of May 11, 2018.This information is current as

Immunity and Immunotherapeutic EffectVaccines That Can Induce Potent Antitumorwith Flagellin-Related Peptides Are Effective Antigen-Containing Liposomes Engrafted

Abdus Faham and Joseph G. Altin

http://www.jimmunol.org/content/185/3/1744doi: 10.4049/jimmunol.10000272010;

2010; 185:1744-1754; Prepublished online 7 JulyJ Immunol 

MaterialSupplementary

7.DC1http://www.jimmunol.org/content/suppl/2010/07/06/jimmunol.100002

Referenceshttp://www.jimmunol.org/content/185/3/1744.full#ref-list-1

, 18 of which you can access for free at: cites 48 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive 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 © 2010 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Antigen-Containing Liposomes Engrafted withFlagellin-Related Peptides Are Effective VaccinesThat Can Induce Potent Antitumor Immunity andImmunotherapeutic Effect

Abdus Faham and Joseph G. Altin

The bacterial protein flagellin can trigger immune responses to infections by interacting with TLR5 on APCs, and Ag-flagellin

fusion proteins can act as effective vaccines. We report that flagellin-related peptides containing a His-tag and sequence related to

conserved N-motif (aa 85–111) of FliC flagellin, purportedly involved in the interaction of flagellin with TLR5, can be used to

target delivery of liposomal Ag to APCs in vitro and in vivo. When engrafted onto liposomes, two flagellin-related peptides,

denoted as 9Flg and 42Flg, promoted strong liposome binding to murine bone marrow-derived dendritic cells and CD11c+

splenocytes, and cell binding correlated with expression of TLR5. Liposomes engrafted with 9Flg or 42Flg induced functional

MyD88-dependent maturation of dendritic cells in vivo. The vaccination of mice with 9Flg liposomes containing OVA induced

OVA-specific T cell priming, increased the number of Ag-responsive IFN-g–producing CD8+ T cells, and increased Ag-specific

IgG1 and IgG2b in serum. Importantly, the vaccination of C57BL/6 mice with syngeneic B16-OVA–derived plasma membrane

vesicles, engrafted with 9Flg or 42Flg, potently inhibited tumor growth/metastasis and induced complete tumor regression in the

majority of mice challenged with the syngeneic B16-OVA melanoma, in the lung and s.c. tumor models. Strong antitumor

responses were also seen in studies using the s.c. P815 tumor model. Therefore, vaccination with Ag-containing liposomes

engrafted with 9Flg or 42Flg is a powerful strategy to exploit the innate and adaptive immune systems for the development of

potent vaccines and cancer immunotherapies. The Journal of Immunology, 2010, 185: 1744–1754.

TLRs are expressed by APCs, such as monocytes, macro-phages, and dendritic cells (DCs), and play a key role intriggering innate immunity and in initiating adaptive im-

mune responses to pathogens (1–3). Several TLRs function assensors of external stimuli, which can recognize specific microbialproducts and structures, such as pathogen-associated molecular pat-terns, which are conserved in pathogens (1–5). One TLR of specialinterest is TLR5, which recognizes the protein flagellin (6, 7), themajor constituent of bacterial flagella involved in motility. TLR5 isa transmembrane protein ∼100 kDa expressed by a variety of cells,including monocytes, DCs, epithelial cells, and mast cells (8–10).Several studies showed that stimulation of TLR5 by monomericflagellin can activate macrophages, DCs, neutrophils, and intestinal

cells to produce inflammatory mediators, such as TNF-a, IL-6, IL-8, and IL-12, to an extent depending on cell type (6, 7, 10–12). Thebinding of flagellin to TLR5 on APCs in vitro activates the MyD88-dependent intracellular signaling pathway, leading to APC matura-tion and secretion of proinflammatory cytokines and chemokines(6, 13–15). Interestingly, when administered in vivo, flagellin indu-ces a proinflammatory response similar to that of LPS; however,flagellin is much more potent than LPS, inducing responses at con-centrations as low as 1 pg/ml (15, 16).The fact that flagellin is one of the first Ags to which DCs are

exposed during an early stage of bacterial infection suggests that itcould play a crucial role in generating protective immunity (17).Consistent with this, a number of studies showed that flagellin hasadjuvant-like properties (17, 18) Vaccination with recombinantflagellin–Ag fusions induced potent T and B cell responses, in-cluding Ag-specific CD8+ T cell and Ab responses to model Ags(19, 20), as well as to a number of disease-relevant Ags (21–23),including influenza M2e Ag (24), Plasmodium vivax merozoitesurface protein-1 (25), and Mycobacterium tuberculosis Ag p27(26). The responses induced were often much greater than thoseinduced by vaccination of the Ag emulsified in CFA (20–26).Also, the linking of flagellin to Ag is essential to inducedetectable Ab responses, because a simple mixing of the two isnot sufficient (20–26). This underscores the notion that TLRactivation and Ag delivery to the same APC, or indeed to thesame location/compartment at the cell surface or within the cell,may be required to elicit an effective adjuvant response.The ability of Ag–flagellin fusion proteins to induce potent Ag-

specific immunity when used as vaccines suggests that the targetingofAg toTLR5 is an effective strategy to elicit Ag-specific immunity.However, the production of fusion proteins can be cumbersome andexpensive and may lack versatility, especially for the delivery of

Division of Biomedical Science and Biochemistry, Research School of Biology,College of Medicine, Biology and Environment, Australian National University,Canberra, Australia

Received for publication January 5, 2010. Accepted for publication May 29, 2010.

This work was supported by Project Grant Number 316949 from the National Healthand Medical Research Council of Australia (to J.G.A.).

Address correspondence and reprint requests to Dr. J.G. Altin, Division of Biomed-ical Science and Biochemistry, Research School of Biology, College of Medicine,Biology and Environment, Australian National University, Canberra, ACT 0200,Australia. E-mail address: Joseph.Altin@anu.edu.au

The online version of this article contains supplemental material.

Abbreviations used in this paper: AF647-DOPE, Alexa Fluor-647–1,2-palmitoyl-oleoyl-sn-glycero-3-phosphatidylethanolamine; a-galcer, a-galactoceramide; ANU, AustralianNational University; BMDC, bone marrow-derived dendritic cell; DC, dendritic cell; F,average fold increase; MFI, mean fluorescence intensity; NTA3-DTDA, 3(nitrilotriaceticacid)-ditetradecylamine; OG488-DHPE, Oregon Green-488–(1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine); pFlg, flagellin-related peptide; PMV, plasma mem-brane vesicle; POPC, 1,2-palmitoyl-oleoyl-sn-glycero-3-phosphatidylcholine; SpDC,splenic semiadherent cell.

Copyright� 2010 by TheAmericanAssociation of Immunologists, Inc. 0022-1767/10/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1000027

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

multipleAgs simultaneously. In this study,we explored the potentialof using relatively short flagellin-related peptides (pFlgs) to manip-ulate immunity by targeting Ag-containing liposomes and mem-brane vesicles to TLR5 on APCs. Using the chelator lipid 3(nitrilotriacetic acid)-ditetradecylamine (NTA3-DTDA) (27, 28),we engrafted Ag-containing liposomes and tumor cell-derivedplasma membrane vesicles (PMVs) with synthetic His-tagged pep-tides that contain a sequence derived from flagellin and explored theability of these to enhance immunity when used as vaccines. Theresults show that peptides having a sequence related to conservedregions in bacterial flagellin can be used to elicit dramatic Ag-specific and antitumor responses in mouse tumor models.

Materials and MethodsReagents

The phospholipids 1-2-disteroyl-phosphatidylcholine, 1,2-palmitoyl-oleoyl-sn-glycero-3-phosphatidylcholine (POPC), and cholesterol were obtainedfrom Sigma-Aldrich (Castle Hill, New South Wales, Australia). NTA3-DTDA was produced in the Research School of Chemistry (AustralianNational University [ANU]), as described previously (27, 29). Disteroyl-phosphatidylethanolamine (polyethylene glycol)-750 was obtained fromAvanti Polar Lipids (Alabaster, AL). The fluorescent lipid Oregon Green-488–(1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine) (OG488-DHPE)was purchased from Invitrogen (Eugene, OR). 1,2-palmitoyl-oleoyl-sn-glycero-3-phosphatidylethanolamine was reacted with Alexa Fluor-647carboxylic acid succinimidyl ester (Invitrogen) and purified to produceAlexa Fluor-647–1,2-palmitoyl-oleoyl-sn-glycero-3-phosphatidylethanolamine(AF647-DOPE), as previously described (30). Paraformaldehyde was fromBDH Chemicals (Kilsyth, Victoria, Australia). RPMI 1640 (Invitrogen)medium was obtained from the Media Unit, John Curtin School of MedicalResearch (ANU).

Cells

Day-6 bone marrow-derived DCs (BMDCs) from C57BL/6 mice were gen-erated as previously described (31, 32). BMDCswere cultured inRPMI 1640medium containing 10% FCS, 2 mM L-glutamine, 50 mM 2-ME, 0.1mg/ml penicillin-streptomycin-neomycin (referred to as complete medium),plus 10 ng/ml recombinant murine GM-CSF (PeproTech, Rocky Hill, NJ).Splenic semiadherent cells, hereafter referred to as splenic DCs (SpDCs),were prepared from RBC-depleted splenocytes by the method of plasticadherence, as described (32). B16-OVA melanoma [C57BL/6 (H-2b)] (ahighly metastatic OVA-secreting tumor cell line) and CHO cells were pro-vided, respectively, by Dr. Mark Hulett and Prof. Ian Young (Division ofMolecular Bioscience, John Curtin School of Medical Research). Humanembryonic kidney cells HEK-293 and mouse P815 mammary mastocytoma[DBA/2J (H-2d)] cells were provided by Dr. Hilary Warren (The CanberraHospital, Garran, Australian Capital Territory). HepG2 cells (humanhepatocellular carcinoma) were obtained from Prof. Chris Parish (Divisionof Immunology and Genetics, John Curtin School of Medical Research). Allcell lines were cultured (in 37˚C incubator with 5% CO2) in RPMI 1640medium with 10% newborn calf serum (plus 0.5 mg/ml geneticin for B16-OVA cells), except for CHO cells, which were cultured in F15 medium(Invitrogen) with 10% newborn calf serum.

