganization of the cell (42).

We are currently studying if the C5a–C5aR axis canactivate upstream signaling pathways of those small Gproteins in cancer cells.

Together with motility stimulation in the C5a-contain-ing culture medium (Fig. 3), increased invasiveness ofC5a-treated C5aR-positive cancer cells in the matrix gel(Fig. 4) and in nude mouse skin (Fig. 5) in the absence ofa C5a concentration gradient suggest that C5a enhancescancer cell random locomotion instead of inducing che-motaxis, which is supported by the checkerboard analysis(Fig. 4E). If C5a were only chemotactic for cancer cells,stimulated cells would be expected to assemble in theprimary cancer site where C5a is released and enriched;thus, C5a would hinder cancer cell spread. In contrast,enhanced random cell locomotion would be more rele-vant for promoting cancer cell invasion and spreading,namely, the phenomenon that cells leave away from thesource of the stimulant. Accordingly, such C5a activity ispresumed to favor cancer cell dissemination from theprimary site (Fig. 5).

This study implies that the C5a–C5aR axis could be atarget for anticancer therapy. For instance, depleting C5with anti-C5 antibodies would suppress C5a generation incancer tissues. Therefore, this suggests that C5aR may alsobecome a possible and feasible target for molecular-basedmedicine by generating specific antagonists. Such agentsmayprovide useful therapeutic options for cancer treatmentin the future.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: H. Nitta, Y. Wada, T. ImamuraDevelopment of methodology: H. Nitta, Y. Wada, A. Irie, G. Yamada,N. ArakiAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): H. Nitta, Y. Kawano, K. Taniguchi, J. Honda,M. Wilson-Morifuji, N. ArakiAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): H. Nitta, Y. Wada, Y. Kawano, A. Irie,K. Kikuchi, N. Araki, T. ImamuraWriting, review, and/or revision of the manuscript: H. Nitta, Y. Wada,Y. Kawano, T. ImamuraAdministrative, technical, or material support (i.e., reporting ororganizing data, constructing databases): Y. Murakami, K. Taniguchi,K. Suzuki, N. ArakiStudy supervision: M. Eto, H. Baba

AcknowledgmentsThe authors thank Ms. T. Kubo for technical assistance with the immu-

nohistochemical staining, and Mr. K. Matsuda, KEYENCE, for the fluores-cence microscope imaging, and analysis.

Grant SupportThis work was supported in part by a Grant-in-Aid for Scientific Research

(B) from the Japanese Ministry of Education and Science (Grant No.18390125 to T. Imamura).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received April 12, 2012; revised November 11, 2012; accepted November13, 2012; published OnlineFirst January 3, 2013.






References1. Carroll MC. The complement system in regulation of adaptive immu-

nity. Nat Immunol 2004;5:981–6.2. Markiewski MM, DeAngelis RA, Benencia F, Ricklin-Lichtsteiner SK,

Koutoulaki A, Gerard C, et al. Modulation of the antitumor immuneresponse by complement. Nat Immunol 2008;9:1225–35.

3. Niculescu F, Rus HG, Retegan M, Vlaicu R. Persistent complementactivation on tumor cells in breast cancer. Am J Pathol 1992;140:1039–43.

4. M€uller-Eberhard HJ. Molecular organization and function of the com-plement system. Annu Rev Biochem 1988;57:321–47.

5. Morgan J, Spendlove I, Durrant L. The role of CD55 in protecting thetumour environment from complement attack. Tissue Antigens2002;60:213–23.

6. Gancz D, Fishelson Z. Cancer resistance to complement-dependentcytotoxicity (CDC): problem-oriented research and development. MolImmunol 2009;46:2794–800.

7. Liu J, Miwa T, Hilliard B, Chen Y, Lambris J, Wells A, et al. Thecomplement inhibitory protein DAF (CD55) suppresses T cell immunityin vivo. J Exp Med 2005:201:567–77.

8. Mikesch JH, Buerger H, Simon R, Brandt B. Decay accelerating factor(CD55): a versatile acting molecule in human malignancies. BiochimBiophys Acta 2006;1766:42–52.

