Enhanced HER-2/neu-specific antitumor immunityby cotransduction of mouse dendritic cells withtwo genes encoding HER-2/neu and alpha tumornecrosis factor

Zhuang Chen,1 Hui Huang,1 Tim Chang,1 Svein Carlsen,1 Anurag Saxena,2 Robert Marr,3

Zhou Xing,3 and Jim Xiang1

Departments of 1Oncology and 2Pathology, Cancer Research Unit, Saskatchewan Cancer Agency, University ofSaskatchewan, Saskatoon, Saskatchewan, Canada; and 3Department of Pathology and Molecular Medicine,McMaster University, Hamilton, Ontario, Canada.

The present study uses an in vivo murine tumor model expressing the human HER-2/neu antigen to evaluate the potential vaccine

using dendritic cells (DCs) infected with adenovirus AdVHER-2. We first investigated whether infected DCs (DCHER - 2 ) engineered

to express HER-2/neu could induce HER-2/neu-specific immune responses. Our data showed that ( i ) AdVHER2- infected DCHER - 2

expressed HER-2/neu by Western blot and flow cytometric analysis, and ( ii ) vaccination of mice with DCHER - 2 induced HER-2/

neu-specific cytotoxic T- lymphocyte (CTL) responses, but protected only 25% of vaccinated mice from challenge of 3�105

MCA26/HER-2 tumor cells. Further, to enhance the efficacy of DCHER - 2 vaccine, we coinfected DCs with both AdVHER-2 and

AdVTNF-a. The infected DCs (DCHER - 2 / TNF -a ) displayed the expression of both HER-2/neu and TNF-a by flow cytometric and

ELISA analysis. We next investigated whether DCHER - 2 / TNF -a could induce stronger HER-2/neu-specific immune responses. We

found that DCHER - 2 / TNF -a displayed up-regulation of immunologically important CD40, CD86, and ICAM-I molecules compared

with DCHER - 2, indicating that the former ones are more mature forms of DCs. Vaccination of DCHER - 2 / TNF -a induced stronger

allogeneic T-cell proliferation and 36% enhanced HER-2/neu-specific T-cell responses in vitro than DCHER - 2 cells. More

importantly, it stimulated the significant anti–HER-2/neu immunity in vivo, which protected 8/8 mice from challenge of 3�105

MCA26/HER-2 tumor cells. Therefore, DCs genetically engineered to express both the tumor antigen and cytokines such as TNF-aas an immunoadjuvant are likely to represent a new direction in DC vaccine of cancer.

Cancer Gene Therapy (2002) 9, 778–786 doi:10.1038/sj.cgt.7700498

Keywords: engineered dendritic cell vaccine; recombinant adenovirus; HER-2; alpha necrosis factor; antitumor immunity

HER-2/neu is a proto-oncogene encoding a 185-kDaprotein with homology to the epidermal growth factor

receptor (EGFR) family.1 This transmembrane proteinconsists of a cystein-rich extracellular domain (ECD) thatfunctions in ligand binding and an intracellular domain(ICD) with kinase activity. It is amplified and overexpressedin several human cancers including breast, ovarian, andgastric carcinomas.2,3 Patients with HER-2/neu-positivetumors are associated with a poor prognosis.4,5 Studies ofpatients with HER2/neu-expressing tumors have demon-strated the existence of humoral and T-cell responses againstHER-2/neu6,7 though insufficient to prevent tumor out-growth. In addition, HER-2/neu-specific T-cell responses

can also be generated in patients8,9 and in animal models10

by immunization with HER-2/neu-derived peptides. Takentogether, these studies provide the support for HER-2/neuas a candidate for tumor vaccines.

The current approved HER-2/neu immunotherapy ispassive transfer of a humanized HER-2/neu monoclonalantibody (Herceptin), which showed inhibition of tumorgrowth in only limited population of HER-2/neu patients.11

Although clinical studies using HER-2/neu peptide vaccineinduced HER-2/neu-specific CD4+ and CD8+ T cells inpatients with breast cancer, it could not result in any tumorrejection.8,9 Therefore, a strategic goal of current HER-2/neu-specific immunotherapy has become the induction ofstronger tumor-specific cytotoxic T- lymphocyte (CTL)responses.

Genetic DNA vaccines using plasmid DNA expressingHER-2/neu have been reported to induce both HER-2/neu-specific humoral and T-cell immune responses, whichpartially protect mice from rechallenge of HER-2/neu-positive tumor cells, and reduce tumor development in rat

Received May 27, 2002.

Address correspondence and reprint requests to: Dr Jim Xiang, Cancer

Research Unit, Saskatchewan Cancer Agency, 20 Campus Drive,Saskatoon, Saskatchewan S7N 4H4, Canada. E-mail: jxiang@scf.sk.ca

Cancer Gene Therapy (2002) 9, 778 – 786D 2002 Nature Publishing Group All rights reserved 0929-1903 /02 $25.00

www.nature.com/cgt

HER-2/neu transgenic mice.12–14 DNA vaccines requirehost antigen-presenting cells (APCs) to present HER-2/neuantigen for priming T-cell responses. However, The hostAPCs in patients and tumor-bearing mice are oftenfunctionally impaired.15–17 As a circumvention strategyaimed at this problem, investigators have begun to employ invitro cytokine-stimulated bone marrow (BM)-deriveddendritic cells (DCs), which are fully functional in primingT-cell responses.18,19

DCs are one of the most potent APCs. Because of thecritical roles of DCs in the generation of primary immuneresponses,20 an important avenue of investigation is theirpotential for modulating immunologic functions such as theinduction of tumor immunity. It has been reported that DCspulsed with MHC class I–restricted tumor peptides wereable to induce CTL-dependent antitumor immune re-sponses in vitro and in vivo.21,22 It has also been reportedthat DCs engineered to express tumor antigens by adeno-virus-mediated gene transfer can elicit tumor-specific CTLresponses.23,24 Furthermore, Lapointe et al25 have demon-strated that these engineered DCs can generate T cellsrecognizing multiple MHC class I and class II tumorepitopes. Of perhaps more importance to cancer immuno-therapy, Labeur et al26 have recently shown that theinduction of antitumor immunity by DC vaccines iscorrelated with the maturation stage(s) of the DCs. In ourprevious report, we have also shown that DCs infected withadenovirus expressing transgene TNF-a underwent aug-mented cellular maturation.27 These mature DCs furtherinduced strong T-cell proliferation/activation and efficientantitumor immunity in animal models.