Mice

Female C57BL/6 (H-2b) and DBA/2J mice were obtained from the AnimalResource Facility (Perth,Western Australia, Australia); ChilcottMyD882/2

knockout mice on a C57BL/6 (H-2b) background were obtained fromAnimal Services (ANU). The mice were kept in the ANU Animal Facilityand used at 6–8 wk of age, in accordance with protocols approved by theANU Animal Experimentation Ethics Committee.

Peptides

pFlgs for engraftment onto liposomes contained a 12His-tag (to enable en-graftment onto lipoplexes) and a GS repeated spacer of length indicated (toreduce steric hindrance), followed by an amino acid sequence to the indi-cated region conserved in bacterial FliC flagellin (Table I). A 12xHispeptide was used as a nontargeted (control) peptide, because it is knownto bind strongly to Ni-NTA3-DTDA and to reduce nonspecific binding ofNi-NTA3-DTDA liposomes to cells. All peptides were synthesized andHPLC purified at the Biomolecular Resource Facility, John Curtin Schoolof Medical Research, ANU. Stock solutions of the peptides were prepared

in distilled water, stored at 220˚C, and thawed and vortexed immediatelybefore use.

Preparation of stealth liposomes and plasma membranevesicles

Stealth liposomes (hereafter referred to as liposomes) were prepared asdescribed (28, 32). Briefly, disteroyl-phosphatidylcholine, cholesterol,disteroyl-phosphatidylethanolamine (polyethylene glycol)-750, NTA3-DTDA, and tracer lipid OG488-DHPE dissolved in ethanol were mixed (mo-lar ratio 46.5:46.5:5:1:1), dried under a stream of nitrogen gas, and rehy-drated in 1 ml PBS (2 mM final lipid) containing Ni2+ (60 mM). Liposomeswere produced by sonication (two 25-s bursts at maximum amplitude) ina TOSCO 100W Ultrasonic Disintegrator (Measuring and Scientific, Lon-don, U.K.). ZetaSizer measurements indicated that liposomes prepared inthis way had an average diameter ∼130 nm and bore a near-neutral surfacecharge (z potential ∼25 mV). Where indicated, the tracer lipid AF647-DOPE was used instead of OG488-DHPE. For immunological experiments,the tracer lipid was omitted, and OVA was encapsulated into the NTA3-DTDA liposomes to give ∼200 mg OVA/mg total lipid, after removal ofunencapsulated OVA by gel filtration, as previously described (28, 32).

PMVs from cultured cells (B16-OVA, B16, or P815) were prepared as de-scribed (32). Cells were washed twice with PBS and suspended in ice-coldPBS (13 107 cells/ml) and lysed by brief sonication (two 15-s bursts) at 4˚C.The lysate was centrifuged (1000 3 g) in 1-ml Eppendorf tubes at 4˚C for10 min; the supernatant was removed from the pelleted debris and centri-fuged (15,0003 g) again for 30 min at 4˚C. The supernatant was discarded,and the pellet (enriched in PMVs) was suspended in PBS and resonicated.Typically, 1 3 107 cell-equivalents of PMVs were suspended in 1 ml PBS.Stock PMVs were stored at 220˚C and briefly resonicated before use inexperiments. PMVswere incorporatedwithNi-NTA3-DTDAand tracer lipid(OG488-DHPE) by mixing with Ni-NTA3-DTDA plus POPC in PBS (finalconcentration 53 106 cell-equivalents PMVs/ml containing 12.5mMNTA3-DTDA plus 37.5 mM NiSO4 and 37.5 mM POPC) and sonicating (two20-s bursts at 4˚C) to promote NTA3-DTDA incorporation. ZetaSizer meas-urements indicated that the resultingPMVshad a diameter∼140 nmandborea near-neutral surface charge.

Liposomes and PMVs were engrafted with peptide by mixing the in-dicated His-tagged peptide dissolved in H2O with a suspension of NTA3-DTDA–containing liposomes or PMVs and incubating for 30 min at roomtemperature. The amount of peptide used for engraftment onto liposomesor PMVs was based on a theoretical calculation of the amount of peptiderequired to be equimolar with anchoring moiety Ni-NTA3-DTDA and anempirical titration to give the optimal binding of the liposomes to mouseBMDCs. Where indicated, the same amount of NTA3-DTDA lipid (∼1.5mg) and approximately the same amount of peptide (3–4.5 mg, dependingon the m.w. of the peptide being used) were used for all vaccinationconditions for each experiment.

In vitro binding of liposomes to cells and flow cytometry

To assess the binding of pFlg-engrafted liposomes to cells, the cells(BMDCs, SpDCs, and splenocytes, at 1 3 105, 0.5 3 105, and 2 3 105

cells/well, respectively) were suspended in 50% FCS in PBS (50% FCS-PBS) and then added to pre-engrafted liposomes in a 96-well polypropy-lene V-bottom Serocluster plate (Costar, Cambridge, MA) and incubatedfor 1 h (at 37˚C or on ice, as indicated) with frequent gentle vortexing ofthe plate (32). Where indicated after this incubation, the cells were washedand blocked for 15 min on ice with mouse IgG before incubating witha fluorochrome-conjugated mAb to the indicated cell surface marker. Cellswere then washed thrice in PBS, fixed with 2% paraformaldehyde in PBS,and analyzed for fluorescence by flow cytometry. Cell fluorescence wasquantified using a flow cytometer (BD Biosciences LSRII; BD Bioscien-ces, San Jose, CA). Typically, 10,000 or 30,000 cells were counted; deadcells were excluded from the analysis by gating based on the forwardscatter versus side scatter. Where indicated, quadrant gates for mAb stain-ing and liposome binding were set to contain.98% of all gated live mock-stained cells (as control) and .98% of all gated live cells stained with12His liposomes (as control), respectively. For detection of intracellularIFN-g (see below), $200,000 cells were counted for each condition.

Measurement of cell proliferation

Pooled splenocytes from each group of vaccinated mice were incubated withPBS, soluble OVA (100 mg/ml), or pFlg-OVA liposomes for 3 h at 37˚C.The cells were washed and incubated in a 96-well flat-bottom plate incomplete RPMI 1640 medium supplemented with IL-2 (50 IU/ml) formeasurement of cell proliferation using a standard [3H]thymidine incorpo-ration assay, as described (32).

The Journal of Immunology 1745

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Intracellular staining for IFN-g and measurement of OVA-specific Ab production

Splenocytes (1 3 106 cells) from the spleens of each group of mice werestimulated for 2 h with 1 mg/ml SIINFEKL; a 1:1000 dilution of BDGolgiStop (BD Biosciences) was added, and the cells were incubated foran additional 3 h before staining with anti–CD8-PE (clone 53-6.7; BDPharmingen, San Diego, CA) and intracellularly with anti–IFN-g-AlexaFluor-647 (clone XMG1.2; BD Pharmingen), as described (32). Flow cyto-metric data were acquired with a BD Biosciences LSR II and analyzedusing FlowJo software (Tree Star, Ashland, OR). CD8+ cells were gatedfrom live cells to give the IFN-g+ CD8+ cells in a two-color dot plot; foreach condition, the background was determined by subtracting the valueobtained by adding PBS instead of peptide.

Ab production in response to vaccinations was analyzed in the sera ofmice by indirect sandwich ELISA, exactly as described (32). Data areexpressed as Ab concentration calculated from standard curves generatedwith OVA-specific IgG1, 1gG2a, and IgG2b mAb standards (BD Pharmin-gen) using KC4 V3.0 Analysis Software.

Vaccination of mice, tumor challenge, and measurement oftumor size

For vaccination, typically 200 ml the indicated vaccine preparation wasadministered i.v. to mice by tail vein injection. The preparations of OVA-bearing liposomes contained ∼10 mg OVA and 50 mg total lipid; vacci-nations with PMVs contained 3 3 105 cell-equivalents of the indicatedPMV preparation. For the tumor-challenge experiments, mice were firstinjected i.v. or s.c. with the indicated tumor cells (2–5 3 105 B16-OVA orP815 cells) in RPMI 1640medium (without serum), and the mice werevaccinated as indicated. Lung metastases of B16-OVA tumor were quan-tified by sacrificing the mice (at day 21 after challenge), removing thelungs, and counting the number of tumor foci in the lungs visually undera dissecting microscope. The size of s.c. tumors was determined from themean of two perpendicular diameters measured using calipers.

Statistical analyses

Data are represented as mean 6 SEM. The Student t test was used forstatistical analysis, and p , 0.05 was considered statistically significant.

