SYMPOSIUM IN WRITING

Rathindra Nath Baral Æ Asim Saha Æ Sunil K. ChatterjeeKenneth A. Foon Æ Arthur M. Krieg Æ George J. Weiner

Malaya Bhattacharya-Chatterjee

Immunostimulatory CpG oligonucleotides enhance the immune responseof anti-idiotype vaccine that mimics carcinoembryonic antigen

Received: 26 August 2002 /Accepted: 3 October 2002 / Published online: 5 March 2003� Springer-Verlag 2003

Abstract We have developed and characterized a mon-oclonal anti-idiotype (Id) antibody, designated 3H1,which mimics a specific epitope of carcinoembryonicantigen (CEA) and can be used as a surrogate for CEA.Anti-Id 3H1 induced anti-CEA immunity in differentspecies of animals as well as humans and showedpromise as a potential vaccine candidate in phase I/IIclinical trials for colorectal cancer patients. One area ofinterest to us has been the development of new immuneadjuvants that may augment the potency of 3H1 as atumor vaccine. Immunostimulatory oliogonucleotidescontaining the unmethylated CpG motif (CpG ODN)are potent inducers of both innate and adaptive immu-nity and can serve as suitable vaccine adjuvants. In thisstudy, using the CEA-transduced MC-38 murine coloncarcinoma model in syngeneic C57BL/6 mice, we as-sessed whether a select CpG ODN (1826) can functionas immune adjuvant in immunization of mice with anti-Id 3H1. Complete Freund’s adjuvant (FA) was used as agold standard in this system. A single immunization of3H1 mixed with CpG ODN 1826 was sufficient to inducemeasurable anti-CEA immunity in naı̈ve mice. However,3 immunizations every other week were necessary toobtain and sustain peak immune reactivity over a longperiod of time. With FA and 3H1, single immunization

was ineffective and multiple immunizations (5 to 6) wereneeded to achieve and sustain peak immunity. Anti-CEA antibody reactivity was comparable in bothgroups, but cellular immune reactivity as measured byimmune splenic lymphocyte T cell proliferation and cy-toxicity assay was slightly higher in the CpG ODNgroup. Mice immunized with 3H1 and either CpG ODNor FA were protected from challenge by lethal doses ofMC-38-CEA cells. However, the degree of protectionwas slightly higher and the duration of survival wassomewhat longer in the group of mice treated with 3H1plus CpG ODN. Thus, CpG ODN 1826 was faster thanFA in increasing anti-tumor immunity induced by anti-Id 3H1 immunization in this prophylactic model.

Keywords Anti-idiotype antibody Æ Cancer vaccine ÆCarcinoembryonic antigen Æ CpG oligonucleotide ÆImmunotherapy

Introduction

Carcinoembryonic antigen (CEA) is a tumor-associatedantigen (TAA) expressed on human colorectal carci-nomas and other adenocarcinomas and is a putativetarget for cancer immunotherapy. Active immunother-apy is an attractive approach for cancer immunother-apy as it promotes help to immune systems to attackmalignant cells in an antigen (Ag)-specific manner. Onearea of active immunotherapy involves the use of TAAwhich is typically weakly immunogenic. The use ofTAA is not always promising, since the immune systemhas become tolerant to the tumor antigens. An effectiveway of breaking tolerance is to present the requiredepitope in a different molecular environment to thetolerant host [28].

Another approach in active immunotherapy involvesthe use of anti-idiotype antibodies. This idea is based onJerne’s network hypothesis [12]. According to the net-work concept, immunization with a given Ag willgenerate the production of antibodies (Ab) against it,

Cancer Immunol Immunother (2003) 52: 317–327DOI 10.1007/s00262-002-0351-x

This article forms part of the Symposi um in Writing on ‘‘Anti-bodies in Cancer Immunotherapy’’, published in this issue (Vol. 52)of the journal

R.N. Baral Æ A. Saha Æ S.K. Chatterjee Æ K.A. FoonM. Bhattacharya-ChatterjeeDepartment of Internal Medicine and the Barrett Cancer Center,University of Cincinnati, Cincinnati, OH 45267, USAE-mail: malaya.chatterjee@uc.edu

A.M. Krieg Æ G.J. WeinerUniversity of Iowa, Iowa City, IA, USA

M. Bhattacharya-Chatterjee (&)The Vontz Center for Molecular Studies, Room 1316,University of Cincinnati, 3125 Eden Avenue,Cincinnati, OH 45267-0509

termed Ab1. This Ab1 can generate a series of anti-idi-otype (Id) antibodies against Ab1 termed Ab2. A subsetof Ab2, termed Ab2b, acts as the internal image of theoriginal antigen recognized by Ab1 and can be used assurrogate Ag. We have generated an anti-Id antibody,3H1, designated CeaVac, that functionally mimics CEA.We have shown previously that 3H1 can induce anti-CEA antibody in small animals and non-human pri-mates and 3H1 was successfully tested earlier in severalclinical trials [6, 7, 8]. In a pre-clinical murine tumormodel, mice immunized with 3H1 conjugated with key-hole limpet hemocyanin (KLH) in combination withFreund’s adjuvant (FA) were protected againstchallenge with lethal doses of CEA-transduced murinetumor cells [19]. In this study, using the same CEA-transduced MC-38 murine colon carcinoma model insyngeneic C57BL/6 mice, we assessed whether a selectCpG ODN (1826) could augment the anti-CEA immu-nity and survival induced by anti-Id 3H1.