In this study, we constructed a recombinant adenovirusAdVHER-2 expressing HER-2/neu tumor antigen andinvestigated the anti–HER-2/neu immune responses byusing AdVHER-2–infected DCs in animal models. Inaddition, we further compared the vaccine efficiency by usingDCs infected with AdVHER-2 alone and DCs coinfectedwith two adenoviruses AdVHER-2 and AdVTNF-a encod-ing two genes, HER-2/neu and TNF-a, respectively.

Materials and methods

Cell lines, antibodies, cytokines, and animals

A mouse colon cancer cell line MCA26 (H-2Kd) and ahuman breast carcinoma cell line T-47D were obtained fromthe American Type Culture Collection (ATCC, Rockville,MD), and maintained in Dulbecco’s modified Eagle’smedium (DMEM; Gibco, Gaithersburg, MD) supplementedwith 10% fetal calf serum (FCS). 293 cell line (adenoviralE1 transformed human embryonic kidney cells ) was alsoobtained from ATCC and maintained in minimum essentialmedium with Earle’s salts (EMEM; Gibco) supplementedwith 10% FCS. Monoclonal rat anti -mouse H-2Kd, Iad,FasL, CD3, CD4, CD8, CD11b, CD11c, CD25, CD40,CD80, CD86, and ICAM-1 antibodies, and the mouse anti -human HER-2/neu antibody were all purchased fromPharmingen (San Diego, CA). The FITC-conjugated goatanti - rat and anti -mouse IgG antibodies were purchasedfrom Bio/Can Scientific (Mississauga, Ontario, Canada).

Recombinant mouse granulocyte -macrophage colony-stim-ulating factor (GM-CSF) and interleukin-4 (IL-4) werepurchased from Endogen (Woburn, MA) and R & DSystems (Minneapolos,MN), respectively. FemaleC57BL/6(H-2Kb) and BALB/c (H-2Kd) mice were obtained fromCharles River Laboratories (St. Laurent, Quebec, Canada).All mice were housed in the animal facility of the Sas-katoon Cancer Center, and all animal experiments werecarried out according to the guidelines of the CanadianCouncil for Animal Care.

Transfection of MCA26 cells with the expression vectorpcDNA-HER2

Total RNA was extracted from T47-D tumor cells usingRNeasy Mini Kit (Qiagen, Mississauga, Canada). A 2.0-kbcDNA fragment coding for the open reading frame ofHER-2 molecule (amino acids 1 to 689) was cloned byreverse transcription polymerase chain reaction (RT-PCR)from a cDNA library of T47-D cells using PFU polymerase(Fig 1A). The cloned HER-2/neu cDNA fragmentincludes the signal peptide (amino acids 1–21) and thefull extracellular (amino acids 22–652) and transmembrane(amino acids 653–675) domains, as well as a short, partialintracellular (amino acids 676–690) domain of HER-2/neu

Figure 1 Expression of HER-2 /neu molecule. A: RT-PCR analysis.RNA was obtained from T47-D and 293 cell lines. The first -strand

cDNA was synthesized from RNA using reverse transcriptase; PCRs

were conducted using primers sets for HER-2 /neu and the control‘‘housekeeping gene,’’ GAPDH. B: Western blot analysis. Protein

extracts were obtained from MCA26, MCA26/HER-2, DCHER - 2, and

DCpLpA and loaded to each well of a polyacrylamide gel under

reducing conditions. The transferred nitrocellulose paper strips wereincubated with the mouse anti–HER-2 /neu antibody followed by the

peroxidase-conjugated goat anti -mouse IgG antibody. Detection of

HER-2 /neu expression was accomplished by using enhanced

chemiluminescence reagent.

Cancer Gene Therapy

HER-2- specific immunity of engineered dendritic cell vaccineZ Chen et al

779

molecule. Two PCR primers were used, namely the senseprimer (50 -CAGTG AGCAC CATGG AGCTG G-30 ) andthe antisense primer (50 -CAGTC TCCGC ATCGT GTACTTCC-3’) with Super Script2 preamplification system(Gibco) according to the manufacturer’s instructions.The cloned cDNA fragment of HER-2/neu was ligatedinto the pCR2.1 vector ( Invitrogene, Carlsbed, CA) toform the pCR2.1/HER-2. The HER-2/neu gene sequencewas then verified by the dideoxy nucleotide sequencingmethod. The HER-2/neu gene fragment was furthercloned into the plasmid pcDNA3.1 (Invitrogen, SanDiego, CA) to form the expression vector pcDNA/HER-2. Twenty million MCA26 cells were resuspendedin 0.7 mL phosphate-buffered saline (PBS) and mixedwith 0.3 mL PBS containing 10 �g pcDNA/HER-2 DNA.Tumor cells were transfected with pcDNA/HER-2 using aBio-Rad gene pulser (Bio-Rad Lab, Mississauga, Ontario,Canada) with parameters of 300 V and 125 �F capa-citance. The transfected cells were selected for growth inmedium containing 4 mg/mL G418 (Geneticin; Gibco).The selected clone MCA26/HER-2 was used for analysisof HER-2/neu expression by flow cytometry.