ResultsDesign of pFlgs for targeting liposomes to APCs

A comparison of the amino acid sequences of flagellins fromdifferent strains of bacteria, facilitated by using the National Centerfor Biotechnology Information’s Conserved Domain Protein Da-tabase, reveals the existence of at least two regions of high ho-mology. The representative flagellin sequence used in this study isthe 494-aa mature flagellin FliC from Salmonella typhimurium(GenBank accession no. D13689; hereafter referred to as flagel-lin). The high-homology regions are defined by aa 79–117 (in Nregion) and aa 420–475 (in C region) (33, 34). The three-dimensional structure, coupled with mutagenesis studies of flagel-lin, showed that the ability of flagellin to interact with TLR5resides in the first 117 aa of the N-terminal region, particularlythe region defined by aa 79–117, but it may require contributionsfrom C-region aa 408–439 (33–38). Peptides spanning differentregions conserved in flagellin and/or that have been implicated in

binding to TLR5 can be designed (Table I), produced syntheti-cally, and examined for their ability to interact with APCs.pFlgs covering aa 79–117 of FliC in the conserved N-region

may contain epitopes that can bind/stimulate TLR5 (34–39).Therefore, two peptides overlapping by 4 aa were designed anddenoted 41Flg (79–101) and 42Flg (98–117); two peptidesdenoted 9Flg (85–111) and 45Flg (98–111) were designed tocover, respectively, the central and N- parts of the conserved N-region (Table I). Two other peptides covering the conserved C-region of flagellin were also produced: 10Flg (405–425) and 11Flg(408–438) (Table I). The His-tag for these was designed to beon the C-terminal end of the peptide sequence to avoid possiblesteric hindrance, as judged by the orientation of this region in thethree-dimensional structure of flagellin and the predicted site ofinteraction with TLR5 (39).

pFlgs promote binding of liposomes to DCs

The ability of pFlg to promote binding of liposomes to APCs wasexplored by incubating (1 h, 37˚C) tracer-containing NTA3-DTDAliposomes engrafted with 12His peptide (as control) or a pFlg pep-tide with day-6 BMDCs before washing, staining the cells withAPC-conjugated CD11c mAb, and analysis by flow cytometry.The results indicated that BMDCs incubated with 12His liposomesexhibited a mean fluorescence intensity (MFI) ∼2-fold above back-ground (data not shown). Preliminary experiments also indicatedthat 5–15%of all BMDCs bound9Flg and 42Flg liposomes (data notshown). Additional studies showed that although 1.00 6 0.05% ofBMDCs that bound 12His liposomes were double positives (boundliposomes and were CD11c+), 9.6 6 1.0% and 6.4 6 0.6% ofBMDCs bound 9Flg and 42Flg liposomes, respectively, and weredouble positives (Fig. 1A). The MFI of the CD11c+ BMDCs wasincreased from a standardized value of 1 (for 12His liposomes) to9.3 6 1.0-fold and 4.0 6 0.5-fold greater for cells incubated with9Flg and 42Flg liposomes, respectively (Fig. 1A).Interestingly, a similar pattern of fluorescence was observed in

binding assays using SpDCs instead of BMDCs (Fig. 1B). Thus,incubation of SpDCs with liposomes caused the proportion ofdouble-positive SpDCs (CD11c+ SpDCs that bound liposomes)to increase from 1% (for 12His liposomes) to 5.6 6 2.2% and2.6 6 0.6% for 9Flg and 42Flg liposomes, respectively; and theMFI of the CD11c+ cells increased 6.2 6 1.4-fold and 3.2 6 0.2-fold, respectively (Fig. 1B). As shown in Fig. 1C, 9Flg and 42Flgliposomes also bound to CD11b+ SpDCs. The proportion of dou-ble-positive cells increased from 1% (12His liposomes) to 10 62.2% and 6.6 6 1.4% for 9Flg and 42Flg liposomes, respectively;the MFI of the CD11b+ cells increased 9.5 6 2.7-fold and 4.1 61.3-fold, respectively (Fig. 1C). Under these conditions, 10Flgand 11Flg liposomes bound weakly and 41Flg and 45Flg liposomesbound very weakly or not at all to BMDCs and SpDCs (Supple-mental Fig. 1), indicating that 9Flg and 42Flg are the most effectiveat promoting binding of liposomes to APCs.

Table I. Name and amino acid sequence of pFlgs designed for engraftment onto liposomes/PMVs

Peptide Name Amino Acid Sequence

9Flg-85-111 HHHHHHHHHHHH-GSGSG-INNNLQRVRELAVQSANSTNSQSDLDS10Flg-405-425 TENPLQKIDAALAQVDTLRSD-GSGSG-HHHHHHHHHHHH11Flg-408-438 PLQKIDAALAQVDTLRSDLGAVQNRFNSAIT-GSGSG-HHHHHHHHHHHH41Flg-79-101 HHHHHHHHHHHH-GSGSGSG-ALNEINNNLQRVRELAVQSANS42Flg-98-117 HHHHHHHHHHHH-GSGSGSG-SANSTNSQSDLDSIQAEITQ45Flg-98-111 HHHHHHHHHHHH-GSGSGSGSGSGSGSGSGS-QSANSTNSQSDLDS

Each pFlg contains a 12His-tag (to enable engraftment onto NTA3-DTDA-liposomes), a spacer region (to reduce sterichindrance), and an amino acid sequence (bold text) derived from the indicated conserved region of FliC flagellin.

1746 LIPOSOMAL Ag WITH FLAGELLIN PEPTIDES ELICITS ANTITUMOR EFFECTS

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The cell-binding specificity of pFlg liposomes was examined byincubating the liposomeswith splenocytes.Preliminary experimentsindicated that ∼1% of splenocytes bound 12His liposomes, but 7–10% of splenocytes bound 9Flg and 42Flg liposomes (data notshown). Further studies showed that although 0.50 6 0.05% ofsplenocytes were double positives (i.e., bound liposomes and wereCD11c+) following incubation with 12His liposomes, the propor-tion of double positives after incubation with 9Flg and 42Flg lip-osomes increased to 3.06 0.2% and 2.56 0.1%, respectively; theMFI of the CD11c+ cells increased 9 6 1-fold and 8.3 6 1.1-fold,respectively (Fig. 2A). Similarly, the double-positive cells (lipo-some positive and CD11b+) increased from 0.5% (for control cells)to 2.66 0.5% and 1.36 0.3% for splenocytes incubated with 9Flgand 42Flg liposomes, respectively; the MFI of the CD11b+ cellsincreased 3.8 6 0.4-fold and 3.1 6 0.7-fold, respectively (Fig.2A). Also, the proportion of double-positive cells (liposome positiveand CD19+) increased from 0.5% (for control cells) to 3.66 0.6%and 1.3 6 0.1% for splenocytes incubated with 9Flg and 42Flgliposomes, respectively; the MFI of the CD19+ cells increased

4.86 0.5-fold and 2.56 0.1-fold, respectively (Fig. 2A). The bind-ing of liposomes engrafted with 10Flg, 11Flg, 41Flg, and 45Flg toCD11c+, CD11b+, and CD19+ splenocytes was generally muchlower than for liposomes engrafted with 9Flg and 42Flg (Supple-mental Fig. 2). There was no significant binding of 9Flg and 42Flgliposomes to CD3+ cells (data not shown).Additionalstudies, inwhich9Flgliposomeswereadministered i.v.

tomice, and the spleenswere removed 1 h later for analysis of tracer-containing splenocytes, showed that 9Flg liposomes had bound to1.7 6 0.2% of the splenocytes subsequently isolated from thespleen; the liposome-associated fluorescence of the CD11c+ cellswas increased 8.76 1.9-fold (Fig. 2B). Interestingly, the liposome+

FIGURE 1. Liposomes engrafted with pFlgs bind murine BMDCs and

splenic DCs. Day-6 BMDCs from C57BL/6 mice were suspended in 50%

FCS-PBS and incubated (1 h, 37˚C) with tracer-containing liposomes

engrafted with peptide 12His (control), 9Flg, or 42Flg. Cells were washed,

blocked with mouse IgG, and stained with APC-conjugated mAb to

CD11c, before analysis by flow cytometry. A, The two-color dot plots

show fluorescence of the cells (OG488-liposomes versus APC-mAb) due

to binding of 12His-, 9Flg-, and 42Flg-engrafted liposomes, as indicated,

followed by staining with APC-mAb to CD11c. B, Two-color dot plots of

similar experiments carried out using murine SpDCs instead of BMDCs.

C, The SpDCs were stained with FITC-conjugated mAb to CD11b (instead

of CD11c) (AF647-liposomes versus FITC-mAb to CD11b). Quadrant gates

were set as detailed in Materials and Methods. For each dot plot, the mean

percentage of gated live cells (BMDCs or SpDCs) that bind engrafted

liposomes and that stain positive for the indicated mAb (i.e., double-

positive cells) is shown. Also shown is the F in MFI of the cells incubated

with liposomes engrafted with the indicated peptide and that stained pos-

itive for the indicated mAb, relative to the respective control cells incu-

bated with 12His liposomes (and defined as having F = 1). Each result

represents the mean 6 SEM obtained from three to five independent

experiments. pp , 0.05; relative to the control cells (12His-liposomes)

for the indicated experiment. F, average fold increase.

FIGURE 2. pFlg liposome splenocytes in vitro and in vivo. Splenocytes

prepared from the spleens of C57BL/6 mice were incubated (1 h, 37˚C) in

50% FCS-PBS, with tracer (OG488-DHPE or AF647-DOPE)-containing

liposomes engrafted with peptide 12His (control), 9Flg, or 42Flg. The cells

were washed, blocked with mouse IgG, and stained with APC-mAb

(CD11c), FITC-mAb (CD11b), or PE-mAb (CD19) before analysis by flow

cytometry. A, The two-color dot plots show APC-mAb (CD11c) versus

OG488 (liposomes), FITC-mAb (CD11b) versus AF647 (liposomes), and

PE-mAb (CD19) versus AF647 (liposomes). B, Mice were injected i.v. with

12His liposomes or 9Flg liposomes containing OG488-DHPE as tracer.