Bacterial DNA and synthetic oligodeoxynucleotidescontaining unmethylated CpG motifs [cytosine-phosp-horothioate-guanine oligodeoxynucleotide (CpG ODN)]are potent immunostimulatory agents and can serve asgood vaccine adjuvants in tumor antigen (idiotype) im-munization [27]. CpG ODN can activate various im-mune cell subsets, including B cells, natural killer (NK)cells and antigen-presenting cells (APC) includingmonocytes, macrophages and dendritic cells (DC) [11,14, 17, 20, 24, 26, 30, 31]. An active immune responsealso involves the production of various cytokines thatwork in an integrated fashion both locally and system-ically. CpG ODN have been shown to induce theproduction of various cytokines including interleukin-6(IL-6), IL-12, IL-18, tumor necrosis factor alpha (TNF-a), interferon alpha (IFN-a) and IFN-c [4, 13, 25, 31].Weiner et al. [27] have reported that a single immuni-zation with idiotype and CpG ODN results in enhancedproduction of Ag-specific IgG2a Ab, which suggest a Thelper 1 (Th1) type immune response, and prevents tu-mor (38C13 B cell lymphoma) growth when adminis-tered by s.c. or i.d. route. The mechanisms responsiblefor CpG ODN-induced immune activation are still un-der investigation. It has been reported that CpG ODNcan trigger the production of reactive oxygen specieswhich activate transcription factor NF-jB [26, 32]. Thisactivation may subsequently activate other genes in-volved in cellular activation. Regardless of the mecha-nism involved, it is clear that select CpG ODN havepowerful immunologic effects and can serve as potentimmune adjuvant in different circumstances, includingcancer immunotherapy.

In this study, we have attempted to elucidate the ef-fectiveness of single and multiple 3H1 immunizationprotocols with CpG ODN 1826 in enhancing anti-CEAand T-cell immunity in C57BL/6 mice. We also tested inthis murine tumor model whether the immunity inducedby 3H1-CpG ODN immunization could protect miceagainst challenge by a lethal dose of CEA-transducedmurine colon carcinoma cells.

Materials and methods

Materials

Purified oligodeoxynucleotide containing CpG motifs (CpG ODN1826) was obtained from the University of Iowa. Human CEA wasobtained commercially from Fitzgerald (Concord, Mass.). Thismaterial was purified from a human colon carcinoma liver metas-tasis and purity amounted to >98% by SDS-PAGE. The mon-oclonal anti-CEA Ab, 8019 (IgG1j) has been described previously[3]. Cell culture media were obtained from Gibco (Grand Island,N.Y.). [3H]-Thymidine, [125I]-K1, and [51Cr]-sodium chromate wereobtained from Perkin Elmer (Boston, Mass.).

Mice and cell lines

Female C57BL/6 mice (H-2b) were obtained from Harlan Labo-ratories (Indianapolis, Ind.). Mice were used at 6–8 weeks of age.Animals were maintained and treated according to guidelines es-tablished by the Animal Care and Use Committee of the Universityof Cincinnati. MC-38 murine colon adenocarcinoma [9] and humanCEA transduced murine carcinoma cells, MC-38-CEA [21] werekindly provided by Dr. Jeffrey Schlom (National Cancer Institute,NIH).

Anti-Id vaccine

Generation, purification and characterization of anti-Id antibody3H1 designated as CeaVac, have been described earlier [3]. Onecontrol anti-Id antibody, 11D10 [2], which mimics an epitope re-lated to a breast cancer-associated human milk fat globule, wasalso used in this study. In most experiments, mice were immunizedwith 3H1 or 11D10 (control) in the presence of CpG ODN 1826 asadjuvant. In some experiments, 3H1 was used along with a carrierand a potent adjuvant (KLH conjugated 3H1+FA).

Immunization

Two immunization protocols were followed, i.e. single and multi-ple. In the single immunization protocol, mice were immunizedwith a mixture of 3H1 (50 lg) and CpG ODN (50 lg) by s.c. routeon day zero. In the multiple immunization protocol, mice wereinjected using the same regimen 4 times every other week, i.e. ondays 0, 14, 28 and 42. In these experiments anti-Id Ab 11D10 wasused as control. To compare the effectiveness of CpG ODN as anadjuvant in 3H1 vaccinations, mice were also injected with KLH-conjugated 3H1 plus FA. Sera were drawn 7 days after eachimmunization by orbital puncture of anesthetized mice and storedat –20�C until use.

Detection of antibodies

Microtiter plates (96-well) were coated with 500 ng/well of a so-lution of 3H1 (Ab2) in PBS overnight at 4�C. Wells were blockedwith 1% bovine serum albumin (BSA) in phosphate-buffered saline(PBS) for 1 h at room temperature. To each well, 50 ll of125I-labeled 8019 (Ab1) containing 100,000 cpm in 1% BSA in PBSwas added simultaneously with 50 ll of diluted murine sera in PBS.After incubation at room temperature for 2 h, the plateswere washed and the radioactivity remaining in the wells was de-termined with a gamma counter (Packard, Meridian, Conn.). Fordetermining the inhibition of binding of 125I-labeled 8019 to CEA,plates were coated with 200 ng/well of a solution of CEA in PBSand blocked with BSA as described. To each well, 50 ll ofdiluted serum and 50 ll of 125I-labeled 8019 were added simulta-neously and incubated for 2 h at room temperature. Afterincubation, the plates were washed and the radioactivity was

318

determined. Assays were performed at least in triplicate, andsamples with a SD of less than 10% were used to calculate themean.

ELISA

Anti-CEA antibodies in murine sera were detected by ELISA.ELISA plates were coated with 50 ll of solution of purified CEA,100 ng/well overnight at 4�C. Wells were blocked with 1% BSA-PBS for 1 h at room temperature. Fifty microliters of diluted serumwere added to each well. After incubation at room temperature for2 h, the plates were washed with PBS containing 0.05% Tween-20.Fifty microliters of goat anti-mouse IgG conjugated with alkalinephosphatase (1:1,500-fold dilution) were added to each well. Afterincubation for 2 h at room temperature, the plates were washedthoroughly. Fifty microliters of substrate dissolved in diethanol-amine buffer (5 mg of substrate in 5 ml of buffer) were added toeach well. Absorbance at 405 nm was determined by ELISA readerafter 60 min at room temperature. Assays were performed at leastin triplicate for each sample, and values within a 10% variancewere included for calculation of the mean.