Recombinant adenovirus AdVHER-2

The cloned cDNA HER-2/neu fragment was ligated into thepLpA plasmid to form the adenoviral vector pLpA-HER-2.The construction of recombinant adenovirus AdVHER-2from the vectors pLpA-HER-2 and pJM17 was conductedas previously reported.28 The recombinant adenovirusesAdVLacZ expressing the E. coli b -galactosidase (a markergene) and AdVpLpA (i.e., with no gene insert used as acontrol adenoviral vector were previously constructed in ourlaboratory28). The AdVGFP and AdVTNF-a expressing thegreen fluorescent protein (GFP) and the mouse TNF-a wasobtained from Dr Z Xing, McMaster University, Ontario,and Dr S Tikoo, University of Saskatchewan, Saskatchewan,Canada, respectively. All these E1-deleted replication-deficient recombinant adenoviruses under the control ofthe cytomegalovirus (CMV) early / immediate promoter /enhancer were amplified in 293 cell line, purified by cesiumchloride ultracentrifugation and stored at �808C.

Preparation of BM-derived DCs

The preparation of BM-derived DCs was previouslydescribed.27 Briefly, BM cells prepared from femora andtibiae of normal BABL/c mice were depleted of red bloodcells with 0.84% ammonium chloride and plated in DC cul-ture medium [DMEM plus 10% FCS, GM-CSF (20 ng/mL)and IL-4 (20 ng/mL)]. On day 3, the nonadherent gra-nulocytes, and T and B cells were gently removed and freshmedia were added, and 2 days later the loosely adherentproliferating DC aggregates were dislodged and replated. Onday 6, the nonadherent cells were harvested. The DCsgenerated in this manner displayed: ( i ) typical morphologicfeatures of DCs (i.e., numerous dendritic processes) and (iiexpression of MHC class I (H-2Kb) and II ( Iab) antigens,costimulatory molecules (CD40, CD80 and CD86), andadhesion molecules ( ICAM-1, CD11b, and CD11c). These

DCs were then used for in vitro AdVHER-2 and/orAdVTNF-a infection.

Adenoviral infection of DCs

To test the amenability of DCs to adenoviral infection, serialdilutions of AdVLacZ stock (2�1010 pfu/mL) were addedto triplicate cultures of DCs in 96-well plates (1�105 cells /well ) to form different multiplicity of infection (MOI). Thecells were incubated with the adenovirus in 293 serum-freemedium (Gibco) for 2 hours at 378C, then the medium wasreplaced with DMEM/10% FCS and the cells incubated foran additional 24 hours at 378C. To assess b -galactosidaseexpression,28 the cells were fixed in formaldehyde/gluta-raldehyde, then stained and counterstained with X-gal andnuclear fast red, respectively. The proportions of positive( i.e., blue staining) cells were determined from triplicatewells and taken as the percentage of transduction. ControlDCs infected with AdVpLpA (DCpLpA) did not exhibit anyintrinsic b -galactosidase activity or false-positive staining.We observed a dose-dependent response to the adenoviralinfections, with maximal staining (82%) at an MOI of�100. Therefore, an MOI of 100 was selected for infectionof DCs with AdVHER-2 and/or AdVTNF-a in this study.DCs infected with AdVHER-2 and AdVHER-2/AdVTNF-a were termed DCHER-2 and DCHER-2 /TNF -a, respectively.To enact this, following viral adsorption for 1 h at 378C insix-well culture plates, the DC culture medium was replacedwith DMEM plus 20% FCS. The cells were incubated for anadditional 24 hours at 378C, and then harvested forphenotypic analysis by flow cytometry and for immunizationof animals. The culture supernatants of the transfectedDCHER-2 /TNF -awere assayed for TNF-a by enzyme- linkedimmunosorbent assay (ELISA) using TNF-a ELISA kit(Endogen, Woburn, MA).

To examine the efficiency of coinfection of DCs with twoindividual adenoviral vectors, we infected DCs with eitherAdVGFP or AdVTNF-a, or both of these two viruses atMOI of 100 for each virus. After 24-hour culture asdescribed above, DCs infected with AdVGFP (DCGFP) weresubjected to flow cytometric analysis. DCs infected withAdVTNF-a (DCTNF-a ) or with both AdVTNF-a andAdVGFP (DCGFP /TNF -a ) were fixed and processed using acommercial kit (Cytofix /CytoPerm Plus with GolgiPlug;Pharmingen), and stained with PE-conjugated anti–TNF-aantibody according to the manufacturer’s protocols. Thesetransfected DCs were then subjected to flow cytometricanalysis.

Flow cytometric analysis

For phenotypic analysis of tumor cells, MCA26 andMCA26/HER-2 cells were stained for 30 minutes on icewith antibodies specific for H-2Kd, Iad, Fas, or HER-2/neu (5 �g/mL each), respectively. For phenotypic analysisof DCs, DCHER-2 and DCHER-2 /TNF -a were incubatedwith antibodies specific for H-2Kd, Iad, CD11b, CD11c,CD40, CD80, CD86, ICAM-1, and HER-2/neu on ice for30 minutes, respectively. For phenotypic analysis of Tcells, naive and MCA26/HER-2–specific T cells asdescribed in the following sections were incubated withantibodies specific for CD3, CD4, CD8, CD25, and FasL

Cancer Gene Therapy

HER-2- specific immunity of engineered dendritic cell vaccineZ Chen et al

780

on ice for 30 minutes, respectively. These cells werewashed three times, then incubated with FITC-conjugatedgoat anti - rat or mouse IgG antibody (1:60). After threemore washes with PBS, these cells were analyzed by flowcytometry. Isotype-matched monoclonal antibodies wereused as controls.