After 1 h, the spleens were removed, and RBC-depleted splenocytes were

prepared, blocked with IgG, stained with APC-mAb (CD11c), and ana-

lyzed by flow cytometry. The two-color dot plots show the fluorescence of

APCs (CD11c mAb) versus OG488 (liposomes). Quadrant gates were set as

detailed in the Materials and Methods. For each dot plot, the mean per-

centage of gated live splenocytes that bound engrafted liposomes and

stained positive for the indicated mAb (i.e., double-positive cells) is

shown. Also shown is the average fold-increase (F) in MFI of the spleno-

cytes incubated with liposomes engrafted with the indicated peptide and

that stained positive for the indicated mAb, relative to the respective con-

trol splenocytes incubated with 12His liposomes (and defined as having

F = 1). Each result represents the mean 6 SEM obtained from three to five

independent experiments. pp , 0.05; relative to the control cells (12His

liposomes) for the indicated experiment.

The Journal of Immunology 1747

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

splenocytes under these conditions accounted for ∼20% of thesplenic CD11c+ cell population, indicating that 9Flg liposomes alsocan bind selectively to these APCs in vivo (Fig. 2B).

pFlg liposomes bind specifically to TLR5-expressing cells andare internalized

To confirm that binding of pFlg liposomes to cells occurs throughinteraction with TLR5, cells were doubly stained with pFlg lip-osomes and TLR5mAb.Many of the commercially available TLR5Abs bind to a cytoplasmic region of TLR5 and, therefore, wereunsuitable for correlating the binding of pFlg liposomes withsurface TLR5 expression. Intracellular TLR5 mAb staining alsocannot be used for the correlation because TLR5 is expressed in-tracellularly and at the cell surface (40, 41). The TLR5 mAb Clone19D759.2 (Mouse IgG2a mAb; Imgenex, San Diego, CA) wasparticularly useful for cell-surface staining of mouse and humanTLR5. This mAb stained BMDCs weakly, and it was more effectiveat staining cells that expressed higher levels of TLR5. Thus, theTLR5 mAb bound HEK-293 cells (widely reported to constitu-tively express TLR5), and it bound particularly well to the humanhepatoma cell line HepG2 (42). Consistent with reports that CHOcells do not express TLR5 (43), we found no significant binding ofTLR5 mAb or pFlg liposomes to CHO cells. Experiments in whichCHO cells were transfected to express mTLR5 also were carried

out, but only very low surface expression of TLR5 expression wasachieved, which did not allow any meaningful comparisons to bemade with these cells (data not shown). Therefore, we conductedbinding experiments with BMDCs and the cell lines HEK-293 andHepG2 that react with the TLR5 mAb.Preliminary experiments indicated that the binding of liposomes

engrafted with the less-active peptides (i.e., 10Flg, 11Flg, 41Flg,and 45Flg) to BMDCs and HEK-293 and HepG2 cells was toolow to permit a reliable correlation of the binding of the engraftedliposomes with the binding of the TLR5 mAb (data not shown).Therefore, experiments were performed to correlate the bindingof 9Flg liposomes with that of TLR5 mAb to these cells. Thetwo-color dot plots in Fig. 3A show that binding of the 9Flg lip-osomes correlates well with staining of TLR5 mAb in each of thethree cell types. This correlation is strongest for HepG2 cells,which exhibit the strongest binding of TLR5 mAb (Fig. 3A, lowerright panel). Other experiments showed a similar binding patternwhen using 42Flg liposomes with HepG2 cells, but the level ofbinding was lower (Fig. 3B). Also, we observed only a smalldifference in the binding of the TLR5 mAb to cells that had beenincubated or not with 9Flg liposomes, indicating no substantialinterference between the binding of the 9Flg liposomes and bind-ing of the TLR5 mAb used (data not shown). The mean percent-age of double-positive cells (AF647

+AF488+; i.e., cells that bound

FIGURE 3. pFlg liposomes bind to TLR5-expressing cells. Murine BMDCs and the cell lines HEK-293 and HepG2 suspended in 50% FCS-PBS were

separately incubated (40 min, 4˚C) with 9Flg liposomes containing tracer AF647-DOPE. The cells were washed twice with PBS and stained for 40 min with

AF488-conjugated mAb to TLR5 in 2% FCS-PBS before analysis by flow cytometry. A, Representative two-color dot plots for the fluorescence due to binding

of AF488-mAb (TLR5) versus 9Flg liposomes (AF647) for each cell type: BMDC, HEK-293, and HepG2. B, Representative dot plots showing the binding of

42Flg liposomes (instead of 9Flg liposomes) to HepG2 cells. Quadrant gates were set as detailed inMaterials and Methods. For each dot plot, the proportion

of cells, as a percentage of live cells, is shown in the indicated quadrant. Also, the MFI of the gated live cell population (BMDCs, HEK-293, or HepG2 cells),

due to the binding of engrafted liposomes (AF647, x-axis) and staining with TLR5 mAb (AF488, y-axis), is shown as the X and Y values, respectively. Each

result in the dot plots represents the mean6 SEM obtained from four to seven independent experiments. C, The bar graph shows the mean percentage6 SEM

of double-positive cells (AF647+AF488

+; i.e., percentage of cells that bound pFlg liposomes and TLR5 mAb) relative to the mean of all AF647+ cells (i.e., the

mean of all cells that bound pFlg liposomes), for conditions in A and B in which the cells were incubated with pFlg liposomes alone, or with pFlg liposomes

then with TLR5 mAb, as indicated. D, BMDC-derived DC2.4 cells were incubated with 9Flg liposomes containing OG488-DHPE as tracer for 2 h at 4˚C or

37˚C. The cells were washed, fixed, and suspended in embedding medium for analysis of OG488 fluorescence by confocal fluorescence microscopy;

a representative image set obtained from three separate experiments is shown (original magnification 31500). pp , 0.05; ppp , 0.02; pppp , 0.01.

1748 LIPOSOMAL Ag WITH FLAGELLIN PEPTIDES ELICITS ANTITUMOR EFFECTS

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

pFlg liposomes and TLR5 mAb), relative to the mean of allAF647

+ cells (i.e., cells that bound pFlg liposomes only), for con-ditions in which the cells were incubated with engrafted liposomesalone or with engrafted liposomes followed by TLR5 mAb, isshown in Fig. 3C. These data show that 9Flg and 42Flg liposomesinteract with TLR5-expressing cells and that binding correlateswith surface expression of TLR5.Other experiments in which BMDCs with bound 9Flg liposomes

were incubated at 37˚C, before an examination of the cells byfluorescence confocal microscopy, showed that cell-bound 9Flgliposomes were actively internalized. Thus, the OG488 tracerwas distributed relatively uniformly at the cell surface after bind-ing 9Flg liposomes at 4˚C (Fig. 3D, left panel), but a 3-h incubationat 37˚C resulted in the majority of the OG488-fluorescence beingintracellular (Fig. 3D, right panel), demonstrating that the 9Flgliposomes are internalized.

Treatment with pFlg liposomes induces DC maturation in vitroand in vivo

The ability of 9Flg liposomes to induce DC maturation was firstassessed using BMDCs. Day-6 BMDCs from C57BL/6 mice wereincubatedwith completemedium, containing added PBS, 12His lip-osomes, 9Flg liposomes, or LPS (10 mg), before washing and cul-turing the cells for 24 h to assess surface expression of the activationmarkers MHC class II and CD86, by flow cytometry. The resultsshowed that the incubation with 9Flg liposomes induced a signifi-cant (60–80%) increase in the relative expression ofMHC class II orCD86, relative to controls cells treatedwith PBSor 12His liposomes(Supplemental Fig. 3A). Analogous studies using BMDCs fromMyD882/2 mice showed no significant effect of 9Flg liposomes(Supplemental Fig. 3B). LPS, a known potent inducer of DC matu-ration, induced significant increases in expression of MHC class IIand CD86 in these cells (Supplemental Fig. 3).The effect of 9Flg liposomes on thematuration ofDCs invivowas

examined by injecting naiveC57BL/6micewith PBS (control), LPS(10 mg in PBS, as positive control), or liposomes engrafted withpeptide 12His or 9Flg and then comparing the level of expression ofthe activation markers CD86 and MHC class II on CD11c+ sple-nocytes after 48 h. Splenocytes from mice for each condition weredoubly stained with a mAb to CD11c and a mAb to CD86 or MHCclass II. As shown in Fig. 4A, the relative expression of CD86 andMHC class II in splenocytes from 12His liposome-injected micewas similar to that seen in the PBS controls (relative expression∼1). As expected, CD11c+ splenocytes from mice treated withLPS exhibited a significant (∼2-fold) increase in the expression ofCD86 and MHC class II (Fig. 4A). Notably, CD11c+ splenocytesfrom mice injected with 9Flg liposomes exhibited ∼2-fold increasein CD86 andMHC class II expression (Fig. 4A). Interestingly, how-ever, treatment with 9Flg liposomes failed to significantly increasethe expression of CD86 and MHC class II in CD11c+ splenocytesfrom MyD882/2 mice (Fig. 4B).

Stimulation of functional Ag presentation and Ag-specificimmunity

Induction of Ag-specific T cell priming. We used liposomes con-taining themodelAgOVAto testwhetherpFlg liposomescanbeusedto enhance Ag-specific responses. Different groups of five naiveC57BL/6micewerevaccinatedondays0,7,and14withPBS,solubleOVA, or OVA-containing liposomes engrafted with 12His (as con-trol), 9Flg, or 42Flg peptide. At day 21, spleens were isolated, andtotal splenocytes prepared from each group were pooled and incu-bated in vitro with soluble OVA to assess cell proliferation by mea-suring [3H]thymidine incorporation. Control mice, immunized withPBS, OVA alone, or 12His-OVA liposomes exhibited little, if any,

proliferative response (Fig. 5A). In contrast, compared with theirrespective controls, splenocytes from mice vaccinated with 9Flg-and 42Flg-engrafted OVA liposomes exhibited ∼10- and ∼20-foldincreases in cell proliferation, respectively (Fig. 5A).