9. Guo RF, Ward PA. Role of C5a in inflammatory responses. Annu RevImmunol 2005;23:821–52.

10. Markiewski MM, Lambris JD. The role of complement in inflammatorydiseases from behind the scenes into the spotlight. Am J Pathol2007;171:715–27.

11. DiScipio RG,SmithCA,M€uller-EberhardHJ,Hugli TE. The activation ofhuman complement component C5 by a fluid phase C5 convertase.J Biol Chem 1983;258:10629–36.

12. Huber-Lang M, Sarma JV, Zetoune FS, Rittirsch D, Neff TA, McGuireSR, et al. Generation of C5a in the absence of C3: a new complementactivation pathway. Nat Med 2006;6:682–7.

13. KasthuriRS, TaubmanMB,MackmanN.Role of tissue factor in cancer.J Clin Oncol 2009;27:4834–8.

14. Huber-LangM, Youkin EM, Sarma JV, Riedermann N, McGuire SR, LuKT, et al. Generation of C5a by phagocytic cells. Am J Pathol2002;161:1849–59.

15. Gerard NP, Gerard C. The chemotactic receptor for human C5aanaphylatoxin. Nature 1991;349:614–7.

16. Zwirner J, Fayyazi A, G€otze O. Expression of the anaphylatoxin C5areceptor in non-myeloid cells. Mol Immunol 1999;36:877–84.

17. Buchner RR, Hugli TE, Ember JA, Morgan EL. Expression of functionalreceptors for human C5a anaphylatoxin (CD88) on the human hepa-tocellular carcinoma cell line HepG2. Stimulation of acute-phaseprotein-specific mRNA and protein synthesis by human C5a anaphy-latoxin. J Immunol 1995;155:308–15.

18. Coussens LM, Werb Z. Inflammation and cancer. Nature 2002;420:860–7.

19. Balkwill F. Cancer and the chemokine network. Nat Rev Cancer2004;4:540–50.

20. RetzMM,SidhuSS,Blaveri E, Kerr SC,DolganovGM,LehmannJ, et al.CXCR4expression reflects tumor progression and regulatesmotility ofbladder cancer cells. Int J Cancer 2005;114:182–9.

21. Lian Z, YoonY, Votaw J,GoodmanMM,William L, ShimH. Silencing ofCXCR4 blocks breast cancer metastasis. Cancer Res 2005;65:967–71.

22. Saur D, Seidler B, Schneider G, Alg€ul H, Beck R, Senekowitsch-Schmidtke R, et al. CXCR4 expression increases liver and lungmetas-tasis in a mouse model of pancreatic cancer. Gastroenterology2005;129:1237–50.

23. Markiewski MM, Lambris JD. Unwelcome complement. Cancer Res2009;69:6367–70.

24. Schraufstatter IU, Trieu K, Sikora L, Sriramarao P, DiScipio R. Com-plement C3a and C5a induce different signal transduction cascades inendothelial cells. J Immunol 2002:169;2102–10.

25. Albini A, Iwamoto Y, Kleinman HK, Martin GR, Aaronso SA, KozlowskiJM, et al. A rapid in vitro assay for quantitating the invasive potential oftumor cells. Cancer Res 1987;47:3239–45.

26. Hanazaki K, Kajikawa S, Shimozawa N, Shimada K, Hiraguri M, KoideN, et al. Prognostic factors of intrahepatic cholangiocarcinoma afterhepatic resection: univariae and multivariate analysis. Hepatogas-troenterology 2002;49:311–6.

27. Banks PN, Barker MD, Burton DR. Recruitment of actin to the cytos-keletons of human monocyte-like cells activated by complementfragment C5a. Is protein kinase C involved? Biochem J 1988;252:765–9.

28. Monk PM, Banks P. Evidence for the involvement of multiple signalingpathways in C5a-induced actin polymerization and nucleation inhuman monocyte-like cells. J Mol Endocrinol 1991;6:241–7.