Western blot analysis

The expression of HER-2/neu in MCA26/HER-2 andDCHER-2 cells was assessed using Western blot. Briefly, theprotein extract was prepared from these cells using theextraction buffer containing 125 mM tris(hydroxymethyl)a-minomethane, 0.05% sodium dodecyl sulfate, and 10%b -mercaptoethanol. Cell extracts were harvested from thesupernatant after centrifugation at 1000�g for 5 minutes.The supernatants containing the protein samples wereelectrophoresed through 10% gels and transferred ontonitrocellulose papers; blots were blocked using PBScontaining 10% bovine serum albumin (BSA) and incubatedwith the anti–HER-2/neu antibody followed by the goatanti -mouse IgG antibody conjugated with peroxidase.Detection of HER-2/neu was accomplished by usingenhanced chemiluminescence reagent according to themanufacturer’s protocol (New England Nuclear Life ScienceProducts, Boston, MA).

Allogeneic mixed lymphocyte reactions

Naive T cells were purified from C57BL/6 mousesplenocytes as nylon wool nonadherent cells.29 The primarymixed lymphocyte reactions MLRs) were performed aspreviously described.29 Briefly, graded doses of irradiatedDCs, DCHER-2, and DCHER-2 /TNF -a (3000 rad) werecocultured in 96-well plates with constant numbers(2�105) of allogeneic T cells from C57BL/6 mouse. After3 days, T-cell proliferation was measured using an overnight[3H]thymidine (1 mCi/mL, Amersham Canada) uptakeassay (1 �Ci/well ). The levels of [3H]thymidine incorpo-ration into the cellular DNA were determined by liquidscintillation counting.

Cytotoxic activity assay

Red blood cell–depleted splenic lymphocytes from micevaccinated with DCHER-2 and DCHER-2 /TNF -a were cocul-tured in 24-well plates with irradiated MCA26/HER-2 cells(40,000 rad), using 5�106 lymphocytes and 2�105

MCA26/HER-2 cells per 2 mL of DMEM plus 10% FCS.After four days, the activated T cells were harvested byFicoll -Hypaque gradient centrifugation and used as ef-fector cells in a chromium-release assay against radio-labeled MCA26 and MCA26/HER-2 target cells. Thetarget cells (104 /well ) were incubated for 8 hours intriplicate cultures with effector cells at various effector /target ratios. The percent specific lysis was calculated usingthe formula:

½ðexperimental CPM�spontaneous CPMÞ=ðmaximal CPM�spontaneous CPMÞ��100

The spontaneous count per minute (CPM) release in theabsence of effector cells was less than 10% of specific lysis;maximal CPM release was effected by adding 1% TritonX-100 to the cells.29

Vaccination of mice with DCHER -2 and DCHER -2 / TNF -a

For evaluation of tumor prevention, mice were vaccinatedsubcutaneously (s.c. ) with 1�106 DCHER-2 and DCHER-2 /

TNF -a cells twice with a 7-day interval. Ten days subsequentto the second vaccination, the mice (n=8 per group) werechallenged by s.c injection of 3�105 MCA26 or MCA26/HER-2 cells. Animal mortality and tumor growth weremonitored daily for up to 60 days; for humanitarian reasons,all mice with tumors that achieved a size of 1.5 cm indiameter were sacrificed.

Results

HER-2 expression on MCA26/HER-2 and DCHER -2 cells

The mouse colon cancer cell line MCA26 and DCs weretransfected and infected with the expression vector pcDNA-HER-2 and the adenoviral vector AdVHER-2, respectively.The HER-2/neu expression of transfected MCA26/HER-2and infected DCHER-2 cells was then examined by Westernblot and flow cytometry. As shown in Figure 1B andFigure 2, MCA26/HER-2 and DCHER-2 cells displayedsignificant amount of cellular and cell -surface HER-2/neuexpression, whereas the original MCA26 tumor cells and the

Figure 2 Flow cytometric analysis of HER-2 /neu expression. The

expression of HER-2/neu in MCA26, MCA26/HER-2, DCHER - 2, and

DCpLpA cells (solid lines) was analyzed by flow cytometry using themouse anti–HER-2 /neu and the FITC-conjugated goat anti -mouse

IgG antibodies. Isotype-matched monoclonal antibodies (dotted

lines) were used as controls.

Cancer Gene Therapy

HER-2- specific immunity of engineered dendritic cell vaccineZ Chen et al

781

control adenovirus- infected DCpLpA cells showed noexpression of HER-2/neu molecule.

Coinfection of DCs with AdVHER-2 and AdVTNF-a

To enhance DC maturation, we coinfected DCs withAdVHER-2 and AdVTNF-a to generate genetically mo-dified DCHER-2 /TNF -a cells. The expression of the cellsurface HER-2/neu on DCHER-2 /TNF -a cells was similar tothat on DCHER-2 cells (data not shown). To examine TNF-aexpression of DCHER-2 /TNF -a, we conducted ELISA assays.Our preliminary data showed that the concentrations ofTNF-a in the supernatants of DCpLpA and DCHER-2 /TNF -acells were approximate 0.1 and 8 ng/mL, respectively. Tofurther examine the efficiency of cotransfection with twoindividual adenoviral vectors, we coinfected DCs withAdVGFP and AdVTNF-a simultaneously. As shown inFigure 3, around 90% of DCGFP and DCTNF-a cells ex-pressed GFP and TNF-a, respectively; and nearly 60% ofDCTNF-a /GFP cells expressed both TNF-a and GFP,indicating that expression of two transgenes in individualDC by coinfection of two different adenoviral vectors isefficient.