Generation of Ag-responsive CD8+ T cells. The number of IFN-g–producing CD8+ T cells from splenocytes of mice vaccinated withpFlg-OVA liposomes (as above) was determined by intracellularstaining for INF-g in response to a 3-h stimulation with 1 mΜSIINFEKL peptide in the presence of monensin. Two-color FACSanalysis of splenocytes from mice vaccinated with PBS, OVAalone, or 12His-, 10Flg-, 11Flg-, 41Flg-, or 45Flg-OVA liposomesshowed no significant IFN-g production by CD8+ T cells (Fig. 5B).Importantly, splenocytes frommicevaccinatedwithOVA liposomesengrafted with 9Flg and 42Flg peptide induced a substantiallyhigher proportion (i.e., 1.0 6 0.1% and 0.7 6 0.2%, respectively)of SIINFEKL-responsive IFN-g–producing CD8+ T cells (Fig. 5B),indicating the induction of a modest CD8+ T cell response to theliposome-associated OVA Ag. Interestingly, ∼50% of the SIIN-FEKL-responsive IFN-g–producing splenocytes from mice vacci-nated with 9Flg liposomes (Fig. 5C) and 42Flg liposomes (data notshown) were CD82.

Induction of Ag-specific Ab. Mice vaccinated with pFlg-OVA lip-osomes also were tested for Ab production. Blood samples from themice vaccinated at day 21 were collected, and sera from the mice ineach group were pooled to assess the production of OVA-specificIgG1, IgG2a, and IgG2b by ELISA. No OVA-specific Ab wasdetected in sera from mice vaccinated with 12His-OVA liposomes(controls), and only low titers of IgG2a (corresponding to IgG2c

in C57BL/6 mice, data not shown) could be detected in serafrom mice vaccinated with 9Flg- or 42Flg-OVA liposomes.

FIGURE 4. Vaccination with pFlg liposomes activates DCs in vivo.

Groups of three mice (C57BL/6) were vaccinated i.v. with PBS alone (as

control) orwithPBScontaining12His liposomes, 9Flg liposomes, orLPS (10

mg). After 48 h, the spleens were removed, and splenocytes were prepared,

blocked with mouse IgG, and stained with allophycocyanin-labeled mAb to

CD11c and FITC-labeled mAb to MHC-II or CD86. A, The relative expres-

sion of each activation marker was expressed as a fold increase: the MFI (of

CD11c+ cells) due to staining for MHC class II or CD86 (for the indicated

condition) was divided by theMFI due to staining forMHC class II or CD86,

respectively, for CD11c+ splenocytes from control (PBS injected) mice. B,

The results for a similar experiment using MyD882/2 C57BL/6 mice. Each

result represents the mean of three independent experiments. pp, 0.002.

The Journal of Immunology 1749

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Interestingly, however, sera from mice vaccinated with 42Flg-OVA liposomes contained a high titer of IgG1 and a moderatelyhigh titer of IgG2b; sera from 9Flg-OVA liposome-vaccinated micehad low/moderate titers of IgG1 and IgG2b (Fig. 5C).

Effect of pFlg-PMV vaccination on tumor growth in tumormodels

B16-OVA tumor growth and metastasis. The potential for pFlgliposomes to be useful as vaccines also can be assessed by theirability to induce antitumor effects. Groups of five C57BL/6 micewere inoculated with B16-OVA cells (2 3 105 cells i.v. at day 0);

vaccinated i.v. on days 2, 8, and 14 by PBS or OVA liposomesengrafted with 12His, 9Flg, 10Flg, 11Flg, or 42Flg; and sacrificedat day 21 to assess the response. As shown in Fig. 6A, the lungs ofcontrol mice vaccinated with PBS and 12His-OVA liposomes con-tained an average of ∼120 tumor foci, whereas the lungs of micevaccinated with OVA liposomes engrafted with 9Flg, 10Flg,11Flg, and 42Flg contained an average of 1 6 1, 2 6 1, 7 6 5,and 2 6 1 tumor foci, respectively.

The antitumor effects of vaccinating with liposomal Agengrafted with pFlg were also explored using tumor cell-derivedPMVs that contain an array of Ags associated with the tumor

FIGURE 5. Vaccination with pFlg liposomes primes T cells in vivo. Groups of five naive C57/BL6mice were vaccinated i.v. on days 0, 7, and 14 with PBS,

solubleOVA, orOVA liposomes engraftedwith peptide 12His, 9Flg, or 42Flg.On day 21, spleenswere removed, and splenocytes from each group ofmicewere

pooled and assessed for induction of OVA-specific T cells. Pooled splenocytes (53 105) from each condition were seeded in a 96-well flat-bottom plate and

incubated (3 h, 37˚C) in complete RPMI 1640medium in the presence of 100mg/mlOVA. The splenocytes werewashed and cultured in completemediumwith

50 IU/ml IL-2 for 72 h before assessing cell proliferation bymeasurement of [3H]thymidine incorporation. A, Mean CPM6 SEM of quadruplicate cultures for

each condition.B, Pooled splenocytes fromvaccinatedmicewere also used to assess thegenerationofCD8+Tcells that produce IFN-g in response to peptideAg

(SIINFEKL). Pooled splenocytes from mice injected once with a mixture of a-galcer (1 mg) and SIINFEKL (0.1 mg) were used as a positive control. The

splenocytes (13 106)were incubated for 5 hwithmonensin in the absence or presence of SIINFEKL (1mg/ml) beforewashing and stainingwith PE-conjugated

mAb to CD8. The cells were fixed, permeabilized, and stained intracellularly with allophycocyanin-labeled mAb to IFN-g before analysis by flow cytometry.

The percentage of CD8+ cells producing IFN-g for each condition was determined from an analysis of the two-color dot plots of the stained cells and by

subtracting the background IFN-g staining (splenocytes stimulated with PBS) from the IFN-g staining induced after incubating splenocytes for each condition

with SIINFEKL. Each bar represents the mean 6 SEM of three independent experiments, except for the results for 41Flg and 45Flg, which are from one

experiment. C, Representative two-color dot plots showing IFN-g staining of CD8+ and CD82 splenocytes from mice treated with a-galcer or liposomes

engrafted with 12His or 9Flg, followed by SIINFEKL stimulation. D, Equal volumes of blood sera from each group of mice were pooled, and serum titers of

OVA-specific IgG1 and IgG2b were measured by indirect ELISA. Each datum point represents the mean Ab concentration6 SD of two independent measure-

ments performed in triplicate at the serum dilution indicated. pp, 0.001; ppp, 0.005. a-galcer, a-galactoceramide.

1750 LIPOSOMAL Ag WITH FLAGELLIN PEPTIDES ELICITS ANTITUMOR EFFECTS

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

and can potentially be used to enhance antitumor immunity (28,32). We prepared PMVs from cultured B16-OVA cells and mod-ified these by incorporating NTA3-DTDA, followed by engraft-ment of pFlg. Binding assays analogous to those in Fig. 1

revealed that the engraftment of pFlgs (particularly 9Flg and42Flg) also mediated the binding of tracer-containing modifiedPMVs to BMDCs (Supplemental Fig. 4). In experiments analo-gous to those above, groups of five mice (C57BL/6) were inocu-lated with B16-OVA cells (2 3 105 cells i.v. at day 0); vaccinatedat days 2, 8, and 14 by i.v. injection of PBS or PMVs engraftedwith 12His, 9Flg, 10Flg, 11Flg, or 42Flg; and sacrificed at day 21to assess the response. As shown in Fig. 6B, the lungs of thecontrol mice vaccinated with PBS and 12His PMVs containedan average of ∼120 tumor foci. In contrast, the lungs of micevaccinated with 9Flg PMVs or 42Flg PMVs contained an averageof 2 6 2 and 4 6 4 tumor foci, respectively, whereas lungs frommice vaccinated with 10Flg PMVs and 11Flg PMVs contained 436 15 and 60 6 12 tumor foci, respectively (Fig. 6B); representa-tive images of the lungs are shown in Fig. 6C.

Effect on established B16-OVA tumors. Experiments also werecarriedoutinmicebearingestablisheds.c.B16-OVAtumors.Micewereinjected s.c. on the backwith 43105B16-OVAcells on day0; after 9dwhen tumors were palpable (∼2 mm in diameter), different groups offivemicewere vaccinated i.v. each week for 5 consecutiveweeks with

FIGURE 6. Vaccination with 9Flg PMVs inhibits tumor growth and

metastasis. C57BL/6 mice were inoculated with 2 3 105 B16-OVA cells

injected i.v. (day 0). Different groups of five mice were vaccinated at days

2, 8, and 14 with OVA liposomes engrafted with peptide 12His, 9Flg,

10Flg, 11Flg, or 42Flg. A, At day 21, the lungs were removed, and the

number of tumor foci was counted. B, C, Analogous experiments in which

C57BL/6 mice were inoculated i.v. with 2 3 105 B16-OVA cells and then

vaccinated at days 2, 8, and 14 with B16-OVA–derived PMVs (3 3 105

cell-equivalents) engrafted with peptide 12His, 9Flg, 10Flg, 11Flg, or

42Flg were also conducted. B, Mean number of tumor foci 6 SEM in

the lungs for each group of mice at day 21. C, Representative images of the

lungs from each group of mice.

