Molecular Immunology 43 (2006) 667–676

Anti-HER2/neu IgG3–(IL-2) and anti-HER2/neu IgG3–(GM-CSF)promote HER2/neu processing and presentation by dendritic cells:

Implications in immunotherapy and vaccination strategies

Jay Soriano Dela Cruza, Kamh Ryan Trinha, Hsiao Wen Chena, Antoni Ribasb,Sherie L. Morrisona, Manuel L. Penicheta,∗

a Department of Microbiology, Immunology, and Molecular Genetics, University of California at Los Angeles (UCLA), Los Angeles, CA 90095, USAb Department of Surgery, Division of Surgical Oncology, University of California at Los Angeles (UCLA), Los Angeles, CA 90095, USA

Received 18 January 2005Available online 23 May 2005

Abstract

notherapy.H riticc theic DCs toi llulard onp d to OVA-si cessingc (GM-CSF)f ns (TAAs).©

K

1

mpte

es

oU

n in ancers

nd

se to

by

e

0d

HER2/neu, a transmembrane glycoprotein overexpressed in several types of human cancers, is a potential target for active immuowever, this protein and especially its extracellular domain (ECDHER2), is weakly immunogenic and is poorly processed by dendells (DCs). Previously, we showed that anti-HER2/neu IgG3–(IL-2) and anti-HER2/neu IgG3–(GM-CSF) fusion proteins can enhancemmunogenicity of ECDHER2 in mice, and that the non-covalent physical association between each antibody fusion proteins and ECDHER2 wasritical to elicit optimal protective immunity against HER2/neu expressing tumors. We now use the professional antigen-presenting

nvestigate the effect of the antibody fusion protein binding to ECDHER2 on its trafficking and presentation. We found that when the extraceomain of HER2/neu fused to ovalbumin (OVA–ECDHER2) is bound by HER2/neu-specific antibody–(IL-2) or antibody–(GM-CSF) fusiroteins, the bound antigen is more efficiently processed by murine bone-marrow-derived dendritic cells (BMDCs) and presentepecific T-cells than the unbound OVA–ECDHER2. We also found that ECDHER2 bound by anti-HER2/neu IgG3–(IL-2) is very efficientlynternalized and that the internalized ECDHER2 is not retained in the early endosomal compartments but traffics to the antigen-proompartments. These results are consistent with our earlier in vivo studies and suggest that both antibody–(IL-2) and antibody–usion proteins can be used to enhance the immune response to poorly immunogenic antigens including tumor-associated antige

2005 Elsevier Ltd. All rights reserved.

eywords: Cancer; HER2/neu; Cytokine; Antibody fusion proteins; Immunotherapy

. Introduction

The ability to potentiate immunity against poorly im-unogenic antigens would have a significant impact on therevention and treatment of many diseases. In malignancy,

umor-associated antigens (TAAs) often fail to elicit anffective immune response. One such TAA is HER2/neu or

Abbreviations: BMDC, bone marrow-derived dendritic cells; ECDHER2,xtracellular domain of HER2/neu; MIIC, MHC-II compartments; ns, non-pecific; DNS, dansyl; OVA, ovalbumin∗ Corresponding author at: UCLA Department of Microbiology, Immunol-gy, and Molecular Genetics, Box 148906, Los Angeles, CA 90095-1489,SA. Tel.: +1 310 206 5127; fax: +1 310 206 5231.

E-mail address: penichet@microbio.ucla.edu (M.L. Penichet).

erbB2, a membrane glycoprotein whose overexpressiosubset of malignancies including breast and ovarian cais correlated with poor prognosis (Slamon et al., 1987).HER2/neu consists of an extracellular domain, ECDHER2,that can exist in circulation as a soluble∼100 kD plasmaprotein (Hayes et al., 2001), a transmembrane domain aan intracellular domain with tyrosine kinase activity (Yardenand Sliwkowski, 2001).

Studies in rats suggested that an immune responrat HER2/neu could be elicited using HER2/neu-derivedpeptides but not with the whole protein (Disis et al., 1996).HER2/neu-specific immunity can also be augmentedimmunizing patients with HER2/neu-derived peptides (Disiset al., 1996, 1999; Salazar et al., 2003). The poor immun

161-5890/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.oi:10.1016/j.molimm.2005.04.007

668 J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676

response to the intact protein and the ability to respond topeptide immunization suggest that antigen-presenting cells(APCs) may not be able to efficiently process ECDHER2

to the peptides required for effective antigen presentation.Indeed, Hiltbold et al., elegantly demonstrated that, althoughECDHER2 is readily internalized by human peripheralblood-derived dendritic cells (DCs), it does not efficientlytraffic to the cellular compartments participating in antigenprocessing and presentation (e.g., lysosomes and MHC-IIcompartments, MIICs); instead, it remains in the earlyendosomes and is subsequently recycled to the cell surfaceand released (Hiltbold et al., 2000).

Previously, we showed that anti-HER2/neu IgG3–(IL-2)and anti-HER2/neu IgG3–(GM-CSF) can enhance theimmunogenicity of ECDHER2 in mice (Dela Cruz et al.,2003). In the present study, we show that murine bonemarrow-derived DCs (BMDCs) are able to effectivelyprocess and present antigens bound by anti-HER2/neuIgG3–(IL-2 or GM-CSF). To demonstrate this, we fused thechicken ovalbumin (OVA) to ECDHER2 (OVA–ECDHER2),and showed that only when OVA–ECDHER2 was boundto anti-HER2/neu IgG3–(IL-2 or GM-CSF) could pulsedBMDCs effectively stimulate OVA-specific T-cells. Confocalmicroscopy demonstrated increased uptake and increasedlysosomal and MHC-II compartment colocalization forECDHER2 bound by anti-HER2/neu IgG3–(IL-2). Ourfi le ina kine-a rate an SF)f ncingt

2

2

edf att osA theM genes ereo bor,M theU C;h gM

2

bio-lII of

the anti-HER2/neu humanized monoclonal antibody, 4D5-8,were described previously (Dela Cruz et al., 2000; Penichetet al., 2001). We used human IL-2 and murine GM-CSF, be-cause human IL-2 exhibits full bioactivity in mice, whereashuman GM-CSF does not. Soluble ECDHER2 was purifiedfrom BHK/erbB2 culture supernatants as described previ-ously (Dela Cruz et al., 2003).