Flow cytometry

CEA-positive MC-38-CEA and CEA-negative MC-38 cells (1·106cells/tube) were incubated with 100 ll of diluted pre-immune sera orsera obtained after each immunization or 100 ll (1 lg) ofAb1 (8019)for 1 h at 4�C.After thorough washing, the cells were incubated withgoat anti-mouse F(ab¢)2 IgG-FITC-labeled antibody (Biosource,Camarillo, Calif.) for 30 min at 4�C. The labeled cells were washedtwice, fixed in 1% paraformaldehyde, and analyzed by flow cytom-etry (EPICS XL/MCL; Beckman-Coulter, Hialeah, Fla.).

ADCC

CEA-positive (MC-38-CEA) and CEA-negative (MC-38) targetcells were labeled by incubation of 2·106 cells/0.5 ml of serum-freeDMEMmedium with 200 ll of [51Cr]-sodium chromate at 37�C for1 h. After thorough washing, the cells were suspended in serum-freeDMEM at a concentration of 1·104 cells/25 ll of medium. Spleensisolated from 3 randomly selected mice, which were either immu-nized with anti-idiotype vaccines or PBS, were pooled, and a singlecell suspension prepared by mechanical dissociation were used aseffector cells. Labeled cells were added in 96-well plates, 1·104 cells/well, followed by the addition of 100 ll of diluted mouse serum (preor post), and 1·106 of effector cells in 50 ll of serum-free DMEM.Plates were incubated at 37�C for 4 h and supernatant of each wellwas harvested by using Skatron Harvesting System (Sterling, Va.)and radioactivity released was counted in gamma counter. Spon-taneous release was determined by assay of radioactivity releasedfrom 51Cr-labeled cells incubated in DMEM medium without ef-fector cells. The maximum or total release was defined as countsfrom labeled target cells incubated with 1%Triton X-100. Sera frommice immunized with PBS or unrelated anti-Id were used as control.Assays were performed in triplicate. Spontaneous release in eachexperiment was around 15% to 20% of the maximum release. Thepercentage specific cytotoxicity was calculated as:

Experiment release� spontaneous release

Total release� spontaneous release� 100

T-cell proliferation assay

Spleen cells pooled from 3 randomly selected mice, which were ei-ther immunized once with 3H1-CpG or 11D10-CpG, were isolatedat different time points post-immunization. These cells (2·105/well)were cultured in a 96-well Falcon plate in the presence of differentstimulants (500–1,000 ng/well) in a tissue culture incubator. On day

4, [3H] thymidine (1 lCi/well) was added, and cells were harvestedafter 18 h. The amount of incorporated [3H] thymidine was ana-lyzed in a liquid scintillation counter (Packard, Meridian, Conn.).Each assay was performed at least in triplicate, and the mean wascalculated from samples with a SD of less than 10%.

Cytotoxicity assay

Splenocytes were harvested and pooled from 3 mice/group, 7 daysafter the fourth immunization. These cells (2·106/ml) were resti-mulated by culture with 3H1 (50 lg/ml) and 30 U/ml rhIL-2. After5 days of culture, these in vitro restimulated splenocytes wereharvested and tested in a standard 51Cr release assay for theirability to lyse the CEA-expressing target cells. 51Cr-labeled tumorcells (1·104 cells/well) were incubated with effector cells (in tripli-cate) at an E:T ratio of 60:1 in round-bottomed 96-well microtiterplates. After 4 h, the supernatant from each well was harvested asdescribed above and the radioactivity was measured with a gammacounter. Specific lysis was calculated by using the formula de-scribed previously.

Murine model for determining the efficacy of 3H1 vaccines

A murine tumor model expressing human CEA was developed byDr. Jeffrey Schlom at the NIH. Murine colorectal carcinoma cells,MC-38, were transduced with human CEA [21]. The transducedcell line, MC-38-CEA, constitutively expresses CEA in culture. Dr.Schlom kindly provided us with the CEA-transduced and non-transduced cell lines. When these cells are injected s.c. (5·105 cells/mouse) into syngeneic C57BL/6 (H-2b) mice, tumors develop in100% of the mice within 10–15 days.

Cell culture

MC-38 and MC-38-CEA were grown in DMEM containing 10%FCS, 1% L-glutamine, 100 U/ml penicillin, and 100 lg/ml strep-tomycin in a humidified atmosphere of 5% CO2 at 37�C. MC-38-CEA cells were cultured in the presence of 200 lg/ml neomycinanalogue G-418. Cells (50–60% confluence) were harvested fromflasks following brief treatment with trypsin [19].

Survival of immunized mice

Mice were immunized with either 3H1-KLH conjugate and FA,3H1-CpG or control vaccine (11D10-CpG) as described above. Asham vaccination with PBS was also performed for comparison.Each group of immunized mice was divided into 2 equal groups fortumor challenge. Ten days after the last immunization, tumorchallenge was performed with 5·105 cells/mouse of MC-38-CEA orMC-38 cells. Growth of tumor and survival were monitored dailyand tumor volume was determined weekly.

Statistical evaluation

Statistical evaluation was performed using SigmaStat software(Jandel, San Rafael, Calif.). P <0.05 was considered to indicatestatistical significance.