DCHER - 2 / TNF -a cells up-regulate expression of CD40,CD86 and ICAM-1

We then analyzed the impact of AdVTNF-a or AdVHER-2infection on the expression of a number of cell surfacemarkers (MHC class I and II antigens, CD11b, CD11c,CD40, CD80, CD86, and ICAM-1) in DCs. As shown inFigure 4, infection of AdVHER-2 led to a mild up-regulation of CD40, CD86, and ICAM-1 expression onDCHER-2 cells. However, the DCHER-2 /TNF -a cells thatexpressed the TNF-a gene in trans were more markedlyaffected, with 2.3- to 3.2- fold greater expression of CD40,CD86, and ICAM-1 being evident in these cells, relative tothe untreated DCs. The expression of MHC class I and IIantigens, CD11b, CD11c, and CD80 remained unchangedon these DCs (data not shown).

DCHER - 2 / TNF -a cells induce enhanced T-cell proliferationin vitro

DCs are potent stimulators of primary mixed lymphocytereaction (MLR) and are able to induce the proliferation ofallogeneic CD8+ T cells in vitro.30 Just as stimulation ofantigen-specific T-cell responses by DCs is strongly af-fected by their maturational status,26 so too is stimulationof primary MLRs by these cells. Thus, we next com-pared the abilities of our DC populations to stimulateprimary MLRs against allogeneic CD8+ T cells. Wefound that DCHER-2 /TNF -a, which had been coinfectedwith AdVHER2/neu and AdVTNF-A, induced stronger

Figure 4 Comparison of the phenotypic changes by flow cytometry.The expression of CD40, CD86 and ICAM-1 in DCHER - 2, DCHER - 2 /

TNF -a and uninfected DCs (solid lines ) was analyzed by flow

cytometry using the rat anti -CD40, CD86 and ICAM-1 antibodies

and the FITC- labeled goat and rat antibody. The values of meanfluorescence intensity are shown in the right upper corners. Isotype-

matched monoclonal antibodies (dotted lines) were used as controls.

One representative experiment of two is shown.

Figure 3 Flow cytometric analysis of double transgene expression.

DCs were infected with either AdVGFP or AdVTNF-a at MOI of 100(1�107 DCs vs 1�109 pfu of either AdVGFP or AdVTNF-a ), or both

of these two adenoviruses at MOI of 100 for each adenovirus (1�107

DCs vs 1�109 pfu of AdVGFP and 1�109 pfu of AdVTNF-a ). The

infected DCs were stained by PE-conjugated anti–TNF-a antibodyas described under Materials and methods. The values of mean

fluorescence intensity in ( i ) upper left, ( ii ) lower right, and ( iii ) upper

right panels represent the percentage of DCs stained with ( i ) red

(PE), ( ii ) green (GFP), and ( iii ) both green (GFP) and red (PE)fluorescences, respectively. They also represent the DCs expressing

( i ) the transgene TNF-a, ( ii ) the transgene GFP and ( iii ) both

TNF-a and GFP, respectively.

Cancer Gene Therapy

HER-2- specific immunity of engineered dendritic cell vaccineZ Chen et al

782

allogeneic T-cell proliferative responses in vitro thanDCHER-2 (P<.05) and DCs (P<.01) (Fig 5).

DCHER -2 / TNF -a significantly enhance tumor-specific T-cellresponses in vitro

Next, we addressed the specific antitumor effector functionsinduced by vaccination of the mice with DCHER-2 /TNF -a.Spleen lymphocytes from immunized mice were coculti-vated with irradiated MCA26/HER-2 tumor cells for 4 daysand harvested for phenotypic analysis by flow cytometry. Asshown in Figure 6, these T lymphocytes comprising bothCD4+ and CD8+ T cells displayed high- level expression ofFasL and CD25 (IL-2R), indicating that they are activated Tcells. To assess the cytotoxic activity, we conducted achromium-release assay by using these activated T cellsas effector cells and the 51Cr- labeled MCA26 and MCA26/HER-2 tumor cells as target cells. As shown in Figure 7,

activated T cells from mice vaccinated with DCHER-2 /TNF -ashowed significant killing activity to MCA26/HER-2 cells.At an effector:target (E:T) ratio of 50, the specific killingactivity for activated T cells from mice with vaccination ofDCHER-2 /TNF -a and DCHER-2 was 53% and 39%, respec-tively. T cells from mice vaccinated with DCHER-2 /TNF -acells displayed substantially (36%) enhanced CTL activityrelative to analogous cells from animals vaccinated with theDCHER-2. This CTL activity was immunologically specific,in as much as none of these populations showed cytotoxicactivities against the HER-2/neu-negative MCA26 tumorcells, and T cells from the mice vaccinated with uninfectedDCs had only little activity against the MCA26/HER-2.

DCHER - 2 / TNF -a strongly induce protective tumor-specificimmunity in vivo

To examine whether DCHER-2 /TNF -a cells were also capableof inducing enhanced antitumor immunity in vivo, we vac-cinated mice twice with DCHER-2 and DCHER-2 /TNF -a, and10 days later challenged the animals with 3�105 MCA26/HER-2 tumor cells. As shown in Figure 8, MCA26/HER-2tumor cell challenges were invariably lethal within 6 weekspostimplantation for the PBS vaccination control mice.Vaccination with DCHER-2 can only protect 25% (2/8) ofthe mice from tumor growth after challenge of MCA26/HER-2. However, vaccination of DCHER-2 /TNF -a showedenhanced antitumor immunity, with 8/8 mice protected fromchallenge of 3�105 MCA26/HER-2 tumor cells. Moreimportantly, vaccination with DCHER-2 /TNF -a can only pro-tect mice from challenge of HER-2/neu-positive MCA26/

Figure 7 DC expressing a TNF-a transgene induce more robust

tumor-specific cytotoxic T-cell responses. Splenic lymphocytes were

harvested from mice that had been vaccinated with DCHER - 2 / TNF -a

(open circles), DCHER - 2 (black circles), or uninfected DCs (open

triangles ) were cocultured with irradiated MCA26/HER-2 cells

(40,000 rad) for 4 days, then used as effector cells in a chromium-

release assay with 51Cr - labeled MCA26/HER-2 tumor cells. Toconfirm that T-cell cytotoxicity was HER-2 /neu tumor specific, we

also included MCA26 cells (black triangles) as a target control and

activated T cells from DCHER - 2 / TNF -a - immunized mice as effector

cells. Each point represents the mean of triplicates. The experimentwas repeated once with consistent results.