Non-HER2/neu-specific (ns) antibody or antibody fusionproteins were used as isotype-matched controls. The nsIgG3is a recombinant human anti-dansyl (DNS) IgG3, whilensIgG3–(IL-2) and nsIgG3–(GM-CSF) are specific forDNS and CEA (carcinoembryonic antigen), respectively.The nsIgG3–(IL-2) fusion proteins and anti-HER2/neuIgG3–(IL-2) are identical in structure except for the dif-ferent variable regions and exhibit comparable level ofIL-2 bioactivity in a CTLL-2 (IL-2-dependent T-cell line)proliferation assay. nsIgG3–(GM-CSF) and anti-HER2/neuIgG3–(GM-CSF) are also identical in structure and exhibitcomparable level of GM-CSF bioactivity in an FDC-P1(GM-CSF-dependent cell line) proliferation assay.

2.3. BMDC generation

BMDCs were generated as described (Ribas et al., 1999).Briefly, bone marrow cells were flushed from femurs ofBALB/c mice with RPMI 1640 (Invitrogen, Carlsbad, CA)a 640wc sors)w etalb ouseI ys-t andh xpo-s a,S -IIm sedi tageo

2O

ncef theE E).P forO e,a -qs for-wC T-TrT sent

ndings suggest that cytokine receptors can play a rontigen uptake, processing and presentation of cytossociated antigens. Thus, these studies demonstovel mechanism of action of antibody–(IL-2 and GM-C

usion proteins and suggest a general approach for enhahe immunogencity of ECDHER2 and other antigens.

. Materials and methods

.1. Mice

The 6- to 8-week-old female BALB/c mice were obtainrom the Experimental Radiation Oncology Vivariumhe University of California at Los Angeles (UCLA, Lngeles, CA). DO11.10 female BALB/c mice carryingHC class-II restricted rearranged T-cell receptor trans

pecific for a chicken ovalbumin peptide (323–339) wbtained from The Jackson Laboratories (Bar HarE). All experiments were performed according toCLA Chancellor’s Animal Research Committee (ARttp://www.oprs.ucla.edu/) Animal Care and Use Traininanual.

.2. Antibody–cytokine fusion proteins and ECDHER2

The construction, purification and analysis of theogical activities of anti-HER2/neu IgG3, anti-HER2/neugG3–(murine GM-CSF) and anti-HER2/neu IgG3–(humanL-2) consisting of human IgG3 with the variable region

nd washed. After an overnight incubation in RPMI 1ith 10% fetal bovine serum in a 100 cm× 20 cm tissueulture plate, non-adherent cells (containing DC precurere harvested and cultured in RPMI 1640 with 10% fovine serum in the presence of 10 ng/ml recombinant m

L-4 and 2 ng/ml recombinant mouse GM-CSF (R&D Sems, Minneapolis, MN). Cells were fed 4 days laterarvested for use after a total of 8 days in culture. Eure to TNF-�′ (Peprotech, Rocky Hill, NJ) or LPS (Sigmt. Louis, MO) for 24 h resulted in upregulation of MHColecules on BMDCs indicating that the day 8 BMDCs u

n the experiments below were mostly in the immature sf DC differentiation.

.4. Construction, purification and characterization ofVA–ECDHER2

EST clone pgr1c.pk001.f5 containing the sequeor chicken OVA was identified and obtained fromST library at the University of Delaware (Newark, DCR was used to amplify the sequence encodingVA and add aKpnI restriction site, Kozak sequencnd the Ig kappa L chain signal peptide 5′ of the seuence encoding OVA and aSmaI restriction site 3′ of theequence encoding for OVA. The primers used were:ard primer, 5′-GGGTACCACCATGGAGACAGACACATCCTGCTATGGGTACTGCTGCTCTGGGTTCCAGGCCACTGGTGACGGC TCCATCGGTGCAGCA-3′ andeverse primer, 5′-CTCTAGACCCGGGGGAAACACATC-GCCAAA GAAG-3′, where the underlined bases repre

J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676 669

the portion of the primer that hybridizes to the sequenceencoding for OVA. The resulting PCR product was clonedinto PCR2.1 vector (Invitrogen) and sequenced. The cDNAclone L-her2/neu-SN (Penichet et al., 1999) was used as atemplate to amplify the sequence encoding ECDHER2. ASmaI restriction site and a (Gly4-Ser)3 linker were added5′ of the ECDHER2 DNA fragment and a stop codon anda XbaI restriction site were added 3′ of this fragment.The primers used were: forward primer, 5′-GCCCGGGGGCCTGGTTCGGGCGGAGGTGGGTCGGGTGGCGG-CGGATCCACCCAAGTGTGCACCGGC-3′ and reverseprimer, 5′-GTCTAGATTAGGACGTCAGAGGGCTGGC-3′, where the underlined bases represent the portion of thesequence that hybridized to the ECDHER2 DNA fragment.Following PCR, the product was cloned into Invitrogen’sPCR2.1 vector (Invitrogen, Carlsbad, CA) and partiallysequenced. The unsequenced portion of the ECDHER2 DNAfragment flanked byScaI and BstBI restriction sites wasreplaced with the sequenced fragment of the cDNA clone, L-HER2/neu-SN. To clone OVA and ECDHER2DNA fragmentsinto a single expression vector, theKpnI/SmaI fragmentcontaining the sequence encoding OVA and theSmaI/XbaIfragment containing the sequence encoding ECDHER2 wereligated into aKpnI and XbaI-digested pCDNA3.1 vector(Invitrogen) that had been modified to carry the eukaryoticselection geneHisD. Murine myeloma NS0/1 cells weret ye tantss with3

ma lublep A).I h orw leb gthOO swnr al,W e’sM ith5 in a3p(

2r

ed,w issuecpb ,

anti-HER2/neu IgG3–(IL-2), nsIgG3–(GM-CSF) or anti-HER2/neu IgG3–(GM-CSF) in antigen excess of 4:1OVA–ECDHER2 to antibody molar ratio. Mixtures wereincubated in RPMI 1640 plus 10% fetal bovine serum for 2 hin a 37◦C water bath to allow complexes to form. The 50�lof the mixtures was added in triplicate to DCs in a final vol-ume of 100�l to give a final OVA–ECDHER2 concentrationof 15�g/ml (the equivalent of 5�g/ml of OVA (Sigma), aconcentration adequate to stimulate T-cell activation by OVApulsed DCs), and placed in the incubator for 24 h. rhIL-2 (agift from Chiron Corporation, Emeryville, CA) or rmGM-CSF (R&D Systems) was added to OVA–ECDHER2 alone ormixtures of OVA–ECDHER2 plus anti-HER2/neu IgG3 usingthe equimolar amount of IL-2 in anti-HER2/neu IgG3–(IL-2)or GM-CSF in anti-HER2/neu IgG3–(GM-CSF). CD4+