FIGURE 7. Vaccination with 9Flg PMVs inhibits the growth of estab-

lished tumors and prolongs survival. A, Subcutaneous B16-OVA tumors

were established by s.c. injection of 4 3 105 B16-OVA cells in the back of

the mice (day 0). At day 9, when tumors were ∼2 mm diameter, separate

groups of five tumor-bearing mice were vaccinated with B16-OVA–derived

PMVs engrafted with 12His, 9Flg, or 42Flg. Tumor size was monitored

every 3 d; the average tumor diameter 6 SEM for each group of mice is

shown pp , 0.0001; 9Flg PMV versus 12His PMV. B, Subcutaneous P815

tumors were established in syngeneic DBA/2J mice by injecting 4 3 105

P815 cells in the back of the mice (day 0). On day 15, when the tumors

reached ∼5 mm diameter, groups of mice were vaccinated with PBS (six

mice) or 12His- or 9Flg-engrafted P815 tumor cell-derived PMVs (seven

mice in each group). The vaccinations were administered again every 5 d

for a total of three consecutive vaccinations. The plot shows the survival

times of the mice in each of the treatment groups.

The Journal of Immunology 1751

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

PBSor 12His, 9Flg, or 42FlgPMVs.As shown inFig. 7A, controlmicevaccinatedwithPBSand12HisPMVsexhibitedaprogressive increasein tumor size, and all mice in these groups were sacrificed by day 30(when tumors reached 15 mm diameter). Surprisingly, however, micevaccinated with 9Flg PMVs and 42Flg PMVs exhibited markedabrogation of tumor growth and a dramatic regression of their tumors(Fig. 7A). Thus, all of the mice vaccinated with 9Flg PMVs, as well as4/5 (i.e., 80%) of the 42Flg PMV-vaccinated mice, had no palpabletumor by day 30; furthermore, these mice showed no signs of tumorand were apparently tumor-free 4 mo after tumor inoculation.

Effect on established P815 tumors. We also examined the effect of9Flg PMV vaccination on s.c. P815 tumors. DBA/2J mice werechallenged s.c. on the back with 0.5 3 106 P815 tumor cells (day0); when tumors reached an average of 5 mm in diameter (day 15),different groups of mice were vaccinated i.v. with syngeneic P815cell-derived PMVs engrafted with 12His (control) or 9Flg peptide.Vaccinations were repeated every 5 d, and tumor size/survival wasmonitored. The results show that control mice exhibited rapid andprogressive tumor growth, with all mice being euthanized by day22 because of tumor growth (tumors .15 mm diameter). In con-trast, 9Flg PMV-vaccinated mice exhibited a markedly slower rateof tumor growth and prolonged survival: 50% of these mice sur-vived at day 22, with some mice surviving to day 25 (Fig. 7B).

DiscussionThis study used a novel approach using stealth liposomes engraftedwith pFlgs to promote delivery of liposome-encapsulated Ag toAPCs in vivo. Previous structure-function studies using truncatedand/or mutated forms of flagellin indicate that the proinflammatoryproperties of flagellin are localized in the conserved N- and C- ter-minal regions (33–37, 44), with conflicting data claiming an essen-tial role for the central hypervariable domain (36). Although notpreviously reported, short peptides making up the regions of flagel-lin implicated in the interaction with TLR5 may also bind TLR5themselves; however, short peptides generally lack significantsecondary/tertiary structure, suggesting that any such interactionis likely to be of low affinity, potentially limiting their use in vaccinedevelopment. However, peptides engrafted onto liposomes promotemultimeric interactions between the liposome-engrafted peptidemolecules with receptors on the apposing cell (45). This attributeprovides a means to reveal and use weakly interacting peptides—anapproach exploited in the present work. In this way, we showed thatthe adjuvant properties of flagellin could be conveniently harnessedto target liposome-associated Ag to APCs, as well as to inducepotent Ag-specific and antitumor immunity.We produced several synthetic His-tagged peptides containing

a sequence of 20–30 aa corresponding to different conservedregions in flagellin considered to be important for interaction withTLR5 (Table I). Of the peptides tested, 9Flg and, to a lesser extent,42Flg promoted substantial binding of NTA3-DTDA liposomes toCD11c+ BMDCs and SpDCs (Fig. 1, Supplemental Fig. 1). Two-color FACS analysis of cells indicates that 9Flg-liposome bindingcorrelates with TLR5 expression (Fig. 3A), with little or no bind-ing to cells lacking the expression of TLR5. Importantly, 9Flgliposomes bound CD11c+ DCs in vitro and in vivo (Fig. 2A,2B), and cell-bound 9Flg liposomes were efficiently internalizedby cultured DCs (Fig. 3B). This indicates that after interactingwith TLR5 on APCs, 9Flg liposomes can be internalized, permit-ting intracellular Ag processing and presentation.Recent evidence suggests that a 14-aa a-helical motif within

the N-terminal domain (aa 95–108) and a conserved C-terminalmotif (around aa 408–438) of flagellin are essential for full acti-vation of TLR5, with the N-region of flagellin (aa 88–97) havingbeen predicted to physically interact with aa 555–559 of TLR5

(33–37, 39). Our finding that liposomes engrafted with 9Flg and42Flg (which contain aa 85–111 and aa 98–112 of flagellin, re-spectively) interact with TLR5-expressing cells, but that lipo-somes engrafted with 10Flg and 11Flg (which contain aa 405–425 and aa 408–438 of flagellin, respectively) exhibit only littlebinding, underscores the importance of the N-terminal aa 85–111region and is consistent with the C-terminal region not playinga major role in the binding of flagellin to TLR5 (15, 17, 33, 39).Although peptides 9Flg (containing aa 85–111 of flagellin) and

45Flg (containing aa 98–111 of flagellin) share considerableoverlap, our results show that compared with 45Flg, peptide 9Flgis considerably more potent at promoting the binding of liposomesto TLR5+ cells and at inducing Ag-specific immunity. The reasonsfor this are unclear. One possible explanation is that 9Flg (and not45Flg) contains the amino acids crucial for interaction with TLR5.Interestingly, however, a 14-aa region defined by aa 95–108 offlagellin is reported to form a hairpin loop, and deletions or site-specific mutations within this region can apparently abolish theproinflammatory activity of flagellin (35). Because 45Flg containsaa 98–111 of flagellin, but 9Flg also contains amino acids pur-ported to be nonessential for functional TLR5 activation (35),another possible explanation for the activity of 9Flg is that, whenanchored onto NTA3-DTDA liposomes, the purported nonessentialregion (aa 85–95) could contribute to overcoming steric con-straints for the interaction between liposome-anchored 9Flg (con-taining the flagellin hairpin region) and TLR5.Interestingly, it was reported that asialo-GM1 could play a cor-

eceptor role in the binding of flagellin to TLR5 in lung (36);however, the precise relevance of this for the flagellin–TLR5 in-teraction remains to be elucidated. The possibility that pFlg lip-osomes can interact with receptors other than TLR5 on cells alsocannot be discounted by the present work. However, our demon-stration of a correlation between the binding of pFlg liposomesand TLR5 mAb to cells known to express TLR5 (Fig. 3A) isconsistent with 9Flg liposomes interacting with TLR5.An important finding from the present work is that the interaction

of pFlg liposomes with APCs can elicit APC activation/maturationin vitro (Supplemental Fig. 4), as well as in vivo (Fig. 4). Theadministration of 9Flg liposomes to mice increased the expressionof activation markers MHC class II and CD86 on CD11c+ on cellsrecovered from spleen 24 h later to an extent comparable to thatseen with the administration of 10 mg LPS (a TLR4 ligand) (Fig.4A). Flagellin is known to elicit DC maturation by binding toTLR5 and signaling through the MyD88 signal-transduction path-way, which involves recruitment of the MyD88 adaptor protein.Therefore, it is not surprising that analogous experiments per-formed using syngeneic MyD882/2 mice failed to show any in-crease in CD86 and MHC class II expression on CD11c+

splenocytes from mice injected with pFlg liposomes (Fig. 4B).In contrast, splenocytes from MyD882/2 mice injected withLPS still elicited a substantial increase in CD86 and MHC classII expression, presumably because LPS (unlike flagellin) can sig-nal DC maturation through the MyD88- and TRIF-dependentpathways (7, 19). These findings show clearly that vaccinationwith 9Flg liposomes is effective at inducing MyD88-dependentmaturation of CD11c+ DCs in vivo.The ability to prime Ag-specific T cells and to generate IFN-g–

producing CD8+ T cells is crucial for the induction of an effectiveCTL response and protective antitumor immunity (23, 47). Furtherconfirmation that pFlg liposomes can be used to elicit Ag-specificresponses was provided by our demonstration that vaccination ofC57BL/6 mice with 9Flg- and 42Flg-OVA liposome induced Ag-specific T cell priming, as assessed by the ability to induce increasedAg-specific cell proliferation (Fig. 5A), as well as to generate a higher