Results

Single immunization with CpG ODN 1826 inducesantigen-specific antibody development

To induce Ab3 production in the sera, mice were im-munized with a single s.c. injection of 50 lg of 3H1 or

319

50 lg of isotype matched anti-idiotype Ab 11D10, inPBS mixed in aqueous solution with 50 lg of CpGODN. Serum was obtained weekly and Ab3 produced inthe sera was determined by radioimmunoassay (RIA)using sera pooled from 5 randomly selected mice. Pre-immune sera were used as control. Data in Fig. 1 showthat Ab3 induced in mice by 3H1 vaccination signifi-cantly inhibited the binding of Ab1 to Ab2, although thelevel of Ab3 induced by single immunization declinedgradually with time. Sera obtained from mice immu-nized with 11D10, which mimics an epitope related tobreast cancer-associated antigen, human milk fat glob-ule [2] did not inhibit the binding of Ab1 (8019) to Ab2(3H1). Also, less than 10% inhibition was obtained withpre-immune sera. Thus, Ab developed in mice immu-nized with 3H1 was truly anti-anti-idiotypic in nature.

The paratope (binding site) of Ab3 developed in miceafter 3H1 immunization was analyzed by inhibition ofbinding of 125I-labeled 8019 to CEA. As described inMaterials and methods, after immunization, sera col-lected from mice were pooled at different time points,diluted 1:10 with PBS and used for analysis. As illus-trated in Fig. 2, single immunization of 3H1 in combi-nation with CpG ODN 1826 was sufficient to induce anAb3 response that could significantly inhibit the bindingof 8019 to CEA, though the titer developed did not lastfor long. Sera obtained from mice immunized with11D10 or pre-immune sera were used as control in thisassay.

Immunization with Ab2b can lead to the generationof anti-anti-Id antibodies (Ab3) that recognize the cor-responding original Ag identified by the Ab1. Because ofthis Ab1-like reactivity, the Ab3 is also called Ab1¢ toindicate that it might differ in its other idiotopes fromAb1. Ab1¢ developed in the Ab3 population is likely tohave an antitumor property. Sera were obtained weeklyafter immunization and the level of Ab1¢ was evaluated

by ELISA. CpG ODN 1826 was effective as an immuneadjuvant at inducing the production of anti-CEA anti-body (Fig. 3). Ab1¢ can be detected in the sera of 3H1immunized mice, even after 35 days post-immunization.ELISA with control anti-Id immunized mice serum at1:10 dilution had an absorbance of 0.19 at 405 nm,which was similar to the value obtained using the pre-immune serum at 1:10 dilution.

Development of cellular immune response

Cellular immunity induced in mice immunized with 3H1or 11D10 in combination with CpG ODN was evaluatedat different time points post-immunization. Spleens fromat least three mice selected at random were pooled foruse in this assay. T-cell proliferation was carried out asdescribed in Materials and methods using 3H1 and CEAas stimulants. One microgram of concanavalin A

Fig. 1 Ab1 (mAb 8019) binding inhibition to Ab2 (3H1) byimmune sera. Induction of anti-anti-idiotype antibody (Ab3) inmice by a single immunization with 3H1-CpG or 11D10-CpG onday zero. Plates were coated with Ab2 (3H1) and the binding of125I-labeled Ab1 (8019) was tested for inhibition in the presence ofimmune sera at 1:10 dilution. Sera were collected at different timepoints after immunization. Pre-immune sera at 1:10 dilution wereused as control which showed less than 10% binding inhibition

Fig. 2 Ab1 (mAb 8019) binding inhibition to hCEA by immunesera. Mice were immunized once with either 3H1-CpG or 11D10-CpG on day zero. Sera were collected at different time points afterimmunization. Plates were coated with CEA and the binding of 125Ilabeled Ab1 (8019) was tested for inhibition in the presence ofimmune sera at 1:10 dilution. Pre-immune sera at 1:10 dilution wereused as control which showed no binding inhibition

Fig. 3 ELISA with CEA-coated plate. Induction of anti-CEA(Abl¢) antibody in mice by single immunization with 3H1-CpG or11D10-CpG on day zero. Plates were coated with CEA andimmune sera were tested at 1:10 dilution. Pre-immune sera at 1:l0dilution were used as control which showed no binding

320

(ConA) was used in these experiments as a positivecontrol. Mean counts from at least triplicate wells, witha SD of less than 10% were used to calculate the mean.Results in Fig. 4 indicate that significant proliferationtakes place using 3H1 or CEA as a stimulant, butstimulation declines sharply after 3 weeks post-immu-nization. There was negligible stimulation in the pres-ence of unrelated anti-Id or antigen (BSA). Thestimulation of spleen cells from mice after immunizationwith 11D10-CpG in the presence of 3H1 or CEA asstimulant in the assay, was significantly lower than thevalue obtained with 3H1-CpG immunization.

Single immunization of mice using CpG ODNas adjuvant confers some survival benefitover control mice

Mice were immunized once with 3H1-CpG or 11D10-CpG or PBS on day zero. On day 10, each group of im-munizedmicewas challengedwithCEA-positive,MC-38-CEA cells. As shown in Fig. 5, CEA-positive tumor grewprogressively in the group of mice immunized with eithercontrol anti-idiotype antibody, 11D10 or mock-vacci-natedwith PBS.MC-38-CEA tumor growthwas slower inthe group ofmice immunizedwith 3H1-CpG.Use of CpGODN 1826 as adjuvant markedly reduced tumor growthwhen compared with control animals (P<0.001). Miceimmunized with control vaccine died within 42 days,whereas some mice (3 out of 10) survived more than 58days when immunization was performed with 3H1-CpG(Fig. 6). From our previous experience, we know thatsingle immunization of anti-Id 3H1 in combination withFA in this murine model was ineffective in inducing anti-CEA humoral and cellular immunity [19].