T-cells were obtained from the splenocytes of DO11.10 mice.Briefly, splenocytes were harvested as described (Dela Cruzet al., 2003) and incubated with an antibody cocktail consist-ing of biotinylated antibodies specific for all cell types exceptCD4+ T-cells (Miltenyi Biotec Inc., Auburn, CA). After abrief incubation with anti-biotin microbeads (microbeadscoupled to monoclocal anti-biotin antibodies; MiltenyiBiotec Inc.), CD4+ T-cells were negatively selected andpurified using the autoMACS Separator following directionssupplied by the manufacturer (Miltenyi Biotec Inc.). LoadedBMDCs were washed twice with RPMI 1640 plus 10% fetalb n1 andt inedu ithc nti-b in-c SanF -p ffer( e for2 ech)w itivec

2B

-i Eu-g lingEI oxi-m ofa ionalC m-b etalbpIC

ransfected with the OVA–ECDHER2 expression vector blectroporation and histidinol-resistant stable transfecelected. Transfectants were biosynthetically labeled5S-methionine (ICN, Irvine, CA). OVA–ECDHER2 was im-unoprecipitated using human anti-HER2/neu followed byrabbit anti-human IgG and a 10% suspension of inso

rotein A (IgGSorb, The Enzyme Center, Malden, Mmmunoprecipitates were analyzed by SDS-PAGE witithout reduction by�-mercaptoethanol (Sigma). A singand of∼150 kDa was observed representing the full-lenVA–ECDHER2 comprised of ECDHER2 (∼100 kD) andVA (∼50 kDa). High OVA–ECDHER2-producing cloneere identified by ELISA. Fusing ECDHER2 to OVA didot affect recognition by either anti-HER2/neu IgG3 or aabbit polyclonal anti-OVA antibody (Accurate Chemicestbury, NY). Transfectants were grown in Iscovodified Dulbecco’s Medium (IMDM) supplemented w% calf serum (Omega Scientific Inc.) and maintained7◦C, 5% CO2 humidified incubator. OVA–ECDHER2 wasurified and quantified as had been described for ECDHER2

Dela Cruz et al., 2003).

.5. OVA–ECDHER2and OVA-specific T-cells as aeporter of antigen processing and presentation

BMDCs from BALB/c mice were harvested, countashed and seeded onto a 96-well round bottom tulture plate at 6× 104 cells in 50�l/well in RPMI 1640lus 10% fetal bovine serum. OVA–ECDHER2 was incu-ated with nsIgG3, anti-HER2/neu IgG3, nsIgG3–(IL-2)

ovine serum and 18× 104 CD4+ T-cells were added i00�l to each well. Supernatants were removed after 48 h

he amount of IL-2 secreted by activated T-cells determsing a sandwich ELISA in a 96-well microtiter plate wapture and detecting (biotinylated) anti-mouse IL-2 aodies (BD Biosciences). After incubation with steptavidonjugated alkaline phosphatase (Zymed, Southrancisco, CA), plates were washed andp-nitrophenyl phoshate disodium (Sigma) dissolved in diethnolamine buSigma) was added. After incubation at room temperaturh, the plates were read at 410 nm. Murine rIL-2 (Peprotas included in all ELISA assays as an internal posontrol.

.6. Analysis of protein uptake and trafficking inMDCs using confocal microscopy

ECDHER2 was labeled with Alexafluor 488 followng the manufacturers instructions (Molecular Probes,ene, OR). An ELISA assay confirmed that labeCDHER2 did not affect recognition by anti-HER2/neu

gG3. BMDCs were harvested, counted and apprately 2× 105 DCs were added to each chambern eight-chambered glass slide (Nalge Nunc Internatorp., Naperville, IL). After 2 h in the incubator, chaers were gently washed with RPMI 1640 plus 2% fovine serum to remove any non-adherent cells. ECDHER2

lus anti-HER2/neu IgG3, ECDHER2 plus anti-HER2/neugG3–(IL-2), ECDHER2 plus anti-HER2/neu IgG3–(GM-SF), ECDHER2 plus nsIgG3–(GM-CSF), ECDHER2 plus

670 J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676

nsIgG3–(IL-2) or ECDHER2 plus nsIgG3 were mixed ata 1:1 molar ratio in RPMI 1640 plus 2% fetal bovineserum and incubated for 2 h at 37◦C to allow complexesto form. A 0.1 ml of ECDHER2 alone, ECDHER2 plus anti-body or ECDHER2 plus antibody–(IL-2) or antibody–(GM-CSF) fusion protein were added to BMDCs to a finalECDHER2 concentration of 50�g/ml. After 5, 15, 30, 60or 90 min, the labeled proteins were removed and 200�l of4% paraformaldehyde in PBS was added to each chamberand incubated for 30 min at 37◦C. After washing with PBS,slides were mounted (Prolong, Molecular Probes) or furtherprocessed.

For immunofluorescence staining 0.15 ml of permeabi-lizing agent (PBS containing 0.2% Tween 20; Sigma) wasadded to each chamber, the slides placed in the incubatorfor 15 min and washed immediately with staining buffer(PBS containing 2% calf serum and 0.1% sodium azide).The remaining incubations were performed in a volume of0.1 ml unless specified. Prior to addition of specific antibody,BMDCs were preincubated with 5�g/ml of monoclonal anti-mouse Fc-receptor (Fc-block, BD Biosciences) for 30 minat 4◦C. Fc-block was replaced with 2.5�g biotinylatedrat monoclonal anti-mouse CD71 (anti-murine transferrinreceptor, TfR, an early endosome marker, BD Biosciences),biotinylated monoclonal rat anti-mouse CD107a (anti-murine LAMP-1, a lysosome marker; BD Biosciences) or1 em r forM at4 oata d ana iths 8( in at4 ng,M 4ESc CA)e ryp-t andp uor4 eta-M n,P

3

3Io

ae dresstO of

Fig. 1. Anti-HER2/neu IgG3–(IL-2) and anti-HER2/neu IgG3–(GM-CSF)promote OVA–ECDHER2 processing and presentation by BMDCs. Thelevel of IL-2 secreted by activated DO11.10 T-cells after a 48 h incubationwith BMDCs loaded with OVA–ECDHER2, or OVA–ECDHER2 mixed withnsIgG3, anti-HER2/neu IgG3, nsIgG3–(IL-2), anti-HER2/neu IgG3–(IL-2),nsIgG3–(GM-CSF) or anti-HER2/neu IgG3–(GM-CSF) was assessed usinga sandwich ELISA. BMDCs loaded with 50 or 5�g/ml of the OVA pro-tein served as a positive control. Values represent the average intensity atOD410 nm of triplicate wells following an incubation at room temperaturefor 2 h. The error bars represent the range of triplicate determinations.