Figure 6 Phenotypic analysis of activated T cells from DCHER - 2 /

TNF -a -vaccinated mice. Naive T cells and activated T cells fromDCHER - 2 / TNF -a - vaccinated mice were analyzed by flow cytometry

using the rat anti -CD3, CD4, CD8, CD25, and FasL antibodies and

the FITC- labeled goat anti - rat antibody (solid lines ). Isotype -

matched monoclonal antibodies (dotted lines) were used as controls.One representative experiment of two is shown.

Figure 5 DC expressing a TNF-a transgene induce stronger mixed

lymphocyte reactions. Irradiated DCHER - 2 / TNF -a (black triangles),DCHER - 2 (black circles), or DCs (open circles), 1�104 cells /well and

reciprocal dilutions thereof, were cocultured for 3 days with 1�105

allogeneic C57BL/6 T cells. The overnight [ 3H]thymidine uptake

seen on day 4 is expressed as the mean of three determinations.Background proliferation of DCs or T cells alone was always below

4000 cpm. One experiment of two is shown.

Cancer Gene Therapy

HER-2- specific immunity of engineered dendritic cell vaccineZ Chen et al

783

HER-2, but not HER-2/neu-negative MCA26 tumor cells,because 8/8 immunized mice died within 6 weeks postimplantation of MCA26 tumor cells. These data furtherconfirm that the antitumor immunity of mice vaccinated withDCHER-2 /TNF -a is HER-2/neu tumor specific.

Discussion

In this study, we constructed a recombinant adenovirusAdVHER-2 expressing the human HER-2/neu antigen. Tolimit the possibility of deleterious consequences, the part ofICD with the most transforming activity of HER-2/neumolecule was deleted.31 We then infected DCs with thisadenoviral vector and investigated the anti -HER2/neuimmune responses by DCHER-2 vaccination in an animalmodel using a murine tumor cell line MCA26/HER-2engineered to express HER-2/neu antigen. We chose to useAd vectors for gene transfer into DCs because these virusescan infect DCs with high efficiency and easily allowintroduction of multiple adenoviral vectors into a singlecell. Our data showed that vaccination of mice withAdVHER-2–transfected DCHER-2 induced (i ) HER-2/neu-specific CTL response with tumor-specific killing of39% at E:T ratio of 50, and (ii ) HER-2/neu-specificantitumor immunity capable of protecting 25% mice fromchallenge of 3�105 MCA26/HER-2 tumor cells.

As a further strategy to enhance the efficacy of DCvaccine, DCs have been genetically engineered to expresscytokines such as IL-12 and GM-CSF. In both mouse andhuman studies, IL-12 expression by engineered DCs canaugment priming to antigens delivered in a variety of ways tothe cells, such that these DCs were powerful catalysts for theinduction of tumor-specific CD4+ Th cells and CD8+

CTLs.32,33 Engineering DCs expressing GM-CSF alsoreportedly increase therapeutic antitumor immunity invivo.34 Recently, it has been further reported that theantitumor efficacy of DC vaccination using DCs engineered

to express human tumor antigens can be greatly enhanced bycotransduction of DCs with the above cytokine genes asimmunoadjuvants.35,36

The maturation processes of DCs are efficiently regulated,such that these cells can achieve different states ofactivation/maturation and thereby different functional prop-erties,26,37 depending on the precise nature of the signalsthey receive from their microenvironment. A number ofcytokines can significantly affect DC function at variouslevels, including their viability, morphology, migration,expression of MHC and accessory molecules, and bindingand processing of antigen peptides.38 For example, TNF-ais known to have substantial effects on DC maturation with astrong T-cell stimulatory potential.39,40 The maturation ofDCs is correlated with the induction of antitumor immunityby DC vaccines.26 More recently, we have demonstrated thatDCs infected with adenovirus expressing transgene TNF-aunderwent augmented cellular maturation.27 These matureDCs further induced strong CD8+ CTL cytotoxicity in vitroand substantially more effective antitumor immunity in vivoin animal models.

To enhance the efficacy of DCHER-2 vaccination, wefurther constructed DCHER-2 /TNF -a expressing both HER-2/neu and TNF-a by coinfection of DCs with AdVHER-2and AdVTNF-a. We found that DCHER-2 /TNF -a cellsdisplayed a significant amount of cell surface HER-2/neumolecule and secreted a significant amount of TNF-a(8 ng/mL) in the supernatant of infected DC cultures. TheseDCHER-2 /TNF -a cells also displayed up-regulation ofimmunologically important CD40, CD86 and ICAM-Imolecules compared with DCHER-2 cells, indicating thatthe former ones are more mature forms of DCs. Vaccinationof DCHER-2 /TNF -a induced stronger allogeneic T-cellproliferation and 36% enhanced HER-2/neu-specific CTLresponses in vitro than DCHER-2 cells. More importantly, itcould stimulate the significant anti–HER-2/neu immunityin vivo, which protected 8/8 mice from challenge of 3�105

MCA26/HER-2 tumor cells, but not HER-2/neu-negativeMCA26 tumor cells.