29. Stetler-Stevenson WG, Aznavoorian S, Liotta LA. Tumor cell interac-tionswith the extracellularmatrix during invasionandmetastasis. AnnuRev Cell Dev Biol 1993;9:541–73.

30. Singh S, Singh UP, Grizzle WE, Lillard JW Jr. CXCL12–CXCR4 inter-actions modulate prostate cancer cell migration, metalloproteinaseexpression and invasion. Lab Invest 2004;84:1666–76.

31. Takafuji S, Ishida A, Miyakuni Y, Nakagawa T. Matrix metalloprotei-nase-9 release from human leukocytes. J Investig Allergol Clin Immu-nol 2003;13:50–5.

32. Nabha SM, dos Santos EB, Yamamoto HA, Belizi A, Dong Z, Meng H,et al. Bonemarrow stromal cells enhance prostate cancer cell invasionthrough type I collagen in anMMP-12-dependentmanner. Int JCancer2008;122:2482–90.

33. Miyata Y, Iwata T, Maruta S, Kanda S, Nishikido M, Kaga S, et al.Expression of matrix metalloproteinase-10 in renal cell carcinoma andits prognostic role. Eur Urol 2007;52:791–7.

34. KessenbrockK,PlaksV,WerbZ.Matrixmetalloproteinases: regulatorsof the tumor microenvironment. Cell 2010;141:52–67.

35. Romualdez AG Jr, Ward PA. A unique complement derived chemo-tactic factor for tumor cells. Proc Natl Acad Sci U S A 1975;72:4128–32.

36. Orr W, Phan SH, Varani J, Ward PA, Kreutzer DL, Webster RO, et al.Chemotactic factor for tumor cells derived from the C5a fragment ofcomplement component C5. Proc Natl Acad Sci U S A 1979;76:1986–9.

37. Schiefeldecker HL, Schlaf G, Koleva M, G€otze O, Jungermann K.Induction of functional anaphylatoxin C5a receptors on hepatocytesby in vivo treatment of rats with IL-6. J Immunol 2000;164:5453–8.

38. Takeyama H, Wakamiya N, O'Hara C, Arthur K, Niloff J, Kufe D, et al.Tumor necrosis factor expression by human ovarian carcinoma in vivo.Cancer Res 1991;51:4476–80.

39. Maeda H, Fang J, Inutsuka T, Kitamoto Y. Vascular permeabilityenhancement in solid tumors: various factors, mechanisms involvedand its implications. Int Immunopharmacol 2003;3:319–28.

40. Hall A. Rho GTPases and the actin cytoskeleton. Science 1998;279:509–14.

41. Nobes CD, Hall A. Rho, Rac, and Cdc42 GTPases regulate theassembly of multimolecular focal complexes associated with actinstress fibers, lamellipodia, and filopodia. Cell 1995;81:53–62.

42. Li Z, Hannigan M, Mo Z, Liu B, Lu W, Wu Y, et al. Directional sensingrequires Gbg-mediated PAK1 and PIXa-dependent activation ofCdc42. Cell 2003;114:215–27.

Nitta et al.





Published OnlineFirst January 3, 2013.Clin Cancer Res Hidetoshi Nitta, Yoshihiro Wada, Yoshiaki Kawano, et al. Expressed C5a Receptor (CD88)Invasiveness by Anaphylatoxin C5a via Aberrantly Enhancement of Human Cancer Cell Motility and

Updated version

10.1158/1078-0432.CCR-12-1204doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2013/01/03/1078-0432.CCR-12-1204.DC1Access the most recent supplemental material at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://clincancerres.aacrjournals.org/content/early/2013/04/03/1078-0432.CCR-12-1204To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-12-1204

http://clincancerres.aacrjournals.org/content/suppl/2013/01/03/1078-0432.CCR-12-1204.DC1

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/early/2013/04/03/1078-0432.CCR-12-1204


Documents

Enhancement of Human Cancer Cell Motility and [email protected] doi: 10.1158/1078-0432.CCR-12-1204 2013 American Association for Cancer Research. Clinical Cancer Research