pulsed BMDCs to stimulate the available OVA-specificT-cells determined. Day 8 BMDCs were incubated withOVA–ECDHER2 alone or OVA–ECDHER2 plus the non-HER2/neu-specific IgG3 [nsIgG3], anti-HER2/neu IgG3, thenon-HER2/neu-specific IL-2 fusion protein [nsIgG3–(IL-2)], the non-HER2/neu-specific GM-CSF fusion protein[nsIgG3–(GM-CSF)], anti-HER2/neu IgG3–(IL-2) or anti-HER2/neu IgG3–(GM-CSF) for 24 h. The BMDCs were thenwashed and incubated with OVA-peptide-specific CD4+ T-cells derived from DO11.10 transgenic mice for an additional48 h and IL-2 secretion by activated T-cells measured (Fig. 1).BMDCs incubated with OVA–ECDHER2, OVA–ECDHER2

plus nsIgG3, OVA–ECDHER2 plus anti-HER2/neu IgG3,OVA–ECDHER2plus nsIgG3–(IL-2), or OVA–ECDHER2plusnsIgG3–(GM-CSF) did not activate the T-cells to secrete IL-2(Fig. 1). In contrast, BMDCs incubated with OVA–ECDHER2

plus anti-HER2/neu IgG3–(IL-2) or anti-HER2/neuIgG3–(GM-CSF) effectively induced T-cell activation.Similar results were observed in three independent exper-iments (data not shown). Therefore, the antibody–cytokinefusion proteins enhance the processing and presentationof OVA–ECDHER2 by DCs and the enhanced processingand presentation required a non-covalent association be-tween OVA–ECDHER2 and the antibody–cytokine fusionproteins.

:100 dilution of rabbit polyclonal anti-H2-DM (H2-DM thurine equivalent of human HLA-DM serves as a markeIICs (Hiltbold et al., 2000), and cells incubated overnight◦C. To detect the presence of anti-H2-DM, biotinylated gnti-rabbit IgG (Sigma) was added and cells incubatedditional 2 h at 4◦C. Finally, slides were washed twice wtaining buffer and 0.1�g of streptavidin–Alexafluor 56Molecular Probes) was added and slides incubated 30 m◦C. After a wash with PBS, slides were mounted (Proloolecular Probes) and examined using a MRC102

onfocal system (Bio-Rad Laboratories, Hercules,quipped with a Nikon E800 microscope and a k

on/argon laser. Quantitation of fluorescence intensityercent colocalization of two fluorescent probes (Alexafl88 and Alexafluor 568) was performed using the Morph Software (Universal Imaging Corp., DowningtowA).

. Results

.1. Anti-HER2/neu IgG3–(IL-2) and anti-HER2/neugG3–(GM-CSF) promote processing and presentationf OVA–ECDHER2 by BMDCs

One question is whether anti-HER2/neu IgG3–(IL-2) ornti-HER2/neu IgG3–(GM-CSF) bound ECDHER2 is morefficiently processed and presented by DCs. To ad

his issue a fusion protein with ECDHER2 joined to chickenVA, OVA–ECDHER2, was constructed and the ability

J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676 671

Fig. 2. Free IL-2 or GM-CSF in combination with OVA–ECDHER2 bound toanti-HER2/neu IgG3 fail to elicit effective antigen presentation by BMDCs.The level of IL-2 secreted by activated DO11.10 T-cells after a 48 h incuba-tion with BMDCs loaded with OVA–ECDHER2, or OVA–ECDHER2 mixedwith rhIL-2, rmGM-CSF, anti-HER2/neu IgG3, anti-HER2/neu IgG3 plusrhIL-2, anti-HER2/neu IgG3 plus rmGM-CSF, anti-HER2/neu IgG3–(IL-2) or anti-HER2/neu IgG3–(GM-CSF) was assessed as described inFig. 1.BMDCs loaded with 50�g/ml of the OVA protein served as a positive con-trol. Values represent the average concentration (pg/ml) of secreted IL-2 intriplicate wells. The error bars represent the range of triplicate determina-tions.

3.2. Free IL-2 or GM-CSF in combination withOVA–ECDHER2 bound to anti-HER2/neu IgG3 fails toelicit effective antigen presentation by BMDCs

GM-CSF and IL-2 are potent activators of DCs (Faulkneret al., 2001; Lutz et al., 1996). In addition, antigens inantibody immune complexes are more efficiently processedand presented by APCs (Anderson and Mosser, 2002;Schuurhuis et al., 2002). Thus, it is possible that BMDCs in-cubated with OVA–ECDHER2 bound to anti-HER2/neu IgG3in the presence of free IL-2 or GM-CSF may stimulate T-cellactivation to the levels observed with BMDCs incubatedwith OVA–ECDHER2 bound to anti-HER2/neu IgG3–(IL-2or GM-CSF). To examine this possibility, we incubatedBMDCs with anti-HER2/neu IgG3 bound OVA–ECDHER2

mixed with free rhIL-2 or rmGM-CSF and assessed T-cellactivation as above. In neither case was potent T-cell acti-vation observed (Fig. 2). Therefore, IL-2 or GM-CSF mustbe attached to anti-HER2/neu IgG3 for effective antigenprocessing and presentation of OVA–ECDHER2 by BMDCs.

3.3. ECDHER2 is principally localized in the earlyendosomes of murine DCs

As described above, ECDHER2 has been reported to traffico

Fig. 3. Early endosomal retention of ECDHER2 in murine BMDCs. MurineBMDCs were allowed to internalize ECDHER2 (green) for 90 min. Afterfixing, BMDCs were permeabilized and stained with Alexafluor 568 (red)-labeled anti-murine transferrin receptor (TfR, early endosomal marker) oranti-murine LAMP-1 (lysosomal marker), mounted and examined using con-focal microscopy. Yellow signals resulting from superimposition of greenand red images indicate colocalization of ECDHER2 with the markers used.BMDCs shown were from a 60× magnification field of acquired confocalimages (Supplementary data, Figs. S1 and S2). The percent of Alexaflour568 colocalizing with Alexaflour 488 was quantified using the MetamorphSoftware (seeTable 1).

2000). To determine if there is also inefficient traffickingof ECDHER2 to the antigen-processing compartments inmurine BMDCs, murine BMDCs were allowed to internalizeECDHER2 for 90 min and its colocalization with markers ofearly endosomes or lysosomes examined by confocal mi-croscopy. Alexafluor 488-labeled ECDHER2 was readily in-ternalized by BMDCs and was found within regions enrichedfor the transferrin receptor, indicating trafficking of ECDHER2

to the early endosomal compartments (Fig. 3; Supplementarydata, Fig. S1; Table 1). However, little to no ECDHER2 couldbe detected within the lysosomal compartments containingLAMP-1 (Fig. 3; Supplementary data, Fig. S2; Table 1).Similar results were obtained, when BMDCs were allowedto internalize ECDHER2 for 30 or 60 min (data not shown).Therefore, internalized ECDHER2 is retained in the early en-dosomes of murine BMDCs and does not efficiently traffic tothe lysosomal compartments similar to the trafficking patternof ECDHER2 observed in human DCs (Hiltbold et al., 2000).