Multiple immunizations are more effective than singleimmunization for induction of immune response usingCpG ODN as adjuvant

Single immunization with 3H1-CpG was able to inducemeasurable anti-CEA immunity in C57BL/6 mice and

mice were protected to some extent from subsequenttumor challenge. We next tested CpG ODN 1826 asadjuvant in multiple immunizations. Mice were immu-nized with 3H1-CpG or 11D10-CpG on day 0 andboosted on days 14, 28, and 42. A group of mice weremock-vaccinated with PBS. Sera were drawn 7 days af-ter each immunization and Ab3 developed in sera weremeasured by RIA. As illustrated in Fig. 7, Ab3 devel-oped in experimental mice after the first immunizationsignificantly inhibited the binding of 125I-labeled 8019 to3H1. The peak value obtained after first immunizationremained almost unchanged after (3 to 4) multiple im-munizations. Sera obtained from mice immunized with11D10-CpG and pre-immune sera were used as controlin this assay. Paratope analysis of Ab3 developed in seraafter multiple immunizations was determined by inhi-bition of binding of 125I-labeled 8019 to CEA. Onceagain, sera from experimental mice inhibited this bind-ing; the latter was significant after the first immuniza-tion and remained same after boosting (Fig. 8).

Fig. 4 T cell proliferation assay. Proliferation of splenocytes ofmice immunized once with 3H1-CpG or 11D10-CpG in thepresence of 3H1 and CEA as stimulants. T-cell proliferation wasassayed by determining [3H] thymidine incorporation

Fig. 5 Tumor progression in immunized mice. Mice were immu-nized once with 3H1-CpG or 11D10-CpG or PBS on day zero andwere challenged with 5·105 MC-38-CEA cells s.c. on day 10.Tumor development was monitored daily in all 3 treatment groups

Fig. 6 Survival curve of mice that received a single immunization.Mice were immunized once with 3H1-CpG or 11D10-CpG. Tendays after immunization, mice were injected s.c. with 5·105 MC-38-CEA cells and survival was monitored daily

321

To further verify the nature of Ab1¢ induced by3H1-CpG immunization, flow cytometric analysis ofpre-immune sera and pooled sera from 3 mice after eachimmunization with 3H1-CpG was performed. Results inFig. 9 demonstrate that sera from mice immunized withthe 3H1 vaccine can bind to the MC-38-CEA cell surface(Fig. 9A), identical to monoclonal anti-CEA Ab 8019after the fourth immunization, whereas binding withpre-immune sera was almost negative. CEA-negativeMC-38 cells did not bind sera from mice immunizedwith 3H1 vaccine or mAb 8019 (Fig. 9B). Sera from11D10-CpG immunized mice also did not bind to eitherMC-38 or MC-38-CEA cells (data not shown).

Direct comparison of CpG ODN as adjuvantwith FA in the murine model

In some experiments, mice were injected 4 times ondays 0, 14, 28 and 42 with either 3H1-CpG or 3H1-

KLH-FA or 11D10-CpG. Blood was collected 7 daysafter the last immunization. To determine whetherAb1¢ generated by 3H1 immunization was cytolytic forCEA-positive tumor cells, ADCC was performed. Re-sults in Fig. 10 illustrate that, sera obtained from miceimmunized with either 3H1-KLH-FA or 3H1-CpGcould lyse CEA-positive tumor cells, MC-38-CEA butnot CEA-negative MC-38 cells. Sera obtained frommice immunized with control vaccine or mock-vacci-nated with PBS could not lyse the target cells, sug-gesting that tumor cell lysis was dependent on Agexpression by target cells.

In a number of murine systems, CpG ODN inducesdevelopment of enhanced CTL activity [5, 15, 22]. Toexplore the cellular mechanism involved, it was im-portant to investigate whether CpG ODN 1826 couldinduce a CTL response following immunization withanti-Id vaccine in vivo. Seven days after the last im-munization, spleens were harvested and CTL assay wasperformed as described in Materials and methods. Ef-fector cells obtained from mice immunized with 3H1-CpG or 3H1-KLH-FA could lyse the CEA-transducedMC-38-CEA but not parent MC-38 cells (Fig. 11).CpG ODN enhanced the ability of 3H1 to induce aCTL response against CEA-positive target cells. Ofnote, effector cells obtained from mice immunized withcontrol vaccine or PBS could not lyse the target cells.Thus, immunity induced in mice immunized with 3H1was CEA-specific.

Protection of immunized mice againsttumor challenge

Experiments described above indicate that 3 to 4 im-munizations with 3H1-CpG were needed to obtain andmaintain peak anti-CEA immunity. If the immunityinduced is effective and CEA-specific, CEA-positive tu-mor cells should be rejected if implanted after this periodand growth of CEA-negative tumors should remainunaffected by this immunization. Ten days after the lastimmunization, each group of mice were further dividedinto 2 equal groups and challenged with lethal doses ofMC-38-CEA or parent MC-38 cells. As shown inFig. 12, control 11D10-CpG immunized mice developedtumor and died within 42 days. Mice immunized with3H1-KLH-FA survived (8 mice out of 10) more than 58days, whereas all mice (10 out of 10) immunized with3H1-CpG, survived more than 58 days. Mice immunizedwith 3H1-CpG or 3H1-KLH-FA had a survival rate thatwas statistically superior to that seen with 11D10-CpGimmunization group (P<0.001). Although the degree ofprotection was slightly higher and the duration of sur-vival of mice immunized with CpG ODN 1826 waslonger compared to mice immunized with FA, the dif-ference was not statistically significant (P=0.465).Regardless of immunization, mice challenged with CEA-negative tumor cells, MC-38, died between 30 to 40 days(data not shown).