1752 LIPOSOMAL Ag WITH FLAGELLIN PEPTIDES ELICITS ANTITUMOR EFFECTS

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

proportion (∼1%) of CD8+ T cells that produced IFN-g in response toAg compared with control vaccinations. These results are consistentwith the reported induction of a SIINFEKL-specific CD8+ T cellIFN-g ELISPOT response to fused flagellin-OVA vaccine (20). Ourresults indicate that vaccination with pFlg-OVA liposomes also in-duced amarked increase in the proportion of IFN-g–producingCD82

cells (Fig. 5C). We did not attempt to characterize the cells or themechanism involved in this response. However, it was reported thatvaccination with flagellin-Ag fusion protein (presumed to elicitresponses through TLR5) also results in a potent induction of CD8+

and CD4+ Ag-specific T cell responses (19) and that the injection offlagellin in the footpadofmice induces the proliferation and activationof NK cells in the draining lymph node (16). Thus, the effects weobserved also align with these previous findings. Interestingly, vacci-nation with 9Flg- and 42Flg-OVA liposomes also elicited the pro-duction of Ag-specific IgG1 and IgG2b Ab (Fig. 5D). The resultsdemonstrate that vaccination with Ag liposomes engrafted with9Flg or 42Flg is effective at inducing humoral and cell-mediatedimmune responses to liposome-associated Ag.Crucial to the effectiveness of pFlg liposomes as a vaccine is their

ability to induce protection against disease.We first inoculatedmicewith B16-OVAmelanoma cells and then vaccinated themwith OVAliposomes engrafted with the different pFlgs to assess the effect ontumor growth and metastasis. We also produced PMVs (known tocontain a battery of tumor Ags) from cultured tumor cells (B16-OVA and P815 cells) and, after incorporation with NTA3-DTDA,engrafted them with 9Flg or 42Flg peptides for use as vaccines. Inboth experimental strategies, the vaccination of mice previouslychallenged with the B16-OVA tumor induced a dramatic inhibitionof tumor growth/metastasis, resulting in the majority of the micehaving a complete absence of tumor foci in their lungs (Fig. 6A, 6B).The antitumor responses induced by vaccinating with OVA lipo-somes engrafted with 10pFlg and 11pFlg (Fig. 6A) were strongerthan for the corresponding vaccinations with PMVs (Fig. 6B). Alikely explanation for this is that, under the conditions used, vacci-nation with OVA liposomes resulted in the surrogate tumor Ag(OVA) being administered at high levels compared with the amountof OVA and other tumor Ags contained in the PMVs. Thus, vacci-nation with OVA liposomes may have resulted in overvaccinationand the consequent induction of an effective antitumor response,even with the less-active peptides 10Flg and 11Flg. However, thesepeptides are clearly less effective at inducing antitumor effectswhenPMVs are used (Fig. 6B, 6C), consistent with them also being lesseffective at mediating the binding of liposomes to DCs (Figs. 1, 2,Supplemental Figs. 1, 2).A major finding is that vaccination with 9Flg PMVs also induced

a strong antitumor response in DBA/2J mice bearing s.c. P815mammary mastocytoma, resulting in inhibited tumor growth andprolonging survival of the animals (Fig. 7B). Importantly, studieswith mice bearing established s.c. B16-OVA tumors (∼2 mm di-ameter) resulted in complete tumor regression; unlike controlmice that exhibited progressive tumor growth (Fig. 7A), the 9FlgPMV-treated mice were tumor-free 4 mo after tumor inoculation.Although the effectiveness can depend on the tumor type, ourresults clearly show that vaccination with 9Flg PMVs can bea highly effective approach to inducing strong antitumor responsesand immunotherapeutic effects in these different tumor models.In summary, this work developed a novel strategy that uses pFlgs

to promote the interaction of liposomal Ag with APCs and that canpotentially circumvent the need to manipulate DCs ex vivo for usein cancer immunotherapies (48). This approach has distinctadvantages over the use of Ag-flagellin fusion proteins to targetAg to TLR5 (19–26) or single-chain variable fragment-engraftedAg-containing liposomes to target Ag-uptake receptors to DCs

(28). First, the relatively short pFlgs can readily be producedsynthetically and do not require the production of recombinantproteins, such as Ag-flagellin fusions and short-chain variablefragments, which can be cumbersome to produce and difficult touse because of poor stability. Second, as demonstrated in thisstudy, pFlg liposomes/PMV vaccines are self-adjuvating, eliminat-ing the need to deliver a concomitant DC maturation signal toinduce an effective immunogenic response (28, 46). This is par-ticularly important, because the absence of a DC-maturation sig-nal can lead to a failure to mount the desired response, or it mightlead to the induction of tolerance to the Ag (47). Third, the ap-proach is versatile, and it can be readily applied to raising immuneresponses to potentially any clinically relevant Ag that can beincorporated in a liposome. Fourth, the use of pFlgs instead ofwild-type flagellin can be expected to reduce the risk for trigger-ing the production of neutralizing Abs in the host, which canadversely affect defense mechanisms against pathogenic flagel-lated bacteria. Our preliminary studies indicate that 9Flg lipo-somes also bind strongly to human monocyte-derived DCs (datanot shown). Therefore, the use of pFlg-engrafted liposomes/PMVscould provide a convenient and attractive means to harness theimmunopotentiating properties of the innate and adaptive immunesystems through TLR5 on APCs for the development of noveleffective vaccines and cancer immunotherapies.

AcknowledgmentsWe thank Lipotek Pty Ltd for providing the NTA3-DTDA for research

purposes. We also thank David Bennett for carrying out some of the pre-

liminary experiments and Prof. S. Akira and S. Uematsu, Department of

Host Defense, Research Institute for Microbial Diseases, Osaka University,

Japan, for making the MYD882/2 mice available for some of the experi-

ments carried out in this study.

DisclosuresJ.G.A. has a commercial interest in Lipotek Pty Ltd.

References1. Akira, S., S. Uematsu, and O. Takeuchi. 2006. Pathogen recognition and innate

immunity. Cell 124: 783–801.2. Gordon, S. 2002. Pattern recognition receptors: doubling up for the innate im-

mune response. Cell 111: 927–930.3. Uematsu, S., and S. Akira. 2006. Toll-like receptors and innate immunity. J. Mol.

Med. 84: 712–725.4. Gay, N. J., and M. Gangloff. 2008. Structure of toll-like receptors. In Handbook

Experimental Pharmacology, Vol. 183. Springer Berlin Heidelberg, New York/Berlin, p. 181–200.

5. Buwitt-Beckmann, U., H. Heine, K. H. Wiesmuller, G. Jung, R. Brock, S. Akira,and A. J. Ulmer. 2006. TLR1- and TLR6-independent recognition of bacteriallipopeptides. J. Biol. Chem. 281: 9049–9057.

6. Hayashi, F., K. D. Smith, A. Ozinsky, T. R. Hawn, E. C. Yi, D. R. Goodlett,J. K. Eng, S. Akira, D. M. Underhill, and A. Aderem. 2001. The innate immuneresponse to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410:1099–1103.

7. Feuillet, V., S. Medjane, I. Mondor, O. Demaria, P. P. Pagni, J. E. Galan,R. A. Flavell, and L. Alexopoulou. 2006. Involvement of Toll-like receptor 5 inthe recognition of flagellated bacteria. Proc. Natl. Acad. Sci. USA 103: 12487–12492.

8. Edwards, A. D., S. S. Diebold, E. M. Slack, H. Tomizawa, H. Hemmi, T. Kaisho,S. Akira, and C. Reis e Sousa. 2003. Toll-like receptor expression in murine DCsubsets: lack of TLR7 expression by CD8 alpha+ DC correlates with unrespon-siveness to imidazoquinolines. Eur. J. Immunol. 33: 827–833.

9. Uematsu, S., M. H. Jang, N. Chevrier, Z. Guo, Y. Kumagai, M. Yamamoto,H. Kato, N. Sougawa, H. Matsui, H. Kuwata, et al. 2006. Detection of patho-genic intestinal bacteria by Toll-like receptor 5 on intestinal CD11c+ laminapropria cells. Nat. Immunol. 7: 868–874.

10. Hawn, T. R., W. R. Berrington, I. A. Smith, S. Uematsu, S. Akira, A. Aderem,K. D. Smith, and S. J. Skerrett. 2007. Altered inflammatory responses in TLR5-deficientmice infectedwithLegionella pneumophila. J. Immunol.179: 6981–6987.

11. Andersen-Nissen, E., T. R. Hawn, K. D. Smith, A. Nachman, A. E. Lampano,S. Uematsu, S. Akira, and A. Aderem. 2007. Cutting edge: Tlr5-/- mice are moresusceptible to Escherichia coli urinary tract infection. J. Immunol. 178: 4717–4720.

The Journal of Immunology 1753

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

12. Uematsu, S., K. Fujimoto, M. H. Jang, B. G. Yang, Y. J. Jung, M. Nishiyama,S. Sato, T. Tsujimura, M. Yamamoto, Y. Yokota, et al. 2008. Regulation ofhumoral and cellular gut immunity by lamina propria dendritic cells expressingToll-like receptor 5. Nat. Immunol. 9: 769–776.

13. Bagchi, A., E. A. Herrup, H. S. Warren, J. Trigilio, H. S. Shin, C. Valentine, andJ. Hellman. 2007. MyD88-dependent and MyD88-independent pathways in syn-ergy, priming, and tolerance between TLR agonists. J. Immunol. 178: 1164–1171.

14. Zaru, R., N. Ronkina, M. Gaestel, J. S. Arthur, and C. Watts. 2007. The MAPK-activated kinase Rsk controls an acute Toll-like receptor signaling response indendritic cells and is activated through two distinct pathways. Nat. Immunol. 8:1227–1235.

15. Tallant, T., A. Deb, N. Kar, J. Lupica, M. J. de Veer, and J. A. DiDonato. 2004.Flagellin acting via TLR5 is the major activator of key signaling pathwaysleading to NF-kappa B and proinflammatory gene program activation in intes-tinal epithelial cells. BMC Microbiol. 4: 33.

16. Tsujimoto, H., T. Uchida, P. A. Efron, P. O. Scumpia, A. Verma, T. Matsumoto,S. K. Tschoeke, R. F. Ungaro, S. Ono, S. Seki, et al. 2005. Flagellin enhancesNK cell proliferation and activation directly and through dendritic cell-NK cellinteractions. J. Leukoc. Biol. 78: 888–897.