Table 1Colocalization analysisa

BMDCs incubated with ECDHER2 %colocalization with

TfR LAMP-1

ECDHER2 48± 6 8 ± 3EE

wH ti-T Thev iza-t es ares

nly to the early endosome of human DCs (Hiltbold et al.,

CDHER2 plus Anti-HER2/neu IgG3 nd 5± 3CDHER2 plus Anti-HER2/neu IgG3–(IL-2) nd 19± 3

a The percent of Alexafluor 488-labeled ECDHER2 in BMDCs incubatedith ECDHER2, ECDHER2 plus anti-HER2/neu IgG3 or ECDHER2 plus anti-ER2/neu IgG3–(IL-2) colocalizing with Alexafluor 568-conjugated anfR or anti-LAMP-1 was quantified using the Metamorph Software.alues represent the average± standard deviation of the percent colocalion of four separate confocal images. Representative of these imaghown inSupplementary data, Figs. S1–S8; nd, not determined.

672 J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676

Fig. 4. Anti-HER2/neu IgG3–(IL-2) prevents retention of ECDHER2 inthe early endosomes of BMDCs and redirects its trafficking to theantigen-processing compartments. BMDCs were allowed to internalizeECDHER2/anti-HER2/neu IgG3–(IL-2) or ECDHER2/ani-HER2/neu IgG3complexes or ECDHER2 for 60 min and fixed. After permeabilization,BMDCs were counter-stained with Alexafluor 568-labeled (A) anti-murineLAMP-1 or (B) anti-murine H2-DM, mounted and examined using con-focal microscopy. BMDCs shown were from a 60× magnification field ofacquired confocal images (Supplementary data, Figs. S3–S8). The percentof Alexafluor 568 colocalizing with Alexafluor 488 for each DC shown wasquantified using the Metamorph Software.

3.4. ECDHER2 bound to anti-HER2/neu IgG3–(IL-2)efficiently traffics to the antigen-processingcompartments

BMDCs were allowed to internalize ECDHER2, ECDHER2

plus anti-HER2/neu IgG3 or ECDHER2 plus anti-HER2/neuIgG3–(IL-2) for 60 min. Little to no trafficking to lysosomalor peptide-loading compartments could be observed inBMDCs incubated with ECDHER2 or ECDHER2 plus anti-HER2/neu IgG3 (Fig. 4A and B;Supplementary data, Figs.S3–S8; Table 1). In contrast, ECDHER2 could be readilydetected within these compartments in BMDCs incubatedwith ECDHER2 plus anti-HER2/neu IgG3–(IL-2) (Fig. 4Aand B;Supplementary data, Figs. S3–S8; Table 1). Similarly,OVA–ECDHER2 that also does not efficiently traffic to

lysosomal compartments of BMDCs (Supplementary data,Fig. S9) could be observed to transit to these compartments,when bound to anti-HER2/neu IgG3–(IL-2) (Supplementarydata, Fig. S9). Thus, ECDHER2 bound to anti-HER2/neuIgG3–(IL-2) is not retained in the early endosomes butinstead moves to the antigen-processing compartments.

The kinetics of internalization of ECDHER2 bound toanti-HER2/neu IgG3–(IL-2) was also examined. After 5 minof incubation, ECDHER2 plus anti-HER2/neu IgG3–(IL-2)and ECDHER2 plus anti-HER2/neu IgG3 were found tocoat the surface of BMDCs with some internalization seen(Fig. 5). Only background staining was observed in BMDCsincubated with free ECDHER2 or with ECDHER2 mixed withnsIgG3 or nsIgG3–(IL-2) (Fig. 5). At the later times, BMDCsincubated with ECDHER2 plus anti-HER2/neu IgG3–(IL-2)showed robust intracellular staining that increased overtime, indicating continuous uptake of the complexes(Figs. 5 and 6A). Although, antibody bound ECDHER2 ismore efficiently taken up by BMDCs than free ECDHER2,the greatest uptake of ECDHER2 was observed in BMDCsloaded with ECDHER2 plus anti-HER2/neu IgG3–(IL-2)(Figs. 5 and 6A). Internalization of free ECDHER2by BMDCsbecame apparent after 15 min of incubation and graduallyincreased over time, although, never reaching the level ofinternalization seen with anti-HER2/neu IgG3–(IL-2) oranti-HER2/neu IgG3 bound ECDHER2 (Figs. 5 and 6A).I ofEw t inBRE pleeI

3a

t ed(I r-f dw1c tedw lit-t r1stTbI -ia he

nterestingly, at the later time points, internalizationCDHER2 in BMDCs incubated with ECDHER2 mixedith either nsIgG3 or nsIgG3–(IL-2) was less than thaMDC incubated with ECDHER2 alone (Figs. 5 and 6A).obust internalization of anti-HER2/neu IgG3–(IL-2) boundCDHER2 by BMDCs was consistently observed in multixperiments. Therefore, ECDHER2 bound to anti-HER2/neugG3–(IL-2) is very efficiently internalized by BMDCs.

.5. Intracellular localization of ECDHER2 bound tonti-HER2/neu IgG3–(GM-CSF)

The kinetics of internalization of ECDHER2 boundo anti-HER2/neu IgG3–(GM-CSF) was also examinFigs. 5 and 6B). After 5 min of incubation, anti-HER2/neugG3–(GM-CSF) bound ECDHER2 coated the outer cell suace of BMDCs. In contrast, ECDHER2 alone or associateith nsIgG3–(GM-CSF) showed no binding (Fig. 5). After5 min, anti-HER2/neu IgG3–(GM-CSF) bound ECDHER2

ould be observed inside BMDCs, while BMDCs incubaith ECDHER2alone or ns IgG3–(GM-CSF) showed very

le ECDHER2 internalization (Figs. 5 and 6B). However, afte5 min, anti-HER2/neu IgG3–(GM-CSF) bound ECDHER2

howed only a slight increase, while ECDHER2 internaliza-ion into BMDCs increased over time (Figs. 5 and 6B).he intracellular accumulation of anti-HER2/neu IgG3ound ECDHER2 was greater than that of anti-HER2/neu

gG3–(GM-CSF) bound ECDHER2 at all time points examned (Figs. 5 and 6B). Thus, although ECDHER2 bound tonti-HER2/neu IgG3–(GM-CSF) is efficiently targeted to t