Fig. 7 Ab1 (mAb 8019) binding inhibition to Ab2 (3H1) byimmune sera. Induction of anti-anti-idiotype antibody (Ab3) inmice after multiple immunizations with 3H1-CpG or 11D10-CpGfollowing the schedule described in Materials and methods. Serawere collected 7 days after each immunization. Plates were coatedwith Ab2 (3H1) and the binding of 125I-labeled Ab1 (8019) wastested for inhibition in the presence of immune sera at 1:10 dilution.Pre-immune sera at 1:l0 dilution were also used as control

Fig. 8 Ab1 (mAb 8019) binding inhibition to hCEA by immunesera. Mice were immunized 4 times with either 3H1-CpG or 11D10-CpG and sera were collected 7 days after each immunization. Plateswere coated with CEA and the binding of 125I-labeled Ab1 (8019)was tested for inhibition in the presence of immune sera at 1:10dilution. Pre-immune sera at 1:10 dilution were also used as control

322

Discussion

An important approach to cancer immunotherapy in-volves the development of an effective immune adjuvant.CFA is the standard adjuvant in animal models, al-though the intensity of local inflammation prevents itsuse clinically. Other adjuvants, such as QS21, Alugel, arecurrently being used in clinical trials. Recombinant cy-tokines also play a certain role as adjuvants. However, asuitable adjuvant that can induce activation of various

immune subsets and production of various cytokineswhich participate in the development of an active re-sponse is thought to be more effective and less toxic thanimmunization using external cytokine as adjuvant.

Bacterial DNA has significant immunostimulatoryeffects on various cell subsets including B cells, NK cellsand monocytes, and can induce the production of manycytokines that are important for the development ofanti-tumor immunity. The results in this study correlatewith the work from other laboratories, which suggests

Fig. 9A, B Immunofluores-cence analysis of CEA-posi-tive MC-38-CEA cells (A)and CEA-negative MC-38cells (B) after reaction withmice sera. Pre- and post-immune sera at 1:l0 dilution,obtained after the Ist, 2nd,3rd and 4th immunizationswith 3HI-CpG, were testedfor binding on both celllines. mAb 8019 at 1 lg wasused as a positive control inA and negative control in B

323

that select synthetic oligodeoxynucleotides containingthe unmethylated CpG motifs (CpG ODN) have im-munologic effects similar to those seen with bacterialDNA. CpG ODN are potent inducers of both innateand adaptive immunity and can serve as suitable vaccineadjuvants. CpG ODN can activate DC in both humanand murine systems as indicated by induction of IL-12production [11, 24]. In addition, possible direct costim-ulatory effects of CpG ODN on T cells have also beenreported [1, 13, 15]. IL-12 production by DC is known to

play a role in the proliferation and activation of T cellsand the induction of IFN-c production by activated Tcells [16]. IFN-c is involved in the development of cell-mediated immune responses [23], which is crucial for theinduction of anti-tumor immunity. However, little isknown about the mechanism, which is most significantfor the development of an enhanced cellular immuneresponse induced by CpG ODN.

The results described in this communication demon-strate that the anti-idiotype antibody 3H1, mimicking

Fig. 9 (Contd.)

324

human CEA, can induce humoral and cellular immuneresponses in C57BL/6 mice, which inhibit the develop-ment of tumor after challenge with lethal doses ofsyngeneic tumor cells. These therapeutic effects are anti-gen-specific, because tumor development was not pre-vented in immunized mice when challenged with the sametumor cells that were CEA negative. The therapeuticeffect was not observed in mice vaccinated with an iso-type-matched unrelated anti-idiotype antibody. In ourprevious study [19], mice were immunized with 3H1-KLHusing FA, and Ab3 developed in mice were detectableafter 2 to 3 immunizations and reached their maximumlevel following 5 to 6 immunizations. Cellular immuneresponse, as measured by T cell proliferation with CEA asstimulant, also peaked after 6 immunizations. Thus, toobtain maximum cellular and humoral immune re-sponses, 5 to 6 weekly immunizations were necessary.

In the present study, we assessed whether a selectCpG ODN (1826) could function as immune adjuvant in

immunization of mice with anti-Id 3H1. Results ob-tained with complete FA have been previously reported[19] and also in this study FA was used as a gold stan-dard with this system. A single immunization of 3H1mixed with CpG ODN 1826 was sufficient to induce adetectable anti-CEA immunity in naı̈ve mice. Ab3 in-duced in mice by 3H1 vaccine inhibited the Ab1-Ab2binding in this system (Fig. 1). Binding of Ab1 to CEAwas also significantly inhibited (Fig. 2). Anti-CEA Abdeveloped in mice was detected by ELISA (Fig. 3).These results suggested the induction of Ab1¢-like anti-bodies in the immunized mice. Proliferation of spleno-cytes in the presence of 3H1 was observed after a singleimmunization (Fig. 4). T cell proliferation in the pres-ence of CEA as stimulant suggests that the cellularimmune response induced by 3H1 immunization wasCEA-specific. The humoral and cellular immuneresponse induced in mice by a single immunization with3H1-CpG gradually declined with time, and immunitydeveloped by this immunization schedule also showedsome tumor protection in mice challenged with CEA-positive tumor cells (Figs. 5, 6).

Next, we investigated whether the anti-CEA immu-nity developed in mice could be maintained by multipleimmunizations of 3H1-CpG. Three to 4 immunizationswith 3H1-CpG were needed to obtain and maintainpeak anti-CEA immunity (Fig. 8). To confirm whetherthe induced Ab1¢ can bind to CEA expressed on thetumor cell surface, we analyzed the sera from immunizedmice by immune flow cytometric analysis using bothMC-38 and MC-38-CEA cells. These results confirmedthat the relevant Ab1¢ had been induced by 3H1 im-munization.