It was originally reported that plasmid DNA encoding atruncated rat HER-2/neu lacking the ICD could induceantitumor immunity that was as strong as the one induced bythe plasmid encoding the full - length HER-2/neu.12 Thisindicates no advantage of ICD in induction of protectiveimmunity. Evidence from clinical trials, however, demon-strated that CD4+ T cells derived from peptide vaccinerecognized peptide epitopes from both ECD and ICD ofHER-2/neu,8,41 suggesting ICD may represent a betterimmune target given its limited exposure to the externalenvironment. It has also been reported that only vaccinationwith DNA encoding ICD or ICD protein of HER-2/neu canelicit antitumor immunity, but not with DNA encoding ECDor ECD protein.14 Conceptually, vaccination with DNAencoding the full length of HER-2/neu or full - length HER-2/neu protein should have the advantage of presenting bothMHC class I and class II epitopes and, therefore, inducedenhanced anti–HER-2/neu immune responses. Therefore,we assume that vaccination of DCs infected with newadenovirus expressing the full - length HER-2/neu proteinwith a single amino acid substitution ( lysine to alanine),42

Figure 8 Vaccination of mice with DCHER - 2 / TNF -a induces strongerantitumor immunity than vaccination with DCHER - 2. BALB/c mice

(n=8) were vaccinated twice with DCHER - 2 and DCHER - 2 / TNF -a, then

s.c. challenged, 10 days subsequent to the second vaccination, with3�105 MCA26/HER-2 or MCA26 tumor cells. Control mice were

treated with PBS and challenged with 3�105 MCA26/HER-2 tumor

cells. The survival time of each mouse was monitored daily. The

animal experiment was repeated once with similar results.

Cancer Gene Therapy

HER-2- specific immunity of engineered dendritic cell vaccineZ Chen et al

784

which can inactivate the transforming activity of ICD, mayfurther enhance the efficiency of DC vaccine strategy asdescribed in this study.

Taken together, we demonstrated that CTL-mediatedanti–HER-2/neu immunity could be induced in animalmodel by vaccination of DCs infected with adenovirusAdVHER-2 expressing the human HER-2/neu antigen.More importantly, we further demonstrated that vaccinationof DCHER-2 /TNF -a cells coinfected with adenovirusesAdVHER-2 and AdVTNF-a could significantly enhanceHER-2/neu-specific CTL responses in vitro and antitumorimmunity in vivo. Therefore, DCs genetically engineered toexpress both the tumor antigen and cytokines such as TNF-aas an immunoadjuvant may represent a new direction in DCvaccine of cancer.

Acknowledgment

This study was supported by a research grant (3-782) fromSaskatchewan Cancer Agency.

References

1. Coussens L, Yang-Feng TL, Liao YC, et al. Tyrosine kinasereceptor with extensive homology to EGF receptor shareschromosomal location with neu oncogene. Science. 1985;230:1132–1139.

2. Slamon DJ, Godolphin W, Jones LA, et al. Studies of theHER-2/neu proto -oncogene in human breast and ovariancancer. Science. 1989;244:707–712.

3. Ishikawa T, Kobayashi M, Mai M, et al. Amplification of thec-erbB-2 (HER-2/neu) gene in gastric cancer cells. Am JPathol. 1997;151:761–768.

4. Paterson MC, Dietrich KD, Danyluk J, et al. Correlationbetween c-erbB-2 amplification and risk of recurrent disease innode-negative breast cancer. Cancer Res. 1991;51:556–567.

5. Andrulis IL, Bull SB, Blackstein ME, et al. neu /erbB-2amplification identifies a poor -prognosis group of women withnode-negative breast cancer. Toronto Breast Cancer StudyGroup. J Clin Oncol. 1998;16:1340–1349.

6. Ioannides CG, Whiteside TL. T cell recognition of humantumors: implications for molecular immunotherapy of cancer.Clin Immunol Immunopathol. 1993;66:91–106.

7. Disis ML, Pupa SM, Gralow JR, et al. High titer HER-2/neu protein -specific antibody can be detected in patientswith early -stage breast cancer. J Clin Oncol. 1997;15:3363–3367.

8. Disis ML, Grabstein KH, Sleath PR, et al. Generation ofimmunity to the HER-2/neu oncogenic protein in patients withbreast and ovarian cancer using a peptide -based vaccine. ClinCancer Res. 1999;5:1289–1297.

9. Zaks TZ, Rosenberg SA. Immunization with a peptide epitope(p369–377 ) from HER-2 /neu leads to peptide - specificcytotoxic T lymphocytes that fail to recognize HER-2/neu+

tumors. Cancer Res. 1998;58:4902–4908.10. Nagata Y, Furugen R, Hiasa A, et al. Peptides derived from

a wild - type murine protooncogene c-erbB-2/HER2/neucan induce CTL and tumor suppression in syngeneic hosts.J Immunol. 1997;159:1336–1343.

11. Weiner LM. Monoclonal antibody therapy of cancer. SeminOncol. 1999;26:43–51.

12. Chen Y, Hu D, Eling DJ, et al. DNA vaccines encoding full -length or truncated Neu induce protective immunity againstNeu-expressingmammary tumors.Cancer Res. 1998;58:1965–1971.

13. Wei WZ, Shi WP, Galy A, et al. Protection against mammarytumor growth by vaccination with full - length, modified humanErbB-2 DNA. Int J Cancer. 1999;81:748–754.

14. Foy TM, Bannink J, Sutherland RA, et al. Vaccination withHER-2/neu DNA or protein subunits protects against growthof a HER-2/neu-expressing murine tumor. Vaccine. 2001;19:2598–2606.

15. Gabrilovich DI, Corak J, Ciernik IF, et al. Decreased antigenpresentation by dendritic cells in patients with breast cancer.Clin Cancer Res. 1997;3:483–490.

16. Troy AJ, Summers KL, Davidson PJ, et al. Minimalrecruitment and activation of dendritic cells within renal cellcarcinoma. Clin Cancer Res. 1998;4:585–593.

17. Chaux P, Favre N, Martin M, et al. Tumor- infiltrating dendriticcells are defective in their antigen-presenting function andinducible B7 expression in rats. Int J Cancer. 1997;72:619–624.