J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676 673

Fig. 5. Time-dependent internalization into murine BMDCs of ECDHER2 under various conditions. Murine BMDCs were incubated with (from left to right)medium alone, ECDHER2, ECDHER2/anti-HER2/neu IgG3 complexes, ECDHER2plus nsIgG3, ECDHER2/anti-HER2/neu IgG3–(IL-2) complexes, ECDHER2plusnsIgG3–(IL-2), ECDHER2/anti-HER2/neu IgG3–(GM-CSF) or ECDHER2plus nsIgG3–(GM-CSF) for 5, 15, 30 or 60 min. BMDCs were then fixed, mounted andexamined using confocal microscopy. Images were collected at 40× magnification and are representative of multiple acquired confocal images. The intensityof Alexafluor 488 present in each panel was measured using the Metamorph Software (seeFig. 6).

surface of BMDCs, active internalization was apparent onlyduring 5 and 15 min of incubation and after 30 and 60 min in-ternalization of the complex became less apparent. DecreasedECDHER2 internalization by BMDCs was consistently ob-served in multiple experiments, when ECDHER2 was boundto anti-HER2/neu IgG3–(GM-CSF). As a result, the levelof internalized ECDHER2 in BMDCs loaded with ECDHER2

plus anti-HER2/neu IgG3–(GM-CSF) was insufficient to ac-curately determine its subcellular location by confocal mi-croscopy.

4. Discussion

The use of conventional adjuvants has not been success-ful in effectively enhancing the immunogenicity of anti-gens, such as ECDHER2 (Disis et al., 1996) that are re-tained in the early endosomes of DCs (Hiltbold et al., 2000)and not efficiently processed to peptides. Consistent withprevious findings demonstrating that antigens covalently ornon-covalently associated with IL-2 or GM-CSF exhibitenhanced immunogenicity (Chen et al., 1994; Harvill etal., 1996; Hazama et al., 1993a,b; Tao and Levy, 1993),mice immunized with ECDHER2 complexed with either anti-HER2/neu IgG3–(IL-2) or anti-HER2/neu IgG3–(GM-CSF)showed a stronger immune response to HER2/neu express-i( rn t

study, we have examined the influence of anti-HER2/neuIgG3–(IL-2) and anti-HER2/neu IgG3–(GM-CSF) onECDHER2 uptake, processing and presentation by BMDCs.

In initial vaccination studies, it was proposed that the en-hanced immunogenicity of ECDHER2 in mice immunizedwith ECDHER2 plus either anti-HER2/neu IgG3–(IL-2 andGM-CSF) merely reflected the immunostimulatory proper-ties of the attached IL-2 and GM-CSF (Dela Cruz et al., 2003).However, an alternative possibility was that anti-HER2/neuIgG3–(IL-2 or GM-CSF) may alter the intracellular traffick-ing of associated antigens like ECDHER2 that are retained inthe early endosomes of DCs and not effectively processed toantigenic peptides (Hiltbold et al., 2000). To test the possibil-ity that the antibody–cytokine fusion proteins lead to the moreeffective presentation of associated antigen, we constructed aunique reagent, OVA–ECDHER2, which allowed us to use theantibody–cytokine fusion proteins specific for ECDHER2 andmonitor the effects on antigen presentation by BMDCs us-ing existing T-cell read-out systems for OVA. Like ECDHER2,OVA–ECDHER2did not traffic to the antigen-processing com-partments of the BMDCs (Supplementary data, Fig. S9) andBMDCs incubated with OVA–ECDHER2 failed to stimulateOVA-specific T-cell activation.

Many studies have shown enhanced uptake and presenta-tion of antigens in immune complexes (Anderson and Mosser,2002; Schuurhuis et al., 2002; Yada et al., 2003). Although,E -cH y

ng tumors than did mice immunized with ECDHER2 aloneDela Cruz et al., 2003) or mixed with nsIgG3–(IL-2) osIgG3–(GM-CSF) (Dela Cruz et al., in press). In the presen

CDHER2 complexed with anti-HER2/neu IgG3 showed inreased internalization of ECDHER2(Figs. 5 and 6A), the anti-ER2/neu IgG3 bound ECDHER2 did not more effectivel

674 J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676

Fig. 6. BMDC-associated ECDHER2 over time. The intensity of Alexafluor 488 present in each panel shown inFig. 5 was measured using the MetamorphSoftware and the integrated intensity (105×) plotted. BMDCs were incubated with (A) ECDHER2 mixed with anti-HER2/neu IgG3–(IL-2), anti-HER2/neuIgG3, ECDHER2, ECDHER2 plus nsIgG3–(IL-2), ECDHER2 plus nsIgG3 or nothing (medium) or (B) ECDHER2 mixed with anti-HER2/neu IgG3–(GM-CSF),nsIgG3–(GM-CSF) or ECDHER2 alone.

traffic to the antigen-processing compartments of BMDCs(Fig. 4; Supplementary data, Figs. S4 and S7), and BMDCsincubated with anti-HER2/neu IgG3 bound OVA–ECDHER2

also did not stimulate OVA-specific T-cell activation (Fig. 1).Addition of free cytokine failed to effectively augment anti-gen presentation by DCs incubated with immune complexes(Fig. 2), indicating a requirement for a physical associationof antigen with the cytokine and suggesting a role forcytokine receptors in the altered trafficking and enhancedpresentation. Indeed, we found that antibodies to the IL-2 re-ceptor interfered with antigen presentation (data not shown).In addition, the trafficking profile we observed for anti-HER2/neu IgG3–(IL-2) bound ECDHER2 resembled that ofthe IL-2 /IL-2 receptor complex that moves to the lysosomalcompartments of T-cells and is degraded (t1/2∼ 70–180 min)(Duprez et al., 1994; Hemar et al., 1995; Lowenthal et al.,1986; Subtil et al., 1994; Yu and Malek, 2001).

Anti-HER2/neu IgG3–(IL-2) and anti-HER2/neuIgG3–(GM-CSF) differed in their influence on ECDHER2

internalization by BMDCs. The internalization of anti-HER2/neu IgG3–(IL-2) bound ECDHER2 was consistent

with the rapid (t1/2∼ 15 min) and continuous internalizationreported for IL-2 via IL-2 receptors (Chang et al., 1996;Lowenthal et al., 1986). In contrast, accumulation ofECDHER2 by BMDCs loaded with ECDHER2 plus anti-HER2/neu IgG3–(GM-CSF) reached a plateau after 15 min(Figs. 5 and 6B). A rapid decrease in cell surface GM-CSFreceptors was observed in WEHI-3BD+ cells incubated withexcess GM-CSF, with the lowest levels detected after 15 minof incubation (Walker and Burgess, 1987). In the presentstudy, the amount of GM-CSF in anti-HER2/neu IgG3–(GM-CSF) used exceeded the amount considered excess by atleast 33-fold, and thus, it is possible that the decrease in theaccumulation of ECDHER2 observed after 15 min reflects adecrease in the number of cell surface GM-CSF receptorson BMDCs. These observations would suggest that only arelatively small amount of internalized antibody–(GM-CSF)bound antigen is required for BMDCs to effectively stim-ulate T-cell activation. The level of internalized ECDHER2

in BMDCs loaded with ECDHER2 plus anti-HER2/neuIgG3–(GM-CSF) was insufficient to accurately determine itssubcellular location by confocal microscopy. Since BMDCs

J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676 675

incubated with either anti-HER2/neu IgG3–(GM-CSF) oranti-HER2/neu IgG3–(IL-2) bound OVA–ECDHER2 elicitcomparable levels of T-cell activation, it is unclear if thelarge amount of antibody–(IL-2) bound antigen internalizedby BMDCs is required for optimal T-cell activation or ifonly a small quantity of antigen targeted to the appropriatecompartment is sufficient.