Several monoclonal antibodies recognizing TAAhave been shown to mediate ADCC [10, 18]. SignificantADCC was invoked in mice immunized with either 3H1-CpG or 3H1-KLH-FA. ADCC could be an additionalimportant mechanism for tumor protection by 3H1immunization. We shall confirm this observation byadoptive transfer of immune sera into naı̈ve mice.Wooldridge et al. [29] have demonstrated that CpGODN can enhance ADCC and improve the in vivo

Fig. 11 Cytotoxic T-cell mediated lysis of CEA positive tumorcells. Mice were immunized 4 times with either 3Hl-CpG or 11D10-CpG or 3H1-KLH and FA (3H1-FA) or PBS. Seven days after thefinal immunization, splenocytes were collected from immune miceand tested against MC-38-CEA and MC-38 cells as targets. Specificcell lysis was determined by the standard 51Cr release assay. Dataare representative of 2 different experiments involving 3 mice foreach group and expressed as means ± SD of triplicate assays

Fig. 12 Survival curve of multiple immunized mice. Mice wereimmunized 4 times with either 3H1-CpG, or 11D10-CpG or 3H1-KLH plus FA (3H1-FA). Ten days after the final immunization,mice were injected s.c. with 5·105 MC-38-CEA cells and survivalwas monitored daily

Fig. 10 ADCC by sera from immunized mice. Sera were obtainedfrom mice immunized 4 times with either 3Hl-CpG, or 11D10-CpG, or 3Hl-KLH plus FA (3H1-FA) or PBS. Specific lysis wasdetermined by in vitro ADCC assay using MC-38-CEA and MC-38cells as targets. Sera were diluted 1:5 with PBS. Pre-immune serawere used as control. Data are representative of 2 differentexperiments involving 3 mice in each group and expressed asmeans ± SD of triplicate assays

325

efficacy of monoclonal antibody therapy in a syngeneicmurine lymphoma model.

Cellular immunity induced in mice by 3H1-CpGvaccination was confirmed by enhanced CTL activity(Fig. 11). Our results correlate with the studies fromother laboratories which have suggested that CpG ODNcan induce the development of enhanced CTL activity[5, 15, 22]. Mice receiving multiple immunizations with3H1-CpG demonstrated survival benefit as compared tothe group that received a single immunization. 3H1-CpG was comparable or slightly better than 3H1-KLH-FA which was used as a gold standard in this model. It isnoteworthy that anti-Id 3H1 did not need any conju-gation with KLH to be used with CpG ODN as opposedto FA. In this regard, CpG ODN was a more usefuladjuvant than FA, and our study demonstrates thatadjuvant CpG ODN 1826 can orchestrate an immuneresponse that leads to the formation of enhanced hu-moral and cellular immunity. Select CpG ODN 1826 cantherefore serve as a potential immune adjuvant in hu-mans by enhancing the efficacy of immunization withincreased anti-tumor immunity.

Acknowledgements This work was supported in part by grants CA86025 and CA 91878 from the National Institutes of Health,Bethesda, Md. We thank Dr. Jeffrey Schlom (NIH) for the MC-38and MC-38-CEA cells and Audrey Morrison for typing themanuscript.

References

1. Bendigs S, Salzer U, Lipford GB, Wagner H, Heeg K (1999)CpG-oligonucleotides co-stimulate primary T cells in the ab-sence of antigen presenting cells. Eur J Immunol 29:1209

2. Bhattacharya-Chatterjee M, Mrozek E, Mukherjee S, CerianiRL, Kohler H, Foon KA (1994) Anti-idiotype antibodies aspotential therapeutic agents for human breast cancer. In: Cer-iani RL (ed.) Antigen and antibody molecular engineering inbreast cancer diagnosis and treatment. Plenum Publishing, NewYork, 139

3. Bhattacharya-Chatterjee M, Mukherjee S, Briddle W, FoonKA, Kohler H (1990) Murine monoclonal anti-idiotype anti-body as a potential network antigen for human carcinoem-bryonic antigen. J Immunol 145:2758

4. Bohle B, Jhan-Schmid B, Maurer D, Kraft D, Ebner C (1999)Oligodeoxynucleotides containing CpG motifs induce IL-12,IL-18 and IFN-c production in cells from allergic individualsand inhibit IgE synthesis in vitro. Eur J Immunol 29:2344

5. Chu RS, Targoni OS, Krieg AM, Lechmann PV, Harding CV(1997) CpG oliogonucleotides act as adjuvants that switch on Thelper 1 (Th1) immunity. J Exp Med 186:1623

6. Foon KA, Chakraborty M, John WJ, Sherratt A, Kohler H,Bhattacharya-Chatterjee M (1995) Immune response to thecarcinoembryonic antigen in patients treated with anti-idiotypevaccine. J Clin Invest 96:334

7. Foon KA, John WJ, Chakraborty M, Das R, Teitelbaum A,Garrison J, Kashala O, Chatterjee SK, Bhattacharya-ChatterjeeM (1999) Clinical and immune responses in resected coloncancer patients treated with anti-idiotype monoclonal antibodyvaccine that mimics the carcinoembryonic antigen. J Clin Oncol17:2889

8. Foon KA, John WJ, Chakraborty M, Sherratt A, Garrison J,Flett M, Bhattacharya-Chatterjee M (1997) Clinical and im-mune responses in advanced colorectal cancer patients treated

with anti-idiotype monoclonal antibody vaccine that mimics thecarcinoembryonic antigen. Clin Cancer Res 3:1267

9. Fox BA, Spiess PJ, Kasid A, Puri R, Mule JJ, Weber JS, Ro-senberg SA (1990) In vitro and in vivo anti-tumor properties ofa T-cell clone generated from murine tumor infiltrating lym-phocytes. J Biol Response Mod 9:499