17. McSorley, S. J., B. T. Cookson, and M. K. Jenkins. 2000. Characterization ofCD4+ T cell responses during natural infection with Salmonella typhimurium. J.Immunol. 164: 986–993.

18. McKee, A. S., M. W. Munks, and P. Marrack. 2007. How do adjuvants work?Important considerations for new generation adjuvants. Immunity 27: 687–690.

19. Cuadros, C., F. J. Lopez-Hernandez, A. L. Dominguez, M. McClelland, andJ. Lustgarten. 2004. Flagellin fusion proteins as adjuvants or vaccines inducespecific immune responses. Infect. Immun. 72: 2810–2816.

20. Huleatt, J. W., A. R. Jacobs, J. Tang, P. Desai, E. B. Kopp, Y. Huang, L. Song,V. Nakaar, and T. J. Powell. 2007. Vaccination with recombinant fusion proteinsincorporating Toll-like receptor ligands induces rapid cellular and humoral im-munity. Vaccine 25: 763–775.

21. Honko, A. N., N. Sriranganathan, C. J. Lees, and S. B. Mizel. 2006. Flagellin isan effective adjuvant for immunization against lethal respiratory challenge withYersinia pestis. Infect. Immun. 74: 1113–1120.

22. McDonald, W. F., J. W. Huleatt, H. G. Foellmer, D. Hewitt, J. Tang, P. Desai,A. Price, A. Jacobs, V. N. Takahashi, Y. Huang, et al. 2007. A West Nile virusrecombinant protein vaccine that coactivates innate and adaptive immunity. J.Infect. Dis. 195: 1607–1617.

23. Le Moigne, V., G. Robreau, and W. Mahana. 2006. Homologous recombinationwith linear DNA to insert antigenic protein in the flagellin: improvement of theTh1 immune response. Microbiol. Immunol. 50: 33–43.

24. Huleatt, J. W., V. Nakaar, P. Desai, Y. Huang, D. Hewitt, A. Jacobs, J. Tang,W. McDonald, L. Song, R. K. Evans, et al. 2008. Potent immunogenicity and ef-ficacy of a universal influenza vaccine candidate comprising a recombinant fusionprotein linking influenza M2e to the TLR5 ligand flagellin. Vaccine 26: 201–214.

25. Bargieri, D. Y., D. S. Rosa, C. J. Braga, B. O. Carvalho, F. T. Costa,N. M. Espındola, A. J. Vaz, I. S. Soares, L. C. Ferreira, and M. M. Rodrigues.2008. New malaria vaccine candidates based on the Plasmodium vivax Mero-zoite Surface Protein-1 and the TLR-5 agonist Salmonella Typhimurium FliCflagellin. Vaccine 26: 6132–6142.

26. Le Moigne, V., G. Robreau, and W. Mahana. 2008. Flagellin as a good carrierand potent adjuvant for Th1 response: study of mice immune response to the p27(Rv2108) Mycobacterium tuberculosis antigen. Mol. Immunol. 45: 2499–2507.

27. Altin, J. G., M. G. Banwell, P. A. Coghlan, C. J. Easton, M. R. Nairn, andD. A. Offermann. 2006. Synthesis of NTA3-DTDA — A Chelator-Lipid thatPromotes Stable Binding of His-Tagged Proteins to Membranes. Aust. J. Chem.59: 302–306.

28. van Broekhoven, C. L., C. R. Parish, C. Demangel, W. J. Britton, and J. G. Altin.2004. Targeting dendritic cells with antigen-containing liposomes: a highly ef-fective procedure for induction of antitumor immunity and for tumor immuno-therapy. Cancer Res. 64: 4357–4365.

29. van Broekhoven, C. L., and J. G. Altin. 2005. The novel chelator lipid 3(nitri-lotriacetic acid)-ditetradecylamine (NTA(3)-DTDA) promotes stable binding ofHis-tagged proteins to liposomal membranes: potent anti-tumor responses

induced by simultaneously targeting antigen, cytokine and costimulatory signalsto T cells. Biochim. Biophys. Acta 1716: 104–116.

30. Herringson, T. P., R. R. Patlolla, and J. G. Altin. 2009. Targeting of plasmidDNA-lipoplexes to cells with molecules anchored via a metal chelator lipid. J.Gene Med. 11: 1048–1063.

31. Inaba, K., M. Inaba, N. Romani, H. Aya, M. Deguchi, S. Ikehara, S. Muramatsu,and R. M. Steinman. 1992. Generation of large numbers of dendritic cells frommouse bonemarrow cultures supplemented with granulocyte/macrophage colony-stimulating factor. J. Exp. Med. 176: 1693–1702.

32. Faham, A., D. Bennett, and J. G. Altin. 2009. Liposomal Ag engrafted withpeptides of sequence derived from HMGB1 induce potent Ag-specific and anti-tumour immunity. Vaccine 27: 5846–5854.

33. Smith, K. D., E. Andersen-Nissen, F. Hayashi, K. Strobe, M. A. Bergman,S. L. Barrett, B. T. Cookson, and A. Aderem. 2003. Toll-like receptor 5 recog-nizes a conserved site on flagellin required for protofilament formation andbacterial motility. Nat. Immunol. 4: 1247–1253.

34. Reichhart, J. M. 2003. TLR5 takes aim at bacterial propeller. Nat. Immunol. 4:1159–1160.

35. Murthy, K. G., A. Deb, S. Goonesekera, C. Szabo, and A. L. Salzman. 2004.Identification of conserved domains in Salmonella muenchen flagellin that areessential for its ability to activate TLR5 and to induce an inflammatory responsein vitro. J. Biol. Chem. 279: 5667–5675.

36. Nempont, C., D. Cayet, M. Rumbo, C. Bompard, V. Villeret, and J. C. Sirard.2008. Deletion of flagellin’s hypervariable region abrogates antibody-mediatedneutralization and systemic activation of TLR5-dependent immunity. J. Immu-nol. 181: 2036–2043.

37. Andersen-Nissen, E., K. D. Smith, R. Bonneau, R. K. Strong, and A. Aderem.2007. A conserved surface on Toll-like receptor 5 recognizes bacterial flagellin.J. Exp. Med. 204: 393–403.

38. Andersen-Nissen, E., K. D. Smith, K. L. Strobe, S. L. Barrett, B. T. Cookson,S. M. Logan, and A. Aderem. 2005. Evasion of Toll-like receptor 5 by flagellatedbacteria. Proc. Natl. Acad. Sci. USA 102: 9247–9252.

39. Samatey, F. A., K. Imada, S. Nagashima, F. Vonderviszt, T. Kumasaka,M. Yamamoto, and K. Namba. 2001. Structure of the bacterial flagellar proto-filament and implications for a switch for supercoiling. Nature 410: 331–337.

40. Wong, C. K., P. F. Cheung, W. K. Ip, and C. W. Lam. 2007. Intracellular sig-naling mechanisms regulating toll-like receptor-mediated activation of eosino-phils. Am. J. Respir. Cell Mol. Biol. 37: 85–96.

41. Fukata, M., K. Breglio, A. Chen, A. S. Vamadevan, T. Goo, D. Hsu, D. Conduah,R. Xu, and M. T. Abreu. 2008. The myeloid differentiation factor 88 (MyD88) isrequired for CD4+ T cell effector function in a murine model of inflammatorybowel disease. J. Immunol. 180: 1886–1894.

42. Xia, C., M. Lu, Z. Zhang, Z. Meng, Z. Zhang, and C. Shi. 2008. TLRs anti-viral effect on hepatitis B virus in HepG2 cells. J. Appl. Microbiol. 105: 1720–1727.

43. Hawn,T.R.,A.Verbon,K.D.Lettinga,L. P. Zhao, S. S. Li, R. J. Laws, S. J. Skerrett,B. Beutler, L. Schroeder, A. Nachman, et al. 2003.A common dominant TLR5 stopcodon polymorphism abolishes flagellin signaling and is associated with suscep-tibility to legionnaires’ disease. J. Exp. Med. 198: 1563–1572.

44. Eaves-Pyles, T. D., H. R. Wong, K. Odoms, and R. B. Pyles. 2001. Salmonellaflagellin-dependent proinflammatory responses are localized to the conservedamino and carboxyl regions of the protein. J. Immunol. 167: 7009–7016.

45. Van Broekhoven, C. L., and J. G. Altin. 2001. A novel system for convenientdetection of low-affinity receptor-ligand interactions: chelator-lipid liposomesengrafted with recombinant CD4 bind to cells expressing MHC class II. Immu-nol. Cell Biol. 79: 274–284.

46. McNamara, N., M. Gallup, A. Sucher, I. Maltseva, D. McKemy, andC. Basbaum. 2006. AsialoGM1 and TLR5 cooperate in flagellin-induced nucle-otide signaling to activate Erk1/2. Am. J. Respir. Cell Mol. Biol. 34: 653–660.

47. Heath, W. R., G. T. Belz, G. M. Behrens, C. M. Smith, S. P. Forehan, I. A. Parish,G. M. Davey, N. S. Wilson, F. R. Carbone, and J. A. Villadangos. 2004. Cross-presentation, dendritic cell subsets, and the generation of immunity to cellularantigens. Immunol. Rev. 199: 9–26.

48. Steinman, R. M., and J. Banchereau. 2007. Taking dendritic cells into medicine.Nature 449: 419–426.

1754 LIPOSOMAL Ag WITH FLAGELLIN PEPTIDES ELICITS ANTITUMOR EFFECTS

by guest on May 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from