A side-by-side comparison of the differentECDHER2–cytokine associations will be required todetermine whether a non-covalent physical linkage betweenECDHER2 and antibody–(IL-2 or GM-CSF) fusion proteinis superior to a direct fusion between ECDHER2 and IL-2or GM-CSF in eliciting an immune response to ECDHER2.However, there are many reasons why antibody–(IL-2 orGM-CSF) bound ECDHER2 may be superior in elicitingan immune response to ECDHER2. The presence of twocytokine molecules in each antibody fusion protein maypromote a more stable interaction with the cytokine receptorson BMDCs due to increased avidity. This may trigger eithermore efficient cytokine receptor-mediated endocytosis ormore effective BMDC activation. In addition, the interactionof the antibody–cytokine fusion protein with the Fc-receptoron BMDCs may stabilize binding to the cytokine receptors,thereby facilitating internalization and efficient antigenprocessing.

In conclusion, antibody–(IL-2 and GM-CSF) fusionp uner y top n byB erentT anti-g s,s a( area M-C igensa ducea L-2a e im-m gestv2r som einsm oeim

A

sonP , CA)f to

Andrew C. Melton for kindly helping with the quantitationof fluorescence intensity and analysis of fluorescence colo-calization (Department of Physiology, UCLA). This workwas supported in part by grants CA86915, CA107023 andCA087990 from NIH/NCI, AI39187 and AI29470, the 2002AACR-California Department of Health Services CareerDevelopment Award in Gender-related Cancer Research,the 2004 Brian D. Novis International Myeloma Foundation(IMF) Senior Grant Award, and the Tumor ImmunologyTraining Grant (5-T32-CA009120-28).

Appendix A. Supplementary data

Supplementary data associated with this article canbe found in the online version atdoi:10.1016/j.molimm.2005.04.007.

References

Anderson, C.F., Mosser, D.M., 2002. Cutting edge: biasing immune re-sponses by directing antigen to macrophage Fc gamma receptors. J.Immunol. 168, 3697–3701.

Chang, D.Z., Wu, Z., Ciardelli, T.L., 1996. A point mutation ininterleukin-2 that alters ligand internalization. J. Biol. Chem. 271,13349–13355.

C einsyte-

787.D C.,

ina-

/

D Re--anti-

D o thehu-

D nera-thancer

D .D.,cinesol.

D raf-f T

F M.,ncesone

H f antion.

H igue,an,

roteins should be effective in eliciting a protective immesponse to a broad range of antigens. Their abilitromote efficient antigen processing and presentatioMDCs suggests they can be used to augment many diffH-cell-dependent immune responses. In particular,ens, such as MUC-1 (Hiltbold et al., 2000) and pathogenuch as HIV-1 (Turville et al., 2004) and mycobacteriPieters, 2001) that are not efficiently processed by APCsppropriate candidates for use with antibody–(IL-2 and GSF) fusion proteins. Indeed, since a broad range of antre easily dansylated, it may not be necessary to prontigen-specific fusion protein but instead anti-DNS-(Ind GM-CSF) fusion proteins may be used to enhance thunogenicity of diverse antigens. Our observations sug

accination with ECDHER2 plus anti-HER2/neu IgG3–(IL-and/or GM-CSF) may elicit a protective HER2/neu-specific

esponse in patients with HER2/neu positive malignancier who are at high-risk to develop HER2/neu expressingalignancies. Moreover, treatment with the fusion protay direct circulating ECDHER2 (Hayes et al., 2001; Visct al., 2000) to DCs and elicit a robust HER2/neu-specific

mmune response in patients with HER2/neu expressingalignancies.

cknowledgements

We are grateful to Dr. Lars Karlsson (Johnson & Johnharmaceutical Research and Development, San Diego

or kindly supplying anti-H2-DM. We are also grateful

hen, T.T., Tao, M.H., Levy, R., 1994. Idiotype–cytokine fusion protas cancer vaccines: relative efficacy of IL-2, IL-4, and granulocmacrophage colony-stimulating factor. J. Immunol. 153, 4775–4

ela Cruz, J.S., Lau, S.Y., Ramirez, E.M., De Giovanni,Forni, G., Morrison, S.L., Penichet, M.L., 2003. Protein vacction with the HER2/neu extracellular domain plus anti-HER2/neuantibody–cytokine fusion proteins induces a protective anti-HER2neuimmune response in mice. Vaccine 21, 1317–1326.

ela Cruz, J.S., Trinh, K.R., Morrison, S.L., Penichet, M.L., 2000.combinant anti-human HER2/neu IgG3–(GM-CSF) fusion protein retains antigen specificity and cytokine function and demonstratestumor activity. J. Immunol. 165, 5112–5121.

ela Cruz, J.S., Morrison, S.L., Penichet, M.L., in press. Insights intmechanism of anti-tumor immunity in mice vaccinated with theman HER2/neu extracellular domain plus anti-HER2/neu IgG3-(IL-2)or anti-HER2/neu IgG3-(GM-CSF) fusion protein. Vaccine.

isis, M.L., Grabstein, K.H., Sleath, P.R., Cheever, M.A., 1999. Getion of immunity to the HER-2/neu oncogenic protein in patients wibreast and ovarian cancer using a peptide-based vaccine. Clin. CRes. 5, 1289–1297.

isis, M.L., Gralow, J.R., Bernhard, H., Hand, S.L., Rubin, WCheever, M.A., 1996. Peptide-based, but not whole protein, vacelicit immunity to HER-2/neu, oncogenic self-protein. J. Immun156, 3151–3158.

uprez, V., Smoljanovic, M., Lieb, M., Dautry-Varsat, A., 1994. Tficking of interleukin 2 and transferrin in endosomal fractions olymphocytes. J. Cell Sci. 107 (Pt. 5), 1289–1295.

aulkner, L., Buchan, G., Lockhart, E., Slobbe, L., Wilson, M., Baird,2001. IL-2 linked to a peptide from influenza hemagglutinin enhaT cell activation by affecting the antigen-presentation function of bmarrow-derived dendritic cells. Int. Immunol. 13, 713–721.

arvill, E.T., Fleming, J.M., Morrison, S.L., 1996. In vivo properties oIgG3–IL-2 fusion protein: a general strategy for immune potentiaJ. Immunol. 157, 3165–3170.

ayes, D.F., Yamauchi, H., Broadwater, G., Cirrincione, C.T., RodrS.P., Berry, D.A., Younger, J., Panasci, L.L., Millard, F., Dugg

676 J.S. Dela Cruz et al. / Molecular Immunology 43 (2006) 667–676

D.B., Norton, L., Henderson, I.C., 2001. Circulating HER-2/erbB-2/c-neu (HER-2) extracellular domain as a prognostic factor in patientswith metastatic breast cancer: cancer and leukemia group B study8662. Clin.Cancer Res. 7, 2703–2711.