10. Herlyn D, Herlyn M, Steplewski Z, Koprowsky H (1979)Monoclonal antibodies in cell-mediated cytotoxicity againsthuman melanoma and colorectal carcinoma. Eur J Immunol9:657

11. Jakob T, Walker PS, Krieg AM, Udey MC, Vogel JC (1998)Activation of cutaneous dendritic cells by CpG-containing oli-godeoxynucleotides: a role for dendritic cells in the augmenta-tion of Th1 responses by immunostimulatory DNA. J Immunol161:3042

12. Jerne NK (1974) Towards a network theory of the immunesystem. Ann Immunol 125C:373

13. Klinman DM, Yi AK, Beaucage SL, Conover J, Krieg AM(1996) CpG motifs present in bacteria DNA rapidly inducelymphocytes to secrete interleukin 6, interleukin-12, and inter-feron-c. Proc Natl Acad Sci USA 93:2879

14. Krieg AM, Yi AK, Matson S, Waldschmidt TJ, Bishop GA,Teasdale R, Koretzky GA, Klinman DM (1995) CpG motifs inbacterial DNA trigger direct B-cell activation. Nature 374:546

15. Lipford GB, Bauer M, Blank G, Reiter R, Wagner H, Heeg K(1997) CpG-containing synthetic oligonucleotides promote Band cytotoxic T cell responses to protein antigen: a new class ofvaccine adjuvants. Eur J Immunol 27:2340

16. Macatonia SE, Hosken NA, Litton M, Vieira P, Hsieh CS,Culpepper JA, Wysocka M, Trinchieri G, Murphy KM,O’Garra A (1995) Dendritic cells produce IL-12 and direct thedevelopment of Th1 cells from naı̈ve CD4+ T cells. J Immunol154:5071

17. Messina JP, Gilkeson GS, Pisetsky DS (1991) Simulation of invitro murine lymphocyte proliferation by bacterial DNA.J Immunol 147:1759

18. Pendurthi TK, Parker R, Schlom J, Primus J (1990) Lympho-kine-activated killer cell cytotoxicity against human colon car-cinomas enhanced by monoclonal antibody D612. Int J Cancer46:1021

19. Pervin S, Chakraborty M, Bhattacharya-Chatterjee M, ZeytinH, Foon KA, Chatterjee SK (1997) Induction of antitumorimmunity by an anti-idiotype antibody mimicking carcinoem-bryonic antigen. Cancer Res 57:728

20. Pisetsky DS, Reich C, Crowley SD, Halpern MD (1995) Im-munological properties of bacterial DNA. Ann NY Acad Sci772:152

21. Robbins PF, Kantor JA, Salgaller M, Hand PH, Fernsten PD,Schlom J (1991) Transduction and expression of human carci-noembryonic antigen gene in a murine colon carcinoma cellline. Cancer Res 51:3657

22. Roman M, Martinorozco E, Goodman S, Nguyen MD, Sato Y,Ronaghy A, Kornbluth RS, Richman DD, Carson DA, Raz E(1997) Immunostimulatory DNA sequences function as Thelper-1-promoting adjuvants. Nat Med 3:849

23. Schmitt E, Hoehn P, Huels C, Goedert S, Palm N, Rude N,Germann T (1994) T helper type 1 development of naı̈ve CD4+

T cells requires the coordinate action of IL-12 and IFN-c and isinhibited by TGF-b. Eur J Immunol 24:793

24. Sparwasser T, Koch ES, Vabulas RM, Heeg K, Lipford GB,Ellwart W, Wagner H (1998) Bacterial DNA and immuno-stimulatory CpG oligonucleotides trigger maturation and acti-vation of murine dendritic cells. Eur J Immunol 28:2045

25. Sparwasser T, Miethke T, Lipford G, Erdmann A, Hacker H,Heeg K, Wagner H (1997) Macrophages sense pathogens viaDNA motifs: induction of tumor necrosis factor-a mediatedshock. Eur J Immunol 27:1671

26. Stacey JJ, Sweet MJ, Hume DA (1996) Macrophages ingest andare activated by bacterial DNA. J Immunol 157:2116

27. Weiner GJ, Liu HM, Wooldridge JE, Dahle CE, Krieg AM(1997) Immunostimulatory oligodeoxynucleotides containing

326

the CpG motif are effective as immune adjuvants in tumorantigen immunization. Proc Natl Acad Sci USA 94:10833

28. Weigle WO (1961) The immune response of rabbits tolerant tobovine serum albumin to the injection of other heterologousserum albumins. J Exp Med 114:111

29. Wooldridge JE, Ballas Z, Krieg AM, Weiner GJ (1997) Im-munostimulatory oligodeoxynucleotides containing CpG motifsenhance the efficacy of monoclonal antibody therapy of lym-phoma. Blood 89:2994

30. Yamamoto S, Kuramoto E, Shimada T, Tokunaga T (1988) Invitro augmentation of natural killer cell activity and production

of interferon-alpha/beta and gamma with deoxyribonucleic acidfraction from Mycobacterium bovis BCG. Jpn J Cancer Res79:866

31. Yamamoto S, Yamamoto T, Shimada T, Kuramoto E, YanoO, Kataoka T, Tokunaga T (1992) DNA from bacteria, but notvertebrates, induces interferons, activates natural killer cells andinhibits tumor growth. Microl Immunol 36:983

32. Yi AK, Klinman DM, Martin TL, Matson S, Krieg AM (1996)Rapid immune activation by CpG motifs in bacterial DNA.Systemic induction of IL-6 transcription through an antioxi-dant-sensitive pathway. J Immunol 157:5394

327