18. Nestle FO, Alijagic S, Gilliet M, et al. Vaccination of mela-noma patients with peptide or tumor lysate -pulsed dendriticcells. Nat Med. 1998;4:328–332.

19. Dhodapkar MV, Steinman RM, Sapp M, et al. Rapid gen-eration of broad T-cell immunity in humans after a singleinjection of mature dendritic cells. J Clin Invest. 1999;104:173–180.

20. Cella M, Sallusto F, Lanzavecchia A. Origin, maturation andantigen presenting function of dendritic cells. Curr OpinImmunol. 1997;9:10–16.

21. Pardoll D. Cancer vaccines. Nat Med. 1998;4:525–530.22. Eggert A, Schreurs M, Boerman O, et al. Biodistribution and

vaccine efficiency of murine dendritic cells are dependenton the route of administration. Cancer Res. 1999;59:3340–3345.

23. Ribas A, Butterfield L, McBride W, et al. Genetic immuniza-tion for the melanoma antigen MART-1/Melan-A usingrecombinant adenovirus - transduced murine dendritic cells.Cancer Res. 1997;57:2865–2869.

24. Butterfield L, Jilani S, Chakraborty N, et al. Generation ofmelanoma-specific cytotoxic T lymphocytes by dendritic cellstransduced with a MART-1 adenovirus. J Immunol. 1998;161:5607–5613.

25. Lapointe R, Royal R, Reeves M, et al. Retrovirally transducedhuman dendritic cells can generate T cells recognizing multipleMHC class I and class II epitopes from the melanoma antigenglycoprotein 100. J Immunol. 2001;167:4758–4764.

26. Labeur M, Roters B, Pers B, et al. Generation of tumorimmunity by bone marrow-derived dendritic cells correlateswith dendritic cell maturation stage. J Immunol. 1999;162:168–175.

27. Zhang W, Chen Z, Li F, et al. TNF- transgene-expressingdendritic cells undergo augmented cellular maturation andinduce more robust T cell activation and antitumor immunitythan dendritic cells generated in recombinant TNF. Immunol-ogy. 2002. In press.

28. Wright P, Braun R, Babiuk L, et al. Adenovirus -mediatedTNF-a gene transfer induces significant tumor regression inmice. Cancer Biother Radiopharm. 1999;14:49–57.

29. Xiang J, Moyana T. Regression of engineered tumor cells secre-ting cytokines is related to a shift in host cytokine profile fromtype 2 to type 1. J Interferon Cytokine Res. 2000;20:349–354.

30. Inaba K, Young J, Steiman R. Direct activation of CD8cytotoxic T lymphocytes by dendritic cells. J Exp Med.1987;166:182–194.

Cancer Gene Therapy

HER-2- specific immunity of engineered dendritic cell vaccineZ Chen et al

785

31. Andrechek E, Hardy W, Siegel P, et al. Amplification of theneu / erbB -2 oncogene in a mouse model of mammarytumorigenesis. Proc Natl Acad Sci USA. 2000;97:3444–3449.

32. Tuting T, Wilson CC, Martin DM, et al. Autologous humanmonocyte -derived dendritic cells genetically modified toexpress melanoma antigens elicit primary cytotoxic T cellresponses in vitro: enhancement by cotransfection of genesencoding the Th1 -biasing cytokines IL -12 and IFN-a.J Immunol. 1998;160:1139–1147.

33. Nishioka Y, Hirao M, Robbins PD, et al. Induction of systemicand therapeutic antitumor immunity using intratumoral injectionof dendritic cells genetically modified to express interleukin 12.Cancer Res. 1999;59:4035–4041.

34. Curiel -Lewandrowski C, Mahnke K, Labeur M, et al.Transfection of immature murine bone marrow-deriveddendritic cells with the granulocyte -macrophage colony-stimulating factor gene potently enhances their in vivoantigen-presenting capacity. J Immunol. 1999;163:174–183.

35. Chen Y, Emtage P, Zhu Q, et al. Induction of ErbB-2 /neu-specific protective and therapeutic antitumor immunity usinggenetically modified dendritic cells: enhanced efficacy bycotransduction of gene encoding IL-12. Gene Ther. 2001;8:316–323.

36. Kass E, Panicali D, Mazzara G, et al. Granulocyte /macro-phage-colony stimulating factor produced by recombinant

avian poxviruses enriches the regional lymph nodes withantigen -presenting cells and acts as an immunoadjuvant.Cancer Res. 2001;61:206–214.

37. Banchereau J, Steinman RM. Dendritic cells and the control ofimmunity. Nature. 1998;392:245–252.

38. Jonuleit H, Knop J, Enk AH. Cytokines and their effects onmaturation, differentiation and migration of dendritic cells.Arch Dermatol Res. 1996;289:1–8.

39. Brunner C, Seiderer J, Schlamp A, et al. Enhanced dendriticcell maturation by TNF-a or cytidine-phosphate -guanosineDNA drives T cell activation in vitro and therapeuticantitumor immune responses in vivo. J Immunol. 2000;165:6278–6286.

40. Cumberbatch M, Kimber I. Tumor necrosis factor -a isrequired for accumulation of dendritic cells in draining lymphnodes and for optimal contact sensitization. Immunology.1995;84:31–35.

41. Kobayashi H, Wood M, Song Y, et al. Defining promiscuousMHC class II helper T-cell epitopes for the HER-2/neu tumorantigen. Cancer Res. 2000;60:5228–5236.

42. Ben-Levy R, Paterson H, Marshall C. A single autophos-phorylation site confers oncogenicity to the Neu/ErbB-2receptor and enables coupling to the MAP kinase pathway.EMBO J. 1994;13:3302–3311.

Cancer Gene Therapy

HER-2- specific immunity of engineered dendritic cell vaccineZ Chen et al

786