Hazama, M., Mayumi-Aono, A., Asakawa, N., Kuroda, S., Hinuma,S., Fujisawa, Y., 1993a. Adjuvant-independent enhanced immune re-sponses to recombinant herpes simplex virus type 1 glycoprotein Dby fusion with biologically active interleukin-2. Vaccine 11, 629–636.

Hazama, M., Mayumi-Aono, A., Miyazaki, T., Hinuma, S., Fujisawa, Y.,1993b. Intranasal immunization against herpes simplex virus infectionby using a recombinant glycoprotein D fused with immunomodulatingproteins, the B subunit ofEscherichia coli heat-labile enterotoxin andinterleukin-2. Immunology 78, 643–649.

Hemar, A., Subtil, A., Lieb, M., Morelon, E., Hellio, R., Dautry-Varsat,A., 1995. Endocytosis of interleukin 2 receptors in human T lympho-cytes: distinct intracellular localization and fate of the receptor alpha,beta, and gamma chains. J. Cell. Biol. 129, 55–64.

Hiltbold, E.M., Vlad, A.M., Ciborowski, P., Watkins, S.C., Finn, O.J.,2000. The mechanism of unresponsiveness to circulating tumor anti-gen MUC1 is a block in intracellular sorting and processing by den-dritic cells. J. Immunol. 165, 3730–3741.

Lowenthal, J.W., MacDonald, H.R., Iacopetta, B.J., 1986. Intracellularpathway of interleukin 2 following receptor-mediated endocytosis.Eur. J. Immunol. 16, 1461–1463.

Lutz, M.B., Assmann, C.U., Girolomoni, G., Ricciardi-Castagnoli, P.,1996. Different cytokines regulate antigen uptake and presentationof a precursor dendritic cell line. Eur. J. Immunol. 26, 586–594.

Penichet, M.L., Challita, P.M., Shin, S.U., Sampogna, S.L., Rosen-blatt, J.D., Morrison, S.L., 1999. In vivo properties of three humanHER2/neu-expressing murine cell lines in immunocompetent mice.Lab. Anim. Sci. 49, 179–188.

P . Aman

P ia in

R A.,.S.,nedcells.

Salazar, L.G., Fikes, J., Southwood, S., Ishioka, G., Knutson, K.L., Goo-ley, T.A., Schiffman, K., Disis, M.L., 2003. Immunization of can-cer patients with HER-2/neu-derived peptides demonstrating high-affinity binding to multiple class II alleles. Clin. Cancer Res. 9, 5559–5565.

Schuurhuis, D.H., Ioan-Facsinay, A., Nagelkerken, B., van Schip, J.J.,Sedlik, C., Melief, C.J., Verbeek, J.S., Ossendorp, F., 2002. Antigen-antibody immune complexes empower dendritic cells to efficientlyprime specific CD8+ CTL responses in vivo. J. Immunol. 168,2240–2246.

Slamon, D.J., Clark, G.M., Wong, S.G., Levin, W.J., Ullrich, A., McGuire,W.L., 1987. Human breast cancer: correlation of relapse and survivalwith amplification of the HER-2/neu oncogene. Science 235, 177–182.

Subtil, A., Hemar, A., Dautry-Varsat, A., 1994. Rapid endocytosis of in-terleukin 2 receptors when clathrin-coated pit endocytosis is inhibited.J. Cell Sci. 107 (Pt. 12), 3461–3468.

Tao, M.H., Levy, R., 1993. Idiotype/granulocyte-macrophage colony-stimulating factor fusion protein as a vaccine for B-cell lymphoma.Nature 362, 755–758.

Turville, S.G., Santos, J.J., Frank, I., Cameron, P.U., Wilkinson, J.,Miranda-Saksena, M., Dable, J., Stossel, H., Romani, N., Piatak Jr.,M., Lifson, J.D., Pope, M., Cunningham, A.L., 2004. Immunodefi-ciency virus uptake, turnover, and 2-phase transfer in human dendriticcells. Blood 103, 2170–2179.

Visco, V., Bei, R., Moriconi, E., Gianni, W., Kraus, M.H., Muraro, R.,2000. ErbB2 immune response in breast cancer patients with solublereceptor ectodomain. Am. J. Pathol. 156, 1417–1424.

Walker, F., Burgess, A.W., 1987. Internalisation and recycling of thegranulocyte-macrophage colony-stimulating factor (GM-CSF) recep-tor on a murine myelomonocytic leukemia. J. Cell. Physiol. 130, 255–

Y a, A.,sen-IIB-nol.

Y et-

Y diated

enichet, M.L., Dela Cruz, J.S., Shin, S.U., Morrison, S.L., 2001recombinant IgG3–(IL-2) fusion protein for the treatment of huHER2/neu expressing tumors. Hum. Antibodies 10, 43–49.

ieters, J., 2001. Entry and survival of pathogenic mycobactermacrophages. Microbes Infect. 3, 249–255.

ibas, A., Butterfield, L.H., McBride, W.H., Dissette, V.B., Koh,Vollmer, C.M., Hu, B., Chen, A.Y., Glaspy, J.A., Economou, J1999. Characterization of antitumor immunization to a defimelanoma antigen using genetically engineered murine dendriticCancer Gene Ther. 6, 523–536.

261.ada, A., Ebihara, S., Matsumura, K., Endo, S., Maeda, T., Nakamur

Akiyama, K., Aiba, S., Takai, T., 2003. Accelerated antigen pretation and elicitation of humoral response in vivo by FcgammaRand FcgammaRI/III-mediated immune complex uptake. Cell Immu225, 21–32.

arden, Y., Sliwkowski, M.X., 2001. Untangling the ErbB signalling nwork. Nat. Rev. Mol. Cell Biol. 2, 127–137.

u, A., Malek, T.R., 2001. The proteasome regulates receptor-meendocytosis of interleukin-2. J. Biol. Chem. 276, 